make check produces 25 failures and hangs (bug #65176).

make check produces 25 failures and hangs (bug #65176).

[Alan Mackenzie]

Hello, Emacs.

For the past couple of days, with a freshly bootstrapped Emacs, I've been
seeing make check on the master branch producing lots of failures and
hanging with 0% processor usage.  After Ctrl-c'ing out of it and grepping
the log files, there are 25 failures noted.  I even raised bug #65176 for
this yesterday.

My system is GNU/Linux, and my .configure command line is:

    $ ./configure --with-gif=no --with-tiff=no --with-gpm \
    --with-native-compilation

Up till now there's been no reaction to this bug, and no signs on
emacs-devel of anybody else having any problems.

Is anybody else seeing these problems?  Has anybody else tried it, and
_not_ seen these problems?

Help would be appreciated.

Thanks!

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: make check produces 25 failures and hangs (bug #65176).

[chad]

[-- Attachment #1: Type: text/plain, Size: 2078 bytes --]

On Thu, Aug 10, 2023 at 3:19 PM Alan Mackenzie <acm@muc.de> wrote:

> [...]
> Is anybody else seeing these problems?  Has anybody else tried it, and
> _not_ seen these problems?
>

I'm also seeing them. If I let "make check" run for a very long time (I was
out of house for 90 minutes), it eventually terminates with 2 checks that
did not finish, in tree-sitter and native-comp. I haven't had the spare
cycles to investigate further yet.

Rerunning "make check" right now, I get a slightly different dnf:

SUMMARY OF TEST RESULTS
> -----------------------
> Files examined: 491
> Ran 7420 tests, 25 failed to run, 7253 results as expected, 9 unexpected,
> 158 skipped
> 1 files did not finish:
>   src/treesit-tests.log
> 3 files contained unexpected results:
>   src/comp-tests.log
>   lisp/progmodes/eglot-tests.log
>   lisp/help-fns-tests.log


In this case, src/treesit-tests.log contains:

 Running 25 tests (2023-08-10 15:49:31-0400, selector ‘(not (or (tag
> :expensive-test) (tag :unstable)))’)
>    passed   1/25  treesit-basic-parsing (0.001179 sec)
>    passed   2/25  treesit-cross-boundary (0.000796 sec)
>    passed   3/25  treesit-cursor-helper-with-missing-node (0.000672 sec)
> Can’t guess python-indent-offset, using defaults: 4
>    passed   4/25  treesit-defun-navigation-nested-1 (0.085175 sec)
>    passed   5/25  treesit-defun-navigation-nested-2 (0.124530 sec)
>    passed   6/25  treesit-defun-navigation-nested-3 (0.004818 sec)
>    passed   7/25  treesit-defun-navigation-nested-4 (0.006185 sec)
> Can’t guess python-indent-offset, using defaults: 4
>    passed   8/25  treesit-defun-navigation-top-level (0.006916 sec)
>    passed   9/25  treesit-indirect-buffer (0.000628 sec)


I've been seeing these failures for a little less than a week. FWIW,
libtreesitter is fluctuating frequently right now; I've rebuilt it from git
HEAD probably 3 times since I started seeing these errors, including this
morning.

Sorry that I can't dig into this much more at this time. Hope this helps,
~Chad

[-- Attachment #2: Type: text/html, Size: 2949 bytes --]

Re: make check produces 25 failures and hangs (bug #65176).

[Alan Mackenzie]

Hello, Chad.

On Thu, Aug 10, 2023 at 15:54:26 -0400, chad wrote:
>    On Thu, Aug 10, 2023 at 3:19â¯PM Alan Mackenzie <[1]acm@muc.de> wrote:

>      [...]
>      Is anybody else seeing these problems?  Has anybody else tried it,
>      and
>      _not_ seen these problems?

>    I'm also seeing them. If I let "make check" run for a very long time (I
>    was out of house for 90 minutes), it eventually terminates with 2
>    checks that did not finish, in tree-sitter and native-comp. I haven't
>    had the spare cycles to investigate further yet.

Thanks!  On the one hand, I'm sorry you're having these problems, on the
other hand, it's a relief that I'm not on my own!

>    Rerunning "make check" right now, I get a slightly different dnf:

>      SUMMARY OF TEST RESULTS
>      -----------------------
>      Files examined: 491
>      Ran 7420 tests, 25 failed to run, 7253 results as expected, 9
>      unexpected, 158 skipped
>      1 files did not finish:
>      Â  src/treesit-tests.log
>      3 files contained unexpected results:
>      Â  src/comp-tests.log
>      Â  lisp/progmodes/eglot-tests.log
>      Â  lisp/help-fns-tests.log

>    In this case, src/treesit-tests.log contains:

>      Â Running 25 tests (2023-08-10 15:49:31-0400, selector â(not (or
>      (tag :expensive-test) (tag :unstable)))â)
>         passed   1/25  treesit-basic-parsing (0.001179 sec)
>         passed   2/25  treesit-cross-boundary (0.000796 sec)
>         passed   3/25  treesit-cursor-helper-with-missing-node
>      (0.000672 sec)
>      Canât guess python-indent-offset, using defaults: 4
>         passed   4/25  treesit-defun-navigation-nested-1 (0.085175
>      sec)
>         passed   5/25  treesit-defun-navigation-nested-2 (0.124530
>      sec)
>         passed   6/25  treesit-defun-navigation-nested-3 (0.004818
>      sec)
>         passed   7/25  treesit-defun-navigation-nested-4 (0.006185
>      sec)
>      Canât guess python-indent-offset, using defaults: 4
>         passed   8/25  treesit-defun-navigation-top-level (0.006916
>      sec)
>         passed   9/25  treesit-indirect-buffer (0.000628 sec)

>    I've been seeing these failures for a little less than a week. FWIW,
>    ....

I've a suspicion (from a frustrating inability to use git bisect
yesterday) that these errors came in with the android support branch.

>    .... libtreesitter is fluctuating frequently right now; I've rebuilt
>    it from git HEAD probably 3 times since I started seeing these
>    errors, including this morning.
>    Sorry that I can't dig into this much more at this time. Hope this
>    helps,

It does, thanks!

>    ~Chad

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: make check produces 25 failures and hangs (bug #65176).

[Alan Mackenzie]

Hello again, Chad.

On Thu, Aug 10, 2023 at 20:14:40 +0000, Alan Mackenzie wrote:
> On Thu, Aug 10, 2023 at 15:54:26 -0400, chad wrote:
> >    On Thu, Aug 10, 2023 at 3:19â¯PM Alan Mackenzie <[1]acm@muc.de> wrote:

> >      [...]
> >      Is anybody else seeing these problems?  Has anybody else tried it,
> >      and
> >      _not_ seen these problems?

> >    I'm also seeing them. If I let "make check" run for a very long time (I
> >    was out of house for 90 minutes), it eventually terminates with 2
> >    checks that did not finish, in tree-sitter and native-comp. I haven't
> >    had the spare cycles to investigate further yet.

> Thanks!  On the one hand, I'm sorry you're having these problems, on the
> other hand, it's a relief that I'm not on my own!

I've reconfigured my Emacs without native compilation and run make
bootstrap.  I now do not see these 25 failures at all.

So the bugs seem to be triggered by native compilation.

[ .... ]

> >    ~Chad

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: make check produces 25 failures and hangs (bug #65176).

[Jens Schmidt]

On 2023-08-10  22:34, Alan Mackenzie wrote:

> So the bugs seem to be triggered by native compilation.

Confirmed.  And the hangs seem to be caused by tests asking for input.

I executed one of these failing tests interactively
(lisp/abbrev-tests.el), and it kept prompting in the minibuffer
where the usual

   (cl-letf (((symbol-function 'read-string) (lambda (&rest _) "<value>")))

should have inhibited that.  Full output of that interactive test to
follow:

Selector: t
Passed:  18
Failed:  3 (3 unexpected)
Skipped: 1
Total:   22/22

Started at:   2023-08-10 23:46:56+0200
Finished.
Finished at:  2023-08-10 23:47:36+0200

.............FFFs.....

F inverse-add-abbrev-skips-trailing-nonword
     Test that adding an inverse abbrev skips trailing nonword characters.
     (ert-test-failed
      ((should (string= (abbrev-expansion "foo" table) "bar")) :form
       (string= "abc" "bar") :value nil :explanation
       (array-elt 0 (different-atoms (97 "#x61" "?a") (98 "#x62" "?b")))))

F inverse-add-abbrev-skips-trailing-nonword/negative-arg
     Test that adding an inverse abbrev skips trailing nonword characters.
     (ert-test-failed
      ((should (string= (abbrev-expansion "text" table) "bar")) :form
       (string= "xxx" "bar") :value nil :explanation
       (array-elt 0 (different-atoms (120 "#x78" "?x") (98 "#x62" "?b")))))

F inverse-add-abbrev-skips-trailing-nonword/positive-arg
     Test that adding an inverse abbrev skips trailing nonword characters.
     (ert-test-failed
      ((should (string= (abbrev-expansion "text" table) "bar")) :form
       (string= "xxx" "bar") :value nil :explanation
       (array-elt 0 (different-atoms (120 "#x78" "?x") (98 "#x62" "?b")))))

Re: make check produces 25 failures and hangs (bug #65176).

[Eli Zaretskii]

> Date: Thu, 10 Aug 2023 23:54:50 +0200
> Cc: emacs-devel@gnu.org
> From: Jens Schmidt <jschmidt4gnu@vodafonemail.de>
> 
> On 2023-08-10  22:34, Alan Mackenzie wrote:
> 
> > So the bugs seem to be triggered by native compilation.
> 
> Confirmed.  And the hangs seem to be caused by tests asking for input.
> 
> I executed one of these failing tests interactively
> (lisp/abbrev-tests.el), and it kept prompting in the minibuffer
> where the usual
> 
>    (cl-letf (((symbol-function 'read-string) (lambda (&rest _) "<value>")))
> 
> should have inhibited that.  Full output of that interactive test to
> follow:
> 
> Selector: t
> Passed:  18
> Failed:  3 (3 unexpected)
> Skipped: 1
> Total:   22/22
> 
> Started at:   2023-08-10 23:46:56+0200
> Finished.
> Finished at:  2023-08-10 23:47:36+0200
> 
> .............FFFs.....
> 
> F inverse-add-abbrev-skips-trailing-nonword
>      Test that adding an inverse abbrev skips trailing nonword characters.
>      (ert-test-failed
>       ((should (string= (abbrev-expansion "foo" table) "bar")) :form
>        (string= "abc" "bar") :value nil :explanation
>        (array-elt 0 (different-atoms (97 "#x61" "?a") (98 "#x62" "?b")))))
> 
> F inverse-add-abbrev-skips-trailing-nonword/negative-arg
>      Test that adding an inverse abbrev skips trailing nonword characters.
>      (ert-test-failed
>       ((should (string= (abbrev-expansion "text" table) "bar")) :form
>        (string= "xxx" "bar") :value nil :explanation
>        (array-elt 0 (different-atoms (120 "#x78" "?x") (98 "#x62" "?b")))))
> 
> F inverse-add-abbrev-skips-trailing-nonword/positive-arg
>      Test that adding an inverse abbrev skips trailing nonword characters.
>      (ert-test-failed
>       ((should (string= (abbrev-expansion "text" table) "bar")) :form
>        (string= "xxx" "bar") :value nil :explanation
>        (array-elt 0 (different-atoms (120 "#x78" "?x") (98 "#x62" "?b")))))

Andrea, could you please look into these test failures?

Re: make check produces 25 failures and hangs (bug #65176).

[Jens Schmidt]

On 2023-08-11  07:37, Eli Zaretskii wrote:

> Andrea, could you please look into these test failures?

Repro-case-golfing, hoping that Andrea hasn't started on this yet:

(require 'cl-lib)
(defun foo ()
   (interactive)
   (cl-letf (((symbol-function 'read-string) (lambda (&rest _) "bar")))
     (message (read-string "Prompt: "))))

M-x foo RET
=> does not prompt

M-: (native-compile 'foo) RET
M-x foo RET
=> prompts

Re: make check produces 25 failures and hangs (bug #65176).

[Po Lu]

Jens Schmidt <jschmidt4gnu@vodafonemail.de> writes:

> On 2023-08-11  07:37, Eli Zaretskii wrote:
>
>> Andrea, could you please look into these test failures?
>
> Repro-case-golfing, hoping that Andrea hasn't started on this yet:
>
> (require 'cl-lib)
> (defun foo ()
>   (interactive)
>   (cl-letf (((symbol-function 'read-string) (lambda (&rest _) "bar")))
>     (message (read-string "Prompt: "))))

Isn't Fread_string a primitive, to which advice has no effect?

Re: make check produces 25 failures and hangs (bug #65176).

[Jens Schmidt]

On 2023-08-12  02:10, Po Lu wrote:

>> Repro-case-golfing, hoping that Andrea hasn't started on this yet:
>>
>> (require 'cl-lib)
>> (defun foo ()
>>    (interactive)
>>    (cl-letf (((symbol-function 'read-string) (lambda (&rest _) "bar")))
>>      (message (read-string "Prompt: "))))
> 
> Isn't Fread_string a primitive, to which advice has no effect?

Not sure about native compilation and primitives, but on Emacs 29 the
natively compiled function `foo' behaves identically to the non-compiled
one (no prompt).  And ERTs rely on that pattern, and have been working
on master until they stopped doing so one or two weeks ago.

Re: make check produces 25 failures and hangs (bug #65176).

[Andrea Corallo]

Eli Zaretskii <eliz@gnu.org> writes:

>> Date: Thu, 10 Aug 2023 23:54:50 +0200
>> Cc: emacs-devel@gnu.org
>> From: Jens Schmidt <jschmidt4gnu@vodafonemail.de>
>> 
>> On 2023-08-10  22:34, Alan Mackenzie wrote:
>> 
>> > So the bugs seem to be triggered by native compilation.
>> 
>> Confirmed.  And the hangs seem to be caused by tests asking for input.
>> 
>> I executed one of these failing tests interactively
>> (lisp/abbrev-tests.el), and it kept prompting in the minibuffer
>> where the usual
>> 
>>    (cl-letf (((symbol-function 'read-string) (lambda (&rest _) "<value>")))
>> 
>> should have inhibited that.  Full output of that interactive test to
>> follow:
>> 
>> Selector: t
>> Passed:  18
>> Failed:  3 (3 unexpected)
>> Skipped: 1
>> Total:   22/22
>> 
>> Started at:   2023-08-10 23:46:56+0200
>> Finished.
>> Finished at:  2023-08-10 23:47:36+0200
>> 
>> .............FFFs.....
>> 
>> F inverse-add-abbrev-skips-trailing-nonword
>>      Test that adding an inverse abbrev skips trailing nonword characters.
>>      (ert-test-failed
>>       ((should (string= (abbrev-expansion "foo" table) "bar")) :form
>>        (string= "abc" "bar") :value nil :explanation
>>        (array-elt 0 (different-atoms (97 "#x61" "?a") (98 "#x62" "?b")))))
>> 
>> F inverse-add-abbrev-skips-trailing-nonword/negative-arg
>>      Test that adding an inverse abbrev skips trailing nonword characters.
>>      (ert-test-failed
>>       ((should (string= (abbrev-expansion "text" table) "bar")) :form
>>        (string= "xxx" "bar") :value nil :explanation
>>        (array-elt 0 (different-atoms (120 "#x78" "?x") (98 "#x62" "?b")))))
>> 
>> F inverse-add-abbrev-skips-trailing-nonword/positive-arg
>>      Test that adding an inverse abbrev skips trailing nonword characters.
>>      (ert-test-failed
>>       ((should (string= (abbrev-expansion "text" table) "bar")) :form
>>        (string= "xxx" "bar") :value nil :explanation
>>        (array-elt 0 (different-atoms (120 "#x78" "?x") (98 "#x62" "?b")))))
>
> Andrea, could you please look into these test failures?

Yes, will do on Monday.

If in the meanwhile someone could git bisect this I guess it would be
make the investigation way quicker.

  Andrea

Re: make check produces 25 failures and hangs (bug #65176).

[Andrea Corallo]

Andrea Corallo <acorallo@gnu.org> writes:

> Eli Zaretskii <eliz@gnu.org> writes:
>
>>> Date: Thu, 10 Aug 2023 23:54:50 +0200
>>> Cc: emacs-devel@gnu.org
>>> From: Jens Schmidt <jschmidt4gnu@vodafonemail.de>
>>> 
>>> On 2023-08-10  22:34, Alan Mackenzie wrote:
>>> 
>>> > So the bugs seem to be triggered by native compilation.
>>> 
>>> Confirmed.  And the hangs seem to be caused by tests asking for input.
>>> 
>>> I executed one of these failing tests interactively
>>> (lisp/abbrev-tests.el), and it kept prompting in the minibuffer
>>> where the usual
>>> 
>>>    (cl-letf (((symbol-function 'read-string) (lambda (&rest _) "<value>")))
>>> 
>>> should have inhibited that.  Full output of that interactive test to
>>> follow:
>>> 
>>> Selector: t
>>> Passed:  18
>>> Failed:  3 (3 unexpected)
>>> Skipped: 1
>>> Total:   22/22
>>> 
>>> Started at:   2023-08-10 23:46:56+0200
>>> Finished.
>>> Finished at:  2023-08-10 23:47:36+0200
>>> 
>>> .............FFFs.....
>>> 
>>> F inverse-add-abbrev-skips-trailing-nonword
>>>      Test that adding an inverse abbrev skips trailing nonword characters.
>>>      (ert-test-failed
>>>       ((should (string= (abbrev-expansion "foo" table) "bar")) :form
>>>        (string= "abc" "bar") :value nil :explanation
>>>        (array-elt 0 (different-atoms (97 "#x61" "?a") (98 "#x62" "?b")))))
>>> 
>>> F inverse-add-abbrev-skips-trailing-nonword/negative-arg
>>>      Test that adding an inverse abbrev skips trailing nonword characters.
>>>      (ert-test-failed
>>>       ((should (string= (abbrev-expansion "text" table) "bar")) :form
>>>        (string= "xxx" "bar") :value nil :explanation
>>>        (array-elt 0 (different-atoms (120 "#x78" "?x") (98 "#x62" "?b")))))
>>> 
>>> F inverse-add-abbrev-skips-trailing-nonword/positive-arg
>>>      Test that adding an inverse abbrev skips trailing nonword characters.
>>>      (ert-test-failed
>>>       ((should (string= (abbrev-expansion "text" table) "bar")) :form
>>>        (string= "xxx" "bar") :value nil :explanation
>>>        (array-elt 0 (different-atoms (120 "#x78" "?x") (98 "#x62" "?b")))))
>>
>> Andrea, could you please look into these test failures?
>
> Yes, will do on Monday.
>
> If in the meanwhile someone could git bisect this I guess it would be
> make the investigation way quicker.
>
>   Andrea

To a very quick look seems to me `comp-subr-trampoline-install' is just
not called as it should.  AFAIR it should be called from Ffset, I'm
really wondering what could have happened.

  Andrea

Re: make check produces 25 failures and hangs (bug #65176).

[Jens Schmidt]

On 2023-08-12  21:09, Andrea Corallo wrote:

> To a very quick look seems to me `comp-subr-trampoline-install' is just
> not called as it should.  AFAIR it should be called from Ffset, I'm
> really wondering what could have happened.

Probably because native-comp-enable-subr-trampolines equals nil in an
"emacs-master -Q"?

When I compare the following snippet from emacs-master/lisp/loadup.el:

         (when (and (featurep 'native-compile)
                    (equal dump-mode "pdump"))
           ;; Don't enable this before bootstrap is completed, as the
           ;; compiler infrastructure may not be usable yet.
           (setq comp-enable-subr-trampolines t))
                 ^^^^

to emacs-29:

       (when (and (featurep 'native-compile)
                  (equal dump-mode "pdump"))
         ;; Don't enable this before bootstrap is completed, as the
         ;; compiler infrastructure may not be usable yet.
         (setq native-comp-enable-subr-trampolines t))
               ^^^^^^^^^^^

it seems that the android merge seems to be the culprit.  But leaving
that to others to judge.

In any case, fixing that setq in emacs-master to refer to
`native-comp-enable-subr-trampolines' again fixes the issue, at least
as far as my "little repro case" is concerned ... running "make check" 
... yep, no more hangs at least on abbrevs-test.el.

Re: make check produces 25 failures and hangs (bug #65176).

[Alan Mackenzie]

Hello, Jens.

On Sun, Aug 13, 2023 at 12:59:26 +0200, Jens Schmidt wrote:
> On 2023-08-12  21:09, Andrea Corallo wrote:

> > To a very quick look seems to me `comp-subr-trampoline-install' is just
> > not called as it should.  AFAIR it should be called from Ffset, I'm
> > really wondering what could have happened.

> Probably because native-comp-enable-subr-trampolines equals nil in an
> "emacs-master -Q"?

> When I compare the following snippet from emacs-master/lisp/loadup.el:

>          (when (and (featurep 'native-compile)
>                     (equal dump-mode "pdump"))
>            ;; Don't enable this before bootstrap is completed, as the
>            ;; compiler infrastructure may not be usable yet.
>            (setq comp-enable-subr-trampolines t))
>                  ^^^^

> to emacs-29:

>        (when (and (featurep 'native-compile)
>                   (equal dump-mode "pdump"))
>          ;; Don't enable this before bootstrap is completed, as the
>          ;; compiler infrastructure may not be usable yet.
>          (setq native-comp-enable-subr-trampolines t))
>                ^^^^^^^^^^^

Excellent debugging work!  Thanks!

> it seems that the android merge seems to be the culprit.  But leaving
> that to others to judge.

There was a variable called comp-enable-subr-trampolines which got
renamed to native-comp-..... on 2023-02-13 by Andrea in commit
5d0912f144.  In the feature/android branch, which was possibly branched
from main _before_ 2023-02-13, this flag was renamed back again to
comp-enable-.... on 2023-02-18 by Po in commit 4ab9fa7507a.  This likely
followed a merge from main onto the android branch.

It seems that native-comp-.... appeared to be a typo in the android
branch, so it got "corrected" to comp-enable-.... to make the software
work.  This "correction" was then forgotten about, an easy mistake to
make.

> In any case, fixing that setq in emacs-master to refer to
> `native-comp-enable-subr-trampolines' again fixes the issue, at least
> as far as my "little repro case" is concerned ... running "make check" 
> ... yep, no more hangs at least on abbrevs-test.el.

I would favour correcting that last use of comp-enable-subr-trampolines.

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: make check produces 25 failures and hangs (bug #65176).

[Eli Zaretskii]

> Date: Sun, 13 Aug 2023 12:59:26 +0200
> Cc: acm@muc.de, yandros@gmail.com, emacs-devel@gnu.org,
>  Po Lu <luangruo@yahoo.com>
> From: Jens Schmidt <jschmidt4gnu@vodafonemail.de>
> 
> On 2023-08-12  21:09, Andrea Corallo wrote:
> 
> > To a very quick look seems to me `comp-subr-trampoline-install' is just
> > not called as it should.  AFAIR it should be called from Ffset, I'm
> > really wondering what could have happened.
> 
> Probably because native-comp-enable-subr-trampolines equals nil in an
> "emacs-master -Q"?
> 
> When I compare the following snippet from emacs-master/lisp/loadup.el:
> 
>          (when (and (featurep 'native-compile)
>                     (equal dump-mode "pdump"))
>            ;; Don't enable this before bootstrap is completed, as the
>            ;; compiler infrastructure may not be usable yet.
>            (setq comp-enable-subr-trampolines t))
>                  ^^^^
> 
> to emacs-29:
> 
>        (when (and (featurep 'native-compile)
>                   (equal dump-mode "pdump"))
>          ;; Don't enable this before bootstrap is completed, as the
>          ;; compiler infrastructure may not be usable yet.
>          (setq native-comp-enable-subr-trampolines t))
>                ^^^^^^^^^^^


These two are supposed to be aliases (the comp-enable-subr-trampolines
one is the old name, the other one is the new name).

If the alias is missing, we should restore it, in addition to fixing
the name we use in startup.el.

Re: make check produces 25 failures and hangs (bug #65176).

[Andrea Corallo]

Eli Zaretskii <eliz@gnu.org> writes:

>> Date: Sun, 13 Aug 2023 12:59:26 +0200
>> Cc: acm@muc.de, yandros@gmail.com, emacs-devel@gnu.org,
>>  Po Lu <luangruo@yahoo.com>
>> From: Jens Schmidt <jschmidt4gnu@vodafonemail.de>
>> 
>> On 2023-08-12  21:09, Andrea Corallo wrote:
>> 
>> > To a very quick look seems to me `comp-subr-trampoline-install' is just
>> > not called as it should.  AFAIR it should be called from Ffset, I'm
>> > really wondering what could have happened.
>> 
>> Probably because native-comp-enable-subr-trampolines equals nil in an
>> "emacs-master -Q"?
>> 
>> When I compare the following snippet from emacs-master/lisp/loadup.el:
>> 
>>          (when (and (featurep 'native-compile)
>>                     (equal dump-mode "pdump"))
>>            ;; Don't enable this before bootstrap is completed, as the
>>            ;; compiler infrastructure may not be usable yet.
>>            (setq comp-enable-subr-trampolines t))
>>                  ^^^^
>> 
>> to emacs-29:
>> 
>>        (when (and (featurep 'native-compile)
>>                   (equal dump-mode "pdump"))
>>          ;; Don't enable this before bootstrap is completed, as the
>>          ;; compiler infrastructure may not be usable yet.
>>          (setq native-comp-enable-subr-trampolines t))
>>                ^^^^^^^^^^^
>
>
> These two are supposed to be aliases (the comp-enable-subr-trampolines
> one is the old name, the other one is the new name).
>
> If the alias is missing, we should restore it, in addition to fixing
> the name we use in startup.el.

Thanks Jens for the analysis and you for the fix.

We have in place the alias, but I guess when the code in loadup is
executed is to early and is maybe not effective?

Best Regards

  Andrea

Re: make check produces 25 failures and hangs (bug #65176).

[Eli Zaretskii]

> From: Andrea Corallo <acorallo@gnu.org>
> Cc: Jens Schmidt <jschmidt4gnu@vodafonemail.de>,  acm@muc.de,
>   yandros@gmail.com,  emacs-devel@gnu.org,  luangruo@yahoo.com
> Date: Mon, 14 Aug 2023 03:51:39 -0400
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> >>          (setq native-comp-enable-subr-trampolines t))
> >>                ^^^^^^^^^^^
> >
> >
> > These two are supposed to be aliases (the comp-enable-subr-trampolines
> > one is the old name, the other one is the new name).
> >
> > If the alias is missing, we should restore it, in addition to fixing
> > the name we use in startup.el.
> 
> Thanks Jens for the analysis and you for the fix.
> 
> We have in place the alias, but I guess when the code in loadup is
> executed is to early and is maybe not effective?

Where is the alias?  I don't think I see it.  What did I miss?

Re: make check produces 25 failures and hangs (bug #65176).

[Andrea Corallo]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Andrea Corallo <acorallo@gnu.org>
>> Cc: Jens Schmidt <jschmidt4gnu@vodafonemail.de>,  acm@muc.de,
>>   yandros@gmail.com,  emacs-devel@gnu.org,  luangruo@yahoo.com
>> Date: Mon, 14 Aug 2023 03:51:39 -0400
>> 
>> Eli Zaretskii <eliz@gnu.org> writes:
>> 
>> >>          (setq native-comp-enable-subr-trampolines t))
>> >>                ^^^^^^^^^^^
>> >
>> >
>> > These two are supposed to be aliases (the comp-enable-subr-trampolines
>> > one is the old name, the other one is the new name).
>> >
>> > If the alias is missing, we should restore it, in addition to fixing
>> > the name we use in startup.el.
>> 
>> Thanks Jens for the analysis and you for the fix.
>> 
>> We have in place the alias, but I guess when the code in loadup is
>> executed is to early and is maybe not effective?
>
> Where is the alias?  I don't think I see it.  What did I miss?

You are right apologies, got confised with the make-obsolete-variable we
have in place.  I added the alias with 922b6490286 in 29.

Should we do the same for `native-comp-deferred-compilation'?

Thanks

  Andrea

Re: make check produces 25 failures and hangs (bug #65176).

[Eli Zaretskii]

> From: Andrea Corallo <acorallo@gnu.org>
> Cc: jschmidt4gnu@vodafonemail.de,  acm@muc.de,  yandros@gmail.com,
>   emacs-devel@gnu.org,  luangruo@yahoo.com
> Date: Mon, 14 Aug 2023 09:38:28 -0400
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> > Where is the alias?  I don't think I see it.  What did I miss?
> 
> You are right apologies, got confised with the make-obsolete-variable we
> have in place.  I added the alias with 922b6490286 in 29.

Thanks.

> Should we do the same for `native-comp-deferred-compilation'?

Yes, please.

Re: make check produces 25 failures and hangs (bug #65176).

[Andrea Corallo]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Andrea Corallo <acorallo@gnu.org>
>> Cc: jschmidt4gnu@vodafonemail.de,  acm@muc.de,  yandros@gmail.com,
>>   emacs-devel@gnu.org,  luangruo@yahoo.com
>> Date: Mon, 14 Aug 2023 09:38:28 -0400
>> 
>> Eli Zaretskii <eliz@gnu.org> writes:
>> 
>> > Where is the alias?  I don't think I see it.  What did I miss?
>> 
>> You are right apologies, got confised with the make-obsolete-variable we
>> have in place.  I added the alias with 922b6490286 in 29.
>
> Thanks.
>
>> Should we do the same for `native-comp-deferred-compilation'?
>
> Yes, please.

Done thanks

  Andrea

Re: make check produces 25 failures and hangs (bug #65176).

[Alan Mackenzie]

Hello, Andrea.

On Mon, Aug 14, 2023 at 09:38:28 -0400, Andrea Corallo wrote:

[ .... ]

> I added the alias with 922b6490286 in 29.

Thanks, I've closed bug #65176.

> Thanks

>   Andrea

-- 
Alan Mackenzie (Nuremberg, Germany).