This file has been updated periodically ever since TeX82 was born;
it has been summarized in "The errors of TeX," Software Practice &
Experience, July 1989. Entries are in chronological order;
thus the most recent news (including all bugfixes made since
that article was published) appears at the bottom of the file.

[Add 519 to these entry numbers to get the corresponding number in
the published article. The article also translates all module
numbers to their final form; what you are about to read is
"authentic source material" from the early days before TeX converged.]

-------------------------------------------------------------------------------

First updates to the TeX82 listing published in September, 1982.
(These changes were included in the original Version 0 of TeX, but they
were discovered after the listing went to press.)

1. Module 943, line 6 (bug discovered 9/28)
change "if cur_cmd=char_num then" to
        if (cur_cmd=letter) or (cur_cmd=other_char) then r:=qi(cur_chr)
        else if cur_cmd=char_num then

2. "pause" changed to "pausing" and "pause_code" to "pausing code", throughout.

3. Module 719, lines 8 and 11 (bug discovered 9/28)
insert "rule_save:=overfull_rule; overfull_rule:=0;" after "save_ptr-2;"
insert "overfull_rule:=rule_save;" before "q:=p+list_offset;"
and insert a declaration of "overfull_rule: scaled" in module 716.

4. Module 1128, lines 6 and following (bug discovered 9/28)
change "while n<>0 do" to "loop"
change "goto done" to
        begin scan_left_brace; new_save_level(false_group);
        goto done;
        end
change "... return 1130>;" to
                ... return 1130>
        else if n=0 then
                begin new_save_level(case_group); goto done;
                end;
and change "done: ... (false_group);" to "done:".

5. Module 182, line 10 (suggestion by DRF on 9/30)
change "0.0" to "?.?"

6. Module 682, new definition of math_spacing (decision of 10/2)
"0234000122*4000133**3**344*0400400*000000234000111*4111112341011"

7. Module 684, new code for case "4" (decision of 10/2)
"4": if cur_style<script_style then x:=thick_mu_skip_code else x:=0;
[Also in module 682 line 7, say "...a conditional thick space (\nonscript..."]

8. Module 11, trie_size changed from 7000 to 8000 because of new
improved (but longer) hyphenation patterns (10/4)

9. Module 453, forgot to change this when 454 changed (noted by DRF 10/6)
format_default_length=20 (not 22)
format_area_length=11 (not 13)

Changes to TEX.WEB made after Version 0 was released in October, 1982.

Since copies of TEX.WEB are not supposed to be edited, there are
two ways to make your version of TeX bug-free:
        1. Get a new copy of TEX.WEB.
        2. Put the corrections into your change file(s).
Some people will find (1) easier than (2), except the main TeX sources at
SCORE won't be updated quite as fast as the sources at SAIL (which have to
be translated into ascii before they are sent to the outside world).
Actually (2) will be quite easy, unless the list of changes becomes quite
long, so it is the recommended procedure. In case (2) it would be useful
to include a comment like "this is fix number xx" (using the numbering
scheme in this file), so that the change could readily be deleted at some
future time.

10. Module 857 line -5 (bug discovered 10/8/82 by HWT)
change it to: if h>0 then decr(h) else h:=trie_op_hash_size;

11. Module 457 line 3 (typo discovered 10/9/82)
change `\.!' to `\.\&'

12. Module 1245 line 8 (bug fixed 10/9/82, discovered by MMD)
insert the following between "begin" and "if":
if format_ident<>0 then initialize; {erase preloaded format}

** Version 0.1 incorporates the above changes.

13. This is an extension to the language, put in to satisfy people who
objected to the fact that \write (and \openout and \closeout) only
caused action after being deferred to the next \shipout. Some applications
call for immediate output, hence a new feature: \immediate followed by
\openout or \write or \closeout causes the output action to take place
without delay. For example, \immediate\write{x} is equivalent to
\shipout\vbox{\write{x}} except that the latter also puts an empty
page into the DVI file.

The extension requires the following new code:
13a. Insert `\immediate' after `\closeout' in module 1248.
13b. Define immediate_code=4 and include the following in module 1252:
        primitive("immediate",extension,immediate_code);
13c. Include the following in module 1254:
        immediate_code:print_esc("immediate");
13d. And, in module 1256:
        immediate_code:@<Implement \.{\\immediate}@>;
13e. Finally, there's a new module inserted after old module 1280.
Here is the WEB coding for this module:

@ The presence of `\.{\\immediate}' causes the |do_extension| procedure
to descend to one level of recursion. Nothing happens unless \.{\\immediate}
is followed by `\.{\\openout}', `\.{\\write}', or `\.{\\closeout}'.
@^recursion@>

@<Implement \.{\\immediate}@>=
begin get_nc_token;
if (cur_cmd=extension)and(cur_chr<=close_node) then
        begin p:=tail; do_extension; {append a whatsit node}
        out_what(tail); {do the action immediately}
        flush_node_list(tail); tail:=p; link(p):=null;
        end
else back_input;
end;

** Version 0.2 incorporates the above changes.

14. Like change 11, this one doesn't affect the program, it just improves
the documentation: Insert the following definitions in module 106:
define set_glue_ratio_zero(#) == #:=0.0 {assign representation of zero ratio}
define set_glue_ratio_one(#) == #:=1.0 {assign representation of unit ratio}

These macros are now introduced in a dozen or so future modules, thereby
eliminating most of the system-dependent changes needed elsewhere for ratios.
(Note: I also changed 0 to 0.0 in two places of module 182, where a glue_ratio
comparision was being made.)

15. Change of module 576 (discovered by HWT, 10/14/82)
"hd:quarterword" should be "hd:eight_bits".
(The same error occurs in module 522, but in that module the remedy is
simply to delete the declaration of hd, since this variable is no longer used.)

16. A most embarrassing bug (discovered by DRF, 10/14/82)
Replace module 531 by:

@ A mild optimization of the output is performed by the |dvi_pop|
routine, which issues a |pop| unless it is possible to cancel a
`|push| |pop|' pair. The parameter to |dvi_pop| is the byte address
following the old |push| that matches the new |pop|.

@p procedure dvi_pop(@!l:integer);
begin if (l=dvi_offset+dvi_ptr)and(dvi_ptr>0) then decr(dvi_ptr)
else dvi_out(pop);
end;

Now we need to make a few changes to subsequent modules:
16a. In 549, after "incr(cur_s)", insert
        if cur_s>0 then dvi_out(push);
and before "decr(cur_s)", insert
        if cur_s>0 then dvi_pop(save_loc);

16b. Delete "dvi_out(push);" and "dvi_pop;" from modules 553, 558, 562, 567.

16c. Change module 559 just as in 16a.

17. Module 605, line 6 (discovered 10/15/82)
The test should be "prev_depth>ignore_depth"

** Version 0.3 incorporates the above changes.

18. Module 11 (noticed by WLS, 10/16/82)
Delete the definition of align_size (it's harmless but never used)

19. (This change and the next cause major changes to the TRIP output;
file TRIP.LOG and its relatives are being kept up to date on area [tug,dek].)
The change avoids error messages when vpackage is called during output.
Such messages can occur when there was no error, because the page is being
boxed without the \skip glue from its insertions; so they should be omitted.
The user who really wants such messages can still get them by saying
"\setbox255=\vbox to 1ht255{\unbox255}".
In module 903, insert the followng before the declaration of "wait":
        save_vbadness:integer; {saved value of |vbadness|}
        save_vfuzz: scaled; {saved value of |vfuzz|}
Then in module 924, insert
        save_vbadness:=vbadness; vbadness:=inf_bad;
        save_vfuzz:=vfuzz; vfuzz:=max_dimen; {inhibit error messages}
before the call on vpackage, and
        vbadness:=save_vbadness; vfuzz:=save_vfuzz;
after.

20. Module 917 computes |page_size| improperly. (Noticed 10/21/82)
Delete the statement "page_so_far[1]:=page_so_far[1]+width(q);"
and change the preceding statement to:
        page_size:=page_size-h-width(q);
(The comment about page_so_far in the first paragraph of module 895 is
correct; I mistakenly introduced a bug in module 917 some months after
writing the first draft of the code, believing that I was making the
algorithm more elegant or something.)

** Version 0.4 incorporates the above changes.

21. Since TeX82 applied to (the "woven" documentation) TEX.TEX uses
about 11500 words of variable-size memory, I'm increasing hi_mem_size
(in module 12) from 12000 to 13000. Actually, I recommend using considerably
larger values for mem_max and hi_mem_size, whenever possible.

22. Addition of the \boxmaxdepth parameter (10/22/82):
This involves renumbering hfuzz_code through dimen_pars, in module 234,
to numbers 9 through 18; inserting the lines
        @d box_max_depth_code=8 {maximum depth of explicit vboxes}
        @d box_max_depth==dimen_par(box_max_depth_code)
        box_max_depth_code:print_esc("boxmaxdepth");
to module 234 and
        primitive("boxmaxdepth",assign_dimen,box_max_depth_code);
to module 235; and changing the call on vpack in module 996 to
        vpackage(link(head),saved(2),saved(1),box_max_depth);

** Version 0.5 incorporates the above changes.

23. Module 1224, line 4 (bug found by GMK/HWT on 10/26/82)
change "(k+x>eqtb_size)" to "(k+x>eqtb_size+1)"

24. Module 247, line 13 (bug found 10/26/82)
(bug was reflected in TRIP.LOG but not noticed)
change "ch_code(p)" to "ch_code(p-single_base)"

25. Module 670, line 18 (bug found 10/27/82)
interchange the statements "fetch(...)" and "math_type(...):=..."
(since the fetch routine sometimes has a side-effect of changing math_type)
A major change (Version 0.6) made on October 28:
The following list of changes counts as "number 26" on the list.
Fonts now have identifiers instead of code numbers; the "\:" primitive
has disappeared; and there are associated new features for "\the".

a. In module 11, delete bad_font_code.
b. In modules 170 and 172, delete print_esc(":"), and change
        `print_int(font_code' to `sprint_cs(font_ident'.
c. In module 205, the comment for set_font is revised.
d. In module 217, there are now five locations for control sequences
        that are perpetually defined; undefined_control_sequence
        is therefore defined to be frozen_control_sequence+5.
e. In module 248, sprint_cs is now included among <Basic printing procedures>.
f. Delete the primitive ":" in modules 250 and 251.
g. Add to module 376: <Declare procedures that scan font-related stuff>
h. In module 377, seven levels are now distinguished; we define
        font_val=4, ident_val=5, tok_val=6.
i. In module 380, delete the previous cases for def_family and set_font,
        and the following takes the place of case assign_toks:
          assign_toks,def_family,set_font,def_font: <Fetch a token list or
                font number or font identifier, provided that level=tok_val>;
j. Module 382 (which now has a new name) has this new ending:
        else if cur_cmd=assign_toks then scanned_result(equiv(m))(tok_val)
        else <Fetch a font number or a font identifier>
k. Module 389 becomes
        <Fetch a font number or a font identifier>=
        if cur_cmd=set_font then scanned_result(cur_chr)(font_val)
        else if cur_cmd=def_font then
                scanned_result(font_ident[cur_font])(ident_val)
        else    begin scan_four_bit_int;
                scanned_result(font_ident[equiv(m+cur_val)])(ident_val);
                end
l. In module 393, the relation `<>tok_val' becomes `<=mu_val'.
m. In module 426, the relation `=tok_val' becomes '>=ident_val'.
        Also add a new case to the case statement:
          font_val: begin print(font_name[cur_val]);
                if font_size[cur_val]<>font_dsize[cur_val] then
                        begin print(" at "); print_scaled(font_size[cur_val]);
                        print("pt");
                        end;
                end;
n. The body of module 427 becomes:
        begin p:=temp_head; link(p):=null;
        if cur_val_level=ident_val then store_new_token(cs_token_flag+cur_val)
        else if cur_val<>null then
                begin r:=link(cur_val); {do not copy the reference count}
                while r<>null do
                        begin store_new_token(info(r)); r:=link(r);
                        end;
                end;
        the_toks:=p;
        end
o. In module 480, delete user_font_code; there's a new comment:
        When the user defines \font\f, say, TeX assigns an internal number
        to the user's font \f. For example, if this internal number is 13,
        we will have font_ident[13]=p and equiv(p)=13, where p is the eqtb
        location of the control sequence \f.
p. In module 481, delete the declaration of font_number, and
        replace the declaration of font_code by
          font_ident:array[internal_font_number] of pointer;
q. New stuff in module 483 (also delete references to font_number, font_code):
        define bad_font_ident=frozen_control_sequence+4 {denotes a null font}
        font_name[undefined_font]:="nullfont";
        font_ident[undefined_font]:=bad_font_ident;
        text(bad_font_ident):="nullfont"; eq_level(bad_font_ident):=level_one;
        eq_type(bad_font_ident):=set_font;
        equiv(bad_font_ident):=undefined_font;
r. In module 490, parameter u is new of type pointer, and the
        read_font_info subroutine is changed to a function that
        returns an internal_font_number. There's a new local variable
          g:internal_font_namber; {the number to return}
        and we set g:=undefined_font immediately upon entering read_font_info.
        Also set read_font_info:=g just before exiting.
s. In module 491, print_int(u) becomes sprint_cs(u).
t. Delete the statements involving font_code and font_number in module 506,
        and set g:=f at the end of that module.
u. The body of module 507 becomes:
        <Declare procedures that scan font-related stuff>=
        procedure scan_font_ident;
        var f:internal_font_number;
        begin <Get the next non-blank non-call...>;
        if cur_cmd=set_font then f:=cur_chr
        else    begin print_nl("! Missing font identifier");
                help2("I was looking for a control sequence whose")
                ("current meaning has been defined by \font.");
                back_error; f:=undefined_font;
                end;
        end;
v. New beginning of module 508:
        The following routine is used to implement `\texinfo f n'.
        The boolean parameter writing is set true if the calling program
        intends to change the parameter value.
        <Declare procedures that scan font-related stuff>=
and "scan_font_number" is changed to "scan_font_ident".
w. In module 509, "print_int(font_code" becomes "sprint_cs(font_ident".
        Also "font code is defined" becomes "\font is loaded" in the help.
x. In module 940, the first line of help is changed to
        "You have to specify a font identifier,"
y. The set_font case in module 1138 reduces to
        set_font: define(cur_font_loc,data,cur_chr);
z. In module 1153, delete all the complicated stuff starting with "scan_int"
        and substitute simply this:
                scan_font_ident; define(p,data,cur_val);
                end;
aa. Change new_font to new_font(a) in module 1169, and add parameter
          (a:small_number)
        in module 1170. Also add the label common_ending, and declare
        variable u to have type pointer. The code beginning with "scan_int"
        is changed to the following:
          get_token;
          if cs_ptr=0 then
                 <Complain about the missing control sequence and return>;
          u:=cs_ptr; scan_optional_equals; scan_file_name;
          <Scan the ``at'' size specification>;
          <If this font has already been loaded, set f to the internal
                font number and goto common_ending>;
          f:=read_font_info(u,cur_name,cur_area,s);
          common_ending: define(u,set_font,f); font_ident[f]:=u;
          exit:end;
bb. New module 1172:
        When the user gives a new identifier to a font that was previously
        loaded, the new value becomes the font_ident of record.
        Font names `xyz' and `XYZ' are considered to be different.
        <If this font has already been loaded...>=
        for f:=font_base+1 to font_ptr do
                if [the test previously in module 1174] then goto common_ending
cc. New module 1173:
        <Complain about the missing control sequence and return>=
        begin print_nl("! A font identifier must be a control sequence");
        help2("You should say, e.g., `\font\f=fontfilename'.")
        ("(I'm going to ignore the \font command you just gave.)");
        back_error; return;
        end
dd. New module 1174:
        <Cases of print_cmd_chr...>=
        set_font:begin print("select font "); print(font_name[chr_code]);
                if font_size[chr_code]<>font_dsize[chr_code] then
                        begin print(" at "); print_scaled(font_size[chr_code]);
                        print("pt");
                        end;
                end;
ee. Delete the comment at the beginning of module 1227, and delete the
        loop for k:=0 to bad_font_code-1, and delete dump_int(undefined_font).
ff. Delete the repeat loop in module 1228.
gg. Change font_code to font_ident in module 1229, and change
        print_int(font_code to sprint_int(font_ident.
hh. In module 1230,
        begin undump(single_base)(undefined_control_sequence)(font_ident[k]);
        undump_qqqq(font_check[k]); [and continue as before]

The changes above have been incorporated into Version 0.6.
The ill-fated Version 0.7

27. (Here's a change that I retracted shortly after making it, since I
discovered that the source was flaky after all; and I also found a reliable
source [NBS Circular 570] confirming my original information. I include
the note here only for historical purposes...)

"After years of searching, I've finally found a definitive definition of
the printer's point; and (unfortunately) my previous conjecture was wrong.
The truth is that 83pc=35cm, exactly; so I am changing TeX to conform.
This means some changes in the comments of modules 101 and 517, and the
following changes to the program:

"27a. In modules 547 and 571, the appropriate DVI numbers are now:
        dvi_four(109375); dvi_four(2039808); {conversion ratio for sp}
"27b. The guts of module 418 become:
        if scan_keyword("in") then set_conversion(63246)(875)
        else if scan_keyword("pc") then set_conversion(12)(1)
        else if scan_keyword("cm") then set_conversion(996)(35)
        else if scan_keyword("mm") then set_conversion(498)(175)
        else if scan_keyword("bp") then set_conversion(10541)(10500)
        else if scan_keyword("dd") then set_conversion(996)(931)
        else if scan_keyword("cc") then set_conversion(11952)(931)
        else..."

*** Version 0.7 of TeX incorporated the changes above (82/10/30);
this version was withdrawn on 82/11/2.
Changes subsequent to Version 0.6

28. The experience with #27 did lead me to one improvement, thanks
to Chuck Bigelow, with respect to Didot points. (11/4/82)
Add the following comment to module 418:
        According to the definitions here, $\rm2660\,dd\approx1000.33297\,mm$;
        this agrees well with the value $\rm1000.333/,mm$ cited by Bosshard
        in {\sl Technische Grundlagen zur Satzherstellung\/} (Bern, 1980).
And change two lines of the code there as follows:
        else if scan_keyword("dd") then set_conversion(1238)(1157)
        else if scan_keyword("cc") then set_conversion(14856)(1157)

29. The new font material is made more robust by ensuring that \the\font
always returns a pointer to a control sequence whose command code is
set_font. (Change 11/4/82)

29a. The code changed in 26aa is changed again, to the following:
        <If the next token isn't a control sequence, issue a complaint and
                return; otherwise set u:=cs_ptr, and set hash_used to the
                location of a ``frozen'' copy of the new font identifier>;
        define(u,set_font,undefined_font);
        scan_optional_equals; scan_file_name;
        <Scan the ``at'' size specification>;
        <If this font has already been loaded, set f to the internal
                font number and goto common_ending>;
        f:=read_font_info(u,cur_name,cur_area,s);
        common_ending: equiv(u):=f; geq_define(hash_used,set_font,f);
        font_ident[f]:=hash_used;
        exit:end;
29b. Instead of 26cc, here's the new module 1173:
        We reserve a special control sequence for the font identifier;
        this one cannot be redefined by the user, so it is safe to
        return it as a value of \the\font.
        <If the next token isn't a control sequence...>=
        get_token;
        if cs_ptr<hash_base then
            begin print_nl(
             "! A font identifier must be a multiletter control sequence");
            help2("You should say, e.g., `\font\ffn=fontfilename'.")
            ("(I'm going to ignore the \font command you just gave.)");
            back_error; return;
            end;
        repeat if hash_is_full then overflow("hash size",hash_size);
        decr(hash_used);
        until text(hash_used)=0; {search for an empty location in hash}
        u:=cs_ptr; text(hash_used):=text(u); {copy the name}
        stat incr(cs_count); tats

30. Module 993, replace lines 5 to 7 (11/5/82):
else    begin if k=vmode then new_save_level(vbox_group)
        else    begin new_save_level(vtop_group); k:=vmode;
                end;
        if looseness<>0 then eq_word_define(int_base+looseness_code,0);
        if hanging_indent<>0 then
                eq_word_define(dimen_base+hanging_indent_code,0);
        if par_shape_ptr<>null then eq_define(par_shape_loc,shape_ref,null);
        end;

31. Module 403, replace line 5 (11/6/82):
if cur_tok<cs_token_flag then
        begin cur_val:=cur_chr;
        if cur_cmd<=right_brace then
                if cur_cmd=right_brace then incr(align_state)
                else decr(align_state);
        end

32. Improvements to error recovery for bad alignments (11/6/82):
In module 1039, replace "begin align_error; goto reswitch; end" by "align_error"
In module 1040, delete "cur_cmd:=left_brace;" and "cur_cmd:=right_brace;"
        and change "error" to "back_error"
In module 1041, delete "cur_cmd:=relax;"

33. More power to \let (11/8/82):
33a. Module 1141 becomes:
        <Assignments>+=
        let:    begin get_token;
                if cs_ptr<>0 then <Carry out the \let operation>
                else    begin print_nl
                         ("! You can use \let only with control sequences");
                        help2("I'm not \let-ting anything change here,")
                                ("since I can only do things like `\let\a=b'.");
                        back_error;
                        end;
                end;
33b. Module 1142 becomes:
        <Carry out the \let...>=
        begin p:=cs_ptr;
        repeat get_token;
        until cur_cmd<>spacer;
        if cur_tok=other_token+"=" then
                begin get_token;
                if cur_cmd=spacer then get_token;
                end;
        if cur_cmd>=call then add_token_ref(cur_chr);
        define(p,cur_cmd,cur_chr);
        end

34. This change helps you see an undefined control sequence
in certain unusual cases (discovered by JDH, 11/8/82):
Add the clause "(base_ptr=input_ptr) or" to module 294, line 2.

35. Here's an improvement in the formula for demerits; previously
more weight was given to minimizing bad spacing on lines with penalties,
so that (slightly loose hyphenated line)(OK line) was considered worse
than (OK hyphenated line)(quite loose line). (fixed 11/8/82)
Change lines 2--7 of module 772 to the following:
        d:=line_penalty+b; d:=d*d;
        if pi<>0 then
                if pi>0 then d:=d+pi*pi
                else if pi>eject_penalty then d:=d-pi*pi;

36. Minor change to save buffer space in non-INITEX (11/10/82):
Enclose the declaration of pool_file in module 50 by init...tini.

37. Minor improvement in format for the context of error messages (11/13/82):
37a. Module 289, type "inserted" split into "backed_up=3", "inserted=4"
and the other type numbers increase: "macro=5", etc.
37b. In module 296, backed_up: if loc=null then print_nl("<recently read>")
                        else print_nl("<to be read again> ");
                inserted: print_nl("<inserted text> ");
37c. Minor changes to comments in modules 289 and 293.
37d. Change "inserted" to "backed_up" in module 294.
37e. Add definition "back_list(#)==begin_token_list(#,backed_up)" in module 305.
37f. In module 306, first test should be ">=backed_up", second "<=inserted".
37g. Change ins_list to back_list in modules 307, 318, 375, 1195.
37h. Add definition to module 308:
        ins_error==begin back_error; token_type:=inserted; end
37i. Change back_error to ins_error in modules 362, 1040, 1046.
37j. Change "<to be read again>" to "<inserted text>" in module 973 help.

38. Module 407, the error message is changed (11/11/82) to:
"! Missing number, treated as zero".

39. Fix anomaly when hbadness or vbadness is small (11/11/82):
Module 593, line 2, "if (-x-total_shrink[normal]>hfuzz) or (hbadness<100)"
Module 603, line 2, "if (-x-total_shrink[normal]>vfuzz) or (vbadness<100)"

40. Added the \tokens primitive (11/13/82):
In module 223, defined tokens_loc and tokens (analogous to every_par).
In module 225, defined the primitive.
In module 1146, changed the comment to include tokens_loc as a possibility.

41. Change get_nc_token to get_token in module 374 (11/13/82).
(In particular \def\foo{...}\foo won't say "undefined c.s." now!)
(This change later retracted during debugging; it was found that
"endv" aborts the job, so this cure was worse than the disease. Then
it was re-established as part of the major change to conditionals,
because endv needed to be more robust anyway.)

42. In module 699, we make \span expand in a preamble (11/13/82).
Change lines -7 through -3 to:
        begin get_nc_token; goto reswitch;
        end;

43. Modules 321 and 322 have been altered for better efficiency (11/12/82)
and better exposition; the three conditions of module 322 are now
in separate if tests.

44. Frozen control sequences are now unredefinable. (11/16/82)
A new procedure get_r_token has been introduced to give uniform error
messages for \def, \let, \read, \font, \mathchardef, and to make them
incapable of changing the frozen equivalents.

Incidentally, get_nc_token is changed to get_x_token; and ch_code
becomes cat_code.
A major change (Version 0.8) made on November 14--16:
Conditional statements are taken from semantics to syntax.

This change, which counts as number 45 on the list, made it necessary
to renumber the modules. And it was such a drastic change, the
differences can only be sketched here, using the old module numbers
for reference. A variety of other things were cleaned up because this
change made it more natural for them to be handled differently.

45a. Module 158, inserted a permanently empty token list called null_list.
45b. Module 204, deleted if_test, case_branch, else_code, convert;
        inserted end_cs_name.
45c. Module 206, inserted if_test, fi_or_else, cs_name, convert, and
        end_template (the latter comes after long_outer_call).
45d. Module 217, there are now nine frozen control sequences; also null_cs.
45e. Modules 247 and 248, must be able to print a null control sequence.
45f. Module 268 (leave_transparent_group) disappears; comments in the
        previous modules also are appropriately simplified.
45g. Module 273, delete endv_token.
45h. Modules 277 and 278, delete references to endv (it no longer appears).
45i. After module 281, a procedure show_cur_cmd_chr takes the
        code for tracing commands from module 937.
45j. Module 317 now forbids \outer control sequences when skipping,
        and inserts \fi in front of them, for error recovery.
45k. Module 334 now uses null_cs if carriage-return is an escape.
45l. Module 343 no longer has endv test.
45m. Module 344, major extensions to expand_calls make it several modules
        longer. It saves global variables like cur_val. It changes
        end_template to endv. It processes cs_name and convert.
        It processes if_test by calling the "conditional" procedure.
        It processes fi_or_else by checking their legality, and (if legal)
        updating the if stack.
45n. Module 373 (pass_block) disappears.
45o. After module 400, radix and cur_val and cur_val_level are initialized.
45p. Module 428 changes so that conv_toks is a procedure rather than a
        function; this procedure is invoked by expand_calls.
45q. Modules 436 and 1032, delete reference to conv_toks.
45r. After module 443, the entire part 48 moves to this part of the program,
        and the parts are renumbered accordingly. \case is changed to
        \ifcase, and it resides under the if_test command.
        New routines are inserted to maintain a linked stack that records
        the current state of conditionals. There's a procedure
        pass_text that skips text while looking for \fi or \else or \or
        on level zero with respect to \if...\fi nesting. The previous
        routines for skip control via handle_right_brace are eliminated,
        and the new ones are somewhat simpler. The new \ifcat is mostly
        combined with the old \if; they no longer ignore spaces.
        The scanner_status is set to normal while processing \ifx.
        \ifx will consider \a to equal 0 after \let\a=0.
45s. Modules 701, 707: endv_token is replaced by a token for a frozen
        control sequence whose command code end_template makes it behave
        like an outer_call. After expand_calls, it is converted to another
        frozen control sequence, whose command code is endv. This two-step
        facilitates error recovery, instead of giving a fatal error stop.
45t. Module 874, remove "pass_block(1); goto done".
45u. After module 1047, routine for end_cs_name prints an error message.
45v. Module 1049, get_nc_token becomes get_token. (else $\ifmmode fails)
45w. Module 1202 gets another case (end_template).
45x. Module 1243, new warning is printed if the if stack isn't empty.
Changes after version 0.8

46. Declare c : ascii_code in module 825 (noted by DRF, 11/21/82)

47. Module 776 lines 4-6 (suggested by DRF, 11/21/82)
threshold:=pretolerance;
if threshold>=0 then
        begin stat if tracing_stats>2 then
                begin begin_diagnostic; print_nl("@@firstpass");
                end; tats
        second_pass:=false;
        end
else    begin threshold:=tolerance; second_pass:=true;
        end;

48. Protection for DVI files (added 11/16/82)
Replace lines 2 and 3 of module 570 by:
        <Update the values of max_h and max_v; but if the page is too large,
                goto done>;
Move the statement "dead_cycles:=0" from module 570 to module 568.
Declare "label done;" in module 568.
And add the following new module after 570:

@ Sometimes the user will generate a huge page because other error messages
are being ignored. Such pages are not output to the \.{dvi} file, since they
may confuse the printing software.

@<Update the values of |max_h| and |max_v|; but if the page is too large...@>=
if (height(p)>max_dimen)or(depth(p)>max_dimen)or(width(p)>max_dimen) then
        begin print_nl("! Huge page cannot be shipped out");
        help2("The page just created is more than 18 feet tall or")@/
         ("more than 18 feet wide, so I suspect something went wrong.");
        error;
        if tracing_output=0 then
                begin begin_diagnostic;
                print_nl("The following box has been deleted:");
                show_box(p);
                end_diagnostic(true);
                end;
        goto done;
        end;
if height(p)+depth(p)>max_v then max_v:=height(p)+depth(p);
if width(p)>max_h then max_h:=width(p)

49. New features \everymath and \everydisplay (12/2/82).
Make changes analogous to those for "\tokens" (see change 40).
Put the following just before the end of module 1050:
  if every_math<>null then begin_token_list(every_math,every_math_text);
And put the following just before the end of module 1056:
  if every_display<>null then
        begin_token_list(every_display,every_display_text);

50. New feature \futurelet (12/2/82):
Module 1141 changes again. (Also, "let" and "futurelet" share cmd code "let".)
let:    begin n:=cur_chr;
        get_r_token; p:=cs_ptr;
        if n=normal then
                begin repeat get_token;
                until cur_cmd<>spacer;
                if cur_tok=other_token+"=" then
                        begin get_token;
                        if cur_cmd=spacer then get_token;
                        end;
                end
        else    begin get_token; q:=cur_tok; get_token; back_input;
                cur_tok:=q; back_input; {look ahead, then back up}
                end; {|back_input| doesn't affect |cur_cmd|, |cur_chr|}
and continue with "if cur_cmd>=call", as before.

*** The changes above have been incorporated into version 0.9 of TeX.
Changes after version 0.9

51. new \endinput primitive (suggested by FY, 12/7/82):
\input and \endinput both use the same command code. The cases in module
955 are deleted and replaced by those in module 965 (which disappears),
because \input is now allowed in any mode. The code for any_mode(input)
is "if cur_chr=0 then start_input else force_eof:=true", and it moves from
module 952 to just before module 1197. The global variable force_eof
is initially false, and module 340 becomes (after first:=start;)
        if not force_eof then
                begin if input_ln.... else force_eof:=true;
                end;
        if force_eof then begin print_char(")"); force_eof:=false;...

52. Module 1005, line 4 (12/8/82)
change "if align_state<0" to "if (mode<0)or(align_state<0)"
(This avoids embarrassing case where TeX says "type a command or say \end"
but when you type \end it says "You can't use \end in restricted horiz mode".)

53. Patch to the new code for \csname (12/21/82)
After eq_type(cs_ptr):=relax, also say equiv(cs_ptr):=256.
(This corrects a bug that would appear only if \csname occurs right
after a file name.)

54. Change 47 introduced a bug when tracingonline=0 (12/20/82)
when omitting firstpass, also do "if tracing_stats>2 then begin_diagnostic;"

55. \hskip -1pt plus 2pt was parsed as \hskip -(1pt plus 2pt)! (12/20/82)
In module 421, before line -4, insert the following:
        if negative then
                begin cur_val:=-cur_val; negative:=false;
                end;
But then realize that "negative" is always false on line -2; simplify.

56. Cosmetic change to paragraph statistics (12/23/82)
tight_fit..very_loose_fit codes have been renumbered very_loose_fit..tight_fit.

57. Module 729 changes to make TeX language more consistent (12/23/82):
else if type(tail)<>glue_node then tail_append(new_penalty(inf_penalty))
else    begin type(tail):=penalty_node; delete_glue_ref(glue_ptr(tail));
        flush_node_list(leader_ptr(tail)); penalty(tail):=inf_penalty;

58. Commas are allowed as alternates to radix points. (12/23/82)
define continental_point_token=other_token+"," {decimal point, Eurostyle}
in module 400, and insert the following twice in 409:
        if cur_tok=continental_point_token then cur_tok:=point_token;

59. \hangindent becomes a normal parameter. (12/23/82)
This simplifies the code in obvious ways; for example, module 1148 disappears.
The command code hang_indent goes away too; what was previously called
hanging_indent is then renamed hang_indent.

60. \prevgraf becomes accessible. (12/23/82)
This involves renaming the "after" field "pg_field" in the nest array;
making a new command code set_prev_graf;
including the following loop into scan_the:
        nest[nest_ptr]:=cur_list; p:=nest_ptr;
        while abs(nest[p].mode_field)<>vmode do decr(p);
        scanned_result(nest[p].pg_field)(int_val);
and including the following active procedure for any_mode(set_prev_graf):
        procedure change_prev_graf;
        var p:0..nest_size; {index into |nest|}
        begin nest[nest_ptr]:=cur_list; p:=nest_ptr;
        while abs(nest[p].mode_field)<>vmode do decr(p);
        scan_optional_equals; scan_int;
        if cur_val<0 then
                begin print_nl("! Bad \prevgraf");
                help1("I allow only nonnegative values here.");
                int_error(cur_val);
                end
        else    begin nest[p].pg_field:=cur_val; cur_list:=nest[nest_ptr];
                end;
        end;

61. \clubpenalty is split off from \widowpenalty. (12/23/82)

62. Bad bug in module 1005 (12/24/82)
(must not go to reswitch if \par is a macro!)
Instead of setting cur_cmd and cur_chr and goto reswitch, just back_input
and set token_type:=inserted.

63. \openin not to prompt if file not present (12/25/82)
Change lines -4 and -3 of module 1183 to:
        if a_open_in(read_file[n]) then read_open[n]:=just_open
        else read_open[n]:=closed;

64. New \jobname primitive (12/25/82)
It is added to the "convert" command in the obvious way.
In conv_toks, the relevant code is
        if job_name=0 then open_log_file
before "selector" is changed and
        print(job_name)
after.

65. Better error recovery for math-only things (12/25/82):
In module 952, don't goto reswitch after insert_dollar_sign.
In module 954, first back_input, then set cur_tok, and
don't bother to set cur_chr or cur_cmd; ins_error instead of back_error.

66. Module 1163, line 3, insert "scan_optional_equals" (12/25/82).
Also make \the\parshape allowed.

67. The location where an \if begins is stacked (12/26/82)
so that a better error message can be given for \end while \if is incomplete.
This means two-word nodes instead of one-word nodes in the if stack.

68. Change 30 is extended to \insert, \vadjust, \valign, \output (12/26/82)
The one-time-only paragraph parameters are now cleared by a
subroutine called normal_paragraph; hang_after is also set to 1.
The essential change being made now is to call normal_paragraph in modules
704, 932, 1009, and 1076.

69. \pagetotal and \pagegoal are added (12/27/82)
The changes are analogous to, but simpler than, those for \prevgraf.

70. Tracing of page-optimization calculations (12/27/82)
A bunch of print commands are added to modules 897, 914, and 918,
activated if tracing_pages>0. Also, \tracingparagraphs is separated
from \tracingstats.

**** The changes above have been incorporated into version 0.91 of TeX

71. The build_page procedure is broken in two parts (Dec 31, 1982)
by making module 919 into a procedure called fire_up.

72. \ifeven1\else is made legal by introducing if_code (Dec 31, 1982)
This improves part of the code in change 45; if_limit has a specific
value that recovers automatically from a former syntax error.

73. Improvement to alignments when columns don't occur (Dec 31, 1982)
Delete module 711; and where it was used in module 708, say this:
        begin fin_col:=true; return;
        end;
Also, in module 717, replace the statement "width(q):=0" by
"<Nullify width(q) and the tabskip glue following the current column>"
where that new module has the following code:
        begin width(q):=0; r:=link(q); s:=glue_ptr(r);
        if s<>zero_glue then
                begin add_glue_rel(zero_glue); delete_glue_ref(s);
                glue_ptr(r):=zero_glue;
                end;
        end

74. Better error message in overfull alignment (Dec 31, 1982)
In module 717, don't set both height and width; set the height zero.
Then in module 719, switch width to height if necessary.  (People didn't
understand the previous error messages, and I couldn't blame them.)

*** The changes above have been incorporated into version 0.92 of TeX82
*** (which was the last version of 1982, completed 11:59pm on December 31)
The first changes after 1982

75. Modules 961 and 962 should be one module. (Jan 3, 1983)
Also, the absolute constant 100 is replaced, and the test becomes
        if ((page_head=page_tail)and(head=tail)and(dead_cycles=0))or
                 (dead_cycles>max_dead_cycles) then

76. Surprise bug: module 1010. (Jan 3, 1983)
The case "if head<>tail" needs an else clause: else pop_nest.
Also remove the "<Scan an optional space>" in that module.

77. Improvement to change 22 (Jan 4, 1983)
In module 996, use box_max_depth from inside the \vbox:
        @!d:scaled; {maximum depth}
        begin d:=box_max_depth; unsave;... vpackage(...,d);

78. \groupbegin and \groupend changed to \begingroup and \endgroup (Jan 4, 83)

79. \deadcycles made accessible (Jan 4, 83)

80. New calculations for split insertions (Jan 4, 83)
In module 918, we now work with natural width, and add in the page depth too:
The line "else w:=x_over_n..." is changed to
        else    begin w:=page_goal-page_total-page_depth;
                if count(n)<>1000 then w:=x_over_n(w,count(n))*1000;
                end;
(Note that cur_page_height and cur_page_depth and page_size have been
renamed page_total, page_depth, and page_goal, in accordance with new syntax.)

*** The changes above have been incorporated into Version 0.93

81. Old bug finally unearthed by PHY (Jan 6, 1983)
insert "incompleat_node:=null;" after "push_nest;" in module 1097.

82. Extension of change 69: \pageshrink, etc. added (Jan 6, 1983)

83. \floatingpenalty and \insertpenalties added (Jan 6, 1983)
Also, the insertion nodes now have a new format, so that the values of
\floatingpenalty, \splittopskip, and \splitmaxdepth can be stored with
each insertion; this requires the obvious changes in several places:
a) Module 136, ins_node has fields float,depth,height,ins_ptr,split_top_ptr
b) Module 184, these fields are displayed
c) Module 198, ins_node case, also delete_glue_ref(split_top_ptr(p))
d) Module 202, ins_node case, also add_glue_ref(split_top_ptr(p))
e) Modules 226-228, new integer parameter \floatingpenalty
f) Module 883, add parameter d to vert_break subroutine
g) Module 885, use d instead of split_max_depth
h) Module 890, argument split_max_depth to call of vert_break
i) Module 894, change width to height (better name)
j) Module 900, initialize insert_penalties here, not in 901
k) Module 914, cost to be awful_bad if insert_penalties>=10000
l) Module 916, add float(p) to insert_penalties if type(r)<>inserting
m) Module 916, subtract page_depth from delta
n) Module 918, subtract page_shrink from delta
o) Module 918, argument depth(p) to call of vert_break
p) Module 921, insert_penalties:=0, save split_top_skip before calling 925
q) Module 921, restore split_top_skip before calling 924
r) Module 928, set split_top_skip:=split_top_ptr(p) before prune_page_top
s) Module 929, after q:=p, incr(insert_penalties)
t) Module 929, before free_node, delete_glue_ref(split_top_ptr(p))
u) Module 933, clear insert_penalties before calling 934
v) Module 978, more local variables needed
w) Module 1010, build the newfangled ins_node

84. Scanner goes to new_line when <return> is category 13 (Jan 7, 1983)
Insert state:=new_line in module 323 just before the reference to module 339;
and delete "state:=new_line;" from modules 328 and 330 (and their names).

85. Distinguish between user \kern and font \kern (Jan 9, 1983)
The nontrivial parts of this are to change "type(s)<>kern_node" to
"(type(s)<>kern_node)or(subtype(s)<>normal)" in modules 809 and 810;
and to say "kern_node: if subtype(s)=explicit then goto done4" in module 812.
The \kern primitive now has "explicit" instead of "normal" in 967.

86. "ignorespace" becomes "ignorespaces" (Jan 9, 1983)

87. Don't omit a blank space after \def, \message, \mark, etc. (Jan 9, 1983)
<Scan an optional space> is removed from modules 431, 848, 874.
"info(r):=space_token;" and "link(r):=get_avail; r:=link(r);" removed from 1278.

*** The above changes appear in Version 0.94.
Version 0.95

88. New active characters in math mode (Jan 12, 1983)
In module 1062, add a label "restart" and change lines -3 and -2 as follows:
fam(p):=(c div 256) mod 16;
if c>=var_code then
  if fam(p)<8 then fam(p):=cur_fam
  else begin <Convert c to an active character whose equivalent is
                        ready to be scanned next>; goto restart;
        end;
In module 1064, the body of the procedure becomes
begin if c>='4000 then <Convert c ...> else <the former body>; end
And there's a new module:
<Convert c...>=
begin cs_ptr:=(c mod 128)+active_base;
cur_cmd:=eq_type(cs_ptr); cur_chr:=equiv(cs_ptr); x_token; back_input;
end

89. Surprise bug: $1-$ treated the - as binary (Jan 15, 1983)
New module <Convert a final bin_noad to an ord_noad>=
        if r_type=bin_noad then type(r):=ord_noad
is called in module 649 before the call of 678, and in module 651
in the place where that code already appears.

90. Another oversight (Jan 15, 1983)
Add "space_factor:=1000;" after "push_nest" in modules 1027, 1029

91. And a more embarrassing one (Jan 16, 1983)
I forgot "spotless:=false;" at the beginning of the procedure in module 80.
But while fixing this, I decided to make it more general since IBMers want
a return code at the end of the job. So there's a history variable that has
four values: spotless, warning_issued, error_message_issued, fatal_error_stop.
a) module 75: declare history, define spotless etc.
b) module 76: initialize history:=spotless
c) module 80: if history<error_mess... then history:=error_mess...
d) 80 and 90: history:=fatal_error_stop before jump_out
e) 92: if history<error_message_issued then
f) 232: if history=spotless then history:=warning_issued
g) 1243: if history=warning_issued then print_nl("(see the log...)")

92. "This can't happen" could happen, so there's new error recovery (Jan 16, 83)
In module 933, replace the call on confusion by a call of the following module:
        @ @<Recover from an unbalanced output routine@>=
        begin print_nl("! Unbalanced \output routine");
        help2("Your sneaky output routine has fewer real {'s than }'s.")@/
        ("I can't handle that very well; good luck."); error;
        repeat get_token;
        until loc=null;
        end
In module 1278, replace the call on confusion by a call of:
        @ @<Recover from an unbalanced write command@>=
        begin print_nl("! Unbalanced \write command");
        help2("On this page there's a \write with fewer real {'s than }'s.")@/
        ("I can't handle that very well; good luck."); error;
        repeat get_token;
        until cur_tok=end_write_token;
        end

93. String overflow clobbered the log file (Jan 18, 1983)
Also, "confusion" before log file open would cause problems.  Also start_input
calling open_log_file calling prompt_file_name calling fatal_error!
To fix these anomalies, open_log_file no longer calls prompt_file_name,
if interaction<scrollmode; instead, it terminates after printing what went
wrong. Also, the fatal_error and overflow and confusion procedures call
the following new subroutine:
procedure normalize_selector;
begin if job_name>0 then selector:=term_and_log else selector:=term_only;
if interaction=batch_mode then decr(selector);
if job_name=0 then open_log_file;
end;

94. \ifeof\fi loops infinitely (Jan 18, discovered by Lamport)
Change 72 converted such a \fi to <space>\fi. Now it is converted to \relax\fi.

95. \limitswitch changed to \displaylimits et al. (Jan 18, 1983)
[Incidentally, this fixes a bug in the former positioning of \int\limitswitch]
a) module 608: subtype of Op can be normal, limits, or no_limits.
b) 620: display the subtype if not normal
c) 667: new logic decides limits before looking at the operand (and the operand
 is now called the nucleus); the italic correction is removed only if
 it should not be put back
d) 1069: subtype(tail):=cur_chr

96. Minor changes to math in unusual cases (Jan 19, 1983)
a:delete "if height(y)<=0 then height(y):=default_rule_thickness" in module 658
b:move "if thickness(q)=default..." from module 664 to module 661
c:delete module 1081, that error message isn't worth the bother
d:in module 1062, char_num case, c:=math_code(cur_val) (if cur_val<128)
e:in module 1064, a similar change

97. Bad spacing from change 6 is corrected (Jan 19, 1983)
underline, overline, radical, vcenter, and accent noads now revert to type
Ord instead of type Inner.
{...} produces type Ord also.
There's a new primitive \mathinner.
The new_noad function now produces an ord_noad (change its calls accordingly).
And the default is changed in module 679 to t:=ord_noad; the fraction_noad
case sets t:=inner_noad, and the (inner_noad,ord_noad) cases swap places.

98. New \mathchoice primitive (Jan 19, 1983)
a) module 204: new command
b) 254: math_choice_group
c) 614: style node three words long, so a choice node can be converted to it
d) 614a: choice node has four subfields: display_mlist, text_mlits, etc.
e) 1081a: routines to build a choice_node like 1027-1029 build discretionaries

99. \input moves to syntax from semantics (Jan 19, 1983)
a) 204, 206: renumber commands
b) When input is to be expanded, if name_in_progress then insert_relax
c) 459,464: name_in_pr:=true; begin_name;...end_name; name_in_pr:=false;
d) 406,461: declare name_in_progress, a global boolean initially false

100. \chardef joins \mathchardef (Jan 19, 1983)
a) 204: math_only becomes math_given; add new command char_given
b) 205: new command char_def
c) 250,251: new primitive char_def
d) 380: char_given,math_given yield integer after \the
e) 401: char_given,math_given yield integer in context of integer
f) 936,943: char_given treated like other_char
g) 1062,1064: same, if cur_chr<128, else assume math_code(x)=x
h) 1143,1144: char_def is analogous to math_char_def

101. \unbox becomes \unhbox,\unvbox; also add \unhcopy (Jan 19, 1983)
Module 1020 changes in the obvious way.

102. \spacefactor, \pagetotal, etc. move to prefixed_command (Jan 20, 1983)

103. \hrule in horizontal mode, \vrule in vertical mode: switch modes (Jan 20)

104. \globaldefs parameter, affects prefixed_command (Jan 20, 1983)

105. After looking at frequency counts, some optimizations made (Jan 21, 1983)
a) "fast_get_avail" and "fast_store_new_token" introduced to speed up the loops
in modules 360 and 431.
b) some procedure call overhead eliminated in begin_token_list, end_token_list,
back_input, and flush_node_list.
c) a few if tests changed from "if a and b then" to "if a then if b then".

106. Changes for space efficiency in math constructions (Jan 22, 1983)
a) In module 1102, mlist_penalties:=(mode>0)
b) The following code is inserted before "free_node" in module 639 (rebox):
        if (is_char_node(p))and(link(p)=null) then
                begin f:=font(p); v:=char_width(f)(char_info(f)(character(p)));
                if v<>width(b) then link(p):=new_kern(width(b)-v);
                end;
c) A new module is called just before clean_box exits:
<Simplify a trivial box>=
q:=list_ptr(x);
if is_char_node(q) then
        begin r:=link(q);
        if r<>null then if link(r)=null then if type(r)=kern_node then
                begin free_node(r,small_node_size); link(q):=null;
                end;
        end

107. Oversight in rebox routine (module 639) corrected (Jan 22, 1983)
if type(b)=vlist_node then b:=hpack(b,natural);

108. Module 217 clobbers eqtb[bad_font_ident] set in change 26q (Jan 22, 1983)
Decided to fix this by making \nullfont a primitive.
This means the procedure missing_font can be deleted, and the test for
undefined font can be removed from the inner loop.
(This reflects a rather dramatic change from TeX80, where
a missing font was a fatal "Whoa" error!)
Note: I thought I could delete module 646, but realized that it still provides
a useful error message. The dump/undump routines (cf. change 26ee) now
dump the null_font information too, as its parameters can be changed.

109. End of program now lists all incomplete \ifs (Jan 24, 1983)

110. Alignment preamble setup to allow \halign\lb (Jan 29, 1983)
The statement "align_state:=-1000000;" is inserted near the beginning
of module 695 (and the comment about align_state=-999999 is deleted).
The constant -999999 is changed to -1000000 in modules 700 and 701.

111. Forgot to test is_char_node(r) (Jan 30, 1983)
in the <Simplify a trivial box> code of change 106c.
By coincidence, this was caught since somebody used font number 11
in the second character of a list of length 2!

112. Improved format for stats at end of run (suggested by DRF, Jan 30, 1983)
Module 1242 changes; nothing subtle.

**** The changes above have been incorporated into version 0.95
Version 0.96

113. space after one-symbol control sequences NOT to be ignored unless
the catcode of that symbol is letter or spacer. (Jan 30, 1983)
The name of module 334 changes slightly; the state is set based
on cat_code(cur_chr), shortly after label start_cs.
A new variable cat is introduced and set to cat_code(cur_chr) in modules
334 and 336.

114. trailing spaces removed on all lines of input (Jan 30, 1983)
last_nonblank is added to input_ln routine (module 31)

115. mmode+accent gives error but assumes math_accent (Feb 3, 1983)
This goes into math_ac procedure (module 1075); module 1035 is eliminated.

116. \iftrue and \iffalse (Feb 5, 1983)
Simple addition of two new conditions.

117. Bad calculation of size for \left and \right (Feb 6, 1983)
In module 680, the first assignment to delta should be
        delta:=(delta1 div 500)*delimiter_factor.

118. \delimitershortfall (new name) replaces \delimiterlimit. (Feb 12)

119. \abovewithdelims.. to be equivalent to \above (Feb 12)
Module 1089 revised so that the delimiters are scanned before the dimension.

120. Remove the kludgy math codes introduced in change 88 (Feb 12)
\fam becomes a normal integer parameter
and \mathcode stored with min_halfword added
and \mathcode allowed to be 32768
and current family is substituted only when it's in range.
Also the initialization of mathcode for letters now specifies family 1.
(This implies a dozen or so obvious revisions.)

121. Bug in module 1152: max spacefactor is 32767 (Feb 12)
also module 1013 gets a restricted range

122. Octal output to be replaced by hexadecimal. (Feb 14)

123. Forgot to include char_given in module 1037, re change 100 (Feb 14)

124. vmode+valign,hmode+halign made legal transitions (Feb 17)

125. \tracingrestores (Feb 18)
This involves a lot of new, rather tedious code that's interspersed with the
eqtb definitions (the body of a procedure called show_eqtb).

126. New error message for hmode+hrule in -hmode. (Feb 25)
head_for_vmode now suggests \leaders in this case. [improves change 103]

127. under/overfull boxes in alignment: Better message. (Feb 27)
par_begin_line becomes pack_begin_line, and it is set (negative) in
module 719, read in modules 590 and 601.

128. New \xcr feature. (suggested by Lamport comments, Mar 4)
align_peek (module 702) ignores it, otherwise it acts like ordinary \cr.

129. In stats, subtract out TeX's own string requirements (Mar 4)
(cf. change 112)
init_pool_ptr and init_str_ptr variables added in obvious way
(declared in module 39, set in module 1240, used in module 1242)

130. \everyhbox and \everyvbox. (Mar 6)
The modifications are obvious; for example, module 1076 gets the statement
        if every_vbox<>null then begin_token_list(every_vbox,every_vbox_text);
and a similar pair of statements goes into module 993.

131. Precaution in module 1105 error recovery (March 9, 1983)
Change the first test to "if (a=null)or danger", in order to
avoid accessing math_quad when the symbol fonts aren't known to be present.

132. Installed float and unfloat to aid portability (suggested by HWT, 3/7/83)

133. \dispskip becomes \abovedisplayskip and \belowdisplayskip (3/9/83)
Also \dispaskip becomes \abovedisplayshortskip; \dispbskip similar.

134. \romannumeral separated from \number (suggested by FY, March 10)
Obvious changes; print_roman_int now accepts negative input but gives
no output in such cases (the test in module 68 becomes "if n<=0 then return").

135. scan_keyword to ignore leading spaces (Mar 12)
In module 375, change "else begin" to
"else if (cur_cmd<>spacer)or(p<>backup_head) then begin".

136. Update to change 112 (Mar 14)
Module 1242 now uses write and write_ln directly (saves space and time).

137. Another change to page-break cost (suggested by Lamport, March 16)
Cf. change 83k above; the relevant lines of module 914 now read
        if b<awful_bad then
                if pi<=eject_penalty then c:=pi
                else    if b<inf_bad then c:=b+pi+insert_penalties
                        else c:=deplorable
        else c:=b;
        if insert_penalties>=10000 then c:=awful_bad;
and a similar change (but without insert_penalties) occurs in module 887.

**** The changes above have been incorporated into version 0.96 (March 16, 1983)
Version 0.97

138. \everyjob (suggested by FY, March 18)
Module 936: main_control inserts every_job as its first action.

139. Improved printout of macro definitions (March 19)
In module 278, simply treat left_brace like right_brace.
(This corresponds to the way the manual describes parameter matching.)

140. Omit blanks as non-delimited parameters (March 19)
In module 360, replace
        else store_new_token(cur_tok);
by
        else @<Store the current token, but |goto continue| if it is
                 a blank space that would become an undelimited parameter@>;
Then define @<Store the current token...@> to be:
begin if cur_tok=space_token then
        if info(r)<=end_match_token then
                if info(r)>=match_token then goto continue;
store_new_token(cur_tok);
end

141. Minor patch in module 386, catches mu_glue (March 21)
begin cur_val:=zero_glue; cur_val_level:=glue_val;
if not is_char_node(tail)and(mode<>0) then
        begin if type(tail)=glue_node then
                begin cur_val:=glue_ptr(tail);
                if subtype(tail)=mu_glue then cur_val_level:=mu_val;
                end;
        end
else if (mode=vmode)and(tail=head)and(last_page_glue<>max_halfword) then
        cur_val:=last_page_glue;
end

142. Patch in module 699, \span expands only one level (March 21)
(See previous change #42.) The "get_x_token" in the code
while (cur_chr=span_code)and(cur_cmd=tab_mark) do get_x_token;
should be replaced by:
        begin get_token;
        if cur_cmd>max_command then
                begin expand_calls; get_token;
                end;
        end

143. Single # in \tokens, \message, etc. (March 22)
(The previous rule was really bad in connection with \uppercase,
or with \write when #'s had to be given four times!)
Module 437 should be prefaced with "if macro_def then".
And get_token should become get_x_token there, if xpand is true.

144. Keyword "to" required in \read. (March 22)
This will avoid the common error of a missing space before the \cs.
Also the stream number can be out of range, for terminal input.
The stream number in a \write can be out of range too, for terminal output.
Modules 1145 and 1258 and 1280, etc., change in the obvious ways.

145. \ifeven<countnumber> replaced by \ifodd<number> (March 26)
This makes the language more consistent.

146. Big surprise bug relating to \if\if aabc\fi (March 26)
(related to change 45; the possibility that cur_if might not be the
correct one when the conditional is evaluated was discovered today)
The main change is to add the change_if_limit procedure, and to add
the variable save_cond_ptr and the code that now depends on it.

147. \if and \ifcat should tolerate primitives (March 28)

148. "absent" becomes "void", a better word (March 28)

149. Module 990: \lastbox to clear the shift_amount (March 28)
since I don't want to figure out what it means in all cases (\vsplit, etc)

150. print_err("...") takes the place of print_nl("! ...") (DRF, March 29)
And wake_up_terminal is introduced in module 34, and used in modules
37, 70, 341, 441, 457, 463, 1210, 1240, 1246.

151. \halign extended to periodic preambles (April 1)
Modules 688--690: cur_loop introduced, stacked/unstacked
Module 695: cur_loop initialized
Module 700: cur_loop set up
Module 708--709: cur_loop used and advanced

152. \leaders to align by the smallest enclosing box. (April 1)
Module 549: new local variable left_edge, initialized to cur_h
Module 557: cur_h:=left_edge+leader_wd*((cur_h-left_edge) div leader_wd)
Module 559: new local variable top_edge, initialized to cur_v-height(this_box)
Module 566: analogous to 557

153. hyphenation after whatsits is OK (April 1)
In module 809, skip past whatsits

154. \par in vertical mode should build_page (April 2)
vmode+par_end moves from 952 to 1005

155. Clear aux to zero in module 703 (April 2)

156. Digits will switch families (April 4)
(initialize their math codes differently in module 225)

157. Refinement to correction 83m (April 7)
the test for not splitting should be
        if ((h<=0)or(h<=delta))and(height(p)+height(r)<=dimen(n)) then

158. (re 128) \xcr renamed to \crcr, at Lamport's request. (April 8)

159. Better error recovery in runaway preamble (April 11)
Module 319, aligning: set align_state:=-1000000

160. \read to get balanced braces (April 12)
Module 440 changed to look more like 431.
Module 443 gives error message if the \read goes off end of file.
Module 319 removes \outer from forbidden control sequence that is \read.

161. Bug found by Jim Sterken (April 14)
Module 798 can make q a char_node, so module 794 needs this patch:
        if not is_char_node(q) then
to be inserted just before "if (type(q)=math_node)..."

162. \uppercase and \lowercase to apply to all characters (April 15)
module 1196 changes; also I put active_base before single_base in eqtb

**** The changes above have been incorporated into version 0.97 (April 16, 1983)
**** except the part of 144 about \read-1 was forgotten and put in 0.98 later
Version 0.98

163. change small_number to 0..65 in module 814 (found by DRF, April 17)

164. improved error recovery in module 1166 (suggested by DRF, April 17)
after error: repeat get_token; until cur_cmd=right_brace; {flush the patterns}

165. improved \read from terminal (suggested by Todd Allen at Yale, May 1)
I had forgotten to implement the extended stream numbers in change 144.
Also, the prompt is now omitted if n<0.

166. \write n writes only to the log file, if n<0. (May 18)

167. Unified syntax for parameters and registers. (May 18)
a) Command code changes: def_font moved to before prefix;
   register eliminated; set_register renamed register and
   moved (with adv_register...div_register) to before prefix;
   min_internal and max_internal defined.
b) scan_the renamed scan_something_internal, and it acts on cur_tok.
c) scan_int, scan_dimen, scan_glue simplified accordingly; instead of
   testing for cur_cmd=the or cur_cmd=register, they now test
   cur_cmd versus min_internal and max_internal.
d) ins_the is removed.
e) \minusthe is removed. Consequently "the_toks" needs no parameter.

168. new parameters \hoffset, \voffset (May 18)

169. \everycr (suggested by Spivak, installed May 24)

170. \countdef, \dimendef, etc. (suggested by DRF long ago, installed May 25)
Straightforward change to internal representation of assign_int and similar
commands, so that the chr part is now a pointer to the eqtb location.

171. \advance, \multiply, \divide (suggested by FY, installed May 25)

172. \hyphenchar (May 25)
A new command assign_font_int is introduced, and incorporated into
scan_something_internal and prefixed_command in the usual way.
The hyphen_char array entries are initialized in modules 483 and 506,
allocated in 482, used in modules 809 [is hyphenation to be suppressed?],
828 [insert the hyphen character], 946 [insert discretionary node
following hyphen], 1027 [implementation of \-].
hyphen_char[k] is dumped and undumped in modules 1229 and 1230.

173. \skewchar (May 25)
Loaded and fetched with the same command code as \hyphenchar.
Module 659 gets a dozen more lines of program, to compute the skew.

174. \noexpand (May 25)
This involves: introduction of no_expand and dont_expand command codes;
frozen_dont_expand as an internal marker; small change to get_next
when that marker is sensed; change to processing of \if and \ifcat;
implementation of no_expand in the expand_calls subroutine.
I also changed the name of expand_calls to "expand".

175. \meaning (May 25)
This adds one case to the "convert" command.
print_meaning is a new subroutine, using code that was in show_whatever.

176. "dm" and "vu" are out, ".5\hsize" is in (May 25)
Straightforward changes to module 416.

177. \texinfo f n becomes \fontdimen n f (May 25)

178. \afterassignment (suggested by ARK, May 27)

179. \chardef\xx=5\xx shouldn't say that \xx is undefined (May 27)

180. \relax to be ignored like spaces in math mode (May 28)
and in a few other places: <Get next non-blank non-relax non-call token>
is now used in modules 379, 988, 994, 1062, 1070, 1134, 1179
(i.e., scan_left_brace, scan_box, scan_math, scan_delimiter, prefixed_command,
do_assignments, and just after \leaders).

181. improve \mathaccent wrt sub/superscripts (sugg by HWT, May 30)

182. its_all_over: remove dead_cycles>max_dead_cycles (May 30)
Modules 961 and 962 are combined and simplified.

*** The changes above were installed into version 0.98 on May 31, 1983 ***
Version 0.99

183. \mark and \insert and \vadjust allowed in restricted hmode (Jun 3)
also in math; this is a comparatively big change
[at present, \mark in a display causes TeX to crash with "can't happen"!]
modules 652 and 679: mark_node,ins_node,adjust_node now permitted
modules 576 and 578: t becomes a global variable adjust_tail
modules 802 and 1110: new calling sequence to get the adjustments
modules 254, 986, 993, 995, 996: adjusted_hbox_group for the new kind of hbox
modules 698--690: cur_head and cur_tail added to alignment stack
modules 703, 712, 715: adjustments gathered and appended during \halign

184. \ht, \wd, and \dp (Jun 6)

185. When displaying noads, use ^ and _ instead of ( and [ (Jun 6)

186. A..F in hex constants could be otherchar as well as letter (Jun 6)

187. Remove <scan optional space> from module 417 (Jun 7)
(it was redundant code)

188. \mkern .5\thinmuskip and \mkern\thinmuskip should be legal (Jun 7)

189. 2.5\space\space\dimen0 should work (Jun 7)
previously it worked after "plus" or "minus" only!

190. <font identifier> to allow also \font for current font (Jun 7)
"if cur_cmd=def_font then f:=cur_font else" added to module 507

191. \gdef not to be global when \globaldefs<0 (Jun 7)
"and (global_defs>=0)" added to module 1139

192. \advance\spaceskip by-\spaceskip should yield zero_glue (Jun 7)
the procedure trap_zero_glue is culled from module 1147

193. \show should work with any token (Jun 7)

194. \tokens to become 256 registers (Jun 8)
\toks and \toksdef added in the straightforward way
(this affects mainly eqtb, scan_something_internal, prefixed_command)

195. allow \indent in math mode (Jun 8)
also, \valign in math mode to give missing $ error
modules 1001 and 1043 disappear; 1004 is generalized slightly.

196. remove redundant code (Jun 8)
In module 1044, there's no need to check cur_group and call off_save.
Similarly in module 1053.

197. new_write_whatsit shouldn't allow \openout-1, \closeout-1 (Jun 8)
(simple change to module 1258)

198. \lastbox should give error in math mode (Jun 8)
(simple change to module 990)

199. \leaders not followed by proper glue should be back_error (Jun 9)
[I made the change; TRIP should test this error message!]

200. Module 702, correction to beginning of \noalign (Jun 9)
if mode=-vmode then normal_paragraph;

201. After alphabetic constant, expand the optional space (Jun 10)

202. Set space_factor:=1000 after rule or constructing an accent (Jun 12)
That's in modules 964 and 1036.

203. blunder in module 783 (caught by Jim Sterken, fixed Jun 14)
disc_width:=0 needs to be set before testing if s=null
(A real bug that existed since the beginning! It showed up on page 37
  of the September 1982 TRIP manual; my hand-checking was incomplete...)

204. Change to optional spaces after <dimen> (Jun 14)
The optional space will now be analogous to that after <number>.

205. Fix conflict between \output and \everydisplay (Jun 14)
In change 49, I should have inserted every_display before calling build_page.

206. Overflow errors to be consistent with statistics reporting (Jun 17)
(See change 129.)

207. \tracing switches to be all positive vs nonpositive (Jun 17)

*** The changes above were installed into version 0.99 on June 19, 1983 ***
Version 0.999

208. \catcode`\%=14 to be done by INITEX (Jun 20)

209. \par in vmode to clear parshape etc. (Jun 21)
Module 1005, which contains vmode+par_end, now calls normal_paragraph.

210. Improvement to change 39 (Jun 21)
Module 593, overfull_rule not appended if solely due to hbadness

211. Alignment bug allows glue_set to be less than -1! (Jun 21)
Modules 723 and 724 need to be patched.

212. Correction to change 134 (found by Debby Clark, Jun 22)
Module 68: n to be declared integer, not nonnegative_integer.

213. "by" to be optional (suggested by Lamport, Jun 22)
Module 1158 disappears

214. Module 1235, slight change in format_ident message (Jun 24)

215. New measures needed to thwart trickery (Jun 25)
Glue set values computed by \span could have been brought into
TeX's registers via, e.g., \valign and \vsplit; so the
"kern" idea of module 722 is insufficient and should be abandoned.
Extra boxes and glue are added; this has additional virtue
of perfect accuracy in alignment of vertical rules.
The main change is to introduce a new module after 722, in which s and t
are updated for spans, and to eliminate the "equivalent kern" code from 719
(the corresponding glue calculations are now done in the new module).

216. leaders to affect height/width of their boxes (Jun 25)
Module 583 splits into two parts (one for hpack, one for vpack).

217. \unskip permitted a little more often (Jun 28)
Module 1018 reports error only if last_page_glue<max_halfword.

218. {\bf A} in math: not to make a box (Jul 1)
Module 1093 changes as follows: info(saved(0)):=fin_mlist(null) becomes
p:=fin_mlist(null); info(saved(0)):=p; if <simple case> then <simplify>.

219. "scaled" feature added to \font input.

220. \mathaccent still not working right? (Jul 2)
Change 181, I forgot to correct delta when box x changed.

221. Remove confusion possible on file of length 1 (found by Lamport)
Correction suggested by DRF, July 2: introduce bypass_eoln

222. Allow things like ^^b. (July 4)
This simplifies modules 332 and 335; also affects 48 and 49.
And (surprise) 58.

223. \escapechar \defaulthyphenchar \defaultskewchar \endlinechar (July 4)
to make TeX less dependent on the character set
A lot of error messages are now broken up so that they use print_esc

224. erstat added for file opening/closing (by DRF, July 7)
That's module 27.

225. \tracingpages output to show total glue too (July 11)
Added a procedure called print_page_totals, used also by show_activities.
Broke modules 214 and 896 into two modules each.

226. Guard against insertion into an hbox (July 11)
A new procedure ensure_vbox is called in modules 917 and 925.

227. <tokenvar>=<tokenvar> allowed (July 11)

228. \errhelp parameter added (July 11)

229. get_preamble_token should look at global_defs (July 11)
Module 699 is patched.

*** The above changes installed in preliminary version 0.999 on July 12, 1983,
but I decided later in the day to do a few more things:

230. \string, \noexpand, \meaning to allow \outer (Todd Allen, July 12)

231. \the to be an expandable control sequence (July 12)
Several things in the language are cleaned up:
a) \the\tenrm replaced by \fontname\tenrm [\fontname<font>]
b) when expanding edef, etc., result of \the still expanded only only level
c) expansion after \def not inhibited, since \noexpand is now present
d) \the\the disallowed.

232. \unhbox and \unhcopy allowed in math mode if the box is void (July 12)
Module 1020 is extended to handle this case.

233. Value of default_rule was incorrectly rounded (July 16)
Module 424: default_rule=26214 {.39999 pt}

234. \mathaccent still not right! (July 16)
I need to make the final height the max of (height of accented letter w/o
superscript, height of unaccented letter w superscript).

235. \newlinechar parameter (July 16)
simple change to print subroutine

236. boxes and rules to be allowed in discretionaries (sugg by HP man, July 16)
in fact this simply requires omitting of the prohibition (in module 1031)
and a few more cases equivalent to kern_node in modules 754, 755, 784.

237. \tracingcommands to show all expandable tokens (July 16)

238. \char to be allowed in \hyphenation list (July 16)
Module 848 changes in the obvious way.

239. \aftergroup (July 16)
A new save_type.

240. Curtail running dimensions inside alignments (sugg by ARK, July 16)
Module 720.

241. Strange pattern data could cause PASCAL error (found July 17)
Put "if hc[1]=127 then hyf[0]:=0;" at beginning of module 878.

242. hc codes for hyphenation are one lower (July 17)
e.g., hc[...]:=lc_code(...)-1. This makes code 127 impossible to match.

243. whatsits also allowed after hyphenatable words (July 17)
Module 812.

244. \/ makes an explicit kern (July 17)
"subtype(tail):=explicit" in module 1023.

*** Version 0.999 installed July 17, 1983.
But later in the day I decided to do a few more things:

245. Lowercase letters allowed in file names (July 18)
In module 452, omit the conversion to uppercase.

246. "No output file" becomes "No pages of output." (July 18)
Module 571.

247. QRS on error leads to confirmation message (sugg by ARK, July 18)
Module 84.

*** The real version 0.999 finally installed July 18.
* Version 1.0

(Changes made after the Version 0.999 listing of TeX82)
(Henceforth the "final" module numbers are used)

248. Module 1215, allow space in \read n to \cs (by FY, July 25, 1983)
@x patch in get_r_token routine
begin restart: get_token;
@y
begin restart: repeat get_token;
until cur_tok<>space_token;
@z

249. Module 498, we must stack the current if type (FY, July 27)
@x patch in conditional routine
begin @<Push the condition stack@>;@+save_cond_ptr:=cond_ptr;@/
@y
@!this_if:small_number; {type of this conditional}
begin @<Push the condition stack@>;@+save_cond_ptr:=cond_ptr;this_if:=cur_chr;@/
@z
Also replace cur_if by this_if in modules 501, 503, 506. The following
patches do only what is necessary to make things work:
@x
  print_cmd_chr(if_test,cur_if);
@y
  print_cmd_chr(if_test,this_if);
@z
@x
if cur_if=if_int_code then scan_int@+else scan_normal_dimen;
@y
if this_if=if_int_code then scan_int@+else scan_normal_dimen;
@z
@x
if cur_if=if_char_code then b:=(n=cur_chr)@+else b:=(m=cur_cmd);
@y
if this_if=if_char_code then b:=(n=cur_chr)@+else b:=(m=cur_cmd);
@z

250. Module 507, \ifx need not put a control sequence in hash table (July 29)
@x
get_token; n:=cs_ptr; p:=cur_cmd; q:=cur_chr;
get_token; if cur_cmd<>p then b:=false
@y
get_next; n:=cs_ptr; p:=cur_cmd; q:=cur_chr;
get_next; if cur_cmd<>p then b:=false
@z

251. Module 86, message is lost (noticed by HWT, July 31)
@x
print("..."); print_ln; return;
@y
print("..."); print_ln; update_terminal; return;
@z

252. Don't put empty at end of \input file! (Aug 1)
[This simplifies the rules and the program, and also gets around a bug
that occurred at the end of files with \endlinechar<0.]
@x Module 362:
@ An empty line is inserted at the end of the file, if the last line
wasn't already empty, because |input_ln| sets |last:=first| when
it discovers an |eof|.
@^empty line at end of file@>

@<Read next line of file into |buffer|, or
  |goto restart| if the file has ended@>=
begin incr(line); first:=start;
if not force_eof then
  begin if input_ln(cur_file,true) then {not end of file}
    firm_up_the_line {this sets |limit|}
  else if limit<>start then firm_up_the_line
    {if |pausing|, the user can add more lines}
  else force_eof:=true;
@y
@ @<Read next line of file into |buffer|, or
  |goto restart| if the file has ended@>=
begin incr(line); first:=start;
if not force_eof then
  begin if input_ln(cur_file,true) then {not end of file}
    firm_up_the_line {this sets |limit|}
  else force_eof:=true;
@z

*** The changes above went into Version 0.9999, which was widely distributed

253. Ridiculous blunder made in change 146 (found by FY, August 16)
@x Correction to module 497
else loop@+begin q:=cond_ptr;
  if link(q)=p then
    begin type(p):=l; return;
    end;
  if q=null then confusion("if");
@:this can't happen if}{\quad if@>
  q:=link(q);
@y
else  begin q:=cond_ptr;
  loop@+  begin if q=null then confusion("if");
@:this can't happen if}{\quad if@>
    if link(q)=p then
      begin type(q):=l; return;
      end;
    q:=link(q);
    end;
@z

254. Minor amendment to stat(s) printing (cf. change 129) (August 16)
@x in module 1334
  wlog_ln(' ',str_ptr-init_str_ptr:1,' strings out of ',
    max_strings-init_str_ptr:1);@/
@y
  wlog(' ',str_ptr-init_str_ptr:1,' string');
  if str_ptr<>init_str_ptr+1 then wlog('s');
  wlog_ln(' out of ',
@z

255. Bug in \xleader computations (found by FY, August 18)
@x in module 592
@!lq,@!lr:integer; {quantities used in calculations for leaders}
@y
@!lq,@!lr,@!lx:integer; {quantities used in calculations for leaders}
@z
@x in module 626
  begin edge:=cur_h+rule_wd;
  @<Let |cur_h| be the position of the first box, and set |leader_wd|
    to the spacing between corresponding parts of boxes@>;
  while cur_h+leader_wd<=edge do
    @<Output a leader box at |cur_h|,
      then advance |cur_h| by |leader_wd|@>;
@y
  begin edge:=cur_h+rule_wd; lx:=0;
  @<Let |cur_h| be the position of the first box, and set |leader_wd+lx|
    to the spacing between corresponding parts of boxes@>;
  while cur_h+leader_wd<=edge do
    @<Output a leader box at |cur_h|,
      then advance |cur_h| by |leader_wd+lx|@>;
@z
@x in module 627
    leader_wd:=leader_wd+lx;
@y
@z
@x in module 628
cur_h:=save_h+leader_wd;
@y
cur_h:=save_h+leader_wd+lx;
@z
@x in module 635
  begin edge:=cur_v+rule_ht;
  @<Let |cur_v| be the position of the first box, and set |leader_ht|
    to the spacing between corresponding parts of boxes@>;
  while cur_v+leader_ht<=edge do
    @<Output a leader box at |cur_v|,
      then advance |cur_v| by |leader_ht|@>;
@y
  begin edge:=cur_v+rule_ht; lx:=0;
  @<Let |cur_v| be the position of the first box, and set |leader_ht+lx|
    to the spacing between corresponding parts of boxes@>;
  while cur_v+leader_ht<=edge do
    @<Output a leader box at |cur_v|,
      then advance |cur_v| by |leader_ht+lx|@>;
@z
@x in module 636
    leader_ht:=leader_ht+lx;
@y
@z
@x in module 637
cur_v:=save_v-height(leader_box)+leader_ht;
@y
cur_v:=save_v-height(leader_box)+leader_ht+lx;
@z
Also insert the following in modules 619 and 629:
@!lx:scaled; {extra space between leader boxes}

256. \/ should apply to ligatures! (August 20)
@x in module 1113
var f:internal_font_number; {the font in the |char_node|}
begin if is_char_node(tail)and(tail<>head) then
  begin f:=font(tail);
  tail_append(new_kern(char_italic(f)(char_info(f)(character(tail)))));
@y
label exit;
var p:pointer; {|char_node| at the tail of the current list}
@!f:internal_font_number; {the font in the |char_node|}
begin if tail<>head then
  begin if is_char_node(tail) then p:=tail
  else if type(tail)=ligature_node then p:=lig_char(tail)
  else return;
  f:=font(p);
  tail_append(new_kern(char_italic(f)(char_info(f)(character(p)))));
@z
@x later in that same module
end;
@y
exit: end;
@z

257. Another debugging hack. (August 27)
@x module 1339
15: begin font_in_short_display:=null_font; short_display(n);
  end;
@y
15: begin font_in_short_display:=null_font; short_display(n);
  end;
16: panicking:=not panicking;
@z

258. Redundant code eliminated (August 27)
Module 531 needn't set and reset name_in_progress [but it's harmless].

259. Bug: \input shouldn't occur during font size spec (Spivak; fixed August 27)
@x module 1258
@ @<Scan the font size specification@>=
@y
@ @<Scan the font size specification@>=
name_in_progress:=true; {this keeps |cur_name| from being changed}
@z
@x module 1258
else s:=-1000
@y
else s:=-1000;
name_in_progress:=false
@z

260. \ifhbox and \ifvbox introduced. (August 27)
@x module 487
@d ifx_code=10 { `\.{\\ifx}' }
@d if_eof_code=11 { `\.{\\ifeof}' }
@d if_true_code=12 { `\.{\\iftrue}' }
@d if_false_code=13 { `\.{\\iffalse}' }
@d if_case_code=14 { `\.{\\ifcase}' }
@y
@d if_hbox_code=10 { `\.{\\ifhbox}' }
@d if_vbox_code=11 { `\.{\\ifvbox}' }
@d ifx_code=12 { `\.{\\ifx}' }
@d if_eof_code=13 { `\.{\\ifeof}' }
@d if_true_code=14 { `\.{\\iftrue}' }
@d if_false_code=15 { `\.{\\iffalse}' }
@d if_case_code=16 { `\.{\\ifcase}' }
@z
@x
primitive("ifx",if_test,ifx_code);
@y
primitive("ifhbox",if_test,if_hbox_code);
@!@:if_hbox_}{\.{\\ifhbox} primitive@>
primitive("ifvbox",if_test,if_vbox_code);
@!@:if_vbox_}{\.{\\ifvbox} primitive@>
primitive("ifx",if_test,ifx_code);
@z
@x module 488
  ifx_code:print_esc("ifx");
@y
  if_hbox_code:print_esc("ifhbox");
  if_vbox_code:print_esc("ifvbox");
  ifx_code:print_esc("ifx");
@z
@x module 501
if_void_code: @<Test if a box is void@>;
@y
if_void_code, if_hbox_code, if_vbox_code: @<Test box register status@>;
@z
@x module 505
@ @<Test if a box is void@>=
begin scan_eight_bit_int; b:=(box(cur_val)=null);
end
@y
@ @<Test box register status@>=
begin scan_eight_bit_int; p:=box(cur_val);
if this_if=if_void_code then b:=(p=null)
else if p=null then b:=false
else if this_if=if_hbox_code then b:=(type(p)=hlist_node)
else b:=(type(p)=vlist_node);
end

261. Serious data structure error (found by Todd Allen, August 29)
@x module 478 (an error introduced in change 231)
  q:=the_toks; link(p):=link(temp_head); p:=q;
@y
  q:=the_toks;
  if link(temp_head)<>null then
    begin link(p):=link(temp_head); p:=q;
    end;
@z

262. Minor patch for efficiency (August 29)
@x module 466
    begin store_new_token(info(r)); r:=link(r);
@y
    begin fast_store_new_token(info(r)); r:=link(r);
@z

263. Minor patch to error message (August 29)
@module 579
  print(" has "); print_int(font_params[f]);
  print(" fontdimen parameters");
@.Font x has n fontdimen...@>
@y
  print(" has only "); print_int(font_params[f]);
  print(" fontdimen parameters");
@.Font x has only...@>
@z

*** The changes above comprise version 0.99999, used only at SAIL (August 29)

264. funny blank spaces are showable (August 30)
@x module 298
spacer: print("blank space");
@y
spacer: chr_cmd("blank space ");
@z

265. \newlinechar [change 235] should affect print_char too (August 31)
@x module 58
procedure print_char(@!c:ascii_code); {prints a single character}
begin case selector of
term_and_log: begin wterm(xchr[c]); write(log_file,xchr[c]);
  incr(term_offset); incr(file_offset);
  if term_offset=max_print_line then
    begin wterm_cr; term_offset:=0;
    end;
  if file_offset=max_print_line then
    begin wlog_cr; file_offset:=0;
    end;
  end;
log_only: begin write(log_file,xchr[c]); incr(file_offset);
  if file_offset=max_print_line then print_ln;
  end;
term_only: begin wterm(xchr[c]); incr(term_offset);
  if term_offset=max_print_line then print_ln;
  end;
no_print: do_nothing;
pseudo: if tally<trick_count then trick_buf[tally mod error_line]:=c;
new_string: begin if pool_ptr<pool_size then append_char(c);
  end; {drop characters if the string space is full}
othercases write(write_file[selector],xchr[c])
endcases;@/
incr(tally);
end;
@y
procedure print_char(@!s:ascii_code); {prints a single character}
label exit;
begin if @<Character |s| is the current new-line character@> then
 if selector<pseudo then
  begin print_ln; return;
  end;
case selector of
term_and_log: begin wterm(xchr[s]); wlog(xchr[s]);
  incr(term_offset); incr(file_offset);
  if term_offset=max_print_line then
    begin wterm_cr; term_offset:=0;
    end;
  if file_offset=max_print_line then
    begin wlog_cr; file_offset:=0;
    end;
  end;
log_only: begin wlog(xchr[s]); incr(file_offset);
  if file_offset=max_print_line then print_ln;
  end;
term_only: begin wterm(xchr[s]); incr(term_offset);
  if term_offset=max_print_line then print_ln;
  end;
no_print: do_nothing;
pseudo: if tally<trick_count then trick_buf[tally mod error_line]:=s;
new_string: begin if pool_ptr<pool_size then append_char(s);
  end; {drop characters if the string space is full}
othercases write(write_file[selector],xchr[s])
endcases;@/
incr(tally);
exit:end;

266. \lastkern,\lastpenalty,\unkern,\unpenalty (September 4)
@x module 208
@d unskip=25 {nullify glue ( \.{\\unskip} )}
@y
@d remove_item=25 {nullify last item ( \.{\\unpenalty},
  \.{\\unkern}, \.{\\unskip} )}
@z
@x module 208
@d last_skip=69 {most recent glue ( \.{\\lastskip} )}
@y
@d last_item=69 {most recent item ( \.{\\lastpenalty},
  \.{\\lastkern}, \.{\\lastskip} )}
@z
@x module 265
primitive("lastskip",last_skip,0);@/
@!@:last_skip_}{\.{\\lastskip} primitive@>
@y
@z
@x module 265
primitive("the",the,0);
@!@:the_}{\.{\\the} primitive@>
primitive("toks",toks_register,0);
@!@:toks_}{\.{\\toks} primitive@>
primitive("unskip",unskip,0);@/
@!@:unskip_}{\.{\\unskip} primitive@>
@y
primitive("the",the,0);@/
@!@:the_}{\.{\\the} primitive@>
primitive("toks",toks_register,0);@/
@!@:toks_}{\.{\\toks} primitive@>
@z
@x module 266
last_skip: print_esc("lastskip");
@y
@z
@x module 266
unskip: print_esc("unskip");
@y
@z
@x module 413
last_skip: @<Fetch the glue in the current node, if any@>;
@y
last_item: @<Fetch an item in the current node, if appropriate@>;
@z
@x module 416 [I also changed the comment]
primitive("dp",set_box_dimen,depth_offset);
@!@:dp_}{\.{\\dp} primitive@>
@y
primitive("dp",set_box_dimen,depth_offset);
@!@:dp_}{\.{\\dp} primitive@>
primitive("lastpenalty",last_item,int_val);
@!@:last_penalty_}{\.{\\lastpenalty} primitive@>
primitive("lastkern",last_item,dimen_val);
@!@:last_kern_}{\.{\\lastkern} primitive@>
primitive("lastskip",last_item,glue_val);
@!@:last_skip_}{\.{\\lastskip} primitive@>
@z
@x module 417
else print_esc("dp");
@y
else print_esc("dp");
last_item: if chr_code=int_val then print_esc("lastpenalty")
else if chr_code=dimen_val then print_esc("lastkern")
else print_esc("lastskip");
@z
@x module 424
@ Here is where \.{\\lastskip} is implemented. The reference count will be
updated later.
@:last_skip_}{\.{\\lastskip} primitive@>

@<Fetch the glue in the current node, if any@>=
begin cur_val:=zero_glue; cur_val_level:=glue_val;
if not is_char_node(tail)and(mode<>0) then
  begin if type(tail)=glue_node then
    begin cur_val:=glue_ptr(tail);
    if subtype(tail)=mu_glue then cur_val_level:=mu_val;
    end;
  end
else if (mode=vmode)and(tail=head)and(last_page_glue<>max_halfword) then
  cur_val:=last_page_glue;
end
@y
@ Here is where \.{\\lastpenalty}, \.{\\lastkern}, and \.{\\lastskip} are
implemented. The reference count for \.{\\lastskip} will be updated later.

@<Fetch an item in the current node...@>=
begin if cur_chr=glue_val then cur_val:=zero_glue@+else cur_val:=0;
cur_val_level:=cur_chr;
if not is_char_node(tail)and(mode<>0) then
  case cur_chr of
  int_val: if type(tail)=penalty_node then cur_val:=penalty(tail);
  dimen_val: if type(tail)=kern_node then cur_val:=width(tail);
  glue_val: if type(tail)=glue_node then
    begin cur_val:=glue_ptr(tail);
    if subtype(tail)=mu_glue then cur_val_level:=mu_val;
    end;
  end {there are no other cases}
else if (mode=vmode)and(tail=head) then
  case cur_chr of
  int_val: cur_val:=last_penalty;
  dimen_val: cur_val:=last_kern;
  glue_val: if last_glue<>max_halfword then cur_val:=last_glue;
  end; {there are no other cases}
end
@z
@x module 982 [also the comment changes]
@!last_page_glue:pointer; {used to implement \.{\\lastskip}}
@y
@!last_glue:pointer; {used to implement \.{\\lastskip}}
@!last_penalty:integer; {used to implement \.{\\lastpenalty}}
@!last_kern:scaled; {used to implement \.{\\lastkern}}
@z
@x module 991
last_page_glue:=max_halfword;
@y
last_glue:=max_halfword; last_penalty:=0; last_kern:=0;
@z
@x module 994
@<Update the value of |last_page_glue|@>;
@y
@<Update the values of |last_glue|, |last_penalty|, and |last_kern|@>;
@z
@x module 996
@ @<Update the value of |last_page_glue|@>=
if last_page_glue<>max_halfword then delete_glue_ref(last_page_glue);
if type(p)=glue_node then
  begin last_page_glue:=glue_ptr(p); add_glue_ref(last_page_glue);
  end
else last_page_glue:=max_halfword
@y
@ @<Update the values of |last_glue|...@>=
if last_glue<>max_halfword then delete_glue_ref(last_glue);
last_penalty:=0; last_kern:=0;
if type(p)=glue_node then
  begin last_glue:=glue_ptr(p); add_glue_ref(last_glue);
  end
else  begin last_glue:=max_halfword;
  if type(p)=penalty_node then last_penalty:=penalty(p)
  else if type(p)=kern_node then last_kern:=width(p);
  end
@z
@x module 1017
if last_page_glue<>max_halfword then delete_glue_ref(last_page_glue);
@<Start a new current page@>; {this sets |last_page_glue:=max_halfword|}
@y
if last_glue<>max_halfword then delete_glue_ref(last_glue);
@<Start a new current page@>; {this sets |last_glue:=max_halfword|}
@z
@x module 1048
vmode+vmove,hmode+hmove,mmode+hmove,any_mode(last_skip),
@y
vmode+vmove,hmode+hmove,mmode+hmove,any_mode(last_item),
@z
@x module 1104
any_mode(unskip): delete_skip;
@y
any_mode(remove_item): delete_last;
@z
@x module 1105
procedure delete_skip;
var p:pointer; {runs through the current list}
begin if (mode=vmode)and(tail=head) then
  @<Apologize for inability to do \.{\\unskip} now,
    unless the previous node was not glue@>
else  begin if not is_char_node(tail) then if type(tail)=glue_node then
@y
procedure delete_last;
var p:pointer; {runs through the current list}
begin if (mode=vmode)and(tail=head) then
  @<Apologize for inability to do the operation now,
    unless \.{\\unskip} follows non-glue@>
else  begin if not is_char_node(tail) then if type(tail)=cur_chr then
@z
@x module 1106
@ @<Apologize for inability to do \.{\\unskip}...@>=
begin if last_page_glue<>max_halfword then
  begin you_cant;
  help2("Sorry...I'm usually unable to take things from the current")@/
    ("page. Try `I\vskip-\lastskip' instead."); error;
@y
@ @<Apologize for inability to do the operation...@>=
begin if (cur_chr<>glue_node)or(last_glue<>max_halfword) then
  begin you_cant;
  help2("Sorry...I'm usually unable to take things from the current")@/
    ("page. Try `I\vskip-\lastskip' instead.");
  if cur_chr=kern_node then help_line[0]:=
    ("page. Try `I\kern-\lastkern' instead.")
  else if cur_chr<>glue_node then help_line[0]:=@|
    ("page. Perhaps you can make the output routine do it.");
  error;
@z
@x module 1107
primitive("unhbox",un_hbox,box_code);@/
@y
primitive("unpenalty",remove_item,penalty_node);@/
@!@:un_penalty_}{\.{\\unpenalty} primitive@>
primitive("unkern",remove_item,kern_node);@/
@!@:un_kern_}{\.{\\unkern} primitive@>
primitive("unskip",remove_item,glue_node);@/
@!@:un_skip_}{\.{\\unskip} primitive@>
primitive("unhbox",un_hbox,box_code);@/
@z
@x module 1108
un_hbox: if chr_code=copy_code then print_esc("unhcopy")
@y
remove_item: if chr_code=glue_node then print_esc("unskip")
  else if chr_code=kern_node then print_esc("unkern")
  else print_esc("unpenalty");
un_hbox: if chr_code=copy_code then print_esc("unhcopy")
@z

*** The above changes installed in version 0.999999 (September 5, 1983)

267. Undo change 29: it was overkill and not needed (Sep 17)
@x module 1257
label exit, common_ending;
var u:pointer; {user's font identifier}
@!s:scaled; {stated ``at'' size, or negative of scaled magnification}
@!f:internal_font_number; {runs through existing fonts}
begin if job_name=0 then open_log_file;
  {avoid confusing \.{texput} with the font name}
@<If the next token isn't a suitable control sequence, issue a complaint
  and |return|; otherwise set |u:=cs_ptr|, and set |hash_used| to the
  location of a ``frozen'' copy of the new font identifier@>;
define(u,set_font,null_font); scan_optional_equals; scan_file_name;
@<Scan the font size specification@>;
@<If this font has already been loaded, set |f| to the internal
  font number and |goto common_ending|@>;
f:=read_font_info(u,cur_name,cur_area,s);
common_ending: equiv(u):=f; geq_define(hash_used,set_font,f);
font_ident[f]:=hash_used;
exit:end;
@y
label common_ending;
var u:pointer; {user's font identifier}
@!s:scaled; {stated ``at'' size, or negative of scaled magnification}
@!f:internal_font_number; {runs through existing fonts}
begin if job_name=0 then open_log_file;
  {avoid confusing \.{texput} with the font name}
get_r_token; u:=cs_ptr;
define(u,set_font,null_font); scan_optional_equals; scan_file_name;
@<Scan the font size specification@>;
@<If this font has already been loaded, set |f| to the internal
  font number and |goto common_ending|@>;
f:=read_font_info(u,cur_name,cur_area,s);
common_ending: equiv(u):=f; font_ident[f]:=u;
end;
@z
@x module 1260
@ We reserve a special control sequence for the font identifier; this one
cannot be redefined by the user, so it is safe to return it as a value of
\.{\\the\\font}.

@<If the next token isn't a suitable control sequence...@>=
get_r_token;
if cs_ptr<hash_base then
  begin print_err("A font identifier must be a multiletter control sequence");
@.A font identifier...@>
  help2("You should say, e.g., `\font\ffn=fontfilename'.")@/
  ("(I'm going to ignore the \font command you just gave.)");
  back_error; return;
  end;
repeat if hash_is_full then overflow("hash size",hash_size);
@:TeX capacity exceeded hash size}{\quad hash size@>
decr(hash_used);
until text(hash_used)=0; {search for an empty location in |hash|}
u:=cs_ptr; text(hash_used):=text(u); {copy the name}
@!stat incr(cs_count);@+tats@;@/
@y
@z
Note: Since module 1260 has disappeared, module 1258 has been split into two.

268. Minor change to diagnostic output format (September 18)
@x module 211 [print_mode]
  2:print("displayed math");
@y
  2:print("display math");
@z

269. Kerns inserted for accents must be explicit (September 20)
@x module 1123
@!p,@!q: pointer; {character and box nodes}
@y
@!p,@!q,@!r:pointer; {character, box, and kern nodes}
@z
@x module 1125
link(tail):=new_kern(delta); link(link(tail)):=p;
link(p):=new_kern(-a-delta); tail:=link(p); p:=q;
@y
r:=new_kern(delta); subtype(r):=explicit; link(tail):=r; link(r):=p;
tail:=new_kern(-a-delta); subtype(tail):=explicit; link(p):=tail; p:=q;
@z

270. "log" changed to "transcript" in a few output messages (Sep 26)
@x module 535
prompt_file_name("log file name",".log");
@y
prompt_file_name("transcript file name",".log");
@z
@x module 1293
      ("lists on your terminal as well as on the log file.");
@y
      ("lists on your terminal as well as in the transcript file.");
@z
@x module 1335
  print_nl("(see the log file for additional information)");
@.see the log file...@>
@y
  print_nl("(see the transcript file for additional information)");
@.see the transcript file...@>
@z

271. Uninitialized variable bug (found by Bernd Schulze, 1 Oct 83)
@x module 944
    begin if trie_op_ptr>=max_quarterword-1 then {overflow}
      begin trie_op_ptr:=max_quarterword;
      new_trie_op:=min_quarterword; return;
@y we allow one more trie op (OK since trie_op_hash_size is big enough)
    begin if trie_op_ptr=max_quarterword then {overflow}
      begin new_trie_op:=min_quarterword; return;
@z

272. Spaces at end of lines ignored also in TEX.POOL (by DRF, 14 Oct 83)
@x module 52
    begin if eoln(pool_file) then bad_pool('! TEX.POOL line too short.');
@.TEX.POOL line too short@>
    read(pool_file,m); append_char(xord[m]);
@y
    begin if eoln(pool_file) then m:=' '@+else read(pool_file,m);
    append_char(xord[m]);
@z

273. |history| updates (by DRF, 14 Oct 83)
@x module 77
deletions_allowed:=true; history:=spotless; error_count:=0;
@y
deletions_allowed:=true; error_count:=0; {|history| is initialized elsewhere}
@z
@x module 1332
@p begin {@!|start_here|}
@y
@p begin {@!|start_here|}
history:=fatal_error_stop; {in case we quit during initialization}
@z
@x ibid.
main_control; {come to life}
@y
history:=spotless; {ready to go!}
main_control; {come to life}
@z

274. improved "runaway" messages (suggested by FY, October 18)
@x module 305 (the introductory comment is changed too)
@d aligning=4 {|scanner_status| when reading an alignment preamble}

@<Glob...@>=
@!scanner_status : normal..aligning; {can a subfile end now?}
@y
@d aligning=4 {|scanner_status| when reading an alignment preamble}
@d absorbing=5 {|scanner_status| when reading a balanced text}

@<Glob...@>=
@!scanner_status : normal..absorbing; {can a subfile end now?}
@z
@x module 306 (string space is also now conserved slightly)
  defining: begin print("definition?"); p:=def_ref;
    end;
  matching: begin print("argument?"); p:=temp_head;
    end;
  aligning: begin print("preamble?"); p:=hold_head;
    end;
  end; {there are no other cases}
  print_ln; show_token_list(link(p),null,error_line-10);
@y
  defining: begin print("definition"); p:=def_ref;
    end;
  matching: begin print("argument"); p:=temp_head;
    end;
  aligning: begin print("preamble"); p:=hold_head;
    end;
  absorbing: begin print("text"); p:=def_ref;
    end;
  end; {there are no other cases}
  print_char("?");print_ln; show_token_list(link(p),null,error_line-10);
@x module 339 (I also changed the module name)
end; {there are no other cases}
@y
absorbing:begin print("text"); info(p):=right_brace_token+"}";
  end;
end; {there are no other cases}
@z
@x module 473
begin scanner_status:=defining; warning_index:=cs_ptr;
def_ref:=get_avail; info(def_ref):=null;
@y
begin if macro_def then scanner_status:=defining
@+else scanner_status:=absorbing;
warning_index:=cs_ptr; def_ref:=get_avail; info(def_ref):=null;
@z

275. similar, but this corrects a real bug (found by FY, October 18)
@x module 1226
  cs_ptr:=q;
  @<Get the next non-blank non-relax non-call token@>;
  if cur_cmd<>left_brace then @<If the righthand side is a token parameter
      or token register, finish the assignment and |goto done|@>;
  back_input; q:=scan_toks(false,false);
@y
  @<Get the next non-blank non-relax non-call token@>;
  if cur_cmd<>left_brace then @<If the righthand side is a token parameter
      or token register, finish the assignment and |goto done|@>;
  back_input; cs_ptr:=q; q:=scan_toks(false,false);
@z

276. Change #119 should have changed module 1090 too (by Barry Smith, Oct 24)
@x module 1183 [which was module 1090 at the time of change 119]
begin if c mod delimited_code=above_code then scan_normal_dimen;
if c>=delimited_code then
  begin scan_delimiter(garbage,false); scan_delimiter(garbage,false);
  end;
@y
begin if c>=delimited_code then
  begin scan_delimiter(garbage,false); scan_delimiter(garbage,false);
  end;
if c mod delimited_code=above_code then scan_normal_dimen;
@z

277. Changes for efficiency, based on empirical frequency data (Nov 9)
@x module 45 [since the result is true almost always]
begin j:=str_start[s]; result:=false;
while j<str_start[s+1] do
  begin if str_pool[j]<>buffer[k] then goto not_found;
@y
begin j:=str_start[s];
while j<str_start[s+1] do
  begin if str_pool[j]<>buffer[k] then
    begin result:=false; goto not_found;
    end;
@z
@x module 380 [many compilers don't handle "while true do" very well]
label done;
begin loop begin get_next;
@^inner loop@>
  if cur_cmd<=max_command then goto done;
  if cur_cmd>=call then
    if cur_cmd<end_template then macro_call
    else  begin cs_ptr:=frozen_endv; cur_cmd:=endv;
      goto done; {|cur_chr=null_list|}
      end
  else expand;
  end;
@y
label restart,done;
begin restart: get_next;
@^inner loop@>
if cur_cmd<=max_command then goto done;
if cur_cmd>=call then
  if cur_cmd<end_template then macro_call
  else  begin cs_ptr:=frozen_endv; cur_cmd:=endv;
    goto done; {|cur_chr=null_list|}
    end
else expand;
goto restart;
@z
@x module 829 [preparation for change to module 852]
label exit,done,continue,deactivate;
@y
label exit,done,done1,continue,deactivate;
@z
@x module 830 [ditto]
@!save_link:pointer; {temporarily holds value of |link(cur_p)|}
@y
@!save_link:pointer; {temporarily holds value of |link(cur_p)|}
@!shortfall:scaled; {used in badness calculations}
@z
@x modue 851 [ditto]
if cur_active_width[1]<line_width then
@y
shortfall:=line_width-cur_active_width[1]; {we're this much too short}
if shortfall>0 then
@z
@x module 852 [avoid calling |badness| in most common case]
else  begin b:=badness(line_width-cur_active_width[1],cur_active_width[2]);
@y
else  begin if shortfall>7230584 then if cur_active_width[2]<1663497 then
    begin b:=inf_bad; fit_class:=very_loose_fit; goto done1;
    end;
  b:=badness(shortfall,cur_active_width[2]);
@z
@x module 852, continued
  else fit_class:=decent_fit;
@y
  else fit_class:=decent_fit;
  done1:
@z
@x module 853 [using |shortfall|, since we now have it]
begin if cur_active_width[1]-line_width>cur_active_width[6] then
  b:=inf_bad+1
else b:=badness(cur_active_width[1]-line_width,cur_active_width[6]);
@y
begin if -shortfall>cur_active_width[6] then b:=inf_bad+1
else b:=badness(-shortfall,cur_active_width[6]);
@z

278. Forgotten |error| call (noticed by Gabi Kuper, December 3, 1983)
@x module 500
    help1("I'm ignoring this; it doesn't match any \if.");
@y
    help1("I'm ignoring this; it doesn't match any \if.");
    error;
@z

*** Version 1.0 released on December 3, 1983 incorporates all of the above.
Version 1.1

279. Problem with change 267 (found by Mike Urban, received 2 Feb 84)
(I had overlooked many problems, e.g. `{\font\a=x \global\a}\the\font'
and `\font\a=x \font\b=x \let\b=\undefined \the\a', etc.
The remedy involves removal of the font_ident array, so there's a
sprinkling of corrections in lots of modules. But basically the change
is quite conservative, so it shouldn't spawn any new bugs (it says here).)

@x module 174 (removing a reference to |font_ident|)
        else sprint_cs(font_ident[font(p)]);
@y
        else @<Print the font identifier for |font(p)|@>;
@z
@x module 176 (removing a reference to |font_ident|)
  else sprint_cs(font_ident[font(p)]);
@y
  else @<Print the font identifier for |font(p)|@>;
@z
@x module 222 (redefining and expanding the `frozen' area)
@d frozen_null_font=frozen_control_sequence+9 {permanent `\.{\\nullfont}'}
@d frozen_dont_expand=frozen_control_sequence+10
  {permanent `\.{\\notexpanded:}'}
@d undefined_control_sequence=frozen_control_sequence+11 {dummy location}
@y
@d frozen_dont_expand=frozen_control_sequence+9
  {permanent `\.{\\notexpanded:}'}
@d frozen_null_font=frozen_control_sequence+10
  {permanent `\.{\\nullfont}'}
@d font_id_base=frozen_null_font-font_base
  {begins table of 257 permanent font identifiers}
@d undefined_control_sequence=frozen_null_font+257 {dummy location}
@z
@x module 234 (removing a reference to |font_ident|, awkwardly)
print_char("="); sprint_cs(font_ident[equiv(n)]);
@y
print_char("=");@/
print_esc(hash[font_id_base+equiv(n)].rh);
  {that's |text(font_id_base+equiv(n))|}
@z
@x module 256 (adding a new macro, font_id_text)
@d hash_is_full == (hash_used=hash_base) {test if all positions are occupied}
@y
@d hash_is_full == (hash_used=hash_base) {test if all positions are occupied}
@d font_id_text(#) == text(font_id_base+#) {a frozen font identifier's name}
@z
@x module 262 (relaxing a former restriction)
else if (text(p)<128)or(text(p)>=str_ptr) then print_esc("NONEXISTENT.")
@y
else if (text(p)<0)or(text(p)>=str_ptr) then print_esc("NONEXISTENT.")
@z
@x module 267 (pick up the changes from 174 and 176)
@!@^Single-character primitives@>
@y
@!@^Single-character primitives@>

Meanwhile, this is a convenient place to catch up on something we were unable
to do before the hash table was defined:

@<Print the font identifier for |font(p)|@>=
print_esc(font_id_text(font(p)))
@z
@x module 415 (removing another reference to font_ident)
  scanned_result(font_ident[cur_val])(ident_val);
@y
  scanned_result(font_id_base+cur_val)(ident_val);
@z
@x module 548 (change to the comment only)
to the user's font~\.{\\f}. For example, if this internal number is 13,
we will have |font_ident[13]=p| and |equiv(p)=13|, where |p| is the |eqtb|
location of the control sequence~\.{\\f}.
@y
to the user's font~\.{\\f}. Adding this number to |font_id_base| gives the
|eqtb| location of a ``frozen'' control sequence that will always select
the font.
@z
@x module 549 (deleting the declaration of font_ident)
@!font_ident:array[internal_font_number] of pointer; {the most recent user
  font identifier corresponding to an internal font number}
@y
@z
@x module 552 (deleting an unnecessary initialization of font_ident)
font_ident[null_font]:=frozen_null_font;
@y
@z
@x module 579 (removing another reference to font_ident)
  begin print_err("Font "); sprint_cs(font_ident[f]);
@y
  begin print_err("Font "); print_esc(font_id_text(f));
@z
@x module 1257 (this is the real change!)
begin if job_name=0 then open_log_file;
  {avoid confusing \.{texput} with the font name}
get_r_token; u:=cs_ptr;
define(u,set_font,null_font); scan_optional_equals; scan_file_name;
@<Scan the font size specification@>;
@<If this font has already been loaded, set |f| to the internal
  font number and |goto common_ending|@>;
f:=read_font_info(u,cur_name,cur_area,s);
common_ending: equiv(u):=f; font_ident[f]:=u;
@y
@!t:str_number; {name for the frozen font identifier}
@!old_setting:0..max_selector; {holds |selector| setting}
begin if job_name=0 then open_log_file;
  {avoid confusing \.{texput} with the font name}
get_r_token; u:=cs_ptr;
if u>=hash_base then t:=text(u)
else if u>=single_base then
  if u=null_cs then t:="FONT"@+else t:=u-single_base
else  begin old_setting:=selector; selector:=new_string;
  print("FONT"); print(u-active_base); selector:=old_setting;
@.FONTx@>
  str_room(1); t:=make_string;
  end;
define(u,set_font,null_font); scan_optional_equals; scan_file_name;
@<Scan the font size specification@>;
@<If this font has already been loaded, set |f| to the internal
  font number and |goto common_ending|@>;
f:=read_font_info(u,cur_name,cur_area,s);
common_ending: equiv(u):=f; eqtb[font_id_base+f]:=eqtb[u]; font_id_text(f):=t;
@z
@x module 1260 (change to the comment only)
the new value becomes the |font_ident| of record. Font names `\.{xyz}' and
@y
the new name becomes the font identifier of record. Font names `\.{xyz}' and
@z
@x module 1322 (no need to dump font_ident)
begin dump_int(font_ident[k]);
dump_qqqq(font_check[k]);
@y
begin dump_qqqq(font_check[k]);
@z
@x module 1322 (or to print from it)
print_nl("\font"); sprint_cs(font_ident[k]); print_char("=");
@y
print_nl("\font"); print_esc(font_id_text(k)); print_char("=");
@z
@x module 1323 (or to undump it)
begin undump(active_base)(undefined_control_sequence)(font_ident[k]);
undump_qqqq(font_check[k]);@/
@y
begin undump_qqqq(font_check[k]);@/
@z

280. Double interrupt possibility (found by Clint Cuzzo, received 9 Feb 84)
@x module 1031
if interrupt>0 then if OK_to_interrupt then
  begin back_input; pause_for_instructions; goto big_switch;
@y
if interrupt<>0 then if OK_to_interrupt then
  begin back_input; check_interrupt; goto big_switch;
@z

281. Improve spacing in $(A,<)$. (12 Feb 84)
@x module 764
"0234000122*4000133**3**344*0400400*000000234000111*4111112341011"
@y
"0234000122*4000133**3**344*0400400*000000234000111*1111112341011"
@z

282. Bad goto! (diagnosis by Clint Cuzzo and George O'Connor, recd 13 Feb 84)
@x module 344 (here we simply change the name of module 346)
any_state_plus(invalid_char): @<Decry the invalid character and
  |goto switch|@>;
@y
any_state_plus(invalid_char): @<Decry the invalid character and
  |goto restart|@>;
@z
@x module 346 (because clear_for_error_prompt might make state=token_list)
goto switch;
@y
goto restart;
@z

283. String pool economy (suggested by DRF, 13 Feb 84)
@x module 537
@<Read the first line of the new file@>;
@y
if name=str_ptr-1 then {we can conserve string pool space now}
  begin flush_string; name:=cur_name;
  end;
@<Read the first line of the new file@>;
@z

284. Nicer scaled output (suggested by METAFONT development, 26 Feb 84)
@x module 103
be reproduced exactly. [{\sl Proof:\/} If round$(x)=\lfloor
x+{1\over2}\rfloor$ and if $\alpha<1$, it is not difficult to verify that
round$(\alpha\,\hbox{round}( \alpha^{-1}n))=n$ for all integers |n|. In
our case $\alpha=2^{16}/10^5$.]

@p procedure print_scaled(@!s:scaled); {prints scaled real, rounded to five
  digits}
begin if s<0 then
  begin print_char("-"); negate(s); {print the sign, if negative}
  end;
print_int(s div unity); {print the integer part}
s:=((s mod unity) * 3125 + 1024) div 2048;
  {now |0<=s<100000| is the fraction part}
print_char(".");
repeat print_char("0"+(s div 10000)); s:=10*(s mod 10000);
until s=0;
end;
@y
be reproduced exactly; the ``simplest'' such decimal number is output,
but there is always at least one digit following the decimal point.

The invariant relation in the \&{repeat} loop is that a sequence of
decimal digits yet to be printed will yield the original number if and only if
they form a fraction~$f$ in the range $s-\delta\L10\cdot2^{16}f<s$.
We can stop if and only if $f=0$ satisfies this condition; the loop will
terminate before $s$ can possibly become zero.

@p procedure print_scaled(@!s:scaled); {prints scaled real, rounded to five
  digits}
var delta:scaled; {amount of allowable inaccuracy}
begin if s<0 then
  begin print_char("-"); negate(s); {print the sign, if negative}
  end;
print_int(s div unity); {print the integer part}
print_char(".");
s:=10*(s mod unity)+5; delta:=10;
repeat if delta>unity then s:=s+@'100000-(delta div 2); {round the last digit}
print_char("0"+(s div unity)); s:=10*(s mod unity); delta:=delta*10;
until s<=delta;
end;
@z

*** Those six changes account for version 1.1 (February 29, 1984).
*** And we put the following change in too, at the last minute:

285. ".00000762939453126pt" didn't round correctly! (March 2, 1984)
@x module 452
  if k<16 then {digits for |k>=16| cannot affect the result}
@y (Note, I also changed the comment in module 102: 16 -> 17)
  if k<17 then {digits for |k>=17| cannot affect the result}
@z

*** Note: I changed "cs_ptr" to "cur_cs", since the name change is more
indicative of the fact that this variable makes a trio with cur_cmd
and cur_chr. The TEX.WEB source changed 78 times, as a result, but
none of the author's change files were affected.
(But note: "save_cs_ptr" was not changed to "save_cur_cs", because
of a name conflict.)

*** And the following two were slipped in at the last second:

286. loophole permitted get_next recursion (March 16, 1984)
@x module 336
  begin @<Back up an outer control sequence so that it can be reread@>;
@y
  begin deletions_allowed:=false;
  @<Back up an outer control sequence so that it can be reread@>;
@z
@x ibid
  end;
@y
  deletions_allowed:=true;
  end;
@z [the changes to deletions_allowed in module 338 can now be removed]

287. terminal not open when the program starts bad (March 24, 1984)
@x module 1332
initialize; {set global variables to their starting values}
@<Check the ``constant'' values...@>@;
if bad>0 then
  begin wake_up_terminal;
  wterm('Ouch---my internal constants have been',
    ' clobbered!---case ',bad:1);
@.Ouch...clobbered@>
  goto final_end;
  end;
@y
@<Check the ``constant'' values...@>@;
if bad>0 then
  begin t_open_out;
  wterm('Ouch---my internal constants have been',
    ' clobbered!---case ',bad:1);
@.Ouch...clobbered@>
  goto final_end;
  end;
initialize; {set global variables to their starting values}
@z

*** Note: "Cosmetic" changes were also made in module 363 ("print_ln"
moved to after "wake_up_terminal") and in some commentary.

*** Another last-minute correction:
288. \patterns{xxx...xxxdxxxdxxx} anomaly found by JRD (Mar 27, 1984)
@x module 962
else  begin hyf[k]:=cur_chr-"0"; digit_sensed:=true;
@y
else  begin hyf[k]:=cur_chr-"0";
  if k<63 then digit_sensed:=true;
@z

*** Yoicks, yet ANOTHER:
289. a whole case of copy_node_list forgotten (Apr 11, 1984)
@x module 206
othercases confusion("copying")
@y
adjust_node: begin r:=get_node(small_node_size);
  adjust_ptr(r):=copy_node_list(adjust_ptr(p));
  end; {|words=1=small_node_size-1|}
othercases confusion("copying")
@z

*** And "finally"...
290. uninitialized variables could be accessed (found by Nick Briggs, June 11)
@x module 552
font_size[null_font]:=0; font_dsize[null_font]:=0;
@y
font_size[null_font]:=0; font_dsize[null_font]:=0;
char_base[null_font]:=0; width_base[null_font]:=0;
height_base[null_font]:=0; depth_base[null_font]:=0;
italic_base[null_font]:=0; lig_kern_base[null_font]:=0;
kern_base[null_font]:=0; exten_base[null_font]:=0;
@z (actual values are immaterial except for debugging safeguards)

291. same, cf. modules 355 and 389 (Nick Briggs, June 11)
@x module 331
scanner_status:=normal; first:=1;
@y
for first:=0 to buf_size do buffer[first]:=0;
scanner_status:=normal; warning_index:=null; first:=1;
@z (error was harmless except that it might trigger debugging checks)

292. missing ligature/kern (found by JRD, June 21)
@x module 1036
  if (cur_cmd=letter)or(cur_cmd=other_char) then r:=qi(cur_chr)
@y
  if (cur_cmd=letter)or(cur_cmd=other_char)or(cur_cmd=char_given) then
    r:=qi(cur_chr)
@z

293. quarterword constraint is made explicit (July 4)
@x module 111
if (buf_size>max_halfword) then bad:=18;
@y
if buf_size>max_halfword then bad:=18;
if max_quarterword-min_quarterword<255 then bad:=19;
@z

294. trivial optimization made for consistency with MF (July 7)
@x module 363
if (pausing>0)and(interaction>nonstop_mode) then
@y
if pausing>0 then if interaction>nonstop_mode then
@z

295. \tracingmacros>1 for more diagnostics (July 8)
@x module 323
  if t=macro then param_start:=param_ptr@+else loc:=link(p);
@y
  if t=macro then param_start:=param_ptr
  else  begin loc:=link(p);
    if tracing_macros>1 then
      begin begin_diagnostic; print_nl("");
      case t of
      mark_text:print_esc("mark");
      write_text:print_esc("write");
      othercases print_cmd_chr(assign_toks,t-output_text+output_routine_loc)
      endcases;@/
      print("->"); token_show(p); end_diagnostic(false);
      end;
    end;
@z

*** The changes above were incorporated in Version 1.1, released July 9, 1984.
Version 1.2

296. "see the transcript file" on offline show commands. (July 27, 1984)
@x module 1293
@ @d show_error==@;@/
  if interaction<error_stop_mode then
    begin help0; decr(error_count);
    end
  else  if tracing_online>0 then
      begin@t@>@;@/
      help3("This isn't an error message; I'm just \showing something.")
        @t\4\4@>@/
      ("Type `I\show...' to show more (e.g., \show\cs,")@/
      ("\showthe\count10, \showbox255, \showlists).");
      end
    else  begin@t@>@;@/
      help5("This isn't an error message; I'm just \showing something.")
        @t\4\4@>@/
      ("Type `I\show...' to show more (e.g., \show\cs,")@/
      ("\showthe\count10, \showbox255, \showlists).")@/
      ("And type `I\tracingonline=1\show...' to show boxes and")@/
      ("lists on your terminal as well as in the transcript file.");
      end;
  error

@<Declare act...@>=
procedure show_whatever;
var p:pointer; {tail of a token list to show}
begin case cur_chr of
show_code: @<Show the current meaning of a token@>;
show_box_code: @<Show the current contents of a box@>;
show_the_code: @<Show the current value of some parameter or register@>;
othercases @<Show the current semantic nest@>
endcases;
end;
@y
@ @<Declare act...@>=
procedure show_whatever;
label common_ending;
var p:pointer; {tail of a token list to show}
begin case cur_chr of
show_lists: begin begin_diagnostic; show_activities;
  end;
show_box_code: @<Show the current contents of a box@>;
show_code: @<Show the current meaning of a token, then |goto common_ending|@>;
othercases @<Show the current value of some parameter or register,
  then |goto common_ending|@>
endcases;@/
@<Complete a potentially long \.{\\show} command@>;
common_ending: if interaction<error_stop_mode then
  begin help0; decr(error_count);
  end
else if tracing_online>0 then
  begin@t@>@;@/
  help3("This isn't an error message; I'm just \showing something.")@/
  ("Type `I\show...' to show more (e.g., \show\cs,")@/
  ("\showthe\count10, \showbox255, \showlists).");
  end
else  begin@t@>@;@/
  help5("This isn't an error message; I'm just \showing something.")
  ("Type `I\show...' to show more (e.g., \show\cs,")@/
  ("\showthe\count10, \showbox255, \showlists).")@/
  ("And type `I\tracingonline=1\show...' to show boxes and")@/
  ("lists on your terminal as well as in the transcript file.");
  end;
error;
end;
@z
@x module 1294
@ @<Show the current meaning of a token@>=
begin get_token; print_nl("> ");
if cur_cs<>0 then
  begin sprint_cs(cur_cs); print_char("=");
  end;
print_meaning; show_error;
@y
@ @<Show the current meaning of a token...@>=
begin get_token; print_nl("> ");
if cur_cs<>0 then
  begin sprint_cs(cur_cs); print_char("=");
  end;
print_meaning; goto common_ending;
@z
@x module 1296
end_diagnostic(true);
print_err("OK");
@.OK@>
show_error;
@y
@z
@x module 1297
flush_list(link(temp_head)); show_error;
@y
flush_list(link(temp_head)); goto common_ending;
@z
@x module 1298
@ @<Show the current semantic nest@>=
begin begin_diagnostic; show_activities; end_diagnostic(true);
print_err("OK");
show_error;
@.OK@>
end
@y
@ @<Complete a potentially long \.{\\show} command@>=
end_diagnostic(true); print_err("OK");
@.OK@>
if selector=term_and_log then if tracing_online<=0 then
  begin selector:=term_only; print(" (see the transcript file)");
  selector:=term_and_log;
  end
@z

297. allow `0' in response to error prompts (October 20, 1984)
@x module 84
"1","2","3","4","5","6","7","8","9": if deletions_allowed then
@y
"0","1","2","3","4","5","6","7","8","9": if deletions_allowed then
@z
Versions 1.3 and 1.4

298. Dirty PASCAL (two quarterword 0's were read as a halfword) (25 Nov 84)
@x module 846
print(" -> @@@@"); print_int(serial(prev_break(passive)));
@y
print(" -> @@@@");
if prev_break(passive)=null then print_char("0")
else print_int(serial(prev_break(passive)));
@z

299. Ditto! (25 Nov 84)
@x module 934
label reswitch, exit, found, not_found;
@y
label reswitch, exit, found, not_found, done;
@z
@x module 939
while info(p)>n-3 do {eliminate hyphens \TeX\ doesn't like}
  begin q:=link(p); free_avail(p); p:=q;
  end;
@<Insert the \(p)pair |(s,p)| into the exception table@>;
@y
loop@+  begin if p=null then goto done;
  if info(p)<n-2 then goto done;
  q:=link(p); free_avail(p); p:=q; {eliminate hyphens that \TeX\ doesn't like}
  end;
done: @<Insert the \(p)pair |(s,p)| into the exception table@>;
@z

300. Major change to memory management (25 Nov 84)
[The following changes have two main merits: (1) TeX should run faster on
virtual-memory systems when the jobs are small; (2) it should be possible
to use the same production version of TeX both for jobs with lots of macros
and few boxes as well as for jobs with lots of boxes and few macros, because
the boundary line between the two types of storage is now chosen dynamically.]
[The user is unaffected, except that there's only one "memory size" overflow
error now instead of two.]
[The TRIP test is now changed slightly: The correct settings of memory
size parameters are now mem_min=mem_bot=1, mem_top=mem_max=3000.]

@x module 11
@!mem_max=30000; {greatest index in \TeX's internal |mem| array;
  must be strictly less than |max_halfword|}
@y
@!mem_max=30000; {greatest index in \TeX's internal |mem| array;
  must be strictly less than |max_halfword|;
  must be equal to |mem_top| in \.{INITEX}, otherwise |>=mem_top|}
@!mem_min=0; {smallest index in \TeX's internal |mem| array;
  must be |min_halfword| or more;
  must be equal to |mem_bot| in \.{INITEX}, otherwise |<=mem_bot|}
@z
@x module 12
@d mem_base=0 {smallest index in the |mem| array; must not be less
  than |min_halfword|}
@d hi_mem_base=13000 {smallest index in the single-word area of |mem|;
  must be substantially larger than |mem_base| and smaller than |mem_max|}
@y
@d mem_bot=0 {smallest index in the |mem| array dumped by \.{INITEX};
  must not be less than |mem_min|}
@d mem_top==30000 {largest index in the |mem| array dumped by \.{INITEX};
  must be substantially larger than |mem_bot|
  and not greater than |mem_max|}
@z
@x module 12
  about |(mem_max-hi_mem_base)/6|, but 2100 is already quite generous}
@y
  about |(mem_max-mem_min)/10|, but 2100 is already quite generous}
@z
@x module 14
if (hi_mem_base<mem_base+100)or(hi_mem_base+100>mem_max) then bad:=4;
@y
if mem_bot+1100>mem_top then bad:=4;
@z
@x module 111
if (min_quarterword>0)or(max_quarterword<127) then bad:=11;
@y
init if (mem_min<>mem_bot)or(mem_max<>mem_top) then bad:=10;@+tini@;@/
if (mem_min>mem_bot)or(mem_max<mem_top) then bad:=10;
if (min_quarterword>0)or(max_quarterword<127) then bad:=11;
@z
@x module 111
if (mem_base<min_halfword)or(mem_max>=max_halfword) then bad:=14;
@y
if (mem_min<min_halfword)or(mem_max>=max_halfword) then bad:=14;
@z
@x module 115
value represents a null pointer.

@d pointer==halfword {a flag or a location in |mem| or |eqtb|}
@d null==mem_base {the null pointer}
@y
value represents a null pointer. \TeX\ does not assume that |mem[null]| exists.

@d pointer==halfword {a flag or a location in |mem| or |eqtb|}
@d null==min_halfword {the null pointer}
@z
@x module 116
@ The |mem| array is divided once and for all into two regions that are
allocated separately. Locations less than |hi_mem_base| are used for storing
@y
@ The |mem| array is divided into two regions that are allocated separately,
but the dividing line between these two regions is not fixed; they grow
together until finding their ``natural'' size in a particular job.
Locations less than or equal to |lo_mem_max| are used for storing
@z
@x module 116
relevant size when a node is freed. The remaining region of |mem| is
allocated in single words using a conventional \.{AVAIL} stack.

Incidentally, it would be feasible to construct implementations of \TeX\ that
are based on 16-bit words instead of 32-bit words, for machines having
comparatively small memories. In such cases it might be desirable to have
two parallel arrays for the upper part of memory, called say \\{mem\_link}|[p]|
and \\{mem\_info}|[p]|,
since the single-word region in the present implementation
consists entirely of |memory_word| items of type |two_halves|.
@^small computers@>
@y
relevant size when a node is freed. Locations greater than or equal to
|hi_mem_min| are used for storing one-word records; a conventional
\.{AVAIL} stack is used for allocation in this region.

Locations of |mem| between |mem_bot| and |mem_top| may be dumped as part
of preloaded format files, by the \.{INITEX} preprocessor.
@.INITEX@>
Production versions of \TeX\ may extend the memory at both ends in order to
provide more space; locations between |mem_min| and |mem_bot| are always
used for variable-size nodes, and locations between |mem_top| and |mem_max|
are always used for single-word nodes.

The key pointers that govern |mem| allocation have a prescribed order:
$$\hbox{|null<=mem_min<=mem_bot<lo_mem_max<
  hi_mem_min<mem_top<=mem_end<=mem_max|.}$$
@z
@x module 116
@!mem : array[mem_base..mem_max] of memory_word; {the big dynamic storage area}
@y
@!mem : array[mem_min..mem_max] of memory_word; {the big dynamic storage area}
@!lo_mem_max : pointer; {the largest location of variable-size memory in use}
@!hi_mem_min : pointer; {the smallest location of one-word memory in use}
@z
@x module 117
@!max_var_used : integer; {how much memory was in use}
@y
@z
@x module 118
occur between |hi_mem_base| and |mem_end|, inclusive, are of type
@y
occur between |hi_mem_min| and |mem_end|, inclusive, are of type
@z
@x module 120
else  begin runaway; {if memory is exhausted, display possible runaway text}
  overflow("macro memory size",mem_max+1-hi_mem_base);
    {quit; all one-word nodes are busy}
@:TeX capacity exceeded macro memory size}{\quad macro memory size@>
  end;
@y
else   begin decr(hi_mem_min); p:=hi_mem_min;
  if hi_mem_min<=lo_mem_max then
    begin runaway; {if memory is exhausted, display possible runaway text}
    overflow("main memory size",mem_max+1-mem_min);
      {quit; all one-word nodes are busy}
@:TeX capacity exceeded main memory size}{\quad main memory size@>
    end;
  end;
@z
@x module 125
label found,exit;
@y
label found,exit,restart;
@z
@x module 125
begin p:=rover; {start at some free node in the ring}
@y
begin restart: p:=rover; {start at some free node in the ring}
@z
@x module 125
overflow("box memory size",hi_mem_base-mem_base);
  {sorry, nothing satisfactory is left}
@:TeX capacity exceeded box memory size}{\quad box memory size@>
found: link(r):=null; {this node is now nonempty}
@!stat var_used:=var_used+s; {maintain usage statistics}
if var_used>max_var_used then max_var_used:=var_used;
@y
if lo_mem_max+2<hi_mem_min then
  @<Grow more variable-size memory and |goto restart|@>;
overflow("main memory size",mem_max+1-mem_min);
  {sorry, nothing satisfactory is left}
@:TeX capacity exceeded main memory size}{\quad main memory size@>
found: link(r):=null; {this node is now nonempty}
@!stat var_used:=var_used+s; {maintain usage statistics}
@z
@x modules 126--127
@ Empirical tests show that the routine in this section performs a
|remove_node| operation about 0.75 times per allocation, on the average,
after which it finds that |r>p+1| about 95\% of the time.

@<Try to allocate...@>=
q:=p+node_size(p); {find the physical successor}
@^inner loop@>
while is_empty(q) do @<Merge node |p| with node |q|@>;
r:=q-s;
if r>p+1 then @<Allocate from the top of node |p| and |goto found|@>;
if r=p then if ((rlink(p)<>rover) or (llink(p)<>rover)) then
  @<Allocate entire node |p| and |goto found|@>;
node_size(p):=q-p {reset the size in case it grew}

@ @<Merge node |p| with node |q|@>=
begin t:=rlink(q);
@^inner loop@>
if q=rover then rover:=t;
llink(t):=llink(q); rlink(llink(q)):=t;@/
q:=q+node_size(q);
end
@y
@ The lower part of |mem| grows by 1000 words at a time, unless
we are very close to going under. When it grows, we simply link
a new node into the available-space list. This method of controlled
growth helps to keep the |mem| usage consecutive when \TeX\ is
implemented on ``virtual memory'' systems.
@^virtual memory@>

@<Grow more variable-size memory and |goto restart|@>=
begin if lo_mem_max+1000<hi_mem_min then t:=lo_mem_max+1000
else t:=(lo_mem_max+hi_mem_min+2) div 2; {|lo_mem_max+2<=t<hi_mem_min|}
p:=llink(rover); q:=lo_mem_max; rlink(p):=q; llink(rover):=q;@/
rlink(q):=rover; llink(q):=p; link(q):=empty_flag; node_size(q):=t-lo_mem_max;@/
lo_mem_max:=t; link(lo_mem_max):=null; info(lo_mem_max):=null;
rover:=q; goto restart;
end

@ Empirical tests show that the routine in this section performs a
node-merging operation about 0.75 times per allocation, on the average,
after which it finds that |r>p+1| about 95\% of the time.

@<Try to allocate...@>=
q:=p+node_size(p); {find the physical successor}
@^inner loop@>
while is_empty(q) do {merge node |p| with node |q|}
  begin t:=rlink(q);
  if q=rover then rover:=t;
  llink(t):=llink(q); rlink(llink(q)):=t;@/
  q:=q+node_size(q);
  end;
r:=q-s;
if r>p+1 then @<Allocate from the top of node |p| and |goto found|@>;
if r=p then if ((rlink(p)<>rover) or (llink(p)<>rover)) then
  @<Allocate entire node |p| and |goto found|@>;
node_size(p):=q-p {reset the size in case it grew}
@z
% module 129 no longer refers to remove_node; I don't recall when it changed!
@x module 134
@d is_char_node(#) == (#>hi_mem_base)
@y
@d is_char_node(#) == (#>=hi_mem_min)
@z
@x module 162
example, locations |mem_base| to |mem_base+3| are always used to store the
specification for glue that is `\.{0pt plus 0pt minus 0pt}'. The
following macro definitions accomplish the static allocation by giving
symbolic names to the fixed positions. Dynamic allocation of variable-size
nodes is restricted to locations |first_mem| through |(hi_mem_base-1)|,
and single-word nodes are dynamically allocated in locations |second_mem|
through |mem_max|, inclusive. It is harmless to let |lig_trick|, |garbage|,
and |backup_head| share the same location of |mem|.

@d zero_glue==mem_base {specification for \.{0pt plus 0pt minus 0pt}}
@y
example, locations |mem_bot| to |mem_bot+3| are always used to store the
specification for glue that is `\.{0pt plus 0pt minus 0pt}'. The
following macro definitions accomplish the static allocation by giving
symbolic names to the fixed positions. Static variable-size nodes appear
in locations |mem_bot| through |lo_mem_stat_max|, and static single-word nodes
appear in locations |hi_mem_stat_min| through |mem_top|, inclusive. It is
harmless to let |lig_trick|, |garbage|, and |backup_head| share the same
location of |mem|.

@d zero_glue==mem_bot {specification for \.{0pt plus 0pt minus 0pt}}
@z
@x module 162
@d first_mem==fil_neg_glue+glue_spec_size {first dynamically allocatable word
  in the variable-size |mem|}
@#
@d page_ins_head==hi_mem_base {list of insertion data for current page}
@d contrib_head==hi_mem_base+1 {vlist of items not yet on current page}
@d page_head==hi_mem_base+2 {vlist for current page}
@d temp_head==hi_mem_base+3 {head of a temporary list of some kind}
@d hold_head==hi_mem_base+4 {head of a temporary list of another kind}
@d adjust_head==hi_mem_base+5 {head of adjustment list returned by |hpack|}
@d active==hi_mem_base+6 {head of active list in |line_break|, needs two words}
@d align_head==hi_mem_base+8 {head of preamble list for alignments}
@d end_span==hi_mem_base+9 {tail of spanned-width lists}
@d omit_template==hi_mem_base+10 {a constant token list}
@d null_list==hi_mem_base+11 {permanently empty list}
@d lig_trick==hi_mem_base+12 {a ligature masquerading as a |char_node|}
@d garbage==hi_mem_base+12 {used for scrap information}
@d backup_head==hi_mem_base+13 {head of token list built by |scan_keyword|}
@d second_mem==hi_mem_base+14 {first dynamically allocatable word in
  the one-word |mem|}
@y
@d lo_mem_stat_max==fil_neg_glue+glue_spec_size-1 {largest statically
  allocated word in the variable-size |mem|}
@#
@d page_ins_head==mem_top {list of insertion data for current page}
@d contrib_head==mem_top-1 {vlist of items not yet on current page}
@d page_head==mem_top-2 {vlist for current page}
@d temp_head==mem_top-3 {head of a temporary list of some kind}
@d hold_head==mem_top-4 {head of a temporary list of another kind}
@d adjust_head==mem_top-5 {head of adjustment list returned by |hpack|}
@d active==mem_top-6 {head of active list in |line_break|, needs two words}
@d align_head==mem_top-8 {head of preamble list for alignments}
@d end_span==mem_top-9 {tail of spanned-width lists}
@d omit_template==mem_top-10 {a constant token list}
@d null_list==mem_top-11 {permanently empty list}
@d lig_trick==mem_top-12 {a ligature masquerading as a |char_node|}
@d garbage==mem_top-12 {used for scrap information}
@d backup_head==mem_top-13 {head of token list built by |scan_keyword|}
@d hi_mem_stat_min==mem_top-13 {smallest statically allocated word in
  the one-word |mem|}
@z
@x module 164
for k:=mem_base+1 to first_mem-1 do mem[k].sc:=0;
  {all glue dimensions are zeroed}
@^data structure assumptions@>
k:=mem_base;@+while k<first_mem do {set first words of glue specifications}
@y
for k:=mem_bot+1 to lo_mem_stat_max do mem[k].sc:=0;
  {all glue dimensions are zeroed}
@^data structure assumptions@>
k:=mem_bot;@+while k<=lo_mem_stat_max do
    {set first words of glue specifications}
@z
@x module 164
rover:=first_mem; link(rover):=empty_flag; {now initialize the dynamic memory}
node_size(rover):=hi_mem_base-rover; {which is one big available node}
llink(rover):=rover; rlink(rover):=rover;@/
link(hi_mem_base):=null; info(hi_mem_base):=null;
for k:=hi_mem_base+1 to second_mem-1 do
  mem[k]:=mem[hi_mem_base];{clear list heads}
@<Initialize the special list heads and constant nodes@>;
avail:=null; mem_end:=second_mem-1; {initialize the one-word memory}
var_used:=first_mem-mem_base; dyn_used:=second_mem-hi_mem_base;
max_var_used:=var_used; {initialize statistics}
@y
rover:=lo_mem_stat_max+1;
link(rover):=empty_flag; {now initialize the dynamic memory}
node_size(rover):=1000; {which is a 1000-word available node}
llink(rover):=rover; rlink(rover):=rover;@/
lo_mem_max:=rover+1000; link(lo_mem_max):=null; info(lo_mem_max):=null;@/
for k:=hi_mem_stat_min to mem_top do
  mem[k]:=mem[lo_mem_max]; {clear list heads}
@<Initialize the special list heads and constant nodes@>;
avail:=null; mem_end:=mem_top;
hi_mem_min:=hi_mem_stat_min; {initialize the one-word memory}
var_used:=lo_mem_stat_max+1-mem_bot; dyn_used:=mem_top+1-hi_mem_stat_min;
  {initialize statistics}
@z
@x module 165
@t\hskip1em@>@!was_mem_end: pointer; {previous |mem_end|}
@y
@t\hskip1em@>@!was_mem_end,@!was_lo_max,@!was_hi_min: pointer;
  {previous |mem_end|, |lo_mem_max|,and |hi_mem_min|}
@z
@x module 166
@!debug was_mem_end:=mem_base; {indicate that everything was previously free}
@y
@!debug was_mem_end:=mem_min; {indicate that everything was previously free}
was_lo_max:=mem_min; was_hi_min:=mem_max;
@z
@x module 167
begin for p:=mem_base to mem_end do free[p]:=false; {you can probably
  do this faster}
@y
begin for p:=mem_min to lo_mem_max do free[p]:=false; {you can probably
  do this faster}
for p:=hi_mem_min to mem_end do free[p]:=false; {ditto}
@z
@x module 167
for p:=0 to mem_end do was_free[p]:=free[p]; {|was_free:=free| might be faster}
was_mem_end:=mem_end;
@y
for p:=mem_min to lo_mem_max do was_free[p]:=free[p];
for p:=hi_mem_min to mem_end do was_free[p]:=free[p];
  {|was_free:=free| might be faster}
was_mem_end:=mem_end; was_lo_max:=lo_mem_max; was_hi_min:=hi_mem_min;
@z
@x module 168
  begin if (p>mem_end)or(p<second_mem) then clobbered:=true
@y
  begin if (p>mem_end)or(p<hi_mem_min) then clobbered:=true
@z
@x module 169
repeat if (p>=hi_mem_base)or(p<first_mem) then clobbered:=true
  else if (rlink(p)>=hi_mem_base)or(rlink(p)<first_mem) then clobbered:=true
  else if  not(is_empty(p))or(node_size(p)<2)or@|
   (p+node_size(p)>hi_mem_base)or@| (llink(rlink(p))<>p) then clobbered:=true;
@y
repeat if (p>=lo_mem_max)or(p<mem_min) then clobbered:=true
  else if (rlink(p)>=lo_mem_max)or(rlink(p)<mem_min) then clobbered:=true
  else if  not(is_empty(p))or(node_size(p)<2)or@|
   (p+node_size(p)>lo_mem_max)or@| (llink(rlink(p))<>p) then clobbered:=true;
@z
@x module 170
p:=mem_base;
while p<=hi_mem_base do {node |p| should not be empty}
@y
p:=mem_min;
while p<=lo_mem_max do {node |p| should not be empty}
@z
@x module 170
  while (p<=hi_mem_base) and not free[p] do incr(p);
  while (p<=hi_mem_base) and free[p] do incr(p);
@y
  while (p<=lo_mem_max) and not free[p] do incr(p);
  while (p<=lo_mem_max) and free[p] do incr(p);
@z
@x module 171
for p:=mem_base to mem_end do
  if not free[p] and ((p>was_mem_end) or was_free[p]) then
@y
for p:=mem_min to lo_mem_max do
  if not free[p] and ((p>was_lo_max) or was_free[p]) then
    begin print_char(" "); print_int(p);
    end;
for p:=hi_mem_min to mem_end do
  if not free[p] and
   ((p<was_hi_min) or (p>was_mem_end) or was_free[p]) then
@z
@x module 172
begin for q:=mem_base to mem_end do
@y
begin for q:=mem_min to lo_mem_max do
@z
@x module 172
@<Search |eqtb| for equivalents equal to |p|@>;
@y
for q:=hi_mem_min to mem_end do
  begin if link(q)=p then
    begin print_nl("LINK("); print_int(q); print_char(")");
    end;
  if info(q)=p then
    begin print_nl("INFO("); print_int(q); print_char(")");
    end;
  end;
@<Search |eqtb| for equivalents equal to |p|@>;
@z
@x module 176
if (p<hi_mem_base)or(p>mem_end) then print_esc("CLOBBERED.")
@y
if (p<hi_mem_min)or(p>mem_end) then print_esc("CLOBBERED.")
@z
@x module 178
begin if (p<mem_base)or(p>=hi_mem_base) then print_char("*")
@y
begin if (p<mem_min)or(p>=lo_mem_max) then print_char("*")
@z
@x module 293
if (p<hi_mem_base) or (p>mem_end) then
@y
if (p<hi_mem_min) or (p>mem_end) then
@z
@x module 639
  print_int(dyn_used); print("; max so far: ");
  print_int(max_var_used); print_char("&");
  print_int(mem_end+1-hi_mem_base); print_ln;
@y
  print_int(dyn_used); print("; still untouched: ");
  print_int(hi_mem_min-lo_mem_max-1); print_ln;
@z
@x module 857
begin save_link:=link(cur_p); {this is OK even if |cur_p=null|}
link(cur_p):=null; print_nl(""); short_display(link(printed_node));
link(cur_p):=save_link; printed_node:=cur_p;
@y
begin print_nl("");
if cur_p=null then short_display(link(printed_node))
else  begin save_link:=link(cur_p);
  link(cur_p):=null; print_nl(""); short_display(link(printed_node));
  link(cur_p):=save_link;
  end;
printed_node:=cur_p;
@z
@x module 1249
if 2*max_halfword<hi_mem_base-mem_base then bad:=41;
@y
if 2*max_halfword<mem_top-mem_min then bad:=41;
@z
@x module 1307
dump_int(hi_mem_base);@/
@y
dump_int(mem_bot);@/
dump_int(mem_top);@/
@z
@x module 1308
if x<>hi_mem_base then goto bad_fmt;
@y
if x<>mem_bot then goto bad_fmt;
undump_int(x);
if x<>mem_top then goto bad_fmt;
@z
@x module 1311
dump_int(rover);
p:=mem_base; q:=rover; x:=0;
@y
dump_int(lo_mem_max); dump_int(rover);
p:=mem_bot; q:=rover; x:=0;
@z
@x module 1311
var_used:=var_used+hi_mem_base-p; dyn_used:=mem_end+1-hi_mem_base;@/
dump_int(mem_end); dump_int(avail);
for k:=p to mem_end do dump_wd(mem[k]);
x:=x+mem_end+1-p;
@y
var_used:=var_used+lo_mem_max-p; dyn_used:=mem_end+1-hi_mem_min;@/
for k:=p to lo_mem_max do dump_wd(mem[k]);
x:=x+lo_mem_max+1-p;
dump_int(hi_mem_min); dump_int(avail);
for k:=hi_mem_min to mem_end do dump_wd(mem[k]);
x:=x+mem_end+1-hi_mem_min;
@z
@x module 1312
undump(min_halfword)(hi_mem_base)(rover);
p:=mem_base; q:=rover; x:=0;
repeat for k:=p to q+1 do undump_wd(mem[k]);
p:=q+node_size(q);
if (p>hi_mem_base)or((q>=rlink(q))and(rlink(q)<>rover)) then goto bad_fmt;
q:=rlink(q);
until q=rover;
undump_size(hi_mem_base)(mem_max)('mem max')(mem_end);
undump(null)(mem_end)(avail);
for k:=p to mem_end do undump_wd(mem[k]);
undump_int(var_used); undump_int(dyn_used);
max_var_used:=var_used
@y
undump(lo_mem_stat_max+1000)(hi_mem_stat_min-1)(lo_mem_max);
undump(lo_mem_stat_max+1)(lo_mem_max)(rover);
p:=mem_bot; q:=rover; x:=0;
repeat for k:=p to q+1 do undump_wd(mem[k]);
p:=q+node_size(q);
if (p>lo_mem_max)or((q>=rlink(q))and(rlink(q)<>rover)) then goto bad_fmt;
q:=rlink(q);
until q=rover;
for k:=p to lo_mem_max do undump_wd(mem[k]);
if mem_min<mem_bot-2 then {make more low memory available}
  begin p:=llink(rover); q:=mem_min+1;
  link(mem_min):=null; info(mem_min):=null; {we don't use the bottom word}
  rlink(p):=q; llink(rover):=q;@/
  rlink(q):=rover; llink(q):=p; link(q):=empty_flag;
  node_size(q):=mem_bot-q;
  end;
undump(lo_mem_max+1)(hi_mem_stat_min)(hi_mem_min);
undump(null)(mem_top)(avail); mem_end:=mem_top;
for k:=hi_mem_min to mem_end do undump_wd(mem[k]);
undump_int(var_used); undump_int(dyn_used)
@z
@x module 1323
undump(min_halfword)(hi_mem_base)(font_glue[k]);
@y
undump(min_halfword)(lo_mem_max)(font_glue[k]);
@z
@x module 1334
  wlog_ln(' ',max_var_used:1,'&',mem_end+1-hi_mem_base:1,@|
    ' words of memory out of ',
    hi_mem_base-mem_base:1,'&',mem_max+1-hi_mem_base:1);@/
@y
  wlog_ln(' ',lo_mem_max-mem_min+mem_end-hi_mem_min+2:1,@|
    ' words of memory out of ',mem_end+1-mem_min:1);@/
@z that's the end of change number 300!

301. node_size field could overflow if low mem was too large (Dec 20, 1984)
[This bug existed from the beginning, if hi_mem_base-first_mem>max_halfword!]
@x module 111
if (mem_min<min_halfword)or(mem_max>=max_halfword) then bad:=14;
@y
if (mem_min<min_halfword)or(mem_max>=max_halfword)or@|
  (mem_bot-mem_min>max_halfword+1) then bad:=14;
@z
@x module 125
@!t:pointer; {temporary register}
@y
@!t:integer; {temporary register}
@z
@x module 125
if lo_mem_max+2<hi_mem_min then
@y
if lo_mem_max+2<hi_mem_min then if lo_mem_max+2<=mem_bot+max_halfword then
@z
@x module 126
rlink(q):=rover; llink(q):=p; link(q):=empty_flag; node_size(q):=t-lo_mem_max;@/
@y
if t>mem_bot+max_halfword then t:=mem_bot+max_halfword;
rlink(q):=rover; llink(q):=p; link(q):=empty_flag; node_size(q):=t-lo_mem_max;@/
@z

302. Improvement to missing-format-file error (DRF, 5 Jan 85)
@x module 524
var j:0..buf_size; {the first space after the file name}
begin if buffer[loc]="&" then
  begin incr(loc); j:=loc; buffer[last]:=" ";
  while buffer[j]<>" " do incr(j);
  pack_buffered_name(0,loc,j-1); {try first without the system file area}
  if w_open_in(fmt_file) then
    begin loc:=j; goto found;
    end;@/
  pack_buffered_name(format_area_length,loc,j-1);
    {now try the system format file area}
  if w_open_in(fmt_file) then
    begin loc:=j; goto found;
    end;
  wake_up_terminal;
  wterm_ln('Sorry, I can''t find that format;',' will try PLAIN.');
@.Sorry, I can't find...@>
  end;
  {now pull out all the stops: try for the system \.{PLAIN} file}
pack_buffered_name(format_default_length-format_ext_length,1,0);
if not w_open_in(fmt_file) then
  begin wake_up_terminal;
  wterm_ln('I can''t find the PLAIN format file!');
@.I can't find PLAIN...@>
@.PLAIN@>
  open_fmt_file:=false; return;
  end;
found:open_fmt_file:=true;
@y
var j:0..buf_size; {the first space after the format file name}
begin j:=loc;
if buffer[loc]="&" then
  begin incr(loc); j:=loc; buffer[last]:=" ";
  while buffer[j]<>" " do incr(j);
  pack_buffered_name(0,loc,j-1); {try first without the system file area}
  if w_open_in(fmt_file) then goto found;
  pack_buffered_name(format_area_length,loc,j-1);
    {now try the system format file area}
  if w_open_in(fmt_file) then goto found;
  wake_up_terminal;
  wterm_ln('Sorry, I can''t find that format;',' will try PLAIN.');
@.Sorry, I can't find...@>
  update_terminal;
  end;
  {now pull out all the stops: try for the system \.{PLAIN} file}
pack_buffered_name(format_default_length-format_ext_length,1,0);
if not w_open_in(fmt_file) then
  begin wake_up_terminal;
  wterm_ln('I can''t find the PLAIN format file!');
@.I can't find PLAIN...@>
@.PLAIN@>
  open_fmt_file:=false; return;
  end;
found:loc:=j; open_fmt_file:=true;
@z

303. Welcoming message appears sooner (DRF, Jan 7)
@x module 61
else  begin print(format_ident); print_ln;
  end;
@y
else  begin print(format_ident); print_ln;
  end;
update_terminal;
@z

304. Minor change to help message in "confusion" (Jan 23)
@x module 95
  help2("One of your earlier faux pas has wounded me deeply,")@/
    ("so I'm barely conscious. Please fix it and try again.");
@y
  help2("One of your faux pas seems to have wounded me deeply...")@/
    ("in fact, I'm barely conscious. Please fix it and try again.");
@z

305. Improved logic for `(see the transcript file...)' (Jan 23)
@x module 245
if history=spotless then history:=warning_issued;
if (tracing_online<=0)and(selector=term_and_log) then decr(selector);
@y
if (tracing_online<=0)and(selector=term_and_log) then
  begin decr(selector);
  if history=spotless then history:=warning_issued;
  end;
@z

306. Change 291 violated standard PASCAL (CET1, Feb 18)
@x module 331
for first:=0 to buf_size do buffer[first]:=0;
@y since some procedures `threaten' |first| globally [BS/6192 sec 6.8.3.9]
first:=buf_size; repeat buffer[first]:=0; decr(first); until first=0;
@z

307. Nonexistent character could be output via ligature hyphenation (April 11)
@x module 582 (slight inefficiency removed while correcting this)
begin if (font_bc[f]<=c)and(font_ec[f]>=c) then
@y
begin if font_bc[f]<=c then if font_ec[f]>=c then
@z
@x module 892 (likewise)
@!hu:array[1..63] of ASCII_code; {like |hc|, before conversion to lowercase}
@y
@!hu:array[1..63] of ASCII_code; {like |hc|, before conversion to lowercase}
@!hyf_char:integer; {hyphen character of the relevant font}
@z
@x module 896 (likewise)
done2: if hyphen_char[hf]<0 then goto done1;
if hyphen_char[hf]>255 then goto done1;
@y
done2: hyf_char:=hyphen_char[hf];
if hyf_char<0 then goto done1;
if hyf_char>255 then goto done1;
@z
@x module 912 and another local variable
@!c:ASCII_code; {character temporarily replaced by a hyphen}
@y
@!c:ASCII_code; {character temporarily replaced by a hyphen}
@!hyf_node:pointer; {the hyphen, if it exists}
@z
@x module 915 (this is the real bugfix)
minor_tail:=null; c:=hu[i+1]; hu[i+1]:=hyphen_char[hf];
repeat l:=reconstitute(l+1,i+1);
if minor_tail=null then pre_break(r):=link(hold_head)
else link(minor_tail):=link(hold_head);
minor_tail:=link(hold_head);
if link(minor_tail)<>null then minor_tail:=link(minor_tail);
until l>i;
hu[i+1]:=c; {restore the character in the hyphen position}
decr(l); hyf[l]:=0
@y
minor_tail:=null; hyf_node:=new_character(hf,hyf_char);
if hyf_node<>null then
  begin incr(i); c:=hu[i]; hu[i]:=hyf_char;
  end;
repeat l:=reconstitute(l+1,i);
if minor_tail=null then pre_break(r):=link(hold_head)
else link(minor_tail):=link(hold_head);
minor_tail:=link(hold_head);
if link(minor_tail)<>null then minor_tail:=link(minor_tail);
until l=i;
if hyf_node<>null then
  begin hu[i]:=c; {restore the character in the hyphen position}
  free_avail(hyf_node); decr(i); l:=i;
  end;
hyf[i]:=0
@z
@x module 918
begin r:=new_disc; pre_break(r):=new_character(hf,hyphen_char[hf]);
@y
begin r:=new_disc; pre_break(r):=new_character(hf,hyf_char);
@z

308. Improper memory usage computed by INITEX in change 300. (DRF, April 15)
@x module 162
@d hi_mem_stat_min==mem_top-13 {smallest statically allocated word in
  the one-word |mem|}
@y
@d hi_mem_stat_min==mem_top-13 {smallest statically allocated word in
  the one-word |mem|}
@d hi_mem_stat_usage=14 {the number of one-word nodes always present}
@z
@x module 164 (text macro without parentheses strikes again!)
var_used:=lo_mem_stat_max+1-mem_bot; dyn_used:=mem_top+1-hi_mem_stat_min;
@y
var_used:=lo_mem_stat_max+1-mem_bot; dyn_used:=hi_mem_stat_usage;
@z

309. A faster |flush_list| routine for the inner loop. (DRF, April 16)
@x module 123
var q:pointer; {the successor of node |p|}
begin while p<>null do
  begin q:=link(p); free_avail(p); p:=q;
  end;
@y
var @!q,@!r:pointer; {list traversers}
begin if p<>null then
  begin r:=p;
  repeat q:=r; r:=link(r); @!stat decr(dyn_used);@+tats@/
  until r=null; {now |q| is the last node on the list}
  link(q):=avail; avail:=p;
  end;
@z

310. Kern for accent positioning could disappear after a line break (April 17)
(Change 269 fixed only part of the problem!)
@x module 155
@d explicit=1 {|subtype| of kern nodes from \.{\\kern} and \.{\\/} and accents}
@y
@d explicit=1 {|subtype| of kern nodes from \.{\\kern} and \.{\\/}}
@d acc_kern=2 {|subtype| of kern nodes from accents}
@z
@x module 191
if subtype(p)<=explicit then
  begin print_esc("kern");
  if subtype(p)=explicit then print_char(" ");
  print_scaled(width(p));
@y
if subtype(p)<>mu_glue then
  begin print_esc("kern");
  if subtype(p)<>normal then print_char(" ");
  print_scaled(width(p));
  if subtype(p)=acc_kern then print(" (for accent)");
@.for accent@>
@z
@x module 837
    math_node,kern_node: break_width[1]:=break_width[1]-width(s);
@y
    math_node,kern_node: if subtype(s)=acc_kern then goto done
      else break_width[1]:=break_width[1]-width(s);
@z
@x module 879
  if non_discardable(q) then goto done1;
@y
  if non_discardable(q) then goto done1;
  if subtype(q)=acc_kern then if type(q)=kern_node then goto done1;
@z
@x module 1125
r:=new_kern(delta); subtype(r):=explicit; link(tail):=r; link(r):=p;
tail:=new_kern(-a-delta); subtype(tail):=explicit; link(p):=tail; p:=q;
@y
r:=new_kern(delta); subtype(r):=acc_kern; link(tail):=r; link(r):=p;
tail:=new_kern(-a-delta); subtype(tail):=acc_kern; link(p):=tail; p:=q;
@z

311. \lastbox and \unkern must not remove discretionary replacements (April 18)
(Change 236 introduced this bug!)
@x module 1079
label exit;
var p:pointer; {runs through the current list}
@y
label exit, done;
var @!p,@!q:pointer; {run through the current list}
@!m:quarterword; {the length of a replacement list}
@z
@x modules 1080 and 1081
  @<Apologize for inability to do \.{\\lastbox} now@>
else  begin if not is_char_node(tail) then
    if (type(tail)=hlist_node)or(type(tail)=vlist_node) then
      begin p:=head; cur_box:=tail; shift_amount(cur_box):=0;
      while link(p)<>tail do p:=link(p);
      tail:=p; link(p):=null;
      end;
  end;
end

@ @<Apologize for inability to do \.{\\lastbox} now@>=
begin you_cant;
help2("Sorry...I'm usually unable to take things from the current")@/
  ("page. This \lastbox will therefore be void."); error;
end
@y
  begin you_cant;
  help2("Sorry...I usually can't take things from the current page.")@/
    ("This \lastbox will therefore be void."); error;
  end
else  begin if not is_char_node(tail) then
    if (type(tail)=hlist_node)or(type(tail)=vlist_node) then
      @<Remove the last box, unless it's part of a discretionary@>;
  end;
end

@ @<Remove the last box...@>=
begin q:=head;
repeat p:=q;
if not is_char_node(q) then if type(q)=disc_node then
  begin for m:=1 to replace_count(q) do p:=link(p);
  if p=tail then goto done;
  end;
q:=link(p);
until q=tail;
cur_box:=tail; shift_amount(cur_box):=0;
tail:=p; link(p):=null;
done: end
@z
@x module 1105
var p:pointer; {runs through the current list}
@y
label exit;
var @!p,@!q:pointer; {run through the current list}
@!m:quarterword; {the length of a replacement list}
@z
@x module 1105
    begin p:=head;
    while link(p)<>tail do p:=link(p);
    link(p):=null; flush_node_list(tail); tail:=p;
    end;
  end;
end;
@y
    begin q:=head;
    repeat p:=q;
    if not is_char_node(q) then if type(q)=disc_node then
      begin for m:=1 to replace_count(q) do p:=link(p);
      if p=tail then return;
      end;
    q:=link(p);
    until q=tail;
    link(p):=null; flush_node_list(tail); tail:=p;
    end;
  end;
exit:end;
@z
@x module 1106
  help2("Sorry...I'm usually unable to take things from the current")@/
    ("page. Try `I\vskip-\lastskip' instead.");
  if cur_chr=kern_node then help_line[0]:=
    ("page. Try `I\kern-\lastkern' instead.")
  else if cur_chr<>glue_node then help_line[0]:=@|
    ("page. Perhaps you can make the output routine do it.");
@y
  help2("Sorry...I usually can't take things from the current page.")@/
    ("Try `I\vskip-\lastskip' instead.");
  if cur_chr=kern_node then help_line[0]:=
    ("Try `I\kern-\lastkern' instead.")
  else if cur_chr<>glue_node then help_line[0]:=@|
    ("Perhaps you can make the output routine do it.");
@z
@x module 1120
if n<128 then replace_count(tail):=n
@y
if n<=max_quarterword then replace_count(tail):=n
@z


*** Note by DEK, 18 Apr 1985
(The code for version 1.3 has been published by Addison-Wesley under the
title "TeX: The Program". It was phototypeset on our APS, but does not
contain changes > 300. Some of those changes have been made in versions
of TeX called 1.3; but the exact definition of version 1.3 is somewhat
vague. Version 1.4 was released on April 18, 1985, after
making change 311. At present I believe Version 1.4 will be identical
to Version 2.0 except perhaps for things like index entries in the
documentation and slightly revised help messages. Version 2.0 will
coincide with the publication of Volume 1 of Computers and Typesetting,
which will come out when METAFONT and Computer Modern are both finished.)
Version 1.5
312. if nonstandard area specified, don't try TEX_area too. (DRF, 26 Apr 85)
@x module 537
  pack_file_name(cur_name,TEX_area,cur_ext);
  if a_open_in(cur_file) then goto done;
@y
  if cur_area="" then
    begin pack_file_name(cur_name,TEX_area,cur_ext);
    if a_open_in(cur_file) then goto done;
    end;
@z

313. length of \write need no longer be limited (cf 235). (NBT, 30 Apr 85)
@x module 1370
show_token_list(link(def_ref),null,buf_size-10); print_ln;
@y
show_token_list(link(def_ref),null,10000000); print_ln;
@z

314. bug introduced in #300, just lucky it didn't cause trouble (CET1, 8 May 85)
@x module 162
@d active==mem_top-6 {head of active list in |line_break|, needs two words}
@y
@d active==mem_top-7 {head of active list in |line_break|, needs two words}
@z

315. Missing eoln after last typeout (Gropp, 11 May 85)
@x module 1332
  wterm('Ouch---my internal constants have been',
@y
  wterm_ln('Ouch---my internal constants have been',
@z

316. Redundancy (CET1, 17 May 85)
[Change 287 was apparently introduced because of a faulty assumption,
although its change to the calling of "initialize" was valid.]
@x module 1332
  begin t_open_out;
  wterm_ln('Ouch---my internal constants have been',
    ' clobbered!---case ',bad:1);
@y
  begin wterm_ln('Ouch---my internal constants have been clobbered!',
    '---case ',bad:1);
@z

317. batchmode to see if log file couldn't be opened (DRF, 22 May 85)
@x module 92
if interaction=batch_mode then decr(selector);
if job_name=0 then open_log_file;
@y
if job_name=0 then open_log_file;
if interaction=batch_mode then decr(selector);
@z

318. If the string pool overflows while reading the command line... (22 May 85)
[This bug was first found in METAFONT, when it could occur more easily]
@x module 525
begin str_room(name_length);
for k:=1 to name_length do append_char(xord[name_of_file[k]]);
make_name_string:=make_string;
@y
begin if (pool_ptr+name_length>pool_size)or(str_ptr=max_strings) then
  make_name_string:="?"
else  begin for k:=1 to name_length do append_char(xord[name_of_file[k]]);
  make_name_string:=make_string;
  end;
@z

319. Bug in \edef\foo{\iffalse\fi\the\toks0} (found by Dan Brotsky, 7 Aug 85)
@x module 478
  if cur_cmd<>the then goto done2;
  q:=the_toks;
  if link(temp_head)<>null then
    begin link(p):=link(temp_head); p:=q;
    end;
@y
  if cur_cmd<=max_command then goto done2;
  if cur_cmd<>the then expand
  else  begin q:=the_toks;
    if link(temp_head)<>null then
      begin link(p):=link(temp_head); p:=q;
      end;
    end;
@z

*** Version 1.5 was released on August 7, 1985.

320. Lowercase `plain' for consistency with the manual
@x module 521
TEX_format_default:='TeXformats:PLAIN.fmt';
@.TeXformats@>
@.PLAIN@>
@y
TEX_format_default:='TeXformats:plain.fmt';
@.TeXformats@>
@.plain@>
@z (index entries also changed to lowercase in modules 524, 1331)

321. Terminal woken up by \show and \showthe (November 27, 1985)
@x module 1294
begin get_token; print_nl("> ");
@y
begin get_token;
if interaction=error_stop_mode then wake_up_terminal;
print_nl("> ");
@z
@x module 1297
begin p:=the_toks; print_nl("> "); token_show(temp_head);
@y
begin p:=the_toks;
if interaction=error_stop_mode then wake_up_terminal;
print_nl("> "); token_show(temp_head);
@z
Changes subsequent to `Version 2.0' as published in C&T, Volume B:

322. Trivial change to a help message.
@x module 1283
  ("Pretend that you're Hercule Poirot, examine all clues,")@/
@y
  ("Pretend that you're Hercule Poirot: Examine all clues,")@/
@z

323. Precaution since |index| can be as high as |max_in_open|. (DRF)
@x module 14
if hash_prime>hash_size then bad:=5;
@y
if hash_prime>hash_size then bad:=5;
if max_in_open>=128 then bad:=6;
@z

324. \kern clobbered at end of pre-break list in discretionary break.
@x module 881
        |disc_break:=true|@>;
    if not is_char_node(q) then
      if (type(q)=math_node)or(type(q)=kern_node) then width(q):=0;
@y
        |disc_break:=true|@>
    else if (type(q)=math_node)or(type(q)=kern_node) then width(q):=0;
@z

325. width after discretionary didn't take account of discarded nodes.
@x module 830 (slight redefinition)
@!t:quarterword; {replacement count, if |cur_p| is a discretionary node}
@y
@!t:integer; {node count, if |cur_p| is a discretionary node}
@z
@x module 837 (slight rearrangement)
if (break_type=unhyphenated)or(cur_p=null) then
  begin s:=cur_p;
  while s<>null do
    begin if is_char_node(s) then goto done;
    case type(s) of
    glue_node:@<Subtract glue from |break_width|@>;
    penalty_node: do_nothing;
    math_node,kern_node: if subtype(s)=acc_kern then goto done
      else break_width[1]:=break_width[1]-width(s);
    othercases goto done
    endcases;@/
    s:=link(s);
    end;
  end
else @<Compute the discretionary |break_width| values@>;
@y
s:=cur_p;
if break_type>unhyphenated then if cur_p<>null then
  @<Compute the discretionary |break_width| values@>;
while s<>null do
  begin if is_char_node(s) then goto done;
  case type(s) of
  glue_node:@<Subtract glue from |break_width|@>;
  penalty_node: do_nothing;
  math_node,kern_node: if subtype(s)=acc_kern then goto done
    else break_width[1]:=break_width[1]-width(s);
  othercases goto done
  endcases;@/
  s:=link(s);
  end;
@z
@x module 840 (this is the main change to fix the bug)
begin t:=replace_count(cur_p); s:=cur_p;
while t>0 do
  begin decr(t); s:=link(s);
  @<Subtract the width of node |s| from |break_width|@>;
  end;
s:=post_break(cur_p);
while s<>null do
  begin @<Add the width of node |s| to |break_width|@>;
@y
begin t:=replace_count(cur_p); v:=cur_p; s:=post_break(cur_p);
while t>0 do
  begin decr(t); v:=link(v);
  @<Subtract the width of node |v| from |break_width|@>;
  end;
while s<>null do
  begin @<Add the width of node |s| to |break_width| and increase |t|,
    unless it's discardable@>;
@z
@x module 840, continued
break_width[1]:=break_width[1]+disc_width;
@y
break_width[1]:=break_width[1]+disc_width;
if t=0 then s:=link(v); {more nodes may also be discardable after the break}
@z
@x module 841 (here we just change |s| to |v|)
@<Subtract the width of node |s|...@>=
if is_char_node(s) then
  begin f:=font(s);
  break_width[1]:=break_width[1]-char_width(f)(char_info(f)(character(s)));
  end
else  case type(s) of
  ligature_node: begin f:=font(lig_char(s));@/
    break_width[1]:=@|break_width[1]-
      char_width(f)(char_info(f)(character(lig_char(s))));
    end;
  hlist_node,vlist_node,rule_node,kern_node:
    break_width[1]:=break_width[1]-width(s);
@y
@<Subtract the width of node |v|...@>=
if is_char_node(v) then
  begin f:=font(v);
  break_width[1]:=break_width[1]-char_width(f)(char_info(f)(character(v)));
  end
else  case type(v) of
  ligature_node: begin f:=font(lig_char(v));@/
    break_width[1]:=@|break_width[1]-
      char_width(f)(char_info(f)(character(lig_char(v))));
    end;
  hlist_node,vlist_node,rule_node,kern_node:
    break_width[1]:=break_width[1]-width(v);
@z
@x module 842
  hlist_node,vlist_node,rule_node,kern_node:
    break_width[1]:=break_width[1]+width(s);
  othercases confusion("disc2")
@:this can't happen disc2}{\quad disc2@>
  endcases
@y
  hlist_node,vlist_node,rule_node:break_width[1]:=break_width[1]+width(s);
  kern_node: if (t=0)and(subtype(s)<>acc_kern) then t:=-1 {discardable}
    else break_width[1]:=break_width[1]+width(s);
  othercases confusion("disc2")
@:this can't happen disc2}{\quad disc2@>
  endcases;
incr(t)
@z

*** Version 2.1 was released on January 26, 1987.

326. Removal of redundant code (found by Pat Monardo, 5 Feb 87)
@x module 1224 [this change need not be made, since it has no effect]
    char_def_code: define(p,char_given,cur_val);
    math_char_def_code: define(p,math_given,cur_val);
@y the cases cannot occur, so we simply don't list them
@z

327. More robustness is needed when debugging (Ronaldo Am\'a, 14 Apr 87)
@x module 174
begin while p>null do
@y maybe mem_min>null (but nodes shouldn't be put in location mem_min exactly)
begin while p>mem_min do
@z
@x module 182
@p procedure show_node_list(@!p:pointer); {prints a node list symbolically}
@y
@p procedure show_node_list(@!p:integer); {prints a node list symbolically}
@z
@x module 182, continued
while p>null do
@y
while p>mem_min do
@z

328. Storage allocation can be more elegant and efficient (4/21/87)
@x module 127
if r=p then if ((rlink(p)<>rover) or (llink(p)<>rover)) then
@y
if r=p then if rlink(p)<>p then
@z

329. Miscalculation of empty-line condition (April 22, 1987)
@x module 360
    begin if limit=start then {previous line was empty}
@y
    begin if (end_line_char<0)or(end_line_char>127) then incr(limit);
    if limit=start then {previous line was empty}
@z

330. A case where we don't have to assume system bookkeeping (April 28, 1987)
@x module 560
done: b_close(tfm_file); read_font_info:=g;
@y [suggested by Jim Sterken]
done: if file_opened then b_close(tfm_file);
read_font_info:=g;
@z [this was not a bug, according to the comment in module 28]

331. jump_out must fix unfinished output (Found by Klaus Gunterman, 3 Aug 87)
@x module 593
doing_leaders:=false; dead_cycles:=0;
@y
doing_leaders:=false; dead_cycles:=0; cur_s:=-1;
@z
@x module 617
cur_s:=-1; ensure_dvi_open;
@y
ensure_dvi_open;
@z
@x module 640
dvi_out(eop); incr(total_pages);
@y
dvi_out(eop); incr(total_pages); cur_s:=-1;
@z
@x module 642
if total_pages=0 then print_nl("No pages of output.")
@y
while cur_s>-1 do
  begin if cur_s>0 then dvi_out(pop)
  else  begin dvi_out(eop); incr(total_pages);
    end;
  decr(cur_s);
  end;
if total_pages=0 then print_nl("No pages of output.")
@z

332. \hangindent=1pt$$\halign{...\cr\noalign{\hrule}}$$ problem (19 Aug 87)
@x module 805
q:=link(head);
@y
q:=link(head); s:=head;
@z
@x module 805, continued
  q:=link(q);
@y
  s:=q; q:=link(q);
@z
@x module 806
if is_running(depth(q)) then depth(q):=depth(p);
@y
if is_running(depth(q)) then depth(q):=depth(p);
if o<>0 then
  begin r:=link(q); link(q):=null; q:=hpack(q,natural);
  shift_amount(q):=o; link(q):=r; link(s):=q;
  end;
@z

333. \hskip 0pt plus 1fil\ifdim problem (found by Alan Guth, 20 Aug 87)
@x module 366
@!cvl_backup,@!radix_backup:small_number; {to save |cur_val_level| and |radix|}
@y
@!cvl_backup,@!radix_backup,@!co_backup:small_number;
  {to save |cur_val_level|, etc.}
@z
@x
backup_backup:=link(backup_head);
@y
co_backup:=cur_order; backup_backup:=link(backup_head);
@z
@x
link(backup_head):=backup_backup;
@y
cur_order:=co_backup; link(backup_head):=backup_backup;
@z

334. leaders too sensitive near exact multiples (M. F. Bridgland, 9 Nov 87)
@x module 626
  begin edge:=cur_h+rule_wd; lx:=0;
@y
  begin rule_wd:=rule_wd+10; {compensate for floating-point rounding}
  edge:=cur_h+rule_wd; lx:=0;
@z
@x ibid
  cur_h:=edge; goto next_p;
@y
  cur_h:=edge-10; goto next_p;
@z
@x module 635
  begin edge:=cur_v+rule_ht; lx:=0;
@y
  begin rule_ht:=rule_ht+10; {compensate for floating-point rounding}
  edge:=cur_v+rule_ht; lx:=0;
@z
@x ibid
  cur_v:=edge; goto next_p;
@y
  cur_v:=edge-10; goto next_p;
@z

335. Glitch in fixed-point multiplication of negatives (W.G. Sullivan, 17Nov87)
@x module 572
begin alpha:=16*z; beta:=16;
while z>=@'40000000 do
  begin z:=z div 2; beta:=beta div 2;
  end;
@y
begin alpha:=16;
while z>=@'40000000 do
  begin z:=z div 2; alpha:=alpha+alpha;
  end;
beta:=256 div alpha; alpha:=alpha*z;
@z

336. If there are no \patterns and some \lccode is 1 (Breitenlohner, 12Dec87)
@x module 952
trie_link(0):=0; trie_char(0):=0; trie_op(0):=0;
@y
trie_link(0):=0; trie_char(0):=0; trie_op(0):=min_quarterword;
@z

337. \csname might encounter undefined_cs in a group (Chris Thompson, 23Dec87)
@x module 372
  begin eqtb[cur_cs]:=eqtb[frozen_relax];
@y
  begin eq_define(cur_cs,relax,256);
@z

338. \outer\def\a0{}\a\a shows temp_head list garbage (Silvio Levy, 20Apr88)
@x module 391
repeat if (info(r)>match_token+127)or(info(r)<match_token) then s:=null
else  begin match_chr:=info(r)-match_token; s:=link(r); r:=s;
  p:=temp_head; link(p):=null; m:=0;
@y
repeat link(temp_head):=null;
if (info(r)>match_token+127)or(info(r)<match_token) then s:=null
else  begin match_chr:=info(r)-match_token; s:=link(r); r:=s;
  p:=temp_head; m:=0;
@z

339. \def\\#1{}\input a\\\z failed (Robert Messer, 24Apr88)
@x module 259
var h:integer; {hash code}
@y
var h:integer; {hash code}
@!d:integer; {number of characters in incomplete current string}
@z
@x module 260
str_room(l);
for k:=j to j+l-1 do append_char(buffer[k]);
text(p):=make_string;
@y
str_room(l); d:=cur_length;
while pool_ptr>str_start[str_ptr] do
  begin decr(pool_ptr); str_pool[pool_ptr+l]:=str_pool[pool_ptr];
  end; {move current string up to make room for another}
for k:=j to j+l-1 do append_char(buffer[k]);
text(p):=make_string; pool_ptr:=pool_ptr+d;
@z

340. Make patterns work when trie_min=0 (Peter Breitenlohner, 10May88)
@x module 951
trie_max:=128; trie_min:=128; trie_link(0):=1; trie_taken[0]:=false;
@y
trie_max:=128; trie_min:=128; trie_link(0):=1; trie_taken[0]:=false;
trie_link(trie_size):=0; trie_back(0):=trie_size; {wrap around}
@z
@x module 953
begin c:=trie_c[p]; {we have |c>0|}
if c<trie_min then trie_min:=c;
z:=trie_link(trie_min-1); {get the first conceivably good hole}
@y
begin c:=trie_c[p];
if c<trie_min then trie_min:=c;
if trie_min=0 then z:=trie_link(trie_size)
else z:=trie_link(trie_min-1); {get the first conceivably good hole}
@z
@x module 966
r:=0; {finally, we will zero out the holes}
@y
r:=trie_size; {finally, we will zero out the holes}
@z

341. Avoid possible trie_pointer out of range (24May88)
@x module 923
hc[0]:=127; hc[hn+1]:=127; hc[hn+2]:=256; {insert delimiters}
@y
hc[0]:=127; hc[hn+1]:=127; hc[hn+2]:=128; {insert delimiters}
@z

342. \input a\romannumeral1 etc.; similar to bug 339. (25May88)
@x module 464
@p function str_toks:pointer; {changes the current string to a token list}
@y
@p function str_toks(@!b:pool_pointer):pointer;
  {changes the string |str_pool[b..pool_ptr]| to a token list}
@z
@x module 464, continued
p:=temp_head; link(p):=null; k:=str_start[str_ptr];
@y
p:=temp_head; link(p):=null; k:=b;
@z
@x module 464, concluded
pool_ptr:=str_start[str_ptr]; str_toks:=p;
@y
pool_ptr:=b; str_toks:=p;
@z
@x module 465
begin get_x_token; scan_something_internal(tok_val,false);
if cur_val_level>=ident_val then @<Copy the token list@>
else begin old_setting:=selector; selector:=new_string;
@y
@!b:pool_pointer; {base of temporary string}
begin get_x_token; scan_something_internal(tok_val,false);
if cur_val_level>=ident_val then @<Copy the token list@>
else begin old_setting:=selector; selector:=new_string; b:=pool_ptr;
@z
@x module 465, continued
  selector:=old_setting; the_toks:=str_toks;
@y
  selector:=old_setting; the_toks:=str_toks(b);
@z
@x module 470
begin c:=cur_chr; @<Scan the argument for command |c|@>;
old_setting:=selector; selector:=new_string;
@<Print the result of command |c|@>;
selector:=old_setting; link(garbage):=str_toks; ins_list(link(temp_head));
@y
@!b:pool_pointer; {base of temporary string}
begin c:=cur_chr; @<Scan the argument for command |c|@>;
old_setting:=selector; selector:=new_string; b:=pool_ptr;
@<Print the result of command |c|@>;
selector:=old_setting; link(garbage):=str_toks(b); ins_list(link(temp_head));
@z

343. **\input\romannumeral6 confusion bypassed (25May88)
@x module 525
begin if (pool_ptr+name_length>pool_size)or(str_ptr=max_strings) then
@y
begin if (pool_ptr+name_length>pool_size)or(str_ptr=max_strings)or
 (cur_length>0) then
@z

344. Avoid negative dividend rounding upward (Chris Thompson, fixed 19Jun88)
@x module 126
else t:=(lo_mem_max+hi_mem_min+2) div 2; {|lo_mem_max+2<=t<hi_mem_min|}
@y
else t:=lo_mem_max+1+(hi_mem_min-lo_mem_max)div 2; {|lo_mem_max+2<=t<hi_mem_min|}
@z

345. Better strategy when near memory overflow (Chris Thompson)
@x module 126
begin if lo_mem_max+1000<hi_mem_min then t:=lo_mem_max+1000
@y
begin if hi_mem_min-lo_mem_max>=1998 then t:=lo_mem_max+1000
@z

346. An omission from change 333 (Tsunetoshi Hayashi, reported 20June88)
@x module 439
cur_val:=0; cur_val_level:=int_val; radix:=0;
@y
cur_val:=0; cur_val_level:=int_val; radix:=0; cur_order:=0;
@z

347. Avoid fatal_error after terminal eof (Tim Morgan, reported 25Oct88)
@x module 71 [serious problem occurred if this was called in open_log_file]
if not input_ln(term_in,true) then fatal_error("End of file on the terminal!");
@y
if not input_ln(term_in,true) then t_open_in;
@z

348. Force terminal output when open_log_file aborts (6Nov88)
@x module 535
  begin print_err("I can't write on file `");
@y
  begin selector:=term_only; print_err("I can't write on file `");
@z

349. By popular request, undo #347 and fix the bug a more complex way.
@x module 71 [this undoes change #347]
if not input_ln(term_in,true) then t_open_in;
@y
if not input_ln(term_in,true) then fatal_error("End of file on the terminal!");
@z
@x module 92 [now we get to the new stuff]
begin if job_name>0 then selector:=term_and_log
@y
begin if log_opened then selector:=term_and_log
@z
@x module 93
  error;
@y
  if log_opened then error;
@z
@x module 527
@!job_name:str_number; {principal file name}
@y
@!job_name:str_number; {principal file name}
@!log_opened:boolean; {has the transcript file been opened?}
@z
@x module 528
@<Initialize the output...@>=job_name:=0; name_in_progress:=false;
@y
@<Initialize the output...@>=
job_name:=0; name_in_progress:=false; log_opened:=false;
@z
@x module 534
selector:=log_only;
@y
selector:=log_only; log_opened:=true;
@z
@x module 535
begin if interaction<scroll_mode then {bypass |fatal_error|}
  begin selector:=term_only; print_err("I can't write on file `");
@.I can't write on file x@>
  print_file_name(cur_name,cur_area,cur_ext); print("'.");@/
  job_name:=0; history:=fatal_error_stop; jump_out;
  end; {abort the program without a log file}
@y
begin selector:=term_only;
@z
@x module 1265 [this change is optional, but it's a slight improvement]
if job_name<>0 then selector:=selector+2;
@y
if log_opened then selector:=selector+2;
@z
@x module 1333
if job_name>0 then
@y
if log_opened then
@z
@x module 1334
if job_name>0 then {the log file is open}
@y
if log_opened then
@z

350. Forgotten update in error recovery (Breitenlohner, 20 Jan 89)
@x module 976
    glue_ptr(p):=r;
@y
    glue_ptr(p):=r; q:=r;
@z
@x module 1004
    glue_ptr(p):=r;
@y
    glue_ptr(p):=r; q:=r;
@z

351. Avoid error \aftergroup\relax\dump (Mittelbach&Schoepf, 17 Feb 89)
@x module 280
begin check_full_save_stack;
save_type(save_ptr):=insert_token; save_level(save_ptr):=level_zero;
save_index(save_ptr):=t; incr(save_ptr);
@y
begin if cur_level>level_one then
  begin check_full_save_stack;
  save_type(save_ptr):=insert_token; save_level(save_ptr):=level_zero;
  save_index(save_ptr):=t; incr(save_ptr);
  end;
@z

352. Cosmetic change to validate comment in 115 (Breitenlohner, 20 Mar 89)
@x module 791
begin q:=link(cur_align);
if (cur_align=null)or(q=null) then confusion("endv");
@y
begin if cur_align=null then confusion("endv");
q:=link(cur_align);@+if q=null then confusion("endv");
@z

353. $$\begingroup\halign{#\cr}$$ caused confusion (Mittelbach, 7 Jun 89)
@x module 1130
mmode+halign: if privileged then init_align;
@y
mmode+halign: if privileged then
  if cur_group=math_shift_group then init_align
  else off_save;
@z

354. Conflict in global array dig with .5\ifdim.6 (Mittelbach, 20 Jun 89)
@x module 450
@!k:small_number; {number of digits in a decimal fraction}
@y
@!k,@!kk:small_number; {number of digits in a decimal fraction}
@!p,@!q:pointer; {top of decimal digit stack}
@z
@x module 452
begin k:=0; get_token; {|point_token| is being re-scanned}
loop@+  begin get_x_token;
  if (cur_tok>zero_token+9)or(cur_tok<zero_token) then goto done1;
  if k<17 then {digits for |k>=17| cannot affect the result}
    begin dig[k]:=cur_tok-zero_token; incr(k);
    end;
  end;
done1: f:=round_decimals(k);
@y
begin k:=0; p:=null; get_token; {|point_token| is being re-scanned}
loop@+  begin get_x_token;
  if (cur_tok>zero_token+9)or(cur_tok<zero_token) then goto done1;
  if k<17 then {digits for |k>=17| cannot affect the result}
    begin q:=get_avail; link(q):=p; info(q):=cur_tok-zero_token;
    p:=q; incr(k);
    end;
  end;
done1: for kk:=k downto 1 do
  begin dig[kk-1]:=info(p); q:=p; p:=link(p); free_avail(q);
  end;
f:=round_decimals(k);
@z

355. String startup problems (Wayne Sullivan, 17 Jul 89)
@x module 31 (Warning: This affects most change files!)
        overflow("buffer size",buf_size);
@:TeX capacity exceeded buffer size}{\quad buffer size@>
@y
        @<Report overflow of the input buffer, and abort@>;
@z
@x module 35
consist of the remainder of the command line, after the part that invoked
\TeX.
@y
consist of the remainder of the command line, after the part that invoked
\TeX.

The first line is special also because it may be read before \TeX\ has
input a format file. In such cases, normal error messages cannot yet
be given. The following code uses concepts that will be explained later.

@<Report overflow of the input buffer, and abort@>=
if format_ident=0 then
  begin write_ln(term_out,'Buffer size exceeded!'); goto final_end;
@.Buffer size exceeded@>
  end
else begin cur_input.loc_field:=first; cur_input.limit_field:=last-1;
  overflow("buffer size",buf_size);
@:TeX capacity exceeded buffer size}{\quad buffer size@>
  end
@z
@x module 1310
k:=pool_ptr-4; undump_four_ASCII
@y
k:=pool_ptr-4; undump_four_ASCII;
init_str_ptr:=str_ptr; init_pool_ptr:=pool_ptr
@z
@x module 1332
tini@/
@y
init_str_ptr:=str_ptr; init_pool_ptr:=pool_ptr; fix_date_and_time;
tini@/
@z
@x module 1332
init_str_ptr:=str_ptr; init_pool_ptr:=pool_ptr;@/
@y
@z

356. Allow integer products to be 31 bits long (Mittelbach, 16 Aug 89)
@x module 105
@p function nx_plus_y(@!n:integer;@!x,@!y:scaled):scaled;
begin if n<0 then
  begin negate(x); negate(n);
  end;
if n=0 then nx_plus_y:=y
else if ((x<=(@'7777777777-y) div n)and(-x<=(@'7777777777+y) div n)) then
  nx_plus_y:=n*x+y
else  begin arith_error:=true; nx_plus_y:=0;
  end;
end;
@y
@d nx_plus_y(#)==mult_and_add(#,@'7777777777)
@d mult_integers(#)==mult_and_add(#,0,@'17777777777)

@p function mult_and_add(@!n:integer;@!x,@!y,@!max_answer:scaled):scaled;
begin if n<0 then
  begin negate(x); negate(n);
  end;
if n=0 then mult_and_add:=y
else if ((x<=(max_answer-y) div n)and(-x<=(max_answer+y) div n)) then
  mult_and_add:=n*x+y
else  begin arith_error:=true; mult_and_add:=0;
  end;
end;
@z
@x module 1240
  if q=multiply then cur_val:=nx_plus_y(eqtb[l].int,cur_val,0)
@y
  if q=multiply then
    if p=int_val then cur_val:=mult_integers(eqtb[l].int,cur_val)
    else cur_val:=nx_plus_y(eqtb[l].int,cur_val,0)
@z

357. Raise the maximum number of tokens normally shown (pointed out by
John Lavagnino in TeXhax on 19 May 89; corrected on 31 Aug 89).
@x module 295
begin if p<>null then show_token_list(link(p),null,1000);
@y
begin if p<>null then show_token_list(link(p),null,10000000);
@z
@x module 1370 (this part of the change is cosmetic only)
show_token_list(link(def_ref),null,10000000); print_ln;
@y
token_show(def_ref); print_ln;
@z

358. Like 353, but with $$\begingroup\eqno$$ (Mittelbach, 25 Jul 89)
@x module 1140
mmode+eq_no: if privileged then start_eq_no;
@y
mmode+eq_no: if privileged then
  if cur_group=math_shift_group then start_eq_no
  else off_save;
@z

359. Major extension to 8-bit character input. (These changes, and a few that
follow, were discussed during the 10th TUG meeting and installed during
the week of 10 Sep 1989. They bring the implementation of TeX82 up to date with
international developments and experience, and I believe they mark the
final development of this program into a completely stable system.)
Several changes to the documentation (not the program) are omitted here.
@x module 18
@!ASCII_code=0..127; {seven-bit numbers}
@y
@!ASCII_code=0..255; {eight-bit numbers}
@z
@x module 19
@d last_text_char=127 {ordinal number of the largest element of |text_char|}

@<Local variables for init...@>=
@!i:0..last_text_char;
@y
@d last_text_char=255 {ordinal number of the largest element of |text_char|}

@<Local variables for init...@>=
@!i:integer;
@z
@x module 21
xchr[0]:=' '; xchr[@'177]:=' ';
  {ASCII codes 0 and |@'177| do not appear in text}
@y
@z
@x module 23
for i:=1 to @'37 do xchr[i]:=' ';
@y changing ' ' to chr(i) here will allow all 8-bit characters to get in
for i:=0 to @'37 do xchr[i]:=' ';
for i:=@'177 to @'377 do xchr[i]:=' ';
@z
@x module 24
for i:=1 to @'176 do xord[xchr[i]]:=i;
@y
for i:=@'200 to @'377 do xord[xchr[i]]:=i;
for i:=0 to @'176 do xord[xchr[i]]:=i;
@z
@x module 38
@<Types...@>=
@!pool_pointer = 0..pool_size; {for variables that point into |str_pool|}
@!str_number = 0..max_strings; {for variables that point into |str_start|}

@ @<Glob...@>=
@!str_pool:packed array[pool_pointer] of ASCII_code; {the characters}
@y [OK to make si(#)==#-128 and so(#)==#+128 (without parens) in change files]
Some \PASCAL\ compilers won't pack integers into a single byte unless the
integers lie in the range |-128..127|. To accommodate such systems
we access the string pool only via macros that can easily be redefined.

@d si(#) == # {convert from |ASCII_code| to |packed_ASCII_code|}
@d so(#) == # {convert from |packed_ASCII_code| to |ASCII_code|}

@<Types...@>=
@!pool_pointer = 0..pool_size; {for variables that point into |str_pool|}
@!str_number = 0..max_strings; {for variables that point into |str_start|}
@!packed_ASCII_code = 0..255; {elements of |str_pool| array}

@ @<Glob...@>=
@!str_pool:packed array[pool_pointer] of packed_ASCII_code; {the characters}
@z
@x module 42
begin str_pool[pool_ptr]:=#; incr(pool_ptr);
@y
begin str_pool[pool_ptr]:=si(#); incr(pool_ptr);
@z
@x module 45
  begin if str_pool[j]<>buffer[k] then
@y
  begin if so(str_pool[j])<>buffer[k] then
@z
@x module 47
var k,@!l:0..127; {small indices or counters}
@y
var k,@!l:0..255; {small indices or counters}
@z
@x module 47
@<Make the first 128 strings@>;
@y
@<Make the first 256 strings@>;
@z
@x module 48
@ @<Make the first 128...@>=
for k:=0 to 127 do
  begin if (@<Character |k| cannot be printed@>) then
    begin append_char("^"); append_char("^");
    if k<@'100 then append_char(k+@'100)
    else append_char(k-@'100);
@y
@ @d app_lc_hex(#)==l:=#;
  if l<10 then append_char(l+"0)@+else append_char(l-10+"a")

@<Make the first 256...@>=
for k:=0 to 255 do
  begin if (@<Character |k| cannot be printed@>) then
    begin append_char("^"); append_char("^");
    if k<@'100 then append_char(k+@'100)
    else if k<@'200 then append_char(k-@'100)
    else begin app_lc_hex(k div 16); app_lc_hex(k mod 16);
      end;
@z
@x module 59
  begin print_char(str_pool[j]); incr(j);
@y
  begin print_char(so(str_pool[j])); incr(j);
@z
@x modules 59 and 60
else if s<128 then
@y
else if s<256 then
@z
@x module 60
  begin print(str_pool[j]); incr(j);
@y
  begin print(so(str_pool[j])); incr(j);
@z
@x module 63
if c>=0 then if c<128 then print(c);
@y
if c>=0 then if c<256 then print(c);
@z
@x module 68
@ In certain situations, \TeX\ prints either a standard visible ASCII
character or its hexadecimal ASCII code.

@p procedure print_ASCII(@!c:integer); {prints a character or its code}
begin if (c>=0) and (c<=127) then print(c)
else  begin print_char("[");
  if c<0 then print_int(c)@+else print_hex(c);
  print_char("]");
  end;
end;
@y
@ Old versions of \TeX\ needed a procedure called |print_ASCII| whose function
is now subsumed by |print|. We retain the old name here as a possible aid to
future software arch\ae ologists.

@d print_ASCII == print
@z
@x module 69
    begin print_char(str_pool[j]); n:=n-v;
    end;
  if n<=0 then return; {nonpositive input produces no output}
  k:=j+2; u:=v div (str_pool[k-1]-"0");
  if str_pool[k-1]="2" then
    begin k:=k+2; u:=u div (str_pool[k-1]-"0");
    end;
  if n+u>=v then
    begin print_char(str_pool[k]); n:=n+u;
    end
  else  begin j:=j+2; v:=v div (str_pool[j-1]-"0");
@y
    begin print_char(so(str_pool[j])); n:=n-v;
    end;
  if n<=0 then return; {nonpositive input produces no output}
  k:=j+2; u:=v div (so(str_pool[k-1])-"0");
  if str_pool[k-1]=si("2") then
    begin k:=k+2; u:=u div (so(str_pool[k-1])-"0");
    end;
  if n+u>=v then
    begin print_char(so(str_pool[k])); n:=n+u;
    end
  else  begin j:=j+2; v:=v div (so(str_pool[j-1])-"0");
@z
@x module 70
  begin print_char(str_pool[j]); incr(j);
@y
  begin print_char(so(str_pool[j])); incr(j);
@z
@x module 222
@d single_base=active_base+128 {equivalents of one-letter control sequences}
@d null_cs=single_base+128 {equivalent of \.{\\csname\\endcsname}}
@y
@d single_base=active_base+256 {equivalents of one-character control sequences}
@d null_cs=single_base+256 {equivalent of \.{\\csname\\endcsname}}
@z
@x module 230
  {table of 128 command codes (the ``catcodes'')}
@d lc_code_base=cat_code_base+128 {table of 128 lowercase mappings}
@d uc_code_base=lc_code_base+128 {table of 128 uppercase mappings}
@d sf_code_base=uc_code_base+128 {table of 128 spacefactor mappings}
@d math_code_base=sf_code_base+128 {table of 128 math mode mappings}
@d int_base=math_code_base+128 {beginning of region 5}
@y
  {table of 256 command codes (the ``catcodes'')}
@d lc_code_base=cat_code_base+256 {table of 256 lowercase mappings}
@d uc_code_base=lc_code_base+256 {table of 256 uppercase mappings}
@d sf_code_base=uc_code_base+256 {table of 256 spacefactor mappings}
@d math_code_base=sf_code_base+256 {table of 256 math mode mappings}
@d int_base=math_code_base+256 {beginning of region 5}
@z
@x module 232
for k:=0 to 127 do
@y
for k:=0 to 255 do
@z
@x module 236
@d del_code_base=count_base+256 {128 delimiter code mappings}
@d dimen_base=del_code_base+128 {beginning of region 6}
@y
@d del_code_base=count_base+256 {256 delimiter code mappings}
@d dimen_base=del_code_base+256 {beginning of region 6}
@z
@x module 240
for k:=0 to 127 do del_code(k):=-1;
@y
for k:=0 to 255 do del_code(k):=-1;
@z
@x module 264
begin if s<128 then cur_val:=s+single_base
@y
begin if s<256 then cur_val:=s+single_base
@z
@x ibid
  for j:=0 to l-1 do buffer[j]:=str_pool[k+j];
@y
  for j:=0 to l-1 do buffer[j]:=so(str_pool[k+j]);
@z
@x module 289
@d cs_token_flag==@'10000 {amount added to the |eqtb| location in a
  token that stands for a control sequence; is a multiple of~256}
@y
@d cs_token_flag==@'7777 {amount added to the |eqtb| location in a
  token that stands for a control sequence; is a multiple of~256, less~1}
@z
@x module 293
  if (info(p)<0)or(c>127) then print_esc("BAD.")
@y
  if info(p)<0 then print_esc("BAD.")
@z
@x module 341
@!cat:0..15; {|cat_code(cur_chr)|, usually}
@y
@!cat:0..15; {|cat_code(cur_chr)|, usually}
@!c,@!cc:ASCII_code; {constituents of a possible expanded code}
@!d:2..3; {number of excess characters in an expanded code}
@z
@x module 352
@<If this |sup_mark| starts a control character...@>=
begin if (cur_chr=buffer[loc])and(loc<limit) then
  begin if buffer[loc+1]<@'100 then cur_chr:=buffer[loc+1]+@'100
  else cur_chr:=buffer[loc+1]-@'100;
  loc:=loc+2; goto reswitch;
  end;
@y
@ Notice that a code like \.{\^\^8} becomes \.x if not followed by a hex digit.
 
@d is_hex(#)==(((#>="0")and(#<="9"))or((#>="a")and(#<="f")))
@d hex_to_cur_chr==
  if c<="9" then cur_chr:=c-"0" @+else cur_chr:=c-"a"+10;
  if cc<="9" then cur_chr:=16*cur_chr+cc-"0"
  else cur_chr:=16*cur_chr+cc-"a"+10

@<If this |sup_mark| starts an expanded character...@>=
begin if cur_chr=buffer[loc] then if loc<limit then
  begin c:=buffer[loc+1]; @+if c<@'200 then {yes we have an expanded char}
    begin loc:=loc+2; 
    if is_hex(c) then if loc<=limit then
      begin cc:=buffer[loc]; @+if is_hex(cc) then
        begin incr(loc); hex_to_cur_chr; goto reswitch;
        end;
      end;
    if c<@'100 then cur_chr:=c+@'100 @+else cur_chr:=c-@'100;
    goto reswitch;
    end;
  end;
@z
@x module 355
  begin cur_chr:=buffer[k+1];
  if cur_chr<@'100 then buffer[k-1]:=cur_chr+@'100
  else buffer[k-1]:=cur_chr-@'100;
  limit:=limit-2; first:=first-2;
  while k<=limit do
    begin buffer[k]:=buffer[k+2]; incr(k);
    end;
  goto start_cs;
  end;
@y
begin c:=buffer[k+1]; @+if c<@'200 then {yes, one is indeed present}
    begin d:=2;
    if is_hex(c) then @+if k+2<=limit then
      begin cc:=buffer[k+2]; @+if is_hex(cc) then incr(d);
      end;
    if d>2 then
      begin hex_to_cur_chr; buffer[k-1]:=cur_chr;
      end
    else if c<@'100 then buffer[k-1]:=c+@'100
    else buffer[k-1]:=c-@'100;
    limit:=limit-d; first:=first-d;
    while k<=limit do
      begin buffer[k]:=buffer[k+d]; incr(k);
      end;
    goto start_cs;
    end;
  end;
@z
@x module 360
There is one more branch.
@y
There is one more branch.

@d end_line_char_inactive == (end_line_char<0)or(end_line_char>255)
@z
@x ibid
    if (end_line_char<0)or(end_line_char>127) then decr(limit)
@y
    if end_line_char_inactive then decr(limit)
@z
@x modules 362 and 483 and 538 and 1337
if (end_line_char<0)or(end_line_char>127) then decr(limit)
@y
if end_line_char_inactive then decr(limit)
@z
@x module 391
if (info(r)>match_token+127)or(info(r)<match_token) then s:=null
@y
if (info(r)>match_token+255)or(info(r)<match_token) then s:=null
@z
@x module 407
   ((cur_chr=str_pool[k])or(cur_chr=str_pool[k]-"a"+"A")) then
@y
   ((cur_chr=so(str_pool[k]))or(cur_chr=so(str_pool[k])-"a"+"A")) then
@z
@x module 414
begin scan_seven_bit_int;
@y
begin scan_char_num;
@z
@x module 432
|scan_something_internal|:

@<Declare procedures that scan restricted classes of integers@>=
procedure scan_seven_bit_int;
begin scan_int;
if (cur_val<0)or(cur_val>127) then
  begin print_err("Bad character code");
@.Bad character code@>
  help2("The numeric code for a character must be between 0 and 127.")@/
    ("I changed this one to zero."); int_error(cur_val); cur_val:=0;
  end;
end;
@y
|scan_something_internal|.
@z (and I'll interchange modules 434 and 435)
@x module 442
if cur_val>127 then
@y
if cur_val>255 then
@z
@x module 464
  begin t:=str_pool[k];
@y
  begin t:=so(str_pool[k]);
@z
@x module 506 (correction needed twice)
if (cur_cmd>active_char)or(cur_chr>127) then
@y
if (cur_cmd>active_char)or(cur_chr>255) then
@z
@x module 519
for j:=str_start[a] to str_start[a+1]-1 do append_to_name(str_pool[j]);
for j:=str_start[n] to str_start[n+1]-1 do append_to_name(str_pool[j]);
for j:=str_start[e] to str_start[e+1]-1 do append_to_name(str_pool[j]);
@y
for j:=str_start[a] to str_start[a+1]-1 do append_to_name(so(str_pool[j]));
for j:=str_start[n] to str_start[n+1]-1 do append_to_name(so(str_pool[j]));
for j:=str_start[e] to str_start[e+1]-1 do append_to_name(so(str_pool[j]));
@z
@x module 526
if (cur_cmd>active_char)or(cur_chr>127) then
@y
if (cur_cmd>active_char)or(cur_chr>255) then
@z
@x module 603 (change to be made twice)
  dvi_out(str_pool[k]);
@y
  dvi_out(so(str_pool[k]));
@z
@x module 617
  for s:=str_start[str_ptr] to pool_ptr-1 do dvi_out(str_pool[s]);
@y
  for s:=str_start[str_ptr] to pool_ptr-1 do dvi_out(so(str_pool[s]));
@z
@x module 766
  begin case str_pool[r_type*8+t+magic_offset] of
@y
  begin case so(str_pool[r_type*8+t+magic_offset]) of
@z
@x module 780
@d span_code=128 {distinct from any character}
@d cr_code=129 {distinct from |span_code| and from any character}
@y
@d span_code=256 {distinct from any character}
@d cr_code=257 {distinct from |span_code| and from any character}
@z
@x module 815
label done,done1,done2,done3,done4;
@y
label done,done1,done2,done3,done4,continue;
@z
@x module 896
  else if (type(s)=kern_node)and(subtype(s)=normal) then c:=128
  else if type(s)=whatsit_node then c:=128
  else goto done1;
  if c<128 then if lc_code(c)<>0 then
    if (lc_code(c)=c)or(uc_hyph>0) then goto done2
    else goto done1;
  s:=link(s);
@y
  else if (type(s)=kern_node)and(subtype(s)=normal) then goto continue
  else if type(s)=whatsit_node then goto continue
  else goto done1;
  if lc_code(c)<>0 then
    if (lc_code(c)=c)or(uc_hyph>0) then goto done2
    else goto done1;
continue: s:=link(s);
@z
@x module 897
    if c>=128 then goto done3;
    if (lc_code(c)=0)or(hn=63) then goto done3;
    hb:=s; incr(hn); hu[hn]:=c; hc[hn]:=lc_code(c)-1;
@y
    if lc_code(c)=0 then goto done3;
    if hn=63 then goto done3;
    hb:=s; incr(hn); hu[hn]:=c; hc[hn]:=lc_code(c);
@z
@x module 898
if c>=128 then goto done3;
if (lc_code(c)=0)or(j=63) then goto done3;
incr(j); hu[j]:=c; hc[j]:=lc_code(c)-1;@/
@y
if lc_code(c)=0 then goto done3;
if j=63 then goto done3;
incr(j); hu[j]:=c; hc[j]:=lc_code(c);@/
@z
@x module 923
hc[0]:=127; hc[hn+1]:=127; hc[hn+2]:=128; {insert delimiters}
@y
hc[0]:=0; hc[hn+1]:=0; hc[hn+2]:=256; {insert delimiters}
@z
@x ibid
  while hc[l]=trie_char(z) do
@y
  while hc[l]=qo(trie_char(z)) do
@z
@x module 931
  repeat if str_pool[u]<hc[j] then goto not_found;
  if str_pool[u]>hc[j] then goto done;
@y
  repeat if so(str_pool[u])<hc[j] then goto not_found;
  if so(str_pool[u])>hc[j] then goto done;
@z
@x module 937
else  begin if (cur_chr>127)or(lc_code(cur_chr)=0) then
@y
else  begin if lc_code(cur_chr)=0 then
@z
@x ibid
    begin incr(n); hc[n]:=lc_code(cur_chr)-1;
@y
    begin incr(n); hc[n]:=lc_code(cur_chr);
@z
@x module 945
@!init @!trie_c:packed array[trie_pointer] of ASCII_code; {characters to match}
@y
@!init @!trie_c:packed array[trie_pointer] of packed_ASCII_code;
  {characters to match}
@z
% In modules 951--954, change "127" to "255" and "128" to "256"
@x module 952
trie_link(0):=0; trie_char(0):=0; trie_op(0):=min_quarterword;
@y
trie_link(0):=0; trie_char(0):=min_quarterword; trie_op(0):=min_quarterword;
@z
@x module 953
begin c:=trie_c[p];
@y
begin c:=so(trie_c[p]);
@z
@x module 955
  begin if trie_link(h+trie_c[q])=0 then goto not_found;
@y
  begin if trie_link(h+so(trie_c[q]))=0 then goto not_found;
@z
@x module 956
repeat z:=h+trie_c[q]; trie_back(trie_link(z)):=trie_back(z);
@y
repeat z:=h+so(trie_c[q]); trie_back(trie_link(z)):=trie_back(z);
@z
@x module 959
  begin q:=trie_l[p]; c:=trie_c[p];
  trie_link(z+c):=trie_ref[q]; trie_char(z+c):=c; trie_op(z+c):=trie_o[p];
@y
  begin q:=trie_l[p]; c:=so(trie_c[p]);
  trie_link(z+c):=trie_ref[q]; trie_char(z+c):=qi(c); trie_op(z+c):=trie_o[p];
@z
@x module 962
  begin if cur_chr="." then cur_chr:=128 {edge-of-word delimiter}
  else  begin cur_chr:=lc_code(cur_chr);
    if cur_chr=0 then
      begin print_err("Nonletter");
@.Nonletter@>
      help1("(See Appendix H.)"); error; cur_chr:=128;
      end;
    end;
  if k<63 then
    begin incr(k); hc[k]:=cur_chr-1; hyf[k]:=0; digit_sensed:=false;
    end;
  end
else  begin hyf[k]:=cur_chr-"0";
  if k<63 then digit_sensed:=true;
  end
@y
  begin if cur_chr="." then cur_chr:=0 {edge-of-word delimiter}
  else  begin cur_chr:=lc_code(cur_chr);
    if cur_chr=0 then
      begin print_err("Nonletter");
@.Nonletter@>
      help1("(See Appendix H.)"); error;
      end;
    end;
  if k<63 then
    begin incr(k); hc[k]:=cur_chr; hyf[k]:=0; digit_sensed:=false;
    end;
  end
  else if k<63 then
    begin hyf[k]:=cur_chr-"0"; digit_sensed:=true;
    end;
  end
@z
@x module 963
  while (p>0)and(c>trie_c[p]) do
@y
  while (p>0)and(c>so(trie_c[p])) do
@z
@x ibid
  if (p=0)or(c<trie_c[p]) then
@y
  if (p=0)or(c<so(trie_c[p])) then
@z
@x module 964
trie_c[p]:=c; trie_o[p]:=min_quarterword;
@y
trie_c[p]:=si(c); trie_o[p]:=min_quarterword;
@z
@x module 965
if hc[1]=127 then hyf[0]:=0;
if hc[k]=127 then hyf[k]:=0;
@y
if hc[1]=0 then hyf[0]:=0;
if hc[k]=0 then hyf[k]:=0;
@z
@x module 966
h.rh:=0; h.b0:=min_quarterword; h.b1:=0; {|trie_link:=0|,
  |trie_op:=min_quarterword|, |trie_char:=0|}
@y
h.rh:=0; h.b0:=min_quarterword; h.b1:=min_quarterword; {|trie_link:=0|,
  |trie_op:=min_quarterword|, |trie_char:=qi(0)|}
@z
@x module 1034
if c<128 then
@y
@z
@x ibid
else space_factor:=1000
@y
@z
@x module 1151
letter,other_char,char_given: if cur_chr>=128 then c:=cur_chr
  else  begin c:=ho(math_code(cur_chr));
@y
letter,other_char,char_given: begin c:=ho(math_code(cur_chr));
@z
@x module 1154
mmode+letter,mmode+other_char,mmode+char_given: if cur_chr<128 then
    set_math_char(ho(math_code(cur_chr)))
  else set_math_char(cur_chr);
mmode+char_num: begin scan_char_num; cur_chr:=cur_val;
  if cur_chr<128 then set_math_char(ho(math_code(cur_chr)))
  else set_math_char(cur_chr);
@y
mmode+letter,mmode+other_char,mmode+char_given:
  set_math_char(ho(math_code(cur_chr)));
mmode+char_num: begin scan_char_num; cur_chr:=cur_val;
  set_math_char(ho(math_code(cur_chr)));
@z
@x module 1232
  p:=cur_chr; scan_seven_bit_int; p:=p+cur_val; scan_optional_equals;
@y
  p:=cur_chr; scan_char_num; p:=p+cur_val; scan_optional_equals;
@z
@x module 1233
else n:=127
@y
else n:=255
@z
@x module 1289
  begin if t>=cs_token_flag then t:=t-active_base;
  c:=t mod 256;
  if c<128 then if equiv(b+c)<>0 then t:=256*(t div 256)+equiv(b+c);
  if t>=cs_token_flag then info(p):=t+active_base
  else info(p):=t;
@y
  begin c:=t mod 256;
  if equiv(b+c)<>0 then info(p):=t-c+equiv(b+c);
@z
@x module 1309
  w.b0:=str_pool[k]; w.b1:=str_pool[k+1];
  w.b2:=str_pool[k+2]; w.b3:=str_pool[k+3];
@y [often qi(so(x))=x, but not e.g. when "quarterwords" are two bytes]
  w.b0:=qi(so(str_pool[k])); w.b1:=qi(so(str_pool[k+1]));
  w.b2:=qi(so(str_pool[k+2])); w.b3:=qi(so(str_pool[k+3]));
@z
@x module 1310
  str_pool[k]:=w.b0; str_pool[k+1]:=w.b1;
  str_pool[k+2]:=w.b2; str_pool[k+3]:=w.b3
@y
  str_pool[k]:=si(qo(w.b0)); str_pool[k+1]:=si(qo(w.b1));
  str_pool[k+2]:=si(qo(w.b2)); str_pool[k+3]:=si(qo(w.b3))
@z
@x module 1368
for k:=str_start[str_ptr] to pool_ptr-1 do dvi_out(str_pool[k]);
@y
for k:=str_start[str_ptr] to pool_ptr-1 do dvi_out(so(str_pool[k]));
@z

360. Major change to allow multiple hyphenation tables.
I've combined this with the code for a minor change:
361. New parameters \lefthyphenmin and \righthyphenmin.
@x module 11 (affects most change files!)
@!trie_size=8000; {space for hyphenation patterns; should be larger for
  \.{INITEX} than it is in production versions of \TeX}
@y
@!trie_size=8000; {space for hyphenation patterns; should be larger for
  \.{INITEX} than it is in production versions of \TeX}
@!trie_op_size=500; {space for ``opcodes'' in the hyphenation patterns}
@z
@x module 212
known as |space_factor|; it holds the current space factor used in spacing
calculations. In math mode, |aux| is also known as |incompleat_noad|; if
@y
known as |space_factor| and |clang|; it holds the current space factor used in
spacing calculations, and the current language used for hyphenation.
(The value of |clang| is undefined in restricted horizontal mode.)
In math mode, |aux| is also known as |incompleat_noad|; if
@z
@x ibid
  @!pg_field,@!aux_field,@!ml_field: integer;
@y
  @!pg_field,@!ml_field: integer;
  @!aux_field: memory_word;
@z
@x module 213
@d prev_depth==aux {the name of |aux| in vertical mode}
@d space_factor==aux {the name of |aux| in horizontal mode}
@d incompleat_noad==aux {the name of |aux| in math mode}
@y
@d prev_depth==aux.sc {the name of |aux| in vertical mode}
@d space_factor==aux.hh.lh {part of |aux| in horizontal mode}
@d clang==aux.hh.rh {the other part of |aux| in horizontal mode}
@d incompleat_noad==aux.int {the name of |aux| in math mode}
@z
@x module 218
@!a:integer; {auxiliary}
@y
@!a:memory_word; {auxiliary}
@z
@x module 219
  if a<=ignore_depth then print("ignored")
  else print_scaled(a);
  if nest[p].pg_field<>0 then
    begin print(", prevgraf ");
    print_int(nest[p].pg_field); print(" line");
    if nest[p].pg_field<>1 then print_char("s");
    end;
  end;
1: begin print_nl("spacefactor "); print_int(a);
  end;
2: if a<>null then
  begin print("this will be denominator of:"); show_box(a);
@y
  if a.sc<=ignore_depth then print("ignored")
  else print_scaled(a.sc);
  if nest[p].pg_field<>0 then
    begin print(", prevgraf ");
    print_int(nest[p].pg_field); print(" line");
    if nest[p].pg_field<>1 then print_char("s");
    end;
  end;
1: begin print_nl("spacefactor "); print_int(a.hh.lh);
  if m>0 then if a.hh.rh>0 then
    begin print(", current language "); print_int(a.hh.rh);
    end;
  end;
2: if a.int<>null then
  begin print("this will be denominator of:"); show_box(a.int);
@z
@x modules 236--238
@d int_pars=50 {total number of integer parameters}
@y
@d language_code=50 {current hyphenation table}
@d left_hyphen_min_code=51 {minimum left hyphenation fragment size}
@d right_hyphen_min_code=52 {minimum right hyphenation fragment size}
@d int_pars=53 {total number of integer parameters}
@z and define/print these new parameters appropriately (mimicking new_line_char)
@x module 418
else begin cur_val:=aux;
  if m=vmode then cur_val_level:=dimen_val@+else cur_val_level:=int_val;
  end
@y
else if m=vmode then
  begin cur_val:=prev_depth; cur_val_level:=dimen_val;
  end
else begin cur_val:=space_factor; cur_val_level:=int_val;
  end
@z
@x module 775
  begin mode:=-vmode; prev_depth:=nest[nest_ptr-2].aux_field;
@y
  begin mode:=-vmode; prev_depth:=nest[nest_ptr-2].aux_field.sc;
@z
@x module 786
begin push_nest; mode:=(-hmode-vmode)-mode; aux:=0;
@y
begin push_nest; mode:=(-hmode-vmode)-mode;
if mode=-hmode then space_factor:=0 @+else prev_depth:=0;
@z
@x module 863
loop@+  begin @<Create an active breakpoint representing the beginning of
    the paragraph@>;
@y
loop@+  begin if threshold>inf_bad then threshold:=inf_bad;
  if second_pass then @<Initialize for hyphenating a paragraph@>;
  @<Create an active breakpoint representing the beginning of the paragraph@>;
@z
@x module 891
passes on only about 5 per cent of the paragraphs.
@y
passes on only about 5 per cent of the paragraphs.

@<Initialize for hyphenating...@>=
begin @!init if trie_not_ready then init_trie; @+tini@;@/
l_hyf:=left_hyphen_min-1;@+if l_hyf<0 then l_hyf:=0;
r_hyf:=right_hyphen_min-1;@+if r_hyf<0 then r_hyf:=0;
min_hyf:=l_hyf+r_hyf+2; cur_lang:=0;
end
@z
@x module 892
@!hyf_char:integer; {hyphen character of the relevant font}
@y
@!hyf_char:integer; {hyphen character of the relevant font}
@!cur_lang:ASCII_code; {current hyphenation table of interest}
@!l_hyf,@!r_hyf,@!min_hyf:integer; {limits on fragment sizes}
@z
@x module 894
begin s:=link(cur_p);
@y
begin if min_hyf>63 then goto done1;
s:=link(cur_p);
@z
@x module 896
  else if type(s)=whatsit_node then goto continue
@y
  else if type(s)=whatsit_node then
    begin @<Advance \(p)past a whatsit node in the \(p)pre-hyphenation loop@>;
    goto continue;
    end
@z
@x module 899 (title of this module also changes appropriately)
if hn<5 then goto done1;
@y
if hn<min_hyf then goto done1;
@z
@x module 902
for j:=2 to hn-3 do if odd(hyf[j]) then goto found1;
@y
for j:=l_hyf+1 to hn-r_hyf-1 do if odd(hyf[j]) then goto found1;
@z
@x module 920
$p_1\ldots p_k$ by setting |@t$z_1$@>:=@t$p_1$@>| and then, for |1<i<=k|,
@y
$p_1\ldots p_k$ by letting $z_0$ be one greater than the relevant language index
and then, for |1<=i<=k|,
@z
@x ibid
the letters in |hc[(l-k+1)..l@,]|, we perform all of the required operations
for this pattern by carrying out the following little program: Set
|v:=trie_op(@t$z_k$@>)|. Then set |hyf[l-hyf_distance[v]]:=@tmax@>(
hyf[l-hyf_distance[v]], hyf_num[v])|, and |v:=hyf_next[v]|; repeat, if
necessary, until |v=min_quarterword|.
@y
the letters in |hc[(l-k+1)..l@,]| of language |t|,
we perform all of the required operations
for this pattern by carrying out the following little program: Set
|v:=trie_op(@t$z_k$@>)|. Then set |v:=v+op_start[t]|,
|hyf[l-hyf_distance[v]]:=@tmax@>(hyf[l-hyf_distance[v]], hyf_num[v])|,
and |v:=hyf_next[v]|; repeat, if necessary, until |v=min_quarterword|.
@z
@x module 921
@!hyf_distance:array[quarterword] of small_number; {position |k-j| of $n_j$}
@!hyf_num:array[quarterword] of small_number; {value of $n_j$}
@!hyf_next:array[quarterword] of quarterword; {continuation of this |trie_op|}
@y
@!hyf_distance:array[1..trie_op_size] of small_number; {position |k-j| of $n_j$}
@!hyf_num:array[1..trie_op_size] of small_number; {value of $n_j$}
@!hyf_next:array[1..trie_op_size] of quarterword; {continuation code}
@!op_start:array[ASCII_code] of 0..trie_op_size; {offset for current language}
@z
@x module 922
@!v:quarterword; {an index into |hyf_distance|, etc.}
@y
@!v:integer; {an index into |hyf_distance|, etc.}
@z
@x module 923
for j:=0 to hn-2 do
  begin z:=hc[j]; l:=j;
@y
if trie_char(cur_lang+1)<>qi(cur_lang) then return; {no patterns for |cur_lang|}
for j:=0 to hn-r_hyf do
  begin z:=trie_link(cur_lang+1)+hc[j]; l:=j;
@z
@x ibid
found: hyf[1]:=0; hyf[hn-2]:=0; hyf[hn-1]:=0; hyf[hn]:=0
@y
found: for j:=0 to l_hyf do hyf[j]:=0;
for j:=0 to r_hyf do hyf[hn-j]:=0
@z
@x module 924
repeat i:=l-hyf_distance[v];
@y
repeat v:=v+op_start[cur_lang]; i:=l-hyf_distance[v];
@z
@x module 930
h:=hc[1];
@y
h:=hc[1]; incr(hn); hc[hn]:=cur_lang;
@z
@x ibid
not_found:
@y
not_found: decr(hn)
@z
@x module 931
  goto found;
@y
  decr(hn); goto found;
@z
@x module 934
@p procedure new_hyph_exceptions; {enters new exceptions}
label reswitch, exit, found, not_found, done;
@y
@d set_cur_lang==if language<=0 then cur_lang:=0
  else if language>255 then cur_lang:=0
  else cur_lang:=language

@p procedure new_hyph_exceptions; {enters new exceptions}
label reswitch, exit, found, not_found;
@z
@x ibid
begin scan_left_brace; {a left brace must follow \.{\\hyphenation}}
@y
begin scan_left_brace; {a left brace must follow \.{\\hyphenation}}
set_cur_lang;
@z
@x module 935
  spacer,right_brace: begin if n>4 then @<Enter a hyphenation exception@>;
@y
  spacer,right_brace: begin if n>1 then @<Enter a hyphenation exception@>;
@z
@x module 938
begin if n>1 then
@y
begin if n<63 then
@z
@x module 939
begin str_room(n); h:=0;
@y
begin incr(n); hc[n]:=cur_lang; str_room(n); h:=0;
@z
@x ibid
loop@+  begin if p=null then goto done;
  if info(p)<n-2 then goto done;
  q:=link(p); free_avail(p); p:=q; {eliminate hyphens that \TeX\ doesn't like}
  end;
done: @<Insert the \(p)pair |(s,p)| into the exception table@>;
@y
@<Insert the \(p)pair |(s,p)| into the exception table@>;
@z
@x module 942
@p@!init @<Declare procedures for preprocessing hyphenation patterns@>@;
@y
@<Declare subprocedures for |line_break|@>=
@!init @<Declare procedures for preprocessing hyphenation patterns@>@;
@z
@x module 943
@d quarterword_diff=max_quarterword-min_quarterword
@d trie_op_hash_size=quarterword_diff+quarterword_diff {double}

@<Glob...@>=
@!init@! trie_op_hash:array[0..trie_op_hash_size] of quarterword;
  {trie op codes for triples}
tini@;@/
@t\hskip1em@>@!trie_op_ptr:quarterword; {highest |trie_op| assigned}
@y
@<Glob...@>=
@!init@! trie_op_hash:array[-trie_op_size..trie_op_size] of 0..trie_op_size;
  {trie op codes for quadruples}
@!trie_used:array[ASCII_code] of quarterword;
  {largest opcode used so far for this language}
@!trie_op_lang:array[1..trie_op_size] of ASCII_code;
  {language part of a hashed quadruple}
@!trie_op_val:array[1..trie_op_size] of quarterword;
  {opcode corresponding to a hashed quadruple}
@!trie_op_ptr:0..trie_op_size; {number of stored ops so far}
tini
@z
@x module 944 should be entirely replaced
@y by the following code:
@ It's tempting to remove the |overflow| stops in the following procedure;
|new_trie_op| could return |min_quarterword| (thereby simply ignoring
part of a hyphenation pattern) instead of aborting the job. However, that would
lead to different hyphenation results on different installations of \TeX\
using the same patterns. The |overflow| stops are necessary for portability
of patterns.

@<Declare procedures for preprocessing hyph...@>=
function new_trie_op(@!d,@!n:small_number;@!v:quarterword):quarterword;
label exit;
var h:-trie_op_size..trie_op_size; {trial hash location}
@!u:quarterword; {trial op code}
@!l:0..trie_op_size; {pointer to stored data}
begin h:=abs(n+313*d+361*v+1009*cur_lang) mod (trie_op_size+trie_op_size)
  - trie_op_size;
loop@+  begin l:=trie_op_hash[h];
  if l=0 then {empty position found for a new op}
    begin if trie_op_ptr=trie_op_size then
      overflow("pattern memory ops",trie_op_size);
    u:=trie_used[cur_lang];
    if u=max_quarterword then
      overflow("pattern memory ops per language",
        max_quarterword-min_quarterword);
    incr(trie_op_ptr); incr(u); trie_used[cur_lang]:=u;
    hyf_distance[trie_op_ptr]:=d;
    hyf_num[trie_op_ptr]:=n; hyf_next[trie_op_ptr]:=v;
    trie_op_lang[trie_op_ptr]:=cur_lang; trie_op_hash[h]:=trie_op_ptr;
    trie_op_val[trie_op_ptr]:=u; new_trie_op:=u; return;
    end;
  if (hyf_distance[l]=d)and(hyf_num[l]=n)and(hyf_next[l]=v)
   and(trie_op_lang[l]=cur_lang) then
    begin new_trie_op:=trie_op_val[l]; return;
    end;
  if h>-trie_op_size then decr(h)@+else h:=trie_op_size;
  end;
exit:end;
@z
@x module 945: move the code from this module into 946, and introduce new code:
@y
@ After |new_trie_op| has compressed the necessary opcode information,
plenty of information is available to unscramble the data into the
final form needed by our hyphenation algorithm.

@<Sort the hyphenation op tables into proper order@>=
op_start[0]:=-min_quarterword;
for j:=1 to 255 do op_start[j]:=op_start[j-1]+qo(trie_used[j-1]);
for j:=1 to trie_op_ptr do
  trie_op_hash[j]:=op_start[trie_op_lang[j]]+trie_op_val[j]; {destination}
for j:=1 to trie_op_ptr do while trie_op_hash[j]>j do
  begin k:=trie_op_hash[j];@/
  t:=hyf_distance[k]; hyf_distance[k]:=hyf_distance[j]; hyf_distance[j]:=t;@/
  t:=hyf_num[k]; hyf_num[k]:=hyf_num[j]; hyf_num[j]:=t;@/
  t:=hyf_next[k]; hyf_next[k]:=hyf_next[j]; hyf_next[j]:=t;@/
  trie_op_hash[j]:=trie_op_hash[k]; trie_op_hash[k]:=k;
  end
@z
@x module 949: Move this code to just after the new module 945!
mentioned so far, let's write a procedure that does the initialization.

@<Declare procedures for preprocessing hyph...@>=
procedure init_pattern_memory; {gets ready to build a linked trie}
var h:0..trie_op_hash_size; {an index into |trie_op_hash|}
@!p:trie_pointer; {an index into |trie_hash|}
begin for h:=0 to trie_op_hash_size do trie_op_hash[h]:=min_quarterword;
trie_op_ptr:=min_quarterword; trie_root:=0; trie_c[0]:=0; trie_ptr:=0;
for p:=0 to trie_size do trie_hash[p]:=0;
end;
@y
mentioned so far, let's write down the code that gets them started.

@<Initialize table entries...@>=
for k:=-trie_op_size to trie_op_size do trie_op_hash[k]:=0;
for k:=0 to 255 do trie_used[k]:=min_quarterword;
trie_op_ptr:=0;
@z
@x module 950
@t\hskip1em@>@!trie_min:trie_pointer;
  {all locations |<=trie_min| are vacant in |trie|}
tini@;@/
@t\hskip1em@>@!trie_max:trie_pointer; {largest location used in |trie|}
@y
@t\hskip10pt@>@!trie_min:array[ASCII_code] of trie_pointer;
  {the first possible slot for each character}
@t\hskip10pt@>@!trie_max:trie_pointer; {largest location used in |trie|}
@t\hskip10pt@>@!trie_not_ready:boolean; {is the trie still in linked form?}
tini
@z
@x modules 951 and 952
@ Here is how these data structures are initialized.

@<Declare procedures for preprocessing hyph...@>=
procedure init_trie_memory; {gets ready to pack into |trie|}
var p:trie_pointer; {index into |trie_ref|, |trie|, |trie_taken|}
begin for p:=0 to trie_ptr do trie_ref[p]:=0;
trie_max:=256; trie_min:=256; trie_link(0):=1; trie_taken[0]:=false;
trie_link(trie_size):=0; trie_back(0):=trie_size; {wrap around}
for p:=1 to 256 do
  begin trie_back(p):=p-1; trie_link(p):=p+1; trie_taken[p]:=false;
  end;
end;

@ Each time \.{\\patterns} appears, it overrides any patterns that were
entered earlier, so the arrays are not initialized until \TeX\ sees
\.{\\patterns}. However, some of the global variables must be
initialized when \.{INITEX} is loaded, in case the user never mentions
any \.{\\patterns}.

@<Initialize table entries...@>=
trie_op_ptr:=min_quarterword;@/
trie_link(0):=0; trie_char(0):=min_quarterword; trie_op(0):=min_quarterword;
for k:=1 to 255 do trie[k]:=trie[0];
trie_max:=255;
@y
@ Each time \.{\\patterns} appears, it contributes further patterns to
the future trie, which will be built only when hyphenation is attempted or
when a format file is dumped. The boolean variable |trie_not_ready|
will change to |false| when the trie is compressed; this will disable
further patterns.

@<Initialize table entries...@>=
trie_not_ready:=true; trie_root:=0; trie_c[0]:=si(0); trie_ptr:=0;

@ Here is how the trie-compression data structures are initialized.
If storage is tight, it would be possible to overlap |trie_op_hash|,
|trie_op_lang|, and |trie_op_val| with |trie|, |trie_hash|, and |trie_taken|,
because we finish with the former just before we need the latter.

@<Get ready to compress the trie@>=
@<Sort the hyphenation...@>;
for p:=0 to trie_size do trie_hash[p]:=0;
trie_root:=compress_trie(trie_root); {identify equivalent subtries}
for p:=0 to trie_ptr do trie_ref[p]:=0;
for p:=0 to 255 do trie_min[p]:=p+1;
trie_link(0):=1; trie_max:=0
@z
@x module 953
begin c:=so(trie_c[p]);
if c<trie_min then trie_min:=c;
if trie_min=0 then z:=trie_link(trie_size)
else z:=trie_link(trie_min-1); {get the first conceivably good hole}
loop@+  begin if z<c then goto not_found;
  h:=z-c;@/
@y
@!l,@!r:trie_pointer; {left and right neighbors}
@!ll:1..256; {upper limit of |trie_min| updating}
begin c:=so(trie_c[p]);
z:=trie_min[c]; {get the first conceivably good hole}
loop@+  begin h:=z-c;@/
@z
@x module 956
repeat z:=h+so(trie_c[q]); trie_back(trie_link(z)):=trie_back(z);
trie_link(trie_back(z)):=trie_link(z); trie_link(z):=0; q:=trie_r[q];
@y
repeat z:=h+so(trie_c[q]); l:=trie_back(z); r:=trie_link(z);
trie_back(r):=l; trie_link(l):=r; trie_link(z):=0;
if l<256 then
  begin if z<256 then ll:=z @+else ll:=256;
  repeat trie_min[l]:=r; incr(l);
  until l=ll;
  end;
q:=trie_r[q];
@z
@x module 958 is entirely replaced
@y by the following new code:
@ When the whole trie has been allocated into the sequential table, we
must go through it once again so that |trie| contains the correct
information. Null pointers in the linked trie will be represented by the
value~0, which properly implements an ``empty'' family.

@<Move the data into |trie|@>=
h.rh:=0; h.b0:=min_quarterword; h.b1:=min_quarterword; {|trie_link:=0|,
  |trie_op:=min_quarterword|, |trie_char:=qi(0)|}
if trie_root=0 then {no patterns were given}
  begin for r:=0 to 256 do trie[r]:=h;
  trie_max:=256;
  end
else begin trie_fix(trie_root); {this fixes the non-holes in |trie|}
  r:=0; {now we will zero out all the holes}
  repeat s:=trie_link(r); trie[r]:=h; r:=s;
  until r>trie_max;
  end;
trie_char(0):=qi("?"); {make |trie_char(c)<>c| for all |c|}
@z
@x module 960
@!r,@!s:trie_pointer; {used to clean up the packed |trie|}
@!h:two_halves; {template used to zero out |trie|'s holes}
begin scan_left_brace; {a left brace must follow \.{\\patterns}}
init_pattern_memory;@/
@<Enter all of the patterns into a linked trie, until coming to a right
  brace@>;
trie_root:=compress_trie(trie_root); {compress the trie}
@<Pack the trie@>;
end;
@y
begin if trie_not_ready then
  begin set_cur_lang; scan_left_brace; {a left brace must follow \.{\\patterns}}
  @<Enter all of the patterns into a linked trie, until coming to a right
  brace@>;
  end
else begin print_err("Too late for "); print_esc("patterns");
  help1("All patterns must be given before typesetting begins.");
  error; link(garbage):=scan_toks(false,false); flush_list(def_ref);
  end;
end;
@z
@x module 963
q:=0;
while l<k do
  begin incr(l); c:=hc[l]; p:=trie_l[q]; first_child:=true;
@y
q:=0; hc[0]:=cur_lang;
while l<=k do
  begin c:=hc[l]; incr(l); p:=trie_l[q]; first_child:=true;
@z
@x module 966 now says that the first call ``takes'' location 1,
@y and the Pascal code in this module is entirely replaced by the following:
@<Declare procedures for preprocessing hyphenation patterns@>=
procedure init_trie;
var @!p:trie_pointer; {pointer for initialization}
@!j,@!k,@!t:integer; {all-purpose registers for initialization}
@!r,@!s:trie_pointer; {used to clean up the packed |trie|}
@!h:two_halves; {template used to zero out |trie|'s holes}
begin @<Get ready to compress the trie@>;
if trie_root<>0 then
  begin first_fit(trie_root); trie_pack(trie_root);
  end;
@<Move the data into |trie|@>;
trie_not_ready:=false;
end;
@z
@x module 1033 gets new code before main_loop_1
@y
if mode>0 then if language<>clang then fix_language;
@z
@x modules 1091 and 1200
push_nest; mode:=hmode; space_factor:=1000;
@y
push_nest; mode:=hmode; space_factor:=1000; clang:=0;
@z
@x module 1324
for k:=0 to trie_max do dump_hh(trie[k]);
dump_int(trie_op_ptr);
for k:=min_quarterword+1 to trie_op_ptr do
  begin dump_int(hyf_distance[k]);
  dump_int(hyf_num[k]);
  dump_int(hyf_next[k]);
  end;
print_ln; print_int(hyph_count); print(" hyphenation exception");
if hyph_count<>1 then print_char("s");
print_nl("Hyphenation trie of length "); print_int(trie_max);
@.Hyphenation trie...@>
print(" has "); print_int(qo(trie_op_ptr)); print(" op");
if trie_op_ptr<>min_quarterword+1 then print_char("s")
@y
print_ln; print_int(hyph_count); print(" hyphenation exception");
if hyph_count<>1 then print_char("s");
if trie_not_ready then init_trie;
dump_int(trie_max);
for k:=0 to trie_max do dump_hh(trie[k]);
dump_int(trie_op_ptr);
for k:=1 to trie_op_ptr do
  begin dump_int(hyf_distance[k]);
  dump_int(hyf_num[k]);
  dump_int(hyf_next[k]);
  end;
print_nl("Hyphenation trie of length "); print_int(trie_max);
@.Hyphenation trie...@>
print(" has "); print_int(trie_op_ptr); print(" op");
if trie_op_ptr<>1 then print_char("s");
print(" out of "); print_int(trie_op_size);
for k:=255 downto 0 do if trie_used[k]>min_quarterword then
  begin print_nl("  "); print_int(qo(trie_used[k]));
  print(" for language "); print_int(k);
  dump_int(k); dump_int(qo(trie_used[k]));
  end
@z
@x module 1325
undump_size(0)(trie_size)('trie size')(trie_max);
for k:=0 to trie_max do undump_hh(trie[k]);
undump(min_quarterword)(max_quarterword)(trie_op_ptr);
for k:=min_quarterword+1 to trie_op_ptr do
  begin undump(0)(63)(hyf_distance[k]); {a |small_number|}
  undump(0)(63)(hyf_num[k]);
  undump(min_quarterword)(max_quarterword)(hyf_next[k]);
  end
@y
undump_size(0)(trie_size)('trie size')(j); {|trie_max|}
for k:=0 to j do undump_hh(trie[k]);
undump_size(0)(trie_op_size)('trie op size')(j); {|trie_op_ptr|}
for k:=1 to j do
  begin undump(0)(63)(hyf_distance[k]); {a |small_number|}
  undump(0)(63)(hyf_num[k]);
  undump(min_quarterword)(max_quarterword)(hyf_next[k]);
  end;
k:=256;
while j>0 do
  begin undump(0)(k-1)(k); undump(1)(j)(x); j:=j-x; op_start[k]:=qo(j);
  end;
@!init trie_not_ready:=false @+tini
@z
@x module 1341 gets two new definitions
@y
@d language_node=4 {|subtype| in whatsits that change the current language}
@d stored_language(#)==mem[#+1].int {language number, in the range |0..255|}
@z
@x module 1344 gets a new definition and a new Pascal statement
@y
@d set_language_code=5 {command modifier for \.{\\setlanguage}}
primitive("setlanguage",extension,set_language_code);@/
@!@:set_language_}{\.{\\setlanguage} primitive@>
@z
@x module 1346 gets a new case
@y
  set_language_code:print_esc("setlanguage");
@z
@x module 1348 gets a new case
@y
set_language_code:@<Implement \.{\\setlanguage}@>;
@z
@x module 1356 gets a new case
@y
language_node:begin print_esc("setlanguage");
  print_int(stored_language(p));
  end;
@z
% in modules 1357 and 1358, change "close_node" to "close_node,language_node".
@x module 1362 becomes two modules
@ @<Advance \(p)past a whatsit node in the |line_break| loop@>=do_nothing
@y
@ @<Advance \(p)past a whatsit node in the \(l)|line_break| loop@>=
if subtype(cur_p)=language_node then cur_lang:=stored_language(cur_p)

@ @<Advance \(p)past a whatsit node in the \(p)pre-hyphenation loop@>=
if subtype(s)=language_node then cur_lang:=stored_language(s)
@z and old module 1367 is moved to just before the old module 1378
@x module 1373 gets a new case
@y
language_node:do_nothing;
@z
@x new modules before the system-dependent changes (i.e. before the old 1376)
@y
@ @<Implement \.{\\setlanguage}@>=
if abs(mode)<>hmode then report_illegal_case
else begin new_whatsit(language_node,small_node_size);
  scan_int;
  if cur_val<=0 then clang:=0
  else if cur_val>255 then clang:=0
  else clang:=cur_val;
  stored_language(tail):=clang;
  end

@ Finally, we need a subroutine that comes into play when a character of
a non-|clang| language is being appended to the current paragraph.

@<Declare action...@>=
procedure fix_language;
var @!l:ASCII_code; {the new current language}
begin if language<=0 then l:=0
else if language>255 then l:=0
else l:=language;
if l<>clang then
  begin new_whatsit(language_node,small_node_size);
  stored_language(tail):=l; clang:=l;
  end;
end;
@z

362. Major extension to ligature capability.
@x module 143
a linked list of character nodes for those characters.
@y
a linked list of character nodes for all original characters that have been
deleted. (This list might be empty if the characters that generated the
ligature were retained in other nodes.)

The |subtype| field is 0, plus 2 and/or 1 if the original source of the
ligature included implicit left and/or right boundaries.
@z
@x in module 144 add the following procedure to the existing code
@y
function new_lig_item(@!c:quarterword):pointer;
var p:pointer; {the new node}
begin p:=get_node(small_node_size); character(p):=c; lig_ptr(p):=null;
new_lig_item:=p;
end;
@z
@x module 193
font_in_short_display:=font(lig_char(p));
short_display(lig_ptr(p)); print_char(")");
@y
if subtype(p)>1 then print_char("|");
font_in_short_display:=font(lig_char(p)); short_display(lig_ptr(p));
if odd(subtype(p)) then print_char("|");
print_char(")");
end
@z
@x modules 208 and 209
@d radical=65 {square root and similar signs ( \.{\\radical} )}
@y
@d no_boundary=65 {suppress boundary ligatures ( \.{\\noboundary} )}
@d radical=66 {square root and similar signs ( \.{\\radical} )}
@z and so on, adding 1 to each definition until getting to max_command=100
@x module 265 gets a new statement
@y
primitive("noboundary",no_boundary,0);@/
@!@:no_boundary_}{\.{\\noboundary} primitive@>
@z
@x module 266 gets a new case
@y
no_boundary:print_esc("noboundary");
@z
@x module 545 is entirely replaced
@y by the following new specifications:
@ The |lig_kern| array contains instructions in a simple programming language
that explains what to do for special letter pairs. Each word in this array is a
|@!lig_kern_command| of four bytes.

\yskip\hang first byte: |skip_byte|, indicates that this is the final program
  step if the byte is 128 or more, otherwise the next step is obtained by
  skipping this number of intervening steps.\par
\hang second byte: |next_char|, ``if |next_char| follows the current character,
  then perform the operation and stop, otherwise continue.''\par
\hang third byte: |op_byte|, indicates a ligature step if less than~128,
  a kern step otherwise.\par
\hang fourth byte: |remainder|.\par
\yskip\noindent
In a kern step, an
additional space equal to |kern[256*(op_byte-128)+remainder]| is inserted
between the current character and |next_char|. This amount is
often negative, so that the characters are brought closer together
by kerning; but it might be positive.

There are eight kinds of ligature steps, having |op_byte| codes $4a+2b+c$ where
$0\le a\le b+c$ and $0\le b,c\le1$. The character whose code is
|remainder| is inserted between the current character and |next_char|;
then the current character is deleted if $b=0$, and |next_char| is
deleted if $c=0$; then we pass over $a$~characters to reach the next
current character (which may have a ligature/kerning program of its own).

If the very first instruction of the |lig_kern| array has |skip_byte=255|,
the |next_char| byte is the so-called right boundary character of this font;
the value of |next_char| need not lie between |bc| and~|ec|.
If the very last instruction of the |lig_kern| array has |skip_byte=255|,
there is a special ligature/kerning program for a left boundary character,
beginning at location |256*op_byte+remainder|.
The interpretation is that \TeX\ puts implicit boundary characters
before and after each consecutive string of characters from the same font.
These implicit characters do not appear in the output, but they can affect
ligatures and kerning.

If the very first instruction of a character's |lig_kern| program has
|skip_byte>128|, the program actually begins in location
|256*op_byte+remainder|. This feature allows access to large |lig_kern|
arrays, because the first instruction must otherwise
appear in a location |<=255|.

Any instruction with |skip_byte>128| in the |lig_kern| array must have
|256*op_byte+remainder<nl|. If such an instruction is encountered during
normal program execution, it denotes an unconditional halt; no ligature
command is performed.

@d stop_flag=qi(128) {value indicating `\.{STOP}' in a lig/kern program}
@d kern_flag=qi(128) {op code for a kern step}
@d skip_byte(#)==#.b0
@d next_char(#)==#.b1
@d op_byte(#)==#.b2
@d rem_byte(#)==#.b3
@z
@x module 549
@<Glob...@>=
@y
@d non_char==qi(256) {a |halfword| code that can't match a real character}
@d non_address==font_mem_size {a spurious |font_index|}

@<Glob...@>=
@z
@x module 548 gets a new type definition
@y
@!font_index=0..font_mem_size;
@z
@x module 549
  {current \.{\\skewchar} values}
@y
  {current \.{\\skewchar} values}
@!bchar_label:array[internal_font_number] of font_index;
  {start of |lig_kern| program for left boundary character,
  |non_address| if there is none}
@!font_bchar:array[internal_font_number] of min_quarterword..non_char;
  {right boundary character, |non_char| if there is none}
@!font_false_bchar:array[internal_font_number] of min_quarterword..non_char;
  {|font_bchar| if it doesn't exist in the font, otherwise |non_char|}
@z
@x module 557
@d char_kern_end(#)==rem_byte(#)].sc
@y NOTE: Optimize kern_base_offset in your change file! It's a constant.
@d char_kern_end(#)==256*op_byte(#)+rem_byte(#)].sc
@d kern_base_offset==256*(kern_flag)
@d lig_kern_restart_end(#)==256*op_byte(#)+rem_byte(#)+32768-kern_base_offset
@d lig_kern_restart(#)==lig_kern_base[#]+lig_kern_restart_end
@z
@x module 560
@!z:scaled; {the design size or the ``at'' size}
@y
@!bch_label:integer; {left boundary start location, or infinity}
@!bchar:0..256; {right boundary character, or 256}
@!z:scaled; {the design size or the ``at'' size}
@z
@x module 566
kern_base[f]:=lig_kern_base[f]+nl;
exten_base[f]:=kern_base[f]+nk;
@y
kern_base[f]:=lig_kern_base[f]+nl-kern_base_offset;
exten_base[f]:=kern_base[f]+kern_base_offset+nk;
@z
@x module 573: The entire module is replaced
@y by the following code.
@ @d check_existence(#)==@t@>@;@/
  begin check_byte_range(#);
  qw:=char_info(f)(#); {N.B.: not |qi(#)|}
  if not char_exists(qw) then abort;
  end

@<Read ligature/kern program@>=
bch_label:=@'77777; bchar:=256;
if nl>0 then
  begin for k:=lig_kern_base[f] to kern_base[f]+kern_base_offset-1 do
    begin store_four_quarters(font_info[k].qqqq);
    if a>128 then
      begin if 256*c+d>=nl then abort;
      if a=255 then if k=lig_kern_base[f] then bchar:=b;
      end
    else begin if b<>bchar then check_existence(b);
      if c<128 then check_existence(d) {check ligature}
      else if 256*(c-128)+d>=nk then abort; {check kern}
      if a<128 then if k-lig_kern_base[f]+a+1>=nl then abort;
      end;
    end;
  if a=255 then bch_label:=256*c+d;
  end;
for k:=kern_base[f]+kern_base_offset to exten_base[f]-1 do
  store_scaled(font_info[k].sc);
@z
@x module 574 (actually a bugfix)
  if a<>0 then check_byte_range(a);
  if b<>0 then check_byte_range(b);
  if c<>0 then check_byte_range(c);
  check_byte_range(d);
@y
  if a<>0 then check_existence(a);
  if b<>0 then check_existence(b);
  if c<>0 then check_existence(c);
  check_existence(d);
@z
@x module 576
font_name[f]:=nom;
@y
if bch_label<nl then bchar_label[f]:=bch_label+lig_kern_base[f]
else bchar_label[f]:=non_address;
font_bchar[f]:=qi(bchar);
font_false_bchar[f]:=qi(bchar);
if bchar<=ec then if bchar>=bc then
  begin qw:=char_info(f)(bchar); {N.B.: not |qi(bchar)|}
  if char_exists(qw) then font_false_bchar[f]:=non_char;
  end;
font_name[f]:=nom;
@z
@x module 708 (slight but optional optimization)
continue: if (qo(y)>=font_bc[g])and(qo(y)<=font_ec[g]) then
  begin q:=char_info(g)(y);
@y
if (qo(y)>=font_bc[g])and(qo(y)<=font_ec[g]) then
  begin continue: q:=char_info(g)(y);
@z
@x module 740 (another bugfix)
  i:=char_info(f)(y);
@y
  i:=char_info(f)(y);
  if not char_exists(i) then goto done;
@z
@x module 741
    repeat cur_i:=font_info[a].qqqq;
    if qo(next_char(cur_i))=skew_char[cur_f] then
      begin if op_bit(cur_i)>=kern_flag then
        s:=char_kern(cur_f)(cur_i);
      goto done1;
      end;
    incr(a);
    until stop_bit(cur_i)>=stop_flag;
@y
    cur_i:=font_info[a].qqqq;
    if skip_byte(cur_i)>stop_flag then
      begin a:=256*(qo(op_byte(cur_i)))+rem_byte(cur_i);
      cur_i:=font_info[a].qqqq;
      end;
    loop begin if qo(next_char(cur_i))=skew_char[cur_f] then
        begin if op_byte(cur_i)>=kern_flag then
          if skip_byte(cur_i)<=stop_flag then s:=char_kern(cur_f)(cur_i);
        goto done1;
        end;
      if skip_byte(cur_i)>=stop_flag then goto done1
      else a:=a+qo(skip_byte(cur_i))+1;
      cur_i:=font_info[a].qqqq;
      end;
@z
@x module 749 (yet another bugfix of the same type!)
@!p,@!v,@!x,@!y,@!z:pointer; {temporary registers for box construction}
@y
@!p,@!v,@!x,@!y,@!z:pointer; {temporary registers for box construction}
@!c:quarterword;@+@!i:four_quarters; {registers for character examination}
@z
@x ibid
    begin cur_c:=rem_byte(cur_i); character(nucleus(q)):=cur_c;
    cur_i:=char_info(cur_f)(cur_c);
@y
    begin c:=rem_byte(cur_i); i:=char_info(cur_f)(c);
    if char_exists(i) then
      begin cur_c:=c; cur_i:=i; character(nucleus(q)):=c;
      end;
@z
@x module 752
@!p:pointer; {temporary register for list manipulation}
@y
@!p,@!r:pointer; {temporary registers for list manipulation}
@z
@x ibid
        repeat cur_i:=font_info[a].qqqq;@/
        @<If instruction |cur_i| is a kern with |cur_c|,
         attach the kern after |q| and |return|;
         or if it is a ligature with |cur_c|, combine
         noads |q| and |p| and |goto restart|@>;
        incr(a);
        until stop_bit(cur_i)>=stop_flag;
@y
        cur_i:=font_info[a].qqqq;
        if skip_byte(cur_i)>stop_flag then
          begin a:=lig_kern_restart(cur_f)(cur_i);
          cur_i:=font_info[a].qqqq;
          end;
        loop@+ begin @<If instruction |cur_i| is a kern with |cur_c|, attach
            the kern after~|q|; or if it is a ligature with |cur_c|, combine
            noads |q| and~|p| appropriately; then |return| if the cursor has
            moved past a noad, or |goto restart|@>;
          if skip_byte(cur_i)>=stop_flag then return;
          a:=a+qo(skip_byte(cur_i))+1;
          cur_i:=font_info[a].qqqq;
          end;
@z
@x The entire code of module 753 is revised
@y and should be replaced by the following:
@ Note that a ligature between an |ord_noad| and another kind of noad
is replaced by an |ord_noad|, when the two noads collapse into one.
But we could make a parenthesis (say) change shape when it follows
certain letters. Presumably a font designer will define such
ligatures only when this convention makes sense.

\chardef\?='174 % vertical line to indicate character retention

@<If instruction |cur_i| is a kern with |cur_c|, ...@>=
if next_char(cur_i)=cur_c then if skip_byte(cur_i)<=stop_flag then
  if op_byte(cur_i)>=kern_flag then
    begin p:=new_kern(char_kern(cur_f)(cur_i));
    link(p):=link(q); link(q):=p; return;
    end
  else  begin check_interrupt; {allow a way out of infinite ligature loop}
      case op_byte(cur_i) of
   qi(1),qi(5): character(nucleus(q)):=rem_byte(cur_i); {\.{=:\?}, \.{=:\?>}}
   qi(2),qi(6): character(nucleus(p)):=rem_byte(cur_i); {\.{\?=:}, \.{\?=:>}}
   qi(3),qi(7),qi(11):begin r:=new_noad; {\.{\?=:\?}, \.{\?=:\?>}, \.{\?=:\?>>}}
       character(nucleus(r)):=rem_byte(cur_i);
       fam(nucleus(r)):=fam(nucleus(q));@/
       link(q):=r; link(r):=p;
       if op_byte(cur_i)<qi(11) then math_type(nucleus(r)):=math_char
       else math_type(nucleus(r)):=math_text_char; {prevent combination}
       end;
     othercases begin link(q):=link(p);
       character(nucleus(q)):=rem_byte(cur_i); {\.{=:}}
       mem[subscr(q)]:=mem[subscr(p)]; mem[supscr(q)]:=mem[supscr(p)];@/
       free_node(p,noad_size);
       end
     endcases;
     if op_byte(cur_i)>qi(3) then return;
     math_type(nucleus(q)):=math_char; goto restart;
     end
@z
@x module 862
@!q,@!r,@!s:pointer; {miscellaneous nodes of temporary interest}
@y
@!q,@!r,@!s,@!prev_s:pointer; {miscellaneous nodes of temporary interest}
@z
@x module 892
nodes $p_a$ and~$p_b$ in the description above are placed into variables
@y
nodes $p_{a-1}$ and~$p_b$ in the description above are placed into variables
@z
@x ibid
@!hc:array[0..65] of halfword; {word to be hyphenated}
@y
@!hc:array[0..65] of 0..256; {word to be hyphenated}
@z
@x ibid
@!hu:array[1..63] of ASCII_code; {like |hc|, before conversion to lowercase}
@y
@!hu:array[0..63] of 0..256; {like |hc|, before conversion to lowercase}
@z
@x module 894
s:=link(cur_p);
@y
prev_s:=cur_p; s:=link(prev_s);
@z
@x module 895
label done,found,not_found,found1,exit;
@y
label common_ending,done,found,found1,not_found,not_found+1,exit;
@z
@x module 896
  else if type(s)=ligature_node then
    begin q:=lig_ptr(s); c:=qo(character(q)); hf:=font(q);
    end
@y
  else if type(s)=ligature_node then
    if lig_ptr(s)=null then goto continue
    else begin q:=lig_ptr(s); c:=qo(character(q)); hf:=font(q);
      end
@z
@x ibid
continue: s:=link(s);
@y
continue: prev_s:=s; s:=link(prev_s);
@z
@x ibid
ha:=s
@y
ha:=prev_s
@z
@x module 898
begin j:=hn; q:=lig_ptr(s);
if font(q)<>hf then goto done3;
repeat c:=qo(character(q));
if lc_code(c)=0 then goto done3;
if j=63 then goto done3;
incr(j); hu[j]:=c; hc[j]:=lc_code(c);@/
q:=link(q);
until q=null;
@y
begin if font(lig_char(s))<>hf then goto done3;
j:=hn; q:=lig_ptr(s);
while q>null do
  begin c:=qo(character(q));
  if lc_code(c)=0 then goto done3;
  if j=63 then goto done3;
  incr(j); hu[j]:=c; hc[j]:=lc_code(c);@/
  q:=link(q);
  end;
@z
@x module 900 gets three new global variables
@y
@!init_list:pointer; {list of punctuation characters preceding the word}
@!init_lig:boolean; {does |init_list| represent a ligature?}
@!init_lft:boolean; {if so, did the ligature involve a left boundary?}
@z
@x module 901
@!q,@!r,@!s:pointer; {temporary registers for list manipulation}
@y
@!p,@!q,@!r,@!s:pointer; {temporary registers for list manipulation}
@!bchar:halfword; {right boundary character of hyphenated word, or |non_char|}
@z
@x module 903 should be entirely replaced
@y by the following:
@ If hyphens are in fact going to be inserted, \TeX\ first deletes the
subsequence of nodes between |ha| and~|hb|. An attempt is made to
preserve the effect that implicit boundary characters and punctuation marks
had on ligatures inside the hyphenated word, by storing a left boundary or
preceding character in |hu[0]| and by storing a possible right boundary
in |bchar|. We set |j:=0| if |hu[0]| is to be part of the reconstruction;
otherwise |j:=1|.
The variable |s| will point to the tail of the current hlist, and
|q| will point to the node following |hb|, so that
things can be hooked up after we reconstitute the hyphenated word.

@<Replace nodes |ha..hb| by a sequence of nodes...@>=
q:=link(hb); link(hb):=null; r:=link(ha); link(ha):=null; bchar:=non_char;
if type(hb)=ligature_node then if odd(subtype(hb)) then
  bchar:=font_bchar[hf];
if is_char_node(ha) then
  begin init_list:=ha; init_lig:=false; hu[0]:=qo(character(ha));
  end
else if type(ha)=ligature_node then
  begin init_list:=lig_ptr(ha); init_lig:=true; init_lft:=(subtype(ha)>1);
  hu[0]:=qo(character(lig_char(ha)));
  if init_list=null then if init_lft then
    begin hu[0]:=256; init_lig:=false;
    end; {in this case a ligature will be reconstructed from scratch}
  free_node(ha,small_node_size);
  end
else goto not_found+1; {no punctuation found}
s:=cur_p; {we have |cur_p<>ha| because |type(cur_p)=glue_node|}
while link(s)<>ha do s:=link(s);
j:=0; goto common_ending;
not_found+1: j:=1; s:=ha; init_list:=null;
if not is_char_node(r) then if type(r)=ligature_node then
 if subtype(r)>1 then
  begin j:=0; hu[0]:=256; init_lig:=false;
  end;
common_ending: flush_node_list(r);
@<Reconstitute nodes for the hyphenated word, inserting discretionary hyphens@>;
flush_list(init_list)
@z
@x modules 905--911 are entirely replaced
@y by the following new code:
Still further complications arise in the presence of ligatures that do not
delete the original characters. When punctuation precedes the word being
hyphenated, \TeX's method is not perfect under all possible scenarios,
because punctuation marks and letters can propagate information back and forth.
For example, suppose the original pre-hyphenation pair
\.{*a} changes to \.{*y} via a \.{\?=:} ligature, which changes to \.{xy}
via a \.{=:\?} ligature; if $p_{a-1}=\.x$ and $p_a=\.y$, the reconstitution
procedure isn't smart enough to obtain \.{xy} again. In such cases the
font designer should include a ligature that goes from \.{xa} to \.{xy}.

@ The processing is facilitated by a subroutine called |reconstitute|. Given
a string of characters $x_j\ldots x_n$, there is a smallest index $m\ge j$
such that the ``translation'' of $x_j\ldots x_n$ by ligatures and kerning
has the form $y_1\ldots y_t$ followed by the translation of $x_{m+1}\ldots x_n$,
where $y_1\ldots y_t$ is some nonempty sequence of character, ligature, and
kern nodes. We call $x_j\ldots x_m$ a ``cut prefix'' of $x_j\ldots x_n$.
For example, if $x_1x_2x_3=\.{fly}$, and if the font contains `fl' as a
ligature and a kern between `fl' and `y', then $m=2$, $y=2$, and $y_1$ will
be a ligature node for `fl' followed by an appropriate kern node~$y_2$.
In the most common case, $x_j$~forms no ligature with $x_{j+1}$ and we
simply have $m=j$, $y_1=x_j$. If $m<n$ we can repeat the procedure on
$x_{m+1}\ldots x_n$ until the entire translation has been found.

The |reconstitute| function returns the integer $m$ and puts the nodes
$y_1\ldots y_t$ into a linked list starting at |link(hold_head)|,
getting the input $x_j\ldots x_n$ from the |hu| array. If $x_j=256$,
we consider $x_j$ to be an implicit left boundary character; in this
case |j| must be strictly less than~|n|. There is a
parameter |bchar|, which is either 256 or an implicit right boundary character
assumed to be present just following~$x_n$. (The value |hu[n+1]| is never
explicitly examined, but the algorithm imagines that |bchar| is there.)

If there exists an index |k| in the range $j\le k\le m$ such that |hyf[k]|
is odd and such that the result of reconstitute would have been different
if $x_{k+1}$ had been |hchar|, then |reconstitute| sets |hyphen_passed|
to the smallest such~|k|. Otherwise it sets |hyphen_passed| to zero.

A special convention is used in the case |j=0|: Then we assume that the
translation of |hu[0]| appears in a special list of charnodes starting at
|init_list|; moreover, if |init_lig| is |true|, then |hu[0]| will be
a ligature character, involving a left boundary if |init_lft| is |true|.
This facility is provided for cases when a hyphenated
word is preceded by punctuation (like single or double quotes) that might
affect the translation of the beginning of the word.

@<Glob...@>=
@!hyphen_passed:small_number; {first hyphen in a ligature, if any}

@ @<Declare the function called |reconstitute|@>=
function reconstitute(@!j,@!n:small_number;@!bchar,@!hchar:halfword):
  small_number;
label continue,done;
var @!p:pointer; {temporary register for list manipulation} 
@!t:pointer; {a node being appended to}
@!q:four_quarters; {character information or a lig/kern instruction}
@!cur_rh:halfword; {hyphen character for ligature testing}
@!test_char:halfword; {hyphen or other character for ligature testing}
@!w:scaled; {amount of kerning}
@!k:font_index; {position of current lig/kern instruction}
begin hyphen_passed:=0; t:=hold_head; w:=0; link(hold_head):=null;
 {at this point |ligature_present=lft_hit=rt_hit=false|}
@<Set up data structures with the cursor following position |j|@>;
continue:@<If there's a ligature or kern at the cursor position, update the data
  structures, possibly advancing~|j|; continue until the cursor moves@>;
@<Append a ligature and/or kern to the translation;
  |goto continue| if the stack of inserted ligatures is nonempty@>;
reconstitute:=j;
end;

@ The reconstitution procedure shares many of the global data structures
by which \TeX\ has processed the words before they were hyphenated.
There is an implied ``cursor'' between characters |cur_l| and |cur_r|;
these characters will be tested for possible ligature activity. If
|ligature_present| then |cur_l| is a ligature character formed from the
original characters following |cur_q| in the current translation list.
There is a ``ligature stack'' between the cursor and character |j+1|,
consisting of pseudo-ligature nodes linked together by their |link| fields.
This stack is normally empty unless a ligature command has created a new
character that will need to be processed later. A pseudo-ligature is
a special node having a |character| field that represents a potential
ligature and a |lig_ptr| field that points to a |char_node| or is |null|.
We have
$$|cur_r|=\cases{|character(lig_stack)|,&if |lig_stack>null|;\cr
  |qi(hu[j+1])|,&if |lig_stack=null| and |j<n|;\cr
  bchar,&if |lig_stack=null| and |j=n|.\cr}$$

@<Glob...@>=
@!cur_l,@!cur_r:halfword; {characters before and after the cursor}
@!cur_q:pointer; {where a ligature should be detached}
@!lig_stack:pointer; {unfinished business to the right of the cursor}
@!ligature_present:boolean; {should a ligature node be made for |cur_l|?}
@!lft_hit,@!rt_hit:boolean; {did we hit a ligature with a boundary character?}

@ @d append_charnode_to_t(#)== begin link(t):=get_avail; t:=link(t);
    font(t):=hf; character(t):=#;
    end
@d set_cur_r==begin if j<n then cur_r:=qi(hu[j+1])@+else cur_r:=bchar;
    if odd(hyf[j]) then cur_rh:=hchar@+else cur_rh:=non_char;
    end

@<Set up data structures with the cursor following position |j|@>=
cur_l:=qi(hu[j]); cur_q:=t;
if j=0 then
  begin ligature_present:=init_lig; p:=init_list;
  if ligature_present then lft_hit:=init_lft;
  while p>null do
    begin append_charnode_to_t(character(p)); p:=link(p);
    end;
  end
else if cur_l<non_char then append_charnode_to_t(cur_l);
lig_stack:=null; set_cur_r

@ We may want to look at the lig/kern program twice, once for a hyphen
and once for a normal letter. (The hyphen might appear after the letter
in the program, so we'd better not try to look for both at once.)

@<If there's a ligature or kern at the cursor position, update...@>=
if cur_l=non_char then
  begin k:=bchar_label[hf];
  if k=non_address then goto done@+else q:=font_info[k].qqqq;
  end
else begin q:=char_info(hf)(cur_l);
  if char_tag(q)<>lig_tag then goto done;
  k:=lig_kern_start(hf)(q); q:=font_info[k].qqqq;
  if skip_byte(q)>stop_flag then
    begin k:=lig_kern_restart(hf)(q); q:=font_info[k].qqqq;
    end;
  end; {now |k| is the starting address of the lig/kern program}
if cur_rh<non_char then test_char:=cur_rh@+else test_char:=cur_r;
loop@+begin if next_char(q)=test_char then if skip_byte(q)<=stop_flag then
    if cur_rh<non_char then
      begin hyphen_passed:=j; hchar:=non_char; cur_rh:=non_char;
      goto continue;
      end
    else begin if hchar<non_char then if odd(hyf[j]) then
        begin hyphen_passed:=j; hchar:=non_char;
        end;
      if op_byte(q)<kern_flag then
      @<Carry out a ligature replacement, updating the cursor structure
        and possibly advancing~|j|; |goto continue| if the cursor doesn't
        advance, otherwise |goto done|@>;
      w:=char_kern(hf)(q); goto done; {this kern will be inserted below}
     end;
  if skip_byte(q)>=stop_flag then
    if cur_rh=non_char then goto done
    else begin cur_rh:=non_char; goto continue;
      end;
  k:=k+qo(skip_byte(q))+1; q:=font_info[k].qqqq;
  end;
done:

@ @d wrap_lig(#)==if ligature_present then
    begin p:=new_ligature(hf,cur_l,link(cur_q));
    if lft_hit then
      begin subtype(p):=2; lft_hit:=false;
      end;
    if # then if lig_stack=null then
      begin incr(subtype(p)); rt_hit:=false;
      end;
    link(cur_q):=p; t:=p; ligature_present:=false;
    end
@d pop_lig_stack==begin if lig_ptr(lig_stack)>null then
    begin link(t):=lig_ptr(lig_stack); {this is a charnode for |hu[j+1]|}
    t:=link(t); incr(j);
    end;
  p:=lig_stack; lig_stack:=link(p); free_node(p,small_node_size);
  if lig_stack=null then set_cur_r@+else cur_r:=character(lig_stack);
  end {if |lig_stack| isn't |null| we have |cur_rh=non_char|}

@<Append a ligature and/or kern to the translation...@>=
wrap_lig(rt_hit);
if w<>0 then
  begin link(t):=new_kern(w); t:=link(t); w:=0;
  end;
if lig_stack>null then
  begin cur_q:=t; cur_l:=character(lig_stack); ligature_present:=true;
  pop_lig_stack; goto continue;
  end

@ @<Carry out a ligature replacement, updating the cursor structure...@>=
begin if cur_l=non_char then lft_hit:=true;
if j=n then if lig_stack=null then rt_hit:=true;
check_interrupt; {allow a way out in case there's an infinite ligature loop}
case op_byte(q) of
qi(1),qi(5):begin cur_l:=rem_byte(q); {\.{=:\?}, \.{=:\?>}}
  ligature_present:=true;
  end;
qi(2),qi(6):begin cur_r:=rem_byte(q); {\.{\?=:}. \.{\?=:>}}
  if lig_stack>null then character(lig_stack):=cur_r
  else begin lig_stack:=new_lig_item(cur_r);
    if j=n then bchar:=non_char
    else begin p:=get_avail; lig_ptr(lig_stack):=p;
      character(p):=qi(hu[j+1]); font(p):=hf;
      end;
    end;
  end;
qi(3):begin cur_r:=rem_byte(q); {\.{\?=:\?}}
  p:=lig_stack; lig_stack:=new_lig_item(cur_r); link(lig_stack):=p;
  end;
qi(7),qi(11):begin wrap_lig(false); {\.{\?=:\?>}, \.{\?=:\?>>}}
  cur_q:=t; cur_l:=rem_byte(q); ligature_present:=true;
  end;
othercases begin cur_l:=rem_byte(q); ligature_present:=true; {\.{=:}}
  if lig_stack>null then pop_lig_stack
  else if j=n then goto done
  else begin append_charnode_to_t(cur_r); incr(j); set_cur_r;
    end;
  end
endcases;
if op_byte(q)>qi(4) then if op_byte(q)<>qi(7) then goto done;
goto continue;
end
@z
@x module 912
@!c:ASCII_code; {character temporarily replaced by a hyphen}
@y
@!c:ASCII_code; {character temporarily replaced by a hyphen}
@!c_loc:0..63; {where that character came from}
@!r_count:integer; {replacement count for discretionary}
@z
@x modules 913--918 are to be completely replaced
@y by the following code:
@ When the following code is performed, |hyf[0]| and |hyf[hn]| will be zero.

@<Reconstitute nodes for the hyphenated word...@>=
repeat l:=j; j:=reconstitute(j,hn,bchar,qi(hyf_char))+1;
if hyphen_passed=0 then
  begin link(s):=link(hold_head);
  while link(s)>null do s:=link(s);
  if odd(hyf[j-1]) then
    begin l:=j; hyphen_passed:=j-1; link(hold_head):=null;
    end;
  end;
if hyphen_passed>0 then
  @<Create and append a discretionary node as an alternative to the
    unhyphenated word, and continue to develop both branches until they
    become equivalent@>;
until j>hn;
link(s):=q

@ @d advance_major_tail==begin major_tail:=link(major_tail); incr(r_count);
    end

@<Create and append a discretionary node as an alternative...@>=
begin r:=get_node(small_node_size);
link(r):=link(hold_head); type(r):=disc_node;
major_tail:=r; r_count:=0;
while link(major_tail)>null do advance_major_tail;
i:=hyphen_passed;
@<Put the \(c)characters |hu[l..i]| and a hyphen into |pre_break(r)|@>;
@<Put the \(c)characters |hu[i+1..@,]| into |post_break(r)|, appending to this
  list and to |major_tail| until synchronization has been achieved@>;
@<Move pointer |s| to the end of the current list, and set |replace_count(r)|
  appropriately@>;
end

@ The new hyphen might combine with the previous character via ligature
or kern. At this point we have |l-1<=i<=j| and |i<hn|.

@<Put the \(c)characters |hu[l..i]| and a hyphen into |pre_break(r)|@>=
minor_tail:=null; pre_break(r):=null; hyf_node:=new_character(hf,hyf_char);
if hyf_node<>null then
  begin incr(i); c:=hu[i]; hu[i]:=hyf_char; free_avail(hyf_node);
  end;
while l<=i do
  begin l:=reconstitute(l,i,font_bchar[hf],non_char)+1;
  if link(hold_head)>null then
    begin if minor_tail=null then pre_break(r):=link(hold_head)
    else link(minor_tail):=link(hold_head);
    minor_tail:=link(hold_head);
    while link(minor_tail)>null do minor_tail:=link(minor_tail);
    end;
  end;
if hyf_node<>null then
  begin hu[i]:=c; {restore the character in the hyphen position}
  l:=i; decr(i);
  end

@ The synchronization algorithm begins with |l=i+1<=j|.

@<Put the \(c)characters |hu[i+1..@,]| into |post_break(r)|...@>=
minor_tail:=null; post_break(r):=null; c_loc:=0;
if bchar_label[hf]<non_address then {put left boundary at beginning of new line}
  begin decr(l); c:=hu[l]; c_loc:=l; hu[l]:=256;
  end;
while l<j do
  begin repeat l:=reconstitute(l,hn,bchar,non_char)+1;
  if c_loc>0 then
    begin hu[c_loc]:=c; c_loc:=0;
    end;
  if link(hold_head)>null then
    begin if minor_tail=null then post_break(r):=link(hold_head)
    else link(minor_tail):=link(hold_head);
    minor_tail:=link(hold_head);
    while link(minor_tail)>null do minor_tail:=link(minor_tail);
    end;
  until l>=j;
  while l>j do
    @<Append characters of |hu[j..@,]| to |major_tail|, advancing~|j|@>;
  end

@ @<Append characters of |hu[j..@,]|...@>=
begin j:=reconstitute(j,hn,bchar,non_char)+1;
link(major_tail):=link(hold_head);
while link(major_tail)>null do advance_major_tail;
end

@ Ligature insertion can cause a word to grow exponentially in size. Therefore
we must test the size of |r_count| here, even though the hyphenated text
was at most 63 characters long.

@<Move pointer |s| to the end of the current list...@>=
if r_count>127 then {we have to forget the discretionary hyphen}
  begin link(s):=link(r); link(r):=null; flush_node_list(r);
  end
else begin link(s):=r; replace_count(r):=r_count;
  end;
s:=major_tail
@z
@x module 1030
@d main_loop=70 {go here to typeset |cur_chr| in the current font}
@d main_loop_1=71 {like |main_loop|, but |(f,c)| = current font and char}
@d main_loop_2=72 {like |main_loop_1|, but |c| is known to be in range}
@d main_loop_3=73 {like |main_loop_2|, but several variables are set up}
@d append_normal_space=74 {go here to append a normal space between words}
@y
@d main_loop=70 {go here to typeset a string of consecutive characters}
@d main_loop_wrapup=80 {go here to finish a character or ligature}
@d main_loop_move=90 {go here to advance the ligature cursor}
@d main_loop_move_lig=95 {same, when advancing past a generated ligature}
@d main_loop_lookahead=100 {go here to bring in another character, if any}
@d main_lig_loop=110 {go here to check for ligatures or kerning}
@d append_normal_space=120 {go here to append a normal space between words}
@z
@x ibid
label big_switch,reswitch,main_loop,main_loop_1,main_loop_2,main_loop_3,
@y
label big_switch,reswitch,main_loop,main_loop_wrapup,
  main_loop_move,main_loop_move+1,main_loop_move+2,main_loop_move_lig,
  main_loop_lookahead,main_loop_lookahead+1,
  main_lig_loop,main_lig_loop+1,main_lig_loop+2,
@z
@x ibid
@<Local variables for the inner loop of |main_control|@>@;
@y
@z
@x ibid
hmode+char_num: begin scan_char_num; cur_chr:=cur_val; goto main_loop;
  end;
@y
hmode+char_num: begin scan_char_num; cur_chr:=cur_val; goto main_loop;@+end;
hmode+no_boundary: begin get_x_token;
  if (cur_cmd=letter)or(cur_cmd=other_char)or(cur_cmd=char_given)or
   (cur_cmd=char_num) then cancel_boundary:=true;
  goto reswitch;
  end;
@z
@x modules 1032--1040 are all to be replaced
@y by the following code:
@ The following part of the program was first written in a structured
manner, according to the philosophy that ``premature optimization is
the root of all evil.'' Then it was rearranged into pieces of
spaghetti so that the most common actions could proceed with little or
no redundancy.

The original unoptimized form of this algorithm resembles the
|reconstitute| procedure, which was described earlier in connection with
hyphenation. Again we have an implied ``cursor`` between characters
|cur_l| and |cur_r|. The main difference is that the |lig_stack| can now
contain a charnode as well as pseudo-ligatures; that stack is now
usually nonempty, because the next character of input (if any) has been
appended to it. In |main_control| we have
$$|cur_r|=\cases{|character(lig_stack)|,&if |lig_stack>null|;\cr
  |font_bchar[cur_font]|,&otherwise.\cr}$$
Several additional global variables are needed.

@<Glob...@>=
@!main_f:internal_font_number; {the current font}
@!main_i:four_quarters; {character information bytes for |cur_l|}
@!main_j:four_quarters; {ligature/kern command}
@!main_k:font_index; {index into |font_info|}
@!main_p:pointer; {temporary register for list manipulation}
@!main_s:integer; {space factor value}
@!bchar:halfword; {right boundary character of current font, or |non_char|}
@!false_bchar:halfword; {nonexistent character matching |bchar|, or |non_char|}
@!cancel_boundary:boolean; {should the left boundary be ignored?}
@!ins_disc:boolean; {should we insert a discretionary node?}

@ The boolean variables of the main loop are normally false, and always reset
to false before the loop is left. That saves us the extra work of initializing
each time.

@<Set init...@>=
ligature_present:=false; cancel_boundary:=false; lft_hit:=false; rt_hit:=false;
ins_disc:=false;

@ We leave |space_factor| unchanged if |sf_code(cur_chr)=0|; otherwise we
set it to |sf_code(cur_chr)|, except that the space factor never changes
from a value less than 1000 to a value exceeding 1000. The most common
case is |sf_code(cur_chr)=1000|, so we want that case to be fast.

The overall structure of the main loop is presented here. Some program labels
are inside the individual sections.

@d adjust_space_factor==@t@>@;@/
  main_s:=sf_code(cur_chr);
  if main_s=1000 then space_factor:=1000
  else if main_s<1000 then
    begin if main_s>0 then space_factor:=main_s;
    end
  else if space_factor<1000 then space_factor:=1000
  else space_factor:=main_s

@<Append character |cur_chr|...@>=
adjust_space_factor;@/
main_f:=cur_font;
bchar:=font_bchar[main_f]; false_bchar:=font_false_bchar[main_f];
if mode>0 then if language<>clang then fix_language;
fast_get_avail(lig_stack); font(lig_stack):=main_f; cur_l:=qi(cur_chr);
character(lig_stack):=cur_l;@/
cur_q:=tail;
if cancel_boundary then
  begin cancel_boundary:=false; main_k:=non_address;
  end
else main_k:=bchar_label[main_f];
if main_k=non_address then goto main_loop_move+2; {no left boundary processing}
cur_r:=cur_l; cur_l:=non_char;
goto main_lig_loop+1; {begin with cursor after left boundary}
@#
main_loop_wrapup:@<Make a ligature node, if |ligature_present|;
  insert a null discretionary, if appropriate@>;
main_loop_move:@<If the cursor is immediately followed by the right boundary,
  |goto reswitch|; if it's followed by an invalid character, |goto big_switch|;
  otherwise move the cursor one step to the right and |goto main_lig_loop|@>;
main_loop_lookahead:@<Look ahead for another character, or leave |lig_stack|
  empty if there's none there@>;
main_lig_loop:@<If there's a ligature/kern command relevant to |cur_l| and
  |cur_r|, adjust the text appropriately; exit to |main_loop_wrapup|@>;
main_loop_move_lig:@<Move the cursor past a pseudo-ligature, then
  |goto main_loop_lookahead| or |main_lig_loop|@>

@ If the current horizontal list is empty, the reference to |character(tail)|
here is not strictly legal, since |tail| will be a node freshly returned by
|get_avail|. But this should cause no problem on most implementations, and we
do want the inner loop to be fast.
@^dirty Pascal@>

A discretionary break is not inserted for an explicit hyphen when we are in
restricted horizontal mode. In particular, this avoids putting discretionary
nodes inside of other discretionaries.

@d pack_lig(#)== {the parameter is either |rt_hit| or |false|}
  begin main_p:=new_ligature(main_f,cur_l,link(cur_q));
  if lft_hit then
    begin subtype(main_p):=2; lft_hit:=false;
    end;
  if # then if lig_stack=null then
    begin incr(subtype(main_p)); rt_hit:=false;
    end;
  link(cur_q):=main_p; tail:=main_p; ligature_present:=false;
  end

@d wrapup(#)==if cur_l<non_char then
  begin if character(tail)=qi(hyphen_char[main_f]) then if link(cur_q)>null then
    ins_disc:=true;
  if ligature_present then pack_lig(#);
  if ins_disc then
    begin ins_disc:=false;
    if mode>0 then tail_append(new_disc);
    end;
  end

@<Make a ligature node, if |ligature_present|;...@>=
wrapup(rt_hit)

@ @<If the cursor is immediately followed by the right boundary...@>=
if lig_stack=null then goto reswitch;
cur_q:=tail; cur_l:=cur_r; {or |character(lig_stack)|}
main_loop_move+1:if not is_char_node(lig_stack) then goto main_loop_move_lig;
main_loop_move+2:if(cur_chr<font_bc[main_f])or(cur_chr>font_ec[main_f]) then
  begin char_warning(main_f,cur_chr); free_avail(lig_stack); goto big_switch;
  end;
main_i:=char_info(main_f)(cur_l);
if not char_exists(main_i) then
  begin char_warning(main_f,cur_chr); free_avail(lig_stack); goto big_switch;
  end;
tail_append(lig_stack) {|main_loop_lookahead| is next}

@ Here we are at |main_loop_move_lig|.
When we begin this code we have |cur_l=character(lig_stack)| and |cur_q=tail|.

@<Move the cursor past a pseudo-ligature...@>=
main_p:=lig_ptr(lig_stack);
if main_p>null then tail_append(main_p);
temp_ptr:=lig_stack; lig_stack:=link(temp_ptr);
free_node(temp_ptr,small_node_size);
main_i:=char_info(main_f)(cur_l); ligature_present:=true;
if lig_stack=null then
  if main_p>null then goto main_loop_lookahead
  else cur_r:=bchar
else cur_r:=character(lig_stack);
goto main_lig_loop

@ The result of \.{\\char} can participate in a ligature or kern, so we must
look ahead for it.

@<Look ahead for another character...@>=
get_next; {set only |cur_cmd| and |cur_chr|, for speed}
if cur_cmd=letter then goto main_loop_lookahead+1;
if cur_cmd=other_char then goto main_loop_lookahead+1;
if cur_cmd=char_given then goto main_loop_lookahead+1;
x_token; {now expand and set |cur_cmd|, |cur_chr|, |cur_tok|}
if cur_cmd=letter then goto main_loop_lookahead+1;
if cur_cmd=other_char then goto main_loop_lookahead+1;
if cur_cmd=char_given then goto main_loop_lookahead+1;
if cur_cmd=char_num then
  begin scan_char_num; cur_chr:=cur_val; goto main_loop_lookahead+1;
  end;
if cur_cmd=no_boundary then bchar:=non_char;
cur_r:=bchar; lig_stack:=null; goto main_lig_loop;
main_loop_lookahead+1: adjust_space_factor;
fast_get_avail(lig_stack); font(lig_stack):=main_f;
cur_r:=qi(cur_chr); character(lig_stack):=cur_r;
if cur_r=false_bchar then cur_r:=non_char {this prevents spurious ligatures}

@ Even though comparatively few characters have a lig/kern program, several
of the instructions here count as part of \TeX's inner loop, since a
potentially long sequential search must be performed. For example, tests with
Computer Modern Roman showed that about 40 per cent of all characters
actually encountered in practice had a lig/kern program, and that about four
lig/kern commands were investigated for every such character.

At the beginning of this code we have |main_i=char_info(main_f)(cur_l)|.

@<If there's a ligature/kern command...@>=
if char_tag(main_i)<>lig_tag then goto main_loop_wrapup;
main_k:=lig_kern_start(main_f)(main_i); main_j:=font_info[main_k].qqqq;
if skip_byte(main_j)<=stop_flag then goto main_lig_loop+2;
main_k:=lig_kern_restart(main_f)(main_j);
main_lig_loop+1:main_j:=font_info[main_k].qqqq;
main_lig_loop+2:if next_char(main_j)=cur_r then
 if skip_byte(main_j)<=stop_flag then
  @<Do ligature or kern command, returning to |main_lig_loop|
  or |main_loop_wrapup| or |main_loop_move|@>;
if skip_byte(main_j)=qi(0) then incr(main_k)
else begin if skip_byte(main_j)>=stop_flag then goto main_loop_wrapup;
  main_k:=main_k+qo(skip_byte(main_j))+1;
  end;
goto main_lig_loop+1

@ When a ligature or kern instruction matches a character, we know from
|read_font_info| that the character exists in the font, even though we
haven't verified its existence in the normal way.

This section could be made into a subroutine, if the code inside
|main_control| needs to be shortened.

\chardef\?='174 % vertical line to indicate character retention

@<Do ligature or kern command...@>=
begin if op_byte(main_j)>=kern_flag then
  begin wrapup(rt_hit);
  tail_append(new_kern(char_kern(main_f)(main_j))); goto main_loop_move;
  end;
if cur_l=non_char then lft_hit:=true
else if lig_stack=null then rt_hit:=true;
check_interrupt; {allow a way out in case there's an infinite ligature loop}
case op_byte(main_j) of
qi(1),qi(5):begin cur_l:=rem_byte(main_j); {\.{=:\?}, \.{=:\?>}}
  main_i:=char_info(main_f)(cur_l); ligature_present:=true;
  end;
qi(2),qi(6):begin cur_r:=rem_byte(main_j); {\.{\?=:}, \.{\?=:>}}
  if lig_stack=null then {right boundary character is being consumed}
    begin lig_stack:=new_lig_item(cur_r); bchar:=non_char;
    end
  else if is_char_node(lig_stack) then {|link(lig_stack)=null|}
    begin main_p:=lig_stack; lig_stack:=new_lig_item(cur_r);
    lig_ptr(lig_stack):=main_p;
    end
  else character(lig_stack):=cur_r;
  end;
qi(3):begin cur_r:=rem_byte(main_j); {\.{\?=:\?}}
  main_p:=lig_stack; lig_stack:=new_lig_item(cur_r);
  link(lig_stack):=main_p;
  end;
qi(7),qi(11):begin wrapup(false); {\.{\?=:\?>}, \.{\?=:\?>>}}
  cur_q:=tail; cur_l:=rem_byte(main_j);
  main_i:=char_info(main_f)(cur_l); ligature_present:=true;
  end;
othercases begin cur_l:=rem_byte(main_j); ligature_present:=true; {\.{=:}}
  if lig_stack=null then goto main_loop_wrapup
  else goto main_loop_move+1;
  end
endcases;
if op_byte(main_j)>qi(4) then
  if op_byte(main_j)<>qi(7) then goto main_loop_wrapup;
if cur_l<non_char then goto main_lig_loop;
main_k:=bchar_label[main_f]; goto main_lig_loop+1;
end
@z
% HINTS for making main_control shorter if you need to conserve space:
% (1) make pack_lig(#) a procedure call instead of a macro call
% (2) make the code of @<Do ligature...@> a function that returns one of
%     four values; you goto a label based on the value returned.
@x module 1042
begin p:=font_glue[cur_font];
if p=null then
  begin f:=cur_font; p:=new_spec(zero_glue); k:=param_base[f]+space_code;
  width(p):=font_info[k].sc; {that's |space(f)|}
  stretch(p):=font_info[k+1].sc; {and |space_stretch(f)|}
  shrink(p):=font_info[k+2].sc; {and |space_shrink(f)|}
  font_glue[f]:=p;
@y
begin main_p:=font_glue[cur_font];
if main_p=null then
  begin main_p:=new_spec(zero_glue); main_k:=param_base[cur_font]+space_code;
  width(main_p):=font_info[main_k].sc; {that's |space(cur_font)|}
  stretch(main_p):=font_info[main_k+1].sc; {and |space_stretch(cur_font)|}
  shrink(main_p):=font_info[main_k+2].sc; {and |space_shrink(cur_font)|}
  font_glue[cur_font]:=main_p;
@z
@x module 1045
any_mode(relax),vmode+spacer,mmode+spacer,mmode+no_boundary:do_nothing;
@y
any_mode(relax),vmode+spacer,mmode+spacer:do_nothing;
@z
@x module 1090
   vmode+ex_space:@t@>@;@/
@y
   vmode+ex_space,vmode+no_boundary:@t@>@;@/
@z
@x module 1322
dump_int(font_glue[k]);@/
@y
dump_int(font_glue[k]);@/
dump_int(bchar_label[k]);
dump_int(font_bchar[k]);
dump_int(font_false_bchar[k]);@/
@z
@x module 1323
undump(min_halfword)(lo_mem_max)(font_glue[k]);@/
@y
undump(min_halfword)(lo_mem_max)(font_glue[k]);@/
undump(0)(font_mem_size)(bchar_label[k]);
undump(min_quarterword)(non_char)(font_bchar[k]);
undump(min_quarterword)(non_char)(font_false_bchar[k]);
@z

363. New \inputlineno feature desired by Spivak
@x module 416 gets a new definition and a new statement
@y
@d input_line_no_code=glue_val+1 {code for \.{\\inputlineno}}
primitive("inputlineno",last_item,input_line_no_code);
@!@:input_line_no_}{\.{\\inputlineno} primitive@>
@z and module 417 changes in the obvious way
@x module 424 gets new code at the beginning
@y
if cur_chr>glue_val then
  begin cur_val:=line; cur_val_level:=int_val;
  end
else
@z

364. New feature \holdinginserts suggested by Mittelbach.
@x modules 236--238 get a new integer parameter
@d int_pars=53 {total number of integer parameters}
@y
@d holding_inserts_code=53 {do not remove insertion nodes from \.{\\box255}}
@d int_pars=54 {total number of integer parameters}
@z and appropriate further lines are added to match all the other parameters
@x module 1014
@<Prepare all the boxes involved in insertions to act as queues@>;
q:=hold_head; link(q):=null; prev_p:=page_head; p:=link(prev_p);
while p<>best_page_break do
  begin if type(p)=ins_node then @<Either insert the material
    specified by node |p| into the appropriate box, or
    hold it for the next page; also delete node |p| from
    the current page@>
@y
if holding_inserts<=0 then
  @<Prepare all the boxes involved in insertions to act as queues@>;
q:=hold_head; link(q):=null; prev_p:=page_head; p:=link(prev_p);
while p<>best_page_break do
  begin if type(p)=ins_node then
    begin if holding_inserts<=0 then
       @<Either insert the material specified by node |p| into the
         appropriate box, or hold it for the next page;
         also delete node |p| from the current page@>;
    end
@z
% also insert begin...end around the Pascal code of module 1018

365. New \badness feature (which I'd been resisting for years)
@x module 416 gets a new definition and a new statement
@y
@d badness_code=glue_val+2 {code for \.{\\badness}}
primitive("badness",last_item,badness_code);
@!@:badness_}{\.{\\badness} primitive@>
@z and module 417 changes in the obvious way
@x module 424 (the new code just added in #363)
  begin cur_val:=line; cur_val_level:=int_val;
@y
  begin if cur_chr=input_line_no_code then cur_val:=line
  else cur_val:=last_badness; {|cur_chr=badness_code|}
  cur_val_level:=int_val;
@z
@x module 646 gets a new global variable
@y
@!last_badness:integer; {badness of the most recently packaged box}
@z
@x module 648
@ @<Set init...@>=adjust_tail:=null;
@y
@ @<Set init...@>=adjust_tail:=null; last_badness:=0;
@z
@x modules 649 and 668
@!b:integer; {badness of the new box}
begin r:=get_node(box_node_size); type(r):=hlist_node;
@y (except it's vlist_node in 668)
begin last_badness:=0; r:=get_node(box_node_size); type(r):=hlist_node;
@z (except it's vlist_node in 668)
% change b to last_badness in modules 660&674 (4 times), 667&678 (3 times)
@x modules 658 and 673
if (hbadness<inf_bad)and(o=normal)and(list_ptr(r)<>null) then
@y (except it's vbadness in 673)
if o=normal then if list_ptr(r)<>null then
@z
@x modules 664 and 676
  begin set_glue_ratio_one(glue_set(r)); {this is the maximum shrinkage}
@y
  begin last_badness:=1000000;
  set_glue_ratio_one(glue_set(r)); {use the maximum shrinkage}
@z
@x ibid
else if (hbadness<100)and(o=normal)and(list_ptr(r)<>null) then
@y (except it's vbadness in 676)
else if o=normal then if list_ptr(r)<>null then
@z

366. New \emergencystretch feature
@x modules 247--248 get a new dimen parameter
@d dimen_pars=20 {total number of dimension parameters}
@y
@d emergency_stretch_code=20 {reduces badnesses on final pass of line-breaking}
@d dimen_pars=21 {total number of dimension parameters}
@z and appropriate further lines are added to match all the other parameters
@x module 828
@!second_pass:boolean; {is this our second attempt to break this paragraph?}
@y
@!second_pass:boolean; {is this our second attempt to break this paragraph?}
@!final_pass:boolean; {is this our final attempt to break this paragraph?}
@z
% change second_pass to final_pass in modules 854 and 873
@x module 863
  second_pass:=false;
  end
else  begin threshold:=tolerance; second_pass:=true;
@y
  second_pass:=false; final_pass:=false;
  end
else  begin threshold:=tolerance; second_pass:=true;
  final_pass:=(emergency_stretch<=0);
@z
@x ibid
  @!stat if tracing_paragraphs>0 then print_nl("@@secondpass");@;@+tats@/
  threshold:=tolerance; second_pass:=true; {if at first you don't
    succeed, \dots}
@y
  if not second_pass then
    begin@!stat if tracing_paragraphs>0 then print_nl("@@secondpass");@;@+tats@/
    threshold:=tolerance; second_pass:=true; final_pass:=(emergency_stretch<=0);
    end {if at first you don't succeed, \dots}
  else begin @!stat if tracing_paragraphs>0 then
      print_nl("@@finalpass");@;@+tats@/
    background[2]:=background[2]+emergency_stretch; final_pass:=true;
    end;
@z

367. New \errorcontextlines feature suggested by a TUG participant
@x modules 236--238 get a new integer parameter
@d int_pars=54 {total number of integer parameters}
@y
@d error_context_lines_code=54 {maximum intermediate line pairs shown}
@d int_pars=55 {total number of integer parameters}
@z and appropriate further lines are added to match all the other parameters
@x module 311
@<Local variables for formatting calculations@>@/
begin base_ptr:=input_ptr; input_stack[base_ptr]:=cur_input;
  {store current state}
loop@+begin cur_input:=input_stack[base_ptr]; {enter into the context}
  @<Display the current context@>;
  if (state<>token_list) then
    if (name>17) or (base_ptr=0) then goto done;
@y
@!nn:integer; {number of contexts shown so far, less one}
@!bottom_line:boolean; {have we reached the final context to be shown?}
@<Local variables for formatting calculations@>@/
begin base_ptr:=input_ptr; input_stack[base_ptr]:=cur_input;
  {store current state}
nn:=-1; bottom_line:=false;
loop@+begin cur_input:=input_stack[base_ptr]; {enter into the context}
  if (state<>token_list) then
    if (name>17) or (base_ptr=0) then bottom_line:=true;
  if (base_ptr=input_ptr)or bottom_line or(nn<error_context_lines) then
    @<Display the current context@>
  else if nn=error_context_lines then
    begin print_nl("..."); incr(nn); {omitted if |error_context_lines<0|}
    end;
  if bottom_line then goto done;
@z
% also insert begin...end around the Pascal code of module 312
% and add the statement "incr(nn)" to that module

368. char_warning inside hyphenation could clobber old_setting
@x module 863
done: @!stat if tracing_paragraphs>0 then end_diagnostic(true);@;@+tats@/
@y
done: @!stat if tracing_paragraphs>0 then
  begin end_diagnostic(true); normalize_selector;
  end;@+tats@/
@z

369. Make ".fmt" more easily switchable (Don Hosek).
@x module 520 gets a new definition
@y
@d format_extension=".fmt" {the extension, as a \.{WEB} constant}
@z now replace ".fmt" by format_extension in modules 529 and 1328.

370. Possible range check on weird nullfont (Breitenlohner, 16 Oct 89).
@x module 565
if (bc>ec+1)or(ec>255) then abort;
@y
if (bc>ec+1)or(ec>255) then abort;
if bc>255 then {|bc=256| and |ec=255|}
  begin bc:=1; ec:=0;
  end;
@z

371. Prevent save_stack conflicts, e.g. in {\hbox\expandafter{\csname
     \endcsname}} (found by Sullivan).
@x module 645
@p procedure scan_spec; {scans a box specification and left brace}
label found;
begin if scan_keyword("to") then saved(0):=exactly
@.to@>
else if scan_keyword("spread") then saved(0):=additional
@.spread@>
else  begin saved(0):=additional; saved(1):=0;
  goto found;
  end;
scan_normal_dimen; saved(1):=cur_val;
found: save_ptr:=save_ptr+2; scan_left_brace;
@y
@p procedure scan_spec(@!c:group_code;@!three_codes:boolean);
  {scans a box specification and left brace}
label found;
var @!s:integer; {temporarily saved value}
@!spec_code:exactly..additional;
begin if three_codes then s:=saved(0);
if scan_keyword("to") then spec_code:=exactly
@.to@>
else if scan_keyword("spread") then spec_code:=additional
@.spread@>
else  begin spec_code:=additional; cur_val:=0;
  goto found;
  end;
scan_normal_dimen;
found: if three_codes then
  begin saved(0):=s; incr(save_ptr);
  end;
saved(0):=spec_code; saved(1):=cur_val; save_ptr:=save_ptr+2;
new_save_level(c); scan_left_brace;
@z
@x module 774
scan_spec; new_save_level(align_group);@/
@y
scan_spec(align_group,false);@/
@z
@x module 1073
  if t=0 then saved(0):=cur_val@+else saved(0):=-cur_val;
  scan_box;
  end;
any_mode(leader_ship): begin saved(0):=leader_flag-a_leaders+cur_chr; scan_box;
  end;
any_mode(make_box): begin saved(0):=0; begin_box;
  end;
@y
  if t=0 then scan_box(cur_val)@+else scan_box(-cur_val);
  end;
any_mode(leader_ship): scan_box(leader_flag-a_leaders+cur_chr);
any_mode(make_box): begin_box(0);
@z
@x module 1075
procedure box_end;
var p:pointer; {|ord_noad| for new box in math mode}
begin if saved(0)<box_flag then @<Append box |cur_box| to the current list,
    shifted by |saved(0)|@>
else if saved(0)<ship_out_flag then @<Store \(c)|cur_box| in a box register@>
else if cur_box<>null then
  if saved(0)>ship_out_flag then @<Append a new leader node that
      uses |cur_box|@>
@y
procedure box_end(@!box_context:integer);
var p:pointer; {|ord_noad| for new box in math mode}
begin if box_context<box_flag then @<Append box |cur_box| to the current list,
    shifted by |box_context|@>
else if box_context<ship_out_flag then @<Store \(c)|cur_box| in a box register@>
else if cur_box<>null then
  if box_context>ship_out_flag then @<Append a new leader node that
      uses |cur_box|@>
@z
@x module 1076
  begin shift_amount(cur_box):=saved(0);
@y
  begin shift_amount(cur_box):=box_context;
@z
@x module 1077
if saved(0)<box_flag+256 then
  eq_define(box_base-box_flag+saved(0),box_ref,cur_box)
else geq_define(box_base-box_flag-256+saved(0),box_ref,cur_box)
@y
if box_context<box_flag+256 then
  eq_define(box_base-box_flag+box_context,box_ref,cur_box)
else geq_define(box_base-box_flag-256+box_context,box_ref,cur_box)
@z
@x module 1078
  begin append_glue; subtype(tail):=saved(0)-(leader_flag-a_leaders);
@y
  begin append_glue; subtype(tail):=box_context-(leader_flag-a_leaders);
@z
@x module 1079
procedure begin_box;
@y
procedure begin_box(@!box_context:integer);
@z
@x ibid
box_end; {in simple cases, we use the box immediately}
@y
box_end(box_context); {in simple cases, we use the box immediately}
@z
@x module 1083
incr(save_ptr); scan_spec;
if k=hmode then
  if (saved(-3)<box_flag)and(abs(mode)=vmode) then
    new_save_level(adjusted_hbox_group)
  else new_save_level(hbox_group)
else  begin if k=vmode then new_save_level(vbox_group)
  else  begin new_save_level(vtop_group); k:=vmode;
@y
saved(0):=box_context;
if k=hmode then
  if (box_context<box_flag)and(abs(mode)=vmode) then
    scan_spec(adjusted_hbox_group,true)
  else scan_spec(hbox_group,true)
else  begin if k=vmode then scan_spec(vbox_group,true)
  else  begin scan_spec(vtop_group,true); k:=vmode;
@z
@x module 1084
procedure scan_box; {the next input should specify a box or perhaps a rule}
begin @<Get the next non-blank non-relax...@>;
if cur_cmd=make_box then begin_box
else if (saved(0)>=leader_flag)and((cur_cmd=hrule)or(cur_cmd=vrule)) then
  begin cur_box:=scan_rule_spec; box_end;
@y
procedure scan_box(@!box_context:integer);
  {the next input should specify a box or perhaps a rule}
begin @<Get the next non-blank non-relax...@>;
if cur_cmd=make_box then begin_box(box_context)
else if (box_context>=leader_flag)and((cur_cmd=hrule)or(cur_cmd=vrule)) then
  begin cur_box:=scan_rule_spec; box_end(box_context);
@z
@x module 1086
pop_nest; box_end;
@y
pop_nest; box_end(saved(0));
@z
@x module 1117
else  begin incr(save_ptr); saved(-1):=0; scan_left_brace;
  new_save_level(disc_group); push_nest; mode:=-hmode; space_factor:=1000;
@y
else  begin incr(save_ptr); saved(-1):=0; new_save_level(disc_group);
  scan_left_brace; push_nest; mode:=-hmode; space_factor:=1000;
@z
@x module 1119
incr(saved(-1)); scan_left_brace; new_save_level(disc_group);
@y
incr(saved(-1)); new_save_level(disc_group); scan_left_brace;
@z
@x module 1167
mmode+vcenter: begin scan_spec; new_save_level(vcenter_group); normal_paragraph;
@y
mmode+vcenter: begin scan_spec(vcenter_group,false); normal_paragraph;
@z
@x module 1172
scan_left_brace; push_math(math_choice_group);
@y
push_math(math_choice_group); scan_left_brace;
@z
@x module 1174
incr(saved(-1)); scan_left_brace; push_math(math_choice_group);
@y
incr(saved(-1)); push_math(math_choice_group); scan_left_brace;
@z
@x module 1241
  if global then saved(0):=box_flag+256+cur_val
  else saved(0):=box_flag+cur_val;
  scan_optional_equals; scan_box;
@y
  if global then n:=256+cur_val@+else n:=cur_val;
  scan_optional_equals; scan_box(box_flag+n);
@z

372. Bugfix 339 didn't go far enough (found by Sch\"opf and Mittelbach).
@x module 516 [NOTE: THIS AFFECTS ALMOST ALL CHANGE FILES!]
@ And here's the second.
@^system dependencies@>

@p function more_name(@!c:ASCII_code):boolean;
begin if c=" " then more_name:=false
else  begin if (c=">")or(c=":") then
    begin area_delimiter:=pool_ptr; ext_delimiter:=0;
    end
  else if (c=".")and(ext_delimiter=0) then ext_delimiter:=pool_ptr;
  str_room(1); append_char(c); {contribute |c| to the current string}
@y
@ And here's the second. The string pool might change as the file name is
being scanned, since a new \.{\\csname} might be entered; therefore we keep
|area_delimiter| and |ext_delimiter| relative to the beginning of the current
string, instead of assigning an absolute address like |pool_ptr| to them.
@^system dependencies@>

@p function more_name(@!c:ASCII_code):boolean;
begin if c=" " then more_name:=false
else  begin str_room(1); append_char(c); {contribute |c| to the current string}
  if (c=">")or(c=":") then
    begin area_delimiter:=cur_length; ext_delimiter:=0;
    end
  else if (c=".")and(ext_delimiter=0) then ext_delimiter:=cur_length;
@z
@x module 517 [NOTE: THIS DOES TOO, AND SO DOES THE NEXT!]
else  begin cur_area:=str_ptr; incr(str_ptr);
  str_start[str_ptr]:=area_delimiter+1;
@y
else  begin cur_area:=str_ptr;
  str_start[str_ptr+1]:=str_start[str_ptr]+area_delimiter; incr(str_ptr);
@z
@x ibid
else  begin cur_name:=str_ptr; incr(str_ptr);
  str_start[str_ptr]:=ext_delimiter; cur_ext:=make_string;
@y
else  begin cur_name:=str_ptr;
  str_start[str_ptr+1]:=str_start[str_ptr]+ext_delimiter-area_delimiter-1;
  incr(str_ptr); cur_ext:=make_string;
@z

373. Allow multiple hyphenmins in the same paragraph (Mike Ferguson).
@x module 212, two new fields for list_state_record
@y
  @!lhm_field,@!rhm_field: quarterword;
@z
@x module 213, two new macros to access those fields
@y
@d lhmin==cur_list.lhm_field {\.{\\lefthyphenmin} at start of paragraph}
@d rhmin==cur_list.rhm_field {\.{\\righthyphenmin} at start of paragraph}
@z
@x module 215, two new initializations
@y
lhmin:=0; rhmin:=0;
@z
@x module 218
  if nest[p].ml_field<0 then print(" (\output routine)");
@y
  if m=hmode then@+if(nest[p].lhm_field<>2)or(nest[p].rhm_field<>3)then
    begin print(" (hyphenmin "); print_int(nest[p].lhm_field); print_char(",");
    print_int(nest[p].rhm_field); print_char(")");
    end;
  if nest[p].ml_field<0 then print(" (\output routine)");
@z
@x module 891
l_hyf:=left_hyphen_min-1;@+if l_hyf<0 then l_hyf:=0;
r_hyf:=right_hyphen_min-1;@+if r_hyf<0 then r_hyf:=0;
min_hyf:=l_hyf+r_hyf+2; cur_lang:=0;
@y
l_hyf:=lhmin; r_hyf:=rhmin; cur_lang:=0;
@z
@x module 892
@!l_hyf,@!r_hyf,@!min_hyf:integer; {limits on fragment sizes}
@y
@!l_hyf,@!r_hyf:integer; {limits on fragment sizes}
@z
@x module 894
begin if min_hyf>63 then goto done1;
prev_s:=cur_p; s:=link(prev_s);
if s<>null then
  begin @<Skip to node |ha|, or |goto done1| if no hyphenation
    should be attempted@>;
  @<Skip to node |hb|, putting letters into |hu| and |hc|@>;
  @<Check that the nodes following |hb| permit hyphenation and that at least
    |min_hyf| letters have been found, otherwise |goto done1|@>;
@y
begin prev_s:=cur_p; s:=link(prev_s);
if s<>null then
  begin @<Skip to node |ha|, or |goto done1| if no hyphenation
    should be attempted@>;
  if l_hyf+r_hyf>63 then goto done1;
  @<Skip to node |hb|, putting letters into |hu| and |hc|@>;
  @<Check that the nodes following |hb| permit hyphenation and that at least
    |l_hyf+r_hyf| letters have been found, otherwise |goto done1|@>;
@z
@x module 899
if hn<min_hyf then goto done1;
@y
if hn<l_hyf+r_hyf then goto done1;
@z
@x module 902
for j:=l_hyf+1 to hn-r_hyf-1 do if odd(hyf[j]) then goto found1;
@y
for j:=l_hyf to hn-r_hyf do if odd(hyf[j]) then goto found1;
@z
@x module 923
for j:=0 to hn-r_hyf do
@y
for j:=0 to hn-r_hyf+1 do
@z
@x ibid
found: for j:=0 to l_hyf do hyf[j]:=0;
for j:=0 to r_hyf do hyf[hn-j]:=0
@y
found: for j:=0 to l_hyf-1 do hyf[j]:=0;
for j:=0 to r_hyf-1 do hyf[hn-j]:=0
@z
@x module 1091 begins with a new subroutine
@y
function norm_min(@!h:integer):small_number;
begin if h<=0 then norm_min:=1@+else if h>=63 then norm_min:=63@+
else norm_min:=h;
end;
@z
@x module 1091 then uses the new subroutine
push_nest; mode:=hmode; space_factor:=1000; clang:=0;
@y
lhmin:=norm_min(left_hyphen_min); rhmin:=norm_min(right_hyphen_min);
push_nest; mode:=hmode; space_factor:=1000; clang:=0;
@z
@x module 1341
@d stored_language(#)==mem[#+1].int {language number, in the range |0..255|}
@y
@d what_lang(#)==link(#+1) {language number, in the range |0..255|}
@d what_lhm(#)==type(#+1) {minimum left fragment, in the range |1..63|}
@d what_rhm(#)==subtype(#+1) {minimum right fragment, in the range |1..63|}
@z
@x module 1356
  print_int(stored_language(p));
@y
  print_int(what_lang(p)); print(" (hyphenmin ");
  print_int(what_lhm(p)); print_char(",");
  print_int(what_rhm(p)); print_char(")");
@z
@x modules 1362 and 1363
@ @<Advance \(p)past a whatsit node in the \(l)|line_break| loop@>=
if subtype(cur_p)=language_node then cur_lang:=stored_language(cur_p)

@ @<Advance \(p)past a whatsit node in the \(p)pre-hyphenation loop@>=
if subtype(s)=language_node then cur_lang:=stored_language(s)
@y
@ @d adv_past(#)==@+if subtype(#)=language_node then
    begin cur_lang:=what_lang(#); l_hyf:=what_lhm(#); r_hyf:=what_rhm(#);@+end

@<Advance \(p)past a whatsit node in the \(l)|line_break| loop@>=@+
adv_past(cur_p)

@ @<Advance \(p)past a whatsit node in the \(p)pre-hyphenation loop@>=@+
adv_past(s)
@z
@x module 1376
  stored_language(tail):=l; clang:=l;
@y
  what_lang(tail):=l; clang:=l;@/
  what_lhm(tail):=norm_min(left_hyphen_min);
  what_rhm(tail):=norm_min(right_hyphen_min);
@z
@x module 1377
  stored_language(tail):=clang;
@y
  what_lang(tail):=clang;
  what_lhm(tail):=norm_min(left_hyphen_min);
  what_rhm(tail):=norm_min(right_hyphen_min);
@z

374. Make \par and end_template definitely non-character (Marc van Leeuwen)
@x module 14 gets a new line
@y
if mem_top<256+11 then bad:=7; {we will want |null_list>255|}
@z
@x module 334
primitive("par",par_end,0); par_loc:=cur_val; par_token:=cs_token_flag+par_loc;
@y
primitive("par",par_end,256); {cf. |scan_file_name|}
par_loc:=cur_val; par_token:=cs_token_flag+par_loc;
@z

375. Alignments must be more robust to prevent crashes (Marc van Leeuwen)
@x module 324
else if token_type=u_template then align_state:=0;
@y
else if token_type=u_template then
  if align_state>500000 then align_state:=0
  else fatal_error("(interwoven alignment preambles are not allowed)");
@.interwoven alignment preambles...@>
@z
@x module 782
if (cur_cmd=assign_glue)and(cur_chr=glue_base+tab_skip_code) then
@y
if cur_cmd=endv then
  fatal_error("(interwoven alignment preambles are not allowed)");
@.interwoven alignment preambles...@>
if (cur_cmd=assign_glue)and(cur_chr=glue_base+tab_skip_code) then
@z
@x module 791
p:=link(q);
@y
if align_state<500000 then
  fatal_error("(interwoven alignment preambles are not allowed)");
@.interwoven alignment preambles...@>
p:=link(q);
@z

376. Avoid kern removal in discretionary breaks (Marc van Leeuwen)
@x module 815
label done,done1,done2,done3,done4,continue;
@y
label done,done1,done2,done3,done4,done5,continue;
@z
@x module 866
glue_node: begin @<If node |cur_p| is a legal breakpoint, call |try_break|@>;
  @<Update the active widths by including the glue in |glue_ptr(cur_p)|@>;
@y
glue_node: begin @<If node |cur_p| is a legal breakpoint, call |try_break|;
  then update the active widths by including the glue in |glue_ptr(cur_p)|@>;
@z
@x ibid
disc_node: @<Try to break after a discretionary fragment@>;
@y
disc_node: @<Try to break after a discretionary fragment, then |goto done5|@>;
@z
@x ibid
end
@y
done5:end
@z
Combine modules 868 and 869 into a single module, with a semicolon between.
@x module 870 (which becomes module 869) gets ... in its title
@y and the following new code just before its final "end":
r:=replace_count(cur_p); s:=link(cur_p);
while r>0 do
  begin @<Add the width of node |s| to |act_width|@>;
  decr(r); s:=link(s);
  end;
prev_p:=cur_p; cur_p:=s; goto done5;
@z
@x Now module 871 becomes 870, and we add a new (very similar) module 871:
@y
@ @<Add the width of node |s| to |act_width|@>=
if is_char_node(s) then
  begin f:=font(s);
  act_width:=act_width+char_width(f)(char_info(f)(character(s)));
  end
else  case type(s) of
  ligature_node: begin f:=font(lig_char(s));
    act_width:=act_width+
      char_width(f)(char_info(f)(character(lig_char(s))));
    end;
  hlist_node,vlist_node,rule_node,kern_node:
    act_width:=act_width+width(s);
  othercases confusion("disc4")
@:this can't happen disc4}{\quad disc4@>
  endcases
@z
@x module 877
@!disc_break:boolean; {was the current break at a discretionary node?}
@y
@!disc_break:boolean; {was the current break at a discretionary node?}
@!post_disc_break:boolean; {and did it have a nonempty post-break part?}
@z
@x ibid
if cur_p<>null then @<Prune unwanted nodes at the beginning of the next line@>;
@y
if cur_p<>null then if not post_disc_break then
  @<Prune unwanted nodes at the beginning of the next line@>;
@z
@x module 881
q:=cur_break(cur_p); disc_break:=false;
@y
q:=cur_break(cur_p); disc_break:=false; post_disc_break:=false;
@z
@x module 883
  if not is_char_node(s) then if next_break(cur_p)<>null then
    if cur_break(next_break(cur_p))=s then s:=r;
@y that subtle bug is no longer possible so we can remove the code
@z
@x module 884
link(s):=r; r:=post_break(q); post_break(q):=null;
@y
link(s):=r; r:=post_break(q); post_break(q):=null; post_disc_break:=true;
@z

377. Pick up a few discretionary hyphens that were lost because of
the new (cautious) algorithm.
@x module 914
@ @d advance_major_tail==begin major_tail:=link(major_tail); incr(r_count);
    end

@<Create and append a discretionary node as an alternative...@>=
begin r:=get_node(small_node_size);
@y
@ In this repeat loop we will insert another discretionary if |hyf[j-1]| is
odd, after both branches of the previous discretionary end at position |j-1|.
Strictly speaking, we aren't justified in doing this, because we don't know
that a hyphen after |j-1| is truly independent of those branches. But in almost
all applications we would rather not lose a potentially valuable hyphenation
point. (Consider the word `difficult', where the letter `c' is in position |j|.)

@d advance_major_tail==begin major_tail:=link(major_tail); incr(r_count);
    end

@<Create and append a discretionary node as an alternative...@>=
repeat r:=get_node(small_node_size);
@z
@x ibid
i:=hyphen_passed;
@y
i:=hyphen_passed; hyf[i]:=0;
@z
@x ibid
end
@y
hyphen_passed:=j-1; link(hold_head):=null;
until not odd(hyf[j-1])
@z

378. Undumped trie must also be dumpable again (Breitenlohner, 11 Dec 89)
@x module 1325
undump_size(0)(trie_size)('trie size')(j); {|trie_max|}
for k:=0 to j do undump_hh(trie[k]);
undump_size(0)(trie_op_size)('trie op size')(j); {|trie_op_ptr|}
@y
undump_size(0)(trie_size)('trie size')(j); @+init trie_max:=j;@+tini
for k:=0 to j do undump_hh(trie[k]);
undump_size(0)(trie_op_size)('trie op size')(j); @+init trie_op_ptr:=j;@+tini
@z
@x ibid
k:=256;
while j>0 do
  begin undump(0)(k-1)(k); undump(1)(j)(x); j:=j-x; op_start[k]:=qo(j);
@y
init for k:=0 to 255 do trie_used[k]:=min_quarterword;@+tini
k:=256;
while j>0 do
  begin undump(0)(k-1)(k); undump(1)(j)(x);@+init trie_used[k]:=qi(x);@+tini
  j:=j-x; op_start[k]:=qo(j);
@z

379. Allow output routine to access page totals. (Suggested by Frank
Mittelbach and Chris Rowley, December 1989)
@x module 421
begin if page_contents=empty then
@y
begin if (page_contents=empty) and (not output_active) then
@z

380. Slightly more robust recovery and detection of \output anomalies.
(Suggested by Chris Thompson, provoked by George Russell, January 1990)
@x module 1026
begin if loc<>null then @<Recover from an unbalanced output routine@>;
@y
begin if (loc<>null) or 
 ((token_type<>output_text)and(token_type<>backed_up)) then
  @<Recover from an unbalanced output routine@>;
@z
@x module 1027
help2("Your sneaky output routine has fewer real {'s than }'s.")@/
@y
help2("Your sneaky output routine has problematic {'s and/or }'s.")@/
@z

381. \eqno\aftergroup*$$ should yield * after the $$ (Michael Downes, 29 Jan 90)
@x module 1194
  begin cur_mlist:=p; cur_style:=text_style; mlist_penalties:=false;
@y
  begin @<Check that another \.\$ follows@>;
  cur_mlist:=p; cur_style:=text_style; mlist_penalties:=false;
@z
@x ibid
else  begin @<Check that another \.\$ follows@>;
@y
else  begin if a=null then @<Check that another \.\$ follows@>;
@z

382. Missed a case in the 8-bit change (Wayne Sullivan, 1 Feb 1990)
@x module 360
    begin if (end_line_char<0)or(end_line_char>127) then incr(limit);
@y
    begin if end_line_char_inactive then incr(limit);
@z

383. Linebreaking needs to be more robust when total demerits get
very high (Frank Mittelbach, 22 Feb 1990)
@x module 830
@!artificial_badness:boolean; {has |b| been forced to zero?}
@y
@!artificial_demerits:boolean; {has |d| been forced to zero?}
@z
@x module 833 (this change is cosmetic only)
@!minimal_demerits:array[very_loose_fit..tight_fit] of scaled; {best total
  demerits known for current line class and position, given the fitness}
@!minimum_demerits:scaled; {best total demerits known for current line class
@y
@!minimal_demerits:array[very_loose_fit..tight_fit] of integer; {best total
  demerits known for current line class and position, given the fitness}
@!minimum_demerits:integer; {best total demerits known for current line class
@z
@x module 836 here we want to ensure that total_demerits < awful_bad
minimum_demerits:=minimum_demerits+abs(adj_demerits);
@y
if abs(adj_demerits)>=awful_bad-minimum_demerits then
  minimum_demerits:=awful_bad-1
else minimum_demerits:=minimum_demerits+abs(adj_demerits);
@z
@x module 851
begin @!stat artificial_badness:=false;@+tats@/
@y
begin artificial_demerits:=false;@/
@z
@x module 854
  begin b:=0; {set badness zero, this break is forced}
  @!stat artificial_badness:=true;@+tats
  end
@y
  artificial_demerits:=true {set demerits zero, this break is forced}
@z
@x module 855
@<Compute the demerits, |d|, from |r| to |cur_p|@>;
@y
if artificial_demerits then d:=0
else @<Compute the demerits, |d|, from |r| to |cur_p|@>;
@z
@x module 856
if artificial_badness then print_char("*")@+else print_int(b);
@.*\relax@>
print(" p="); print_int(pi);
print(" d="); print_int(d);
@y
if b>inf_bad then print_char("*")@+else print_int(b);
@.*\relax@>
print(" p="); print_int(pi); print(" d=");
if artificial_demerits then print_char("*")@+else print_int(d);
@z
@x module 859
d:=line_penalty+b; d:=d*d;
@y
begin d:=line_penalty+b;
if abs(d)>=10000 then d:=100000000@+else d:=d*d;
@z
@x ibid
if abs(fit_class-fitness(r))>1 then d:=d+adj_demerits
@y
if abs(fit_class-fitness(r))>1 then d:=d+adj_demerits;
end
@z

384. Math fonts may disappear outside of \eqno (Marc van Leeuwen).
@x module 1194
  if danger then flush_math;
@y
  danger:=false;
  @<Check that the necessary fonts for math symbols are present;
    if not, flush the current math lists and set |danger:=true|@>;
@z

385. Forgot to rule out charnode when testing for node type (Marc van Leeuwen).
@x module 805
  begin if type(q)=unset_node then
    @<Set the unset box |q| and the unset boxes in it@>
  else if type(q)=rule_node then
    @<Make the running dimensions in rule |q| extend to the
      boundaries of the alignment@>;
@y
  begin if not is_char_node(q) then
    if type(q)=unset_node then
      @<Set the unset box |q| and the unset boxes in it@>
    else if type(q)=rule_node then
      @<Make the running dimensions in rule |q| extend to the
        boundaries of the alignment@>;
@z
@x module 903
if type(hb)=ligature_node then if odd(subtype(hb)) then
@y
if not is_char_node(hb) then
 if type(hb)=ligature_node then if odd(subtype(hb)) then
@z
@x module 1202
  if p<>null then if type(p)=glue_node then d:=0;
@y
  if p<>null then if not is_char_node(p) then if type(p)=glue_node then d:=0;
@z

386. Don't change the font of punctuation before a hyphenated word
     (Scott Allendorf, reported 7 Mar 90)
@x module 895
label common_ending,done,found,found1,not_found,not_found+1,exit;
@y
label common_ending,done,found,found1,found2,not_found,exit;
@z
@x module 903
if is_char_node(ha) then
  begin init_list:=ha; init_lig:=false; hu[0]:=qo(character(ha));
  end
else if type(ha)=ligature_node then
  begin init_list:=lig_ptr(ha); init_lig:=true; init_lft:=(subtype(ha)>1);
  hu[0]:=qo(character(lig_char(ha)));
  if init_list=null then if init_lft then
    begin hu[0]:=256; init_lig:=false;
    end; {in this case a ligature will be reconstructed from scratch}
  free_node(ha,small_node_size);
  end
else goto not_found+1; {no punctuation found}
s:=cur_p; {we have |cur_p<>ha| because |type(cur_p)=glue_node|}
while link(s)<>ha do s:=link(s);
j:=0; goto common_ending;
not_found+1: j:=1; s:=ha; init_list:=null;
if not is_char_node(r) then if type(r)=ligature_node then
 if subtype(r)>1 then
  begin j:=0; hu[0]:=256; init_lig:=false;
  end;
@y
if is_char_node(ha) then
  if font(ha)<>hf then goto found2
  else begin init_list:=ha; init_lig:=false; hu[0]:=qo(character(ha));
    end
else if type(ha)=ligature_node then
  if font(lig_char(ha))<>hf then goto found2
  else begin init_list:=lig_ptr(ha); init_lig:=true; init_lft:=(subtype(ha)>1);
    hu[0]:=qo(character(lig_char(ha)));
    if init_list=null then if init_lft then
      begin hu[0]:=256; init_lig:=false;
      end; {in this case a ligature will be reconstructed from scratch}
    free_node(ha,small_node_size);
    end
else begin {no punctuation found; look for left boundary}
  if not is_char_node(r) then if type(r)=ligature_node then
   if subtype(r)>1 then goto found2;
  j:=1; s:=ha; init_list:=null; goto common_ending;
  end;
s:=cur_p; {we have |cur_p<>ha| because |type(cur_p)=glue_node|}
while link(s)<>ha do s:=link(s);
j:=0; goto common_ending;
found2: s:=ha; j:=0; hu[0]:=256; init_lig:=false; init_list:=null;
@z

387. Balance the parens showing on the terminal (for Lispers).
@x module 304
@!in_open : 0..max_in_open; {the number of lines in the buffer, less one}
@y
@!in_open : 0..max_in_open; {the number of lines in the buffer, less one}
@!open_parens : 0..max_in_open; {the number of open text files}
@z
@x module 331
in_open:=0; max_buf_stack:=0;
@y
in_open:=0; open_parens:=0; max_buf_stack:=0;
@z
@x module 362
  begin print_char(")"); force_eof:=false;
  update_terminal; {show user that file has been read}
@y
  begin print_char(")"); decr(open_parens);
  update_terminal; {show user that file has been read}
  force_eof:=false;
@z
@x module 537
print_char("("); print(name); update_terminal; state:=new_line;
@y
print_char("("); incr(open_parens); print(name); update_terminal;
state:=new_line;
@z
@x module 1334
if job_name=0 then open_log_file;
@y
if job_name=0 then open_log_file;
while open_parens>0 do
  begin print(" )"); decr(open_parens);
  end;
@z

388. Optimize \ifx\p\q after \let\p=\q (Marc van Leeuwen says that
  AmS-TeX uses this a lot).
@x module 508
  while (p<>null)and(q<>null) do
  if info(p)<>info(q) then p:=null
  else  begin p:=link(p); q:=link(q);
    end;
b:=((p=null)and(q=null));
@y
if p=q then b:=true
else begin while (p<>null)and(q<>null) do
    if info(p)<>info(q) then p:=null
    else  begin p:=link(p); q:=link(q);
      end;
  b:=((p=null)and(q=null));
  end;
@z
@x module 538 (while we're at it, might as well optimize this too)
begin if not input_ln(cur_file,false) then do_nothing;
@y
begin if input_ln(cur_file,false) then do_nothing;
@z
@x module 1020 (and this)
else  begin wait:=false; s:=ins_ptr(p);
  link(last_ins_ptr(r)):=s; s:=last_ins_ptr(r);
@y
else  begin wait:=false; s:=last_ins_ptr(r); link(s):=ins_ptr(p);
@z

389. Treat migration properly in displays (Marc van Leeuwen).
@x module 1199 (a bad bug that really has bitten)
adjust_tail:=adjust_head; b:=hpack(p,natural);
@y
adjust_tail:=adjust_head; b:=hpack(p,natural); p:=list_ptr(b);
@z
@x module 1204
shift_amount(b):=s+d; append_to_vlist(b);
if t<>adjust_head then
  begin link(tail):=link(adjust_head); tail:=t;
   end
@y
shift_amount(b):=s+d; append_to_vlist(b)
@z
@x module 1205
  tail_append(new_penalty(post_display_penalty));
  end
else  begin tail_append(new_penalty(post_display_penalty));
  tail_append(new_param_glue(g2));
  end
@y
  g2:=0;
  end;
if t<>adjust_head then {migrating material comes after equation number}
  begin link(tail):=link(adjust_head); tail:=t;
  end;
tail_append(new_penalty(post_display_penalty));
if g2>0 then tail_append(new_param_glue(g2))
@z

-----------Here I draw the line with respect to further changes

390. Uninitialized nullfont parameters (found by Lance Carnes, 11 May 90).
@x module 552
hyphen_char[null_font]:="-"; skew_char[null_font]:=-1;
@y
hyphen_char[null_font]:="-"; skew_char[null_font]:=-1;
bchar_label[null_font]:=non_address;
font_bchar[null_font]:=non_char; font_false_bchar[null_font]:=non_char;
@z

391. Disable \write{\the\prevgraf} (B. Jackowski, July 1990).
@x module 422
begin nest[nest_ptr]:=cur_list; p:=nest_ptr;
while abs(nest[p].mode_field)<>vmode do decr(p);
scanned_result(nest[p].pg_field)(int_val);
end
@y
if mode=0 then scanned_result(0)(int_val) {|prev_graf=0| within \.{\\write}}
else begin nest[nest_ptr]:=cur_list; p:=nest_ptr;
  while abs(nest[p].mode_field)<>vmode do decr(p);
  scanned_result(nest[p].pg_field)(int_val);
  end
@z

392. Report correct line number when buffer overflows (George Russell).
@x module 538
begin if input_ln(cur_file,false) then do_nothing;
firm_up_the_line;
if end_line_char_inactive then decr(limit)
else  buffer[limit]:=end_line_char;
first:=limit+1; loc:=start; line:=1;
@y
begin line:=1;
if input_ln(cur_file,false) then do_nothing;
firm_up_the_line;
if end_line_char_inactive then decr(limit)
else  buffer[limit]:=end_line_char;
first:=limit+1; loc:=start;
@z

393. (I sincerely hope that there won't be any more)

* Possibly nice ideas that will not be implemented

. Classes of marks analogous to classes of insertions

. \showcontext to show the current location without stopping for error

. \everyeof to insert tokens before an \input file ends
  (strange example: \everyeof{\noexpand} will allow things
    like \xdef\a{\input foo}!)

. generalize \leftskip and \rightskip to token lists (problems with
   displayed math then)

. generalize \widowline and \clubline to go further into a paragraph

* Bad ideas that will not be implemented

. several people want to be able to remove arbitrary elements of lists,
   but that must never be done because some of those elements (e.g. kerns
   for accents) depend on floating point arithmetic

. if anybody wants letter spacing desperately they should put it in their own
   private version (e.g. generalize the hpack routine) and NOT call it TeX.