bug-gnu-emacs
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

bug#43725: 28.0.50; Include feature/native-comp into master


From: Andrea Corallo
Subject: bug#43725: 28.0.50; Include feature/native-comp into master
Date: Tue, 06 Oct 2020 16:39:11 +0000
User-agent: Gnus/5.13 (Gnus v5.13) Emacs/28.0.50 (gnu/linux)

Andrea Corallo via "Bug reports for GNU Emacs, the Swiss army knife of
text editors" <bug-gnu-emacs@gnu.org> writes:

> Lars Ingebrigtsen <larsi@gnus.org> writes:
>
>> Eli Zaretskii <eliz@gnu.org> writes:
>>
>>>> Mmmh, I think technically we could, similarly to what we do for the
>>>> advice, synthesize compile and install a trampoline.  This would read
>>>> the symbol-function and calls what's in inside.
>>>> 
>>>> This trampoline installation would be triggered inside Ffset.
>>>> 
>>>> So yeah I think we could, if that's the preferred way I can try this
>>>> way.
>>>
>>> Lars, Stefan: do you agree that this is the preferred way?
>>
>> I'm not really qualified to have an opinion here, but if this allows
>> redefining primitives, I'm all for it.  Redefining primitives is a
>> useful tool.
>>
>> Would these trampolines be installed only if the primitives are
>> redefined, so there'd be no performance impact on code running normally?
>
> That's correct.
>
> I did some experimentation today and also the implementation was easy as
> the trampoline to be synthesized is exactly the same to what we
> synthesize already for advising, essentially I just had to add the
> proper trigger in fset.
>
> At this point I'm also for going this way as it just reduce the
> incompatibly surface and I don't see considerable downsides.
>
>   Andrea

As my understanding is that that's the consensus I've pushed the revert
of those changes in the testsuite and made Ffset effective also for
redefining primitives.

FYI with the last tweaks `make check` is clean.

  Andrea





reply via email to

[Prev in Thread] Current Thread [Next in Thread]