Re: Emacs 31 release-blocking bugs

> From: Sean Whitton <[email protected]>
> Date: Tue, 28 Jul 2026 17:02:43 +0100
> 
> These are the open bugs, at the time of writing, that I think we need to
> either fix or decide not to fix before we can prepare a release
> candidate.

Btw, we can manage this by using the "blocking" feature of Debbugs,
see the "RELEASE-BLOCKING BUGS" section in admin/release-process.

> - #80903 31.0.50; vc-next-action fails irregularly for git
> - #80996 31; Libraries that still need to be assigned to packages
> - #81018 macOS 15 crash build as of
>    876a1db6ee00f1d1b2af0329236acc8bdcceda5b
> - #81038 31.0.50; In org-mode, unfill-paragraph doesn't work at the very
>    end of a paragraph

Is this due to some change in Emacs 31?  If so, which change?

> - #81077 31.0.60; Sometimes only upper parts of text are unexpectedly
>    displayed

What are we supposed to do about this?  AFAIU, we don't really
understand why the problem happens, and why it happens to so few
users.

> - #81091 package-selected-packages not updated after init
> - #81105 32.0.50; installed ELPA package info missing on some distros
>    after commit 7c22a7d

IMO, this is not serious enough, as an easy workaround exists, and a
comprehensive solution is not easy.

> - #81112 31.0.60; Package.el error when rejecting during review
> - #81151 32.0.50; Make c-ts-mode-indent-style work with setopt-local

I don't think this is a regression in Emacs 31.

> - #81193 31.0.90; Wrong offsets in 'set-frame-size-and-position'
> - #81211 31.0.90; Android build fails with current Java compilers
> - #81213 31.0.90; Package list is colorized after upgrading packages

A simple fix for this one is to remove the code which highlights
packages in color.

> - #81254 work-buffer--release incorrectly resets current-buffer on first
>    release, clearing mark-active
> - #81373 31.0.90; package-upgrade and package-retention-policy
> - #81394 31.0.90; unexpected buffer completion behavior
> - #81407 31.0.90; Crash in display_line: stale glyph_row after
>    window-config change from menu-bar :enable eval during redisplay
> - #81408 31.0.90; gnus-icalendar-event--find-attendee no longer matches
>    regexp quoted strings
> - #81424 32.0.50; Build time dependency on TeX

This problem doesn't exist on the emacs-31 branch, only on master.

> - #81443 31.0.90; excessive backup files created when deleting from
>    vc-dir
> - #81460 31.0.90; tsx-ts-mode matches html tag's opening < with > in
>    arrow function =>

Is this really new in Emacs 31?

> - #81489 32.0.50; w32--terminal-is-conhost is nil in gui emacs.
> - #81490 32.0.50; wrong-type-argument internal to package-install

This is said to be a problem only on the master branch.

> - #81503 32.0.50; crash/assert in garbage_collect

I very much doubt we can do anything with this report, since the
problem happened in a very special and complicated scenario (client
frames, suspend/resume events, a problem with the minibuffer resolved
by some manual action).

> - #81509 31.0.91; Calling dired on nonexistent directory changes current
>    buffer
> - #81510 31.0.90; NS (macOS): clicking a disabled (:enable nil) tab-bar
>    item crases Emacs
> - The problem with icalendar tests failing in certain timezones:
>   <https://https://lists.gnu.org/archive/html/emacs-devel/2026-07/msg00445.html>
> 
> I used
> <https://debbugs.gnu.org/cgi/pkgreport.cgi?bug-rev=on;ordering=oldview;package=emacs>
> to prepare this.
> 
> I ignored anything that is reproducible in Emacs 30 and not serious.
> I focused on things that are about new features added or significantly
> modified in Emacs 31, and started from the date of the feature freeze,
> both at Eli's suggestion.

Thanks.