[Patch, Fortran, OOP] PR 45933:[4.6 regression] ICE in gfc_add_component_ref, at fortran/class.c:77

[Janus Weil]

Mark,

> Thank you for jumping on this so quickly.

you're welcome. Would be great if you could update to the latest
revision and check if your code triggers any further problems (and in
particular whether you get correct run-time behaviour).

Cheers,
Janus



> On Thu, 2010-10-07 at 18:26 +0200, Janus Weil wrote:
>> Hi all,
>>
>> this patch fixes the code which resolves type-bound operator calls: It
>> was only working if the passed-object is a simple variable, but failed
>> on more complicated expressions like derived-type components. The
>> reason for this being a regression is probably that we modified a lot
>> of the polymorphic TBP handling in 4.6.
>>
>> The patch was regtested on x86_64-unknown-linux-gnu. I think it's
>> pretty obvious. Ok for trunk?
>>
>> Cheers,
>> Janus
>>
>>
>> 2010-10-07  Janus Weil  <janus@gcc.gnu.org>
>>
>>       PR fortran/45933
>>       * resolve.c (resolve_typebound_function): Use correct declared type
>>       for type-bound operators.
>>
>>
>> 2010-10-07  Janus Weil  <janus@gcc.gnu.org>
>>
>>       PR fortran/45933
>>       * gfortran.dg/typebound_operator_5.f03: New.
>
>
>