Skip to content

ci: drop the glib2 and gtk4 hold on Windows #331

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
mishan merged 1 commit into from
Oct 3, 2026
Merged

ci: drop the glib2 and gtk4 hold on Windows #331

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
31 changes: 0 additions & 31 deletions .github/workflows/ci.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1201,37 +1201,6 @@ jobs:
adwaita-icon-theme:p
librsvg:p

# glib 2.90.0 (in ucrt64 since 2026-09-09) and gtk4 4.24.0 (2026-09-11)
# moved GEmblem, GDBusActionGroup and GdkCursor over to
# G_DECLARE_FINAL_TYPE. That macro expands to
#
# typedef struct { GObjectClass parent_class; } GEmblemClass;
#
# -- an *unnamed* struct -- while giomm and gdkmm still hand-write
# `using GEmblemClass = struct _GEmblemClass;'. Two system headers, two
# different types by the same name, and every translation unit that
# includes <gtkmm.h> dies on three `conflicting declaration' errors.
# Nothing we compile reaches it, and there is nothing to upgrade into:
# ucrt64's newest glibmm (2.86.0) and gtkmm-4.0 (4.20.0) both predate
# the change.
#
# So hold the last glib2 and gtk4 that agree with them. pacman -U after
# the install above, because pacboy would otherwise pull the new ones
# straight back in. Both downgrades are safe to mix with what stays
# current: gtkmm 4.20.0 and glibmm 2.86.0-4 were built long before
# either pinned version, and librsvg 2.63.0 -- the one package rebuilt
# the same hour glib 2.90 landed -- imports nothing that 2.88.3 does
# not export.
#
# Drop this step once ucrt64 ships a glibmm/gtkmm built against glib
# 2.90; the build will simply stop needing it.
- name: Hold glib2 and gtk4 where gtkmm still compiles
run: |
set -e
pacman -U --noconfirm \
https://repo.msys2.org/mingw/ucrt64/mingw-w64-ucrt-x86_64-glib2-2.88.3-1-any.pkg.tar.zst \
https://repo.msys2.org/mingw/ucrt64/mingw-w64-ucrt-x86_64-gtk4-4.22.4-1-any.pkg.tar.zst

# cairo 1.18.6 (in ucrt64 since 2026-09-21) aborts the process shortly
# after anything draws text. It moved the Direct2D factory that the
# DirectWrite path needs into per-thread data, and that data is set up
Expand Down
12 changes: 0 additions & 12 deletions .github/workflows/release.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -326,18 +326,6 @@ jobs:
adwaita-icon-theme:p
librsvg:p

# The same pin the ci workflow carries, for the same reason: glib 2.90
# and gtk4 4.24 declare GEmblemClass, GDBusActionGroupClass and
# GdkCursorClass one way, ucrt64's giomm/gdkmm another, and <gtkmm.h>
# stops compiling. Hold the last pair that agrees with them, and take
# this out when the packages catch up.
- name: Hold glib2 and gtk4 where gtkmm still compiles
run: |
set -e
pacman -U --noconfirm \
https://repo.msys2.org/mingw/ucrt64/mingw-w64-ucrt-x86_64-glib2-2.88.3-1-any.pkg.tar.zst \
https://repo.msys2.org/mingw/ucrt64/mingw-w64-ucrt-x86_64-gtk4-4.22.4-1-any.pkg.tar.zst

# RelWithDebInfo is already CMakeLists.txt's default. Spelled out
# anyway, so that a release does not quietly change if that changes.
- name: Build
Expand Down
Loading