Re: [PATCH] Fix PR125481 Wrong generic resolution between procedure pointer and variable actuals of same TKR

Hi Jerry!

Am 01.06.26 um 7:31 PM schrieb Jerry D:
> On 5/31/26 3:28 PM, Jerry D wrote:
>> On 5/31/26 1:28 PM, Harald Anlauf wrote:
>>> Hi Jerry!
>>>
>>> Am 31.05.26 um 5:46 AM schrieb Jerry D:
>>>> The attached patch regression tested on x86_64.
>>>
>>> Are you sure that the reference to F18:15.5.2.4 is correct?
>>> I cannot find your quote there, at least not verbosely.
>>
>> I am working off drafts here. I will hunt this down in 2023 which is 
>> more current.
> 
> I found this in my draft 2023:  F23:15.5.2.5, para 2

I looked at this subclause a long time.  I was probably too boneheaded
to understand your comment in the code.  It is rather an interpretation
than a quote, because it says (J3/24-007):

"(1) The requirements in this subclause apply to actual arguments that
correspond to nonallocatable nonpointer dummy data objects.

(2) The dummy argument shall be type compatible with the actual 
argument. ..."

So not a halluzination of the LLM...

>>>
>>> Furthermore, can you provide a testcase where the new error message
>>> will be tested?
>>>
> 
> New test case added to exercise two possible errors.

Thanks, this is great.

> I have attached here a revised patch file. I have rerun regression tests 
> on my worktree.
> 
> If this is OK I will apply to trunk and rerun all of testsuite there 
> before I commit. I noticed some unrelated gfortran/gomp failures this 
> morning. I dont know what that is about.

Yes, it is OK for mainline and backport after a grace period.

Thanks for the patch!

Harald