bug#54248: 29.0.50; Uninterned symbols malcompiling

bug#54248: 29.0.50; Uninterned symbols malcompiling

[Michael Heerdegen]

Hello,

this must be a few days old at maximum.  Random example:

Byte compile a file containing

  (set '#:my-toggle-invisible 3)

Load the compiled file ==> (wrong-type-argument symbolp (nil))

Interpreted code works.

CC'ing Alan because this might be related to his latest changes to the
compiler, I didn't investigate, sorry if not.


TIA,

Michael.


In GNU Emacs 29.0.50 (build 41, x86_64-pc-linux-gnu, GTK+ Version 3.24.24, cairo version 1.16.0)
 of 2022-03-04 built on drachen
Repository revision: e46bb68ac27625e4bc466da842f74baea63e835a
Repository branch: master
Windowing system distributor 'The X.Org Foundation', version 11.0.12011000
System Description: Debian GNU/Linux 11 (bullseye)

Configured using:
 'configure --with-native-compilation --with-xwidgets'

bug#54248: 29.0.50; Uninterned symbols malcompiling

[Alan Mackenzie]

Hello, Michael.

On Sat, Mar 05, 2022 at 02:24:41 +0100, Michael Heerdegen wrote:

> Hello,

> this must be a few days old at maximum.  Random example:

> Byte compile a file containing

>   (set '#:my-toggle-invisible 3)

> Load the compiled file ==> (wrong-type-argument symbolp (nil))

> Interpreted code works.

> CC'ing Alan because this might be related to his latest changes to the
> compiler, I didn't investigate, sorry if not.

Thanks!  A bit of bisecting shows that the bug was introduced by the
following commit:

commit 6092ee1c3ff503fbe8087e13b7eae2f904c4af3b
Author: Alan Mackenzie <acm@muc.de>
Date:   Thu Feb 24 17:30:39 2022 +0000

    Amend byte-run-strip-symbol-positions so that an unexec build builds

..  I'll have a look at this and see if I can work out how this commit
causes the bug.

> TIA,

> Michael.


> In GNU Emacs 29.0.50 (build 41, x86_64-pc-linux-gnu, GTK+ Version 3.24.24, cairo version 1.16.0)
>  of 2022-03-04 built on drachen
> Repository revision: e46bb68ac27625e4bc466da842f74baea63e835a
> Repository branch: master
> Windowing system distributor 'The X.Org Foundation', version 11.0.12011000
> System Description: Debian GNU/Linux 11 (bullseye)

> Configured using:
>  'configure --with-native-compilation --with-xwidgets'

-- 
Alan Mackenzie (Nuremberg, Germany).

bug#54248: 29.0.50; Uninterned symbols malcompiling

[Alan Mackenzie]

Hello, Michael.

On Sat, Mar 05, 2022 at 02:24:41 +0100, Michael Heerdegen wrote:

> Hello,

> this must be a few days old at maximum.  Random example:

> Byte compile a file containing

>   (set '#:my-toggle-invisible 3)

> Load the compiled file ==> (wrong-type-argument symbolp (nil))

> Interpreted code works.

> CC'ing Alan because this might be related to his latest changes to the
> compiler, I didn't investigate, sorry if not.

Yes, it was a bug in the byte compiler I introduced recently.  I've just
committed a fix.  Could you please update your repository, and let me
know whether this fix fixes the bug fully.  Thanks!

Could you also let me know whether this fix has any positive impact on
the other bug, #54079.  Thanks again!

> TIA,

> Michael.


> In GNU Emacs 29.0.50 (build 41, x86_64-pc-linux-gnu, GTK+ Version 3.24.24, cairo version 1.16.0)
>  of 2022-03-04 built on drachen
> Repository revision: e46bb68ac27625e4bc466da842f74baea63e835a
> Repository branch: master
> Windowing system distributor 'The X.Org Foundation', version 11.0.12011000
> System Description: Debian GNU/Linux 11 (bullseye)

> Configured using:
>  'configure --with-native-compilation --with-xwidgets'

-- 
Alan Mackenzie (Nuremberg, Germany).

bug#54248: 29.0.50; Uninterned symbols malcompiling

[Michael Heerdegen]

Alan Mackenzie <acm@muc.de> writes:

> Yes, it was a bug in the byte compiler I introduced recently.  I've
> just committed a fix.  Could you please update your repository, and
> let me know whether this fix fixes the bug fully.  Thanks!

Looks good, I don't see this issue any more.

> Could you also let me know whether this fix has any positive impact on
> the other bug, #54079.  Thanks again!

I don't see any impact unfortunately, that issue looks unchanged.

Anyway, thanks for fixing this one, I'm just closing it...

Michael.