---- Brad King <[EMAIL PROTECTED]> wrote:
>
> FYI, there is a bug report for this problem:
>
> http://www.cmake.org/Bug/bug.php?op=show&bugid=3109
>
> I briefly looked at fixing it but got stuck on the free-format versus
> fixed-format parsing. Unfortunately I have no time for this bug now.
>
> Please add any conclusion for this discussion to that bug report.
>
Brad,
I played around with it a bit this weekend.
Something like this could be done in cmDependsFortranLexer.in.l
...
%option reentrant
%option noyywrap
%pointer
%s free_fmt fixed_fmt
%x str_sq str_dq
%%
%{
bool fixed = cmDependsFortranParser_isFixedFormat(yyextra);
if (fixed)
BEGIN(fixed_fmt);
%}
\" {
cmDependsFortranParser_StringStart(yyextra);
BEGIN(str_dq);
}
...
!.*\n { return EOSTMT; } /* Treat comments like */
<fixed_fmt>^[cC*dD].*\n { return EOSTMT; } /* empty lines */
cmDependsFortranParser_isFixedFormat would be a function that we could add to
cmDependsFortran.cxx which returnss true if a given file is fixed format.
The question really becomes: How does cmDependsFortranParser_isFixedFormat
determine whether the file is fixed or free format? I believe I could add a
check for the file extension fairly easily. Adding a check for command line
flags or to see whether SET_SOURCE_FILE_PROPERTIES(FORTRAN_FIXED_FORMAT) (as
Alan suggested) looks to be more involved, as cmDependsFortran or the parser
does not appear to be plumbed to access that type of information.
If I submiited a patch with the changes to the parser and a
cmDependsFortranParser_isFixedFormat function that is based on file extension,
could we get it incoporated? We could then worry about more robust ways about
determining if a file is free or fixed format later.
Thanks,
Bryan Walsh
_______________________________________________
CMake mailing list
[email protected]
http://www.cmake.org/mailman/listinfo/cmake
