[Patch, Fortran] PR34899 - Support <tab><digit> as continuation line marker (fixed form)

[Tobias Burnus]

Jerry DeLisle wrote:
>> PS: The patch was motivated to get to build 
>> http://www.chara.gsu.edu/~gudehus/fits_library_package.html. However, 
>> due to other problems it still won't build. gfortran stops first for 
>> integer_var = .true. (unless -std=legacy) and then for "real_var = ' 
>> '". Replacing them by transfer it stops at the same line as ifort.
>
> The patch is OK. See my rant posted with the PR.  Lets not make 
> gfortran a catch all compiler.  The application you are wanting here 
> really needs to be fixed and done right.  Neither gfortran or ifort 
> should be expecte to compile crap.
Well, actually I do not need this application badly  - I was more 
curious to see whether gfortran can compile it.

I think the <tab><digit> makes sense since it is widely supported - 
including by g77 (which makes it a kind of regression) and NAG f95 
(which contains almost no vendor extensions).

Otherwise, I am (negatively) impressed that one can write a program 
which is so badly written that none of my 6 compilers can compile it 
while it is seemingly still compilable with some compilers. And no, I do 
not intent to enhance gfortran to be able to make it compile.

Tobias