At 02:37 AM 8/14/2006, Arjen Markus wrote:
>[EMAIL PROTECTED] wrote:
>
>>Hi Alan,
>>
>>My knowledgebase of fortran lies with fixed format f77 fortran, so your right
>>the solution below only works for fixed format Fortran. There is currently
>>nothing in the .l file to set the <fixed_fmt> state, so the line:
>>
>><fixed_fmt>^[cC*dD].*\n { return EOSTMT; }
>>never gets triggered. Having only worked with fixed format fortran and not
>>having a newer fortran book on my book shelf... How does the compiler
>>determine whether a file is fixed or free format? A test needs to be put in
>>the .l file to determine whether fixed or free format fortran is being used.
>>
>All Fortran compilers I know of use two methods to select between free form or
>fixed form source:
>- The extension of the file (.f and .for are used for fixed form, .f90 for
>free form)
>- A command-line option which tells the compiler to ignore the first rule and
>accept
> the source file as either fixed form or free form.
>
>That said, similar rules could be adopted for CMake. (Of course the
>command-line option
>makes it a bit awkward, but then anybody who wants to use that will somehow
>have
>to specify that he/she wants that anyway)
Can we just have the rule for both? Basically, anything after a c should be a
comment?
If it is old fixed format fortran it won't have modules anyway, and should have
no depends.
So, if it finds too many comments because of a stray c in the code it won't
hurt.
-Bill
_______________________________________________
CMake mailing list
[email protected]
http://www.cmake.org/mailman/listinfo/cmake
