bug#48221: 28.0.50; defsubsts aren't native compiled?

bug#48221: 28.0.50; defsubsts aren't native compiled?

[Lars Ingebrigtsen]

The following test has been failing the past few days.

1 unexpected results:
   FAILED  help-fns-test-lisp-defsubst

And, indeed, `C-h f' on the subst in question says:

posn-window is a compiled Lisp function in ‘subr.el’.

While a regular function is native compiled:

last is a native compiled Lisp function in ‘subr.el’.



In GNU Emacs 28.0.50 (build 1, x86_64-pc-linux-gnu, GTK+ Version 3.24.24, cairo version 1.16.0)
 of 2021-04-30 built on xo
Repository revision: 7c901d90e620b4d3651b86c13faf1e81eeb3db10
Repository branch: master
Windowing system distributor 'The X.Org Foundation', version 11.0.12010000
System Description: Debian GNU/Linux bullseye/sid


-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#48221: 28.0.50; defsubsts aren't native compiled?

[Andrea Corallo via Bug reports for GNU Emacs, the Swiss army knife of text editors]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> The following test has been failing the past few days.
>
> 1 unexpected results:
>    FAILED  help-fns-test-lisp-defsubst
>
> And, indeed, `C-h f' on the subst in question says:
>
> posn-window is a compiled Lisp function in ‘subr.el’.
>
> While a regular function is native compiled:
>
> last is a native compiled Lisp function in ‘subr.el’.

Hi Lars,

AFAIR having the defsubst non native compiled is what we want as we need
the bytecode definition to have the inline to be performed.  Actually
the test is verifying that the defsubst is *not* native compiled.

I tried to reproduce here (0c993ed1d3) but for me the test is passing,
that's odd...

Thanks

  Andrea

bug#48221: 28.0.50; defsubsts aren't native compiled?

[Lars Ingebrigtsen]

Andrea Corallo <akrl@sdf.org> writes:

> AFAIR having the defsubst non native compiled is what we want as we need
> the bytecode definition to have the inline to be performed.  Actually
> the test is verifying that the defsubst is *not* native compiled.
>
> I tried to reproduce here (0c993ed1d3) but for me the test is passing,
> that's odd...

Hm, odd.  The two tests here are:

----
(ert-deftest help-fns-test-lisp-defun ()
  (let ((regexp (if (boundp 'comp-ctxt)
                    "a native compiled Lisp function in .+subr\\.el"
                  "a compiled Lisp function in .+subr\\.el"))
        (result (help-fns-tests--describe-function 'last)))
    (should (string-match regexp result))))

(ert-deftest help-fns-test-lisp-defsubst ()
  (let ((regexp (if (boundp 'comp-ctxt)
                    "a native compiled Lisp function in .+subr\\.el"
                  "a compiled Lisp function in .+subr\\.el"))
        (result (help-fns-tests--describe-function 'posn-window)))
    (should (string-match regexp result))))
----

Unless I'm really misreading something here, they both seem to expect
the native-compiled results?

(A related question -- is (boundp 'comp-ctxt) really the best way to
test for whether we have a native-compiled Emacs?  That's just a defvar
in comp-cstr.el...)

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#48221: 28.0.50; defsubsts aren't native compiled?

[Eli Zaretskii]

> From: Lars Ingebrigtsen <larsi@gnus.org>
> Date: Wed, 05 May 2021 11:36:02 +0200
> Cc: 48221@debbugs.gnu.org
> 
> (A related question -- is (boundp 'comp-ctxt) really the best way to
> test for whether we have a native-compiled Emacs?  That's just a defvar
> in comp-cstr.el...)

We could perhaps provide 'native-compilation in comp.c?  WDYT?

bug#48221: 28.0.50; defsubsts aren't native compiled?

[Andrea Corallo via Bug reports for GNU Emacs, the Swiss army knife of text editors]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Lars Ingebrigtsen <larsi@gnus.org>
>> Date: Wed, 05 May 2021 11:36:02 +0200
>> Cc: 48221@debbugs.gnu.org
>> 
>> (A related question -- is (boundp 'comp-ctxt) really the best way to
>> test for whether we have a native-compiled Emacs?  That's just a defvar
>> in comp-cstr.el...)
>
> We could perhaps provide 'native-compilation in comp.c?  WDYT?

We can use since a while (featurep 'nativecomp), I probably missed this
substitution, I'm preparing a patch.

  Andrea

bug#48221: 28.0.50; defsubsts aren't native compiled?

[Andrea Corallo via Bug reports for GNU Emacs, the Swiss army knife of text editors]

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

> Eli Zaretskii <eliz@gnu.org> writes:
>
>>> From: Lars Ingebrigtsen <larsi@gnus.org>
>>> Date: Wed, 05 May 2021 11:36:02 +0200
>>> Cc: 48221@debbugs.gnu.org
>>> 
>>> (A related question -- is (boundp 'comp-ctxt) really the best way to
>>> test for whether we have a native-compiled Emacs?  That's just a defvar
>>> in comp-cstr.el...)
>>
>> We could perhaps provide 'native-compilation in comp.c?  WDYT?
>
> We can use since a while (featurep 'nativecomp), I probably missed this
> substitution, I'm preparing a patch.

3481903d13

  Andrea

bug#48221: 28.0.50; defsubsts aren't native compiled?

[Eli Zaretskii]

> From: Andrea Corallo <akrl@sdf.org>
> Cc: Lars Ingebrigtsen <larsi@gnus.org>, 48221@debbugs.gnu.org
> Date: Wed, 05 May 2021 15:03:02 +0000
> 
> >> We could perhaps provide 'native-compilation in comp.c?  WDYT?
> >
> > We can use since a while (featurep 'nativecomp), I probably missed this
> > substitution, I'm preparing a patch.
> 
> 3481903d13

Thanks.  I think the feature should be renamed to native-comp, like
the other variables we talked about.

bug#48221: 28.0.50; defsubsts aren't native compiled?

[Eli Zaretskii]

> Date: Wed, 05 May 2021 18:10:58 +0300
> From: Eli Zaretskii <eliz@gnu.org>
> Cc: larsi@gnus.org, 48221@debbugs.gnu.org
> 
> Thanks.  I think the feature should be renamed to native-comp

Sorry, I meant native-compile.

bug#48221: 28.0.50; defsubsts aren't native compiled?

[Andrea Corallo via Bug reports for GNU Emacs, the Swiss army knife of text editors]

Eli Zaretskii <eliz@gnu.org> writes:

>> Date: Wed, 05 May 2021 18:10:58 +0300
>> From: Eli Zaretskii <eliz@gnu.org>
>> Cc: larsi@gnus.org, 48221@debbugs.gnu.org
>> 
>> Thanks.  I think the feature should be renamed to native-comp
>
> Sorry, I meant native-compile.

NP, BTW in the configure we use `native-compilation' shouldn't we use
this?

Thanks

  Andrea

bug#48221: 28.0.50; defsubsts aren't native compiled?

[Lars Ingebrigtsen]

Andrea Corallo <akrl@sdf.org> writes:

>>> Thanks.  I think the feature should be renamed to native-comp
>>
>> Sorry, I meant native-compile.
>
> NP, BTW in the configure we use `native-compilation' shouldn't we use
> this?

"native-compilation" is a handful...  would it make sense to change that
to "--with-native-compile" instead?  (I don't really have an opinion
here, though.)

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#48221: 28.0.50; defsubsts aren't native compiled?

[Andrea Corallo via Bug reports for GNU Emacs, the Swiss army knife of text editors]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> Andrea Corallo <akrl@sdf.org> writes:
>
>>>> Thanks.  I think the feature should be renamed to native-comp
>>>
>>> Sorry, I meant native-compile.
>>
>> NP, BTW in the configure we use `native-compilation' shouldn't we use
>> this?
>
> "native-compilation" is a handful...  would it make sense to change that
> to "--with-native-compile" instead?  (I don't really have an opinion
> here, though.)

FWIW I've really no preference here, I was asked for
"--with-native-compilation", as long as we decide something it's all
good by me.

  Andrea

bug#48221: 28.0.50; defsubsts aren't native compiled?

[Eli Zaretskii]

> From: Andrea Corallo <akrl@sdf.org>
> Cc: Eli Zaretskii <eliz@gnu.org>, 48221@debbugs.gnu.org
> Date: Wed, 05 May 2021 15:46:02 +0000
> 
> > "native-compilation" is a handful...  would it make sense to change that
> > to "--with-native-compile" instead?  (I don't really have an opinion
> > here, though.)
> 
> FWIW I've really no preference here, I was asked for
> "--with-native-compilation", as long as we decide something it's all
> good by me.

I'm not sure it is worth our while to make small changes in this name,
since I believe we will want to release Emacs 28.1 with that option
being the default (if libgccjit is available).

bug#48221: 28.0.50; defsubsts aren't native compiled?

[Eli Zaretskii]

> From: Andrea Corallo <akrl@sdf.org>
> Cc: larsi@gnus.org, 48221@debbugs.gnu.org
> Date: Wed, 05 May 2021 15:23:47 +0000
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> >> Date: Wed, 05 May 2021 18:10:58 +0300
> >> From: Eli Zaretskii <eliz@gnu.org>
> >> Cc: larsi@gnus.org, 48221@debbugs.gnu.org
> >> 
> >> Thanks.  I think the feature should be renamed to native-comp
> >
> > Sorry, I meant native-compile.
> 
> NP, BTW in the configure we use `native-compilation' shouldn't we use
> this?

I thought native-compile was better because we talked about renaming a
bunch of variables to have that prefix, so it would be consistent.

bug#48221: 28.0.50; defsubsts aren't native compiled?

[Andrea Corallo via Bug reports for GNU Emacs, the Swiss army knife of text editors]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Andrea Corallo <akrl@sdf.org>
>> Cc: larsi@gnus.org, 48221@debbugs.gnu.org
>> Date: Wed, 05 May 2021 15:23:47 +0000
>> 
>> Eli Zaretskii <eliz@gnu.org> writes:
>> 
>> >> Date: Wed, 05 May 2021 18:10:58 +0300
>> >> From: Eli Zaretskii <eliz@gnu.org>
>> >> Cc: larsi@gnus.org, 48221@debbugs.gnu.org
>> >> 
>> >> Thanks.  I think the feature should be renamed to native-comp
>> >
>> > Sorry, I meant native-compile.
>> 
>> NP, BTW in the configure we use `native-compilation' shouldn't we use
>> this?
>
> I thought native-compile was better because we talked about renaming a
> bunch of variables to have that prefix, so it would be consistent.

Okay, I've the impression we all agreed on native-compile for the
feature name so I did the renaming with 40736357b8.  The configure flag
in any case is a different subject.

Thanks

  Andrea

bug#48221: 28.0.50; defsubsts aren't native compiled?

[Eli Zaretskii]

> From: Andrea Corallo <akrl@sdf.org>
> Cc: larsi@gnus.org, 48221@debbugs.gnu.org
> Date: Wed, 05 May 2021 20:33:49 +0000
> 
> Okay, I've the impression we all agreed on native-compile for the
> feature name so I did the renaming with 40736357b8.

Thanks.

Btw, I'm waiting for the other renames to begin updating the
documentation.  I hope it will happen soon, okay?

bug#48221: 28.0.50; defsubsts aren't native compiled?

[Andrea Corallo via Bug reports for GNU Emacs, the Swiss army knife of text editors]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Andrea Corallo <akrl@sdf.org>
>> Cc: larsi@gnus.org, 48221@debbugs.gnu.org
>> Date: Wed, 05 May 2021 20:33:49 +0000
>> 
>> Okay, I've the impression we all agreed on native-compile for the
>> feature name so I did the renaming with 40736357b8.
>
> Thanks.
>
> Btw, I'm waiting for the other renames to begin updating the
> documentation.  I hope it will happen soon, okay?

Sure, I've already started this, I'll be pushing commits today.

Regards

  Andrea

bug#48221: 28.0.50; defsubsts aren't native compiled?

[Eli Zaretskii]

> From: Andrea Corallo <akrl@sdf.org>
> Cc: larsi@gnus.org, 48221@debbugs.gnu.org
> Date: Thu, 06 May 2021 07:59:50 +0000
> 
> > Btw, I'm waiting for the other renames to begin updating the
> > documentation.  I hope it will happen soon, okay?
> 
> Sure, I've already started this, I'll be pushing commits today.

Great, that will pave the way for adding the necessary documentation
to the manuals.

bug#48221: 28.0.50; defsubsts aren't native compiled?

[Lars Ingebrigtsen]

Eli Zaretskii <eliz@gnu.org> writes:

> We could perhaps provide 'native-compilation in comp.c?  WDYT?

Yes, good idea.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#48221: 28.0.50; defsubsts aren't native compiled?

[Andrea Corallo via Bug reports for GNU Emacs, the Swiss army knife of text editors]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> Andrea Corallo <akrl@sdf.org> writes:
>
>> AFAIR having the defsubst non native compiled is what we want as we need
>> the bytecode definition to have the inline to be performed.  Actually
>> the test is verifying that the defsubst is *not* native compiled.
>>
>> I tried to reproduce here (0c993ed1d3) but for me the test is passing,
>> that's odd...
>
> Hm, odd.  The two tests here are:

Sorry, dunno why I was convinced your were talking about
`comp-test-defsubst' :/

31f64d862b fixes `help-fns-test-lisp-defsubst'.

Thanks

  Andrea

bug#48221: 28.0.50; defsubsts aren't native compiled?

[Lars Ingebrigtsen]

Andrea Corallo <akrl@sdf.org> writes:

> Sorry, dunno why I was convinced your were talking about
> `comp-test-defsubst' :/
>
> 31f64d862b fixes `help-fns-test-lisp-defsubst'.

Thanks; that fixes the problem here, too.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no