[gfortran] Cleanup handling of sym vs. sym->result was: Re: [Patch/gfortran] Fix for PR16939

[Tobias Schlüter]

Tobias Schlüter wrote:
> [ my original reply didn't make it to the list ]
> 
> THOMAS Paul Richard 169137 wrote:
> 
>>>Oh, and I'm wondering if I should have added 
>>>
>>>>!sym->attr.pointer to that
>>
>>
>>I think that your next remark is right.
> 
> 
> I'll look into this later today.

This is fixed in the attached patch together with a number of other cleanups,
made possible by my previous patch.  The combination of both implements what I
was saying here:

> Bleh, we need to make sure that every function has sym->result's properties
> set consistently with sym's properties.  That way we could do away with a lot
> of special cases.  There is code for this in resolve_formal_arglist but it
> doesn't handle everything, because all those special cases that are scattered
> around the code generation passes wouldn't be necessary then.

The necessary change to resolve.c was in the previous patch.  What is below is
the ensuing cleanup.  I also attached two more testcases which verify that
-ff2c works reliably.  I would have liked to test these testcases with both
-ff2c and -fno-f2c but I couldn't find a way of making the testsuite do this
without duplicating the testcases, so I left that to someone doing future work
on the testsuite.

Bubblestrapped and regtested, ok?

- Tobi

2005-05-11  Tobias Schl"uter  <tobias.schlueter@physik.uni-muenchen.de>

	* trans-expr.c (gfc_conv_variable): POINTER results don't need f2c
	calling conventions.  Look at sym instead of sym->result.
	* trans-types.c (gfc_sym_type): Remove workaround for frontend bug.
	Remove condition which is always false with workaround removed.
	(gfc_return_by_reference): Always look at sym, never at sym->result.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: followup.diff
Type: text/x-patch
Size: 2885 bytes
Desc: not available
URL: <http://gcc.gnu.org/pipermail/fortran/attachments/20050511/ddfbe264/attachment.bin>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: f2c_4.f90
URL: <http://gcc.gnu.org/pipermail/fortran/attachments/20050511/ddfbe264/attachment.f90>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: f2c_5.f90
URL: <http://gcc.gnu.org/pipermail/fortran/attachments/20050511/ddfbe264/attachment-0001.f90>