[Patch, Fortran] PR fortran/37425: Finish type-bound operators

[Daniel Kraft]

Hi Tobi and Paul,

Paul Richard Thomas wrote:
>> So, I think you could construct a testcase that breaks if you use EXEC_CALL
>> instead EXEC_ASSIGN_CALL.  This should also go into the testsuite.
> 
> EXEC_ASSIGN_CALL was introduced to fix PR25746, I think that you will
> find that the associated elemental_subroutine_3.f90 tests the
> dependency checking.

thanks for the hints on EXEC_ASSIGN_CALL!  I derived a type-bound 
operator version of elemental_subroutine_3.f90, but unfortunatly I came 
across a general type-bound procedures bug (calling ELEMENTAL type-bound 
procedures on an array is rejected including an ICE; I think this is 
valid, isn't it?) -- I will open a PR for this and work on it.

However, with a quick hack to temporarily fix that problem, my derived 
test really aborted; I changed my patch accordingly so that 
EXPR_COMPCALLs do have an assign-flag now (I think that's a better 
solution than adding another EXPR/EXEC_ASSIGN_COMPCALL) and generate 
EXEC_ASSIGN_CALLs when necessary, and this made the dependency-check 
work correctly.

Attached is the most recent version of my patch, now from my point of 
view every issues are resolved.  Ok for trunk after a fresh regression-test?

I will then work on the general ELEMENTAL problem and commit a test-case 
for the dependency-check when that is fixed.

Yours,
Daniel

-- 
Done:  Arc-Bar-Cav-Ran-Rog-Sam-Tou-Val-Wiz
To go: Hea-Kni-Mon-Pri
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: patch.changelog
URL: <http://gcc.gnu.org/pipermail/fortran/attachments/20090825/a1602097/attachment.ksh>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: patch.diff
URL: <http://gcc.gnu.org/pipermail/fortran/attachments/20090825/a1602097/attachment-0001.ksh>