[nautilus/wip/oholy/ci-failures: 6/6] progress-info: Prevent abortion caused by missing locks
- From: António Fernandes <antoniof src gnome org>
- To: commits-list gnome org
- Cc:
- Subject: [nautilus/wip/oholy/ci-failures: 6/6] progress-info: Prevent abortion caused by missing locks
- Date: Wed, 5 Feb 2020 12:21:52 +0000 (UTC)
commit f3aadd68fdc1a3eec8d90a092674632443a0c85e
Author: Ondrej Holy <oholy redhat com>
Date: Mon Feb 3 10:15:29 2020 +0100
progress-info: Prevent abortion caused by missing locks
The test-file-operations-move-files and test-file-operations-copy-files
test cases randomly crash with "GLib-FATAL-CRITICAL: g_source_is_destroyed:
assertion 'g_atomic_int_get (&source->ref_count) > 0' failed", which causes
GitLab CI pipeline failures among others, which is a shame.
I am not able to reproduce those crashes locally, so I can't easily verify
the call stack, but only place where `g_source_destroy()` is called is
when handling progress infos. I see two potential issues which might cause
this. Let's try to fix them.
src/nautilus-progress-info.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
---
diff --git a/src/nautilus-progress-info.c b/src/nautilus-progress-info.c
index 45f2842a9..e44d34dcc 100644
--- a/src/nautilus-progress-info.c
+++ b/src/nautilus-progress-info.c
@@ -191,10 +191,10 @@ idle_callback (gpointer data)
gboolean cancelled_at_idle;
GSource *source;
- source = g_main_current_source ();
-
G_LOCK (progress_info);
+ source = g_main_current_source ();
+
/* Protect agains races where the source has
* been destroyed on another thread while it
* was being dispatched.
@@ -309,8 +309,8 @@ set_details_in_thread (GTask *task,
{
if (!g_cancellable_is_cancelled (cancellable))
{
- set_details (info, _("Canceled"));
G_LOCK (progress_info);
+ set_details (info, _("Canceled"));
info->cancel_at_idle = TRUE;
g_timer_stop (info->progress_timer);
queue_idle (info, TRUE);
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]