diff options
| author | Gary Kramlich <grim@reaperworld.com> | 2022-10-28 01:32:23 -0500 |
|---|---|---|
| committer | Gary Kramlich <grim@reaperworld.com> | 2022-10-28 01:32:23 -0500 |
| commit | fe0135a77ed801645310b6d208744466715a50d5 (patch) | |
| tree | cde295557db1266bb49ce4939ac5153c7bb51d6e /libpurple/core.c | |
| parent | ed2ea783b9fa0201aedb392398d3276b6b99dbd3 (diff) | |
| download | pidgin-fe0135a77ed801645310b6d208744466715a50d5.tar.gz | |
Fix an issue where gio was bombing out early and causing our unit tests to fail
This was accidentally introduced in /r/1921 when the unit tests moved to a
GMemorySettingsBackend. This was fixed upstream in 2.73.3. The upstream merge
request is here https://gitlab.gnome.org/GNOME/glib/-/merge_requests/2813/.
Testing Done:
Used convey to build under debian, ubuntu, and fedora.
Reviewed at https://reviews.imfreedom.org/r/1986/
Diffstat (limited to 'libpurple/core.c')
| -rw-r--r-- | libpurple/core.c | 9 |
1 files changed, 9 insertions, 0 deletions
diff --git a/libpurple/core.c b/libpurple/core.c index e37c397233..b13e719730 100644 --- a/libpurple/core.c +++ b/libpurple/core.c @@ -112,6 +112,15 @@ purple_core_init(PurpleUiInfo *ui_info) { ops = purple_core_get_ui_ops(); + /* This monster is to work around a bug that was fixed in glib 2.73.3. Once + * we require glib 2.74.0 this should be removed. + */ + if(TRUE) { + char *path = g_build_filename(purple_config_dir(), "dummy.ini", NULL); + g_object_unref(g_keyfile_settings_backend_new(path, "/dummy/", NULL)); + g_free(path); + } + /* The signals subsystem is important and should be first. */ purple_signals_init(); |