dtos copies its standard input to standard output, converting Fortran double-precision constants, built-in functions, and type declarations to single precision.
Floating-point FORMAT specifications are left intact; on some ancient systems, they may require modifications. They do not under the rules of Fortran 77.
Leading tabs are correctly interpreted according to common extended Fortran rules.
dtos recognizes all of the standard Fortran 77 single- and double-precision functions, as well as the pair rand/drand (UNIX pseudo-random number generators), and the pair r1mach/d1mach from the PORT library framework.
dtos's other purpose is to demonstrate a modest lex(1) program.
All options in this program are diagnostic, and suppress processing of the input stream. Execution terminates with a success return code after processing one or more options, but unrecognized options cause immediate termination with a failure return code.
Text beyond column 72 is discarded when lines are collected into Fortran statements.
dtos does not handle embedded ASCII tab characters correctly when long lines are to be broken. A Fortran-sensitive detabbing utility should be applied first if the input file possibly contains embedded tabs. Note that expand(1) cannot be used to do this job correctly!
Mixed-precision code may not be converted correctly. For example, DBLE(FLOAT(N)) will become SNGL(FLOAT(N)), which is syntactically incorrect.
Functions and variables of type COMPLEX are not converted, because Fortran 77 does not define a double precision complex type. Complex constants will be converted, however, since their real and imaginary parts look like normal floating-point values.
Nelson H. F. Beebe Department of Mathematics University of Utah Salt Lake City, UT 84112 Tel: +1 801 581 5254 FAX: +1 801 581 4148 Email: beebe@math.utah.edu, beebe@acm.org, beebe@computer.org