> Fixed-form, known as "reference format", is still more-or-less > required by IBM. Forced in gcobol with the option "-ffixed-form". > Can be controlled inside a source code file with the compiler > directive ">> SOURCE FORMAT IS FREE" > > Columns 1-6 ignored > Column 7 * for comment, - for continuation, and a few other > special things > Column 8 Area A -- labels started here > Column 12 Area B -- statements go here > Column 73 and beyond are ignored > > Free-form, is much more forgiving. Forced with "-ffree-form", or from > inside the program with ">> SOURCE FORMAT IS FIXED" > > There is no line limit in free form; the entire program could fit on a > single line. > > GCOBOL uses a heuristic when the format is unspecified. It looks at > first line of source code. If the first six characters are digits or > blanks, it switches to "extended reference format", where the first > six characters are ignored, column seven is the indicator column, *> > comments can start anywhere, and there is no line length limit.
This gives three reference-formats: "fixed" "free" and "extended". For two of those we have seen the flags -ffixed-form and -ffree-form, so I'd _guess_ the last one would be -fextended-form. Question: Is there a reason to have multiple flags for that? We dropped this in GnuCOBOL, finding that there are even more reference formats and use now a single one: -fformat=fixed/free/extended/.../auto I'd like to suggest to drop those current three flags to one flag for choosing the reference-format with a value telling the frontend which one to use. Side-note: auto-choosing "extended" was at least confusing for me (and the NIST suite initial compile-try). That likely confused me most, because of not knowing another compiler choosing _that_ format automatically. I'd recommend to choose "fixed" if there are line numbers recognized, but much more, I'd suggest to move from multiple format flags to a single one. Simon
