Re: [PATCH] Pass const frame_info_ptr reference for skip_[language_]trampoline

[Tom de Vries <tdevries@suse.de>]

On 5/3/23 17:11, Mark Wielaard wrote:
> On Wed, May 03, 2023 at 08:50:07AM -0600, Tom Tromey wrote:
>>>>>>> "Mark" == Mark Wielaard <mark@klomp.org> writes:
>>
>> Mark> It isn't clear why only in this one place g++ sees the issue (probably
>> Mark> because it can inlined enough code in this specific case).
>>
>> s/inlined/inline/
>>
>> Could you add a "Bug:" trailer with a link to the bugzilla entry?
> 
> Fixed typo and added Bug link.
> 
>      PR build/30413
>      Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=30413
> 
>> Anyway, other than that nit, this looks ok to me.  It's a little
>> unfortunate we don't know why exactly this particular spot warns, but on
>> the other hand, the fix is harmless (actually an improvement) and we've
>> worked around plenty of compiler oddities before.
>>
>> Reviewed-By: Tom Tromey <tom@tromey.com>
> 
> Thanks, pushed with that added. So that the code compiles again.
> 
> In the bug Tom de Vries has some ideas to "fix" this more generically
> by changing the frame_info_ptr destructor to give the compiler even
> more visibility into what is happening, which might help prevent
> similar issues in the future if you don't want to change a pass by
> value by a pass by reference.
> 

Hi Mark,

thanks for fixing this.

I've submitted a patch at 
https://sourceware.org/pipermail/gdb-patches/2023-May/199342.html .

Indeed it's not a "fix" for the warning as such.

It's more a follow-up on the observation that the code that triggers the 
warning was introduced for a reason that's no longer valid, so we can 
just remove it and simplify the code, which then also "fixes" the warning.

Thanks,
- Tom