bug#56744: 29.0.50; "Autoloaded" information lost in native compilation?

[Andrea Corallo]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Andrea Corallo <acorallo@gnu.org>
>> Cc: 56744@debbugs.gnu.org
>> Date: Sun, 20 Aug 2023 16:45:51 -0400
>> 
>> Eli Zaretskii <eliz@gnu.org> writes:
>> 
>> > Andrea,
>> >
>> > Could you please take a look at this year-old bug report?
>> >
>> >> Date: Sun, 24 Jul 2022 19:10:03 +0300
>> >> From: Eli Zaretskii <eliz@gnu.org>
>> >> 
>> >> 
>> >> To reproduce:
>> >> 
>> >>   emacs -Q
>> >>   C-h f rmail-movemail-variant-p RET
>> >> 
>> >> (or any other autoloaded function).  Assuming that the function's .el
>> >> file was already natively compiled to produce a .eln file, the
>> >> "autoloaded" part will not be shown in the *Help* buffer, Emacs will
>> >> only say that this is a native-compiled function.
>> 
>> On current master before requiring rmail on C-h f
>> rmail-movemail-variant-p RET I get:
>> 
>> ===
>> rmail-movemail-variant-p is an autoloaded byte-compiled Lisp function
>> in ‘rmail.el’.
>> 
>> (rmail-movemail-variant-p &rest VARIANTS)
>> ===
>> 
>> After requiring rmail (and having it native compiled) I get:
>> 
>> ===
>> rmail-movemail-variant-p is an autoloaded native-compiled Lisp
>> function in ‘rmail.el’.
>> 
>> (rmail-movemail-variant-p &rest VARIANTS)
>> 
>> Inferred type: (function (&rest t) boolean)
>> ===
>> 
>> So I guess the issue has already been fixed (or I'm not reproducing it
>> correctly)?
>
> No, you are reproducing correctly.  It's also seems to be fixed on the
> emacs-29 branch, so I will close the bug.  Sorry for not
> double-checking before I pinged you.

No problem, a bug that's already solved is the best outcome we could
desire :)

Best Regards

  Andrea