[Bug fortran/50981] [4.4/4.5/4.6 Regression] Wrong-code for scalarizing ELEMENTAL call with absent OPTIONAL argument

["dominiq at lps dot ens.fr" <gcc-bugzilla at gcc dot gnu dot org>]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50981

--- Comment #30 from Dominique d'Humieres <dominiq at lps dot ens.fr> 2012-01-26 16:12:13 UTC ---
(In reply to comment #28)
> Created attachment 26468 [details]
> better patch
> 
> This one should work.

It does;-) 

I have applied the patch on revision 183541 on top of the Paul's patch at
http://gcc.gnu.org/ml/fortran/2012-01/msg00214.html (pr51870) and Tobias' ones
for pr51970 and pr51434 comment #12. 

It regtested without regression and all the failures reported in comments #25
and #26 are gone. It  passes the full comment #13 test fixed as in comment #22
(I blindly trust that it is the "right thing to do") and it fixes pr51514 and
the test in pr41587 comment #1 (at least part of it: it now prints "          0
          0          42" while I'ld expect "         42          42         
42"(?); without the patch I get an ICE
pr41587_db.f90:10:0: internal compiler error: in gfc_conv_descriptor_data_get,
at fortran/trans-array.c:147).

The test in comment #23 gives an ICE with/without the patch:

pr50981_4.f90: In function 'MAIN__':
pr50981_4.f90:16:0: internal compiler error: in fold_convert_loc, at
fold-const.c:2016

i.e., the same ICE as for pr51977 (I did not tested the patchlet in comment
#27).

Thanks for the work!