Bug 734898 – crash after ssl warning

After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.

Bug 734898 - crash after ssl warning


Summary:	crash after ssl warning


Status:	RESOLVED DUPLICATE of bug 735706

Product:	epiphany
Classification:	Core
Component:	General
Version:	git master
Hardware:	Other Linux

Importance:	Normal critical
Target Milestone:	---
Assigned To:	Epiphany Maintainers
QA Contact:	Epiphany Maintainers

URL:
Whiteboard:

Depends on:
Blocks:

Reported:	2014-08-16 08:08 UTC by Frederic Peters
Modified:	2014-09-10 22:36 UTC

See Also:
GNOME target:	---
GNOME version:	---

Attachments
ephy-web-view: fix crash using Load Anyway button (1.01 KB, patch) 2014-08-25 15:11 UTC, Michael Catanzaro	none	Details \| Review

Description Frederic Peters 2014-08-16 08:08:40 UTC

A site caused the display of the "Look out!" page, I clicked "Load Anyway" and Epiphany crashed; this happened several times.

Core was generated by `epiphany'.
Program terminated with signal SIGSEGV, Segmentation fault.

+ Trace 233963

#0 g_type_check_instance_cast
at gtype.c line 4014
#0 g_type_check_instance_cast
at gtype.c line 4014
#1 allow_tls_certificate_cb
at ephy-web-view.c line 709
#2 ffi_call_unix64
from /usr/lib/x86_64-linux-gnu/libffi.so.6
#3 ffi_call
from /usr/lib/x86_64-linux-gnu/libffi.so.6
#4 g_cclosure_marshal_generic
at gclosure.c line 1448
#5 g_closure_invoke
at gclosure.c line 768
#6 signal_emit_unlocked_R
at gsignal.c line 3553
#7 g_signal_emit_valist
at gsignal.c line 3309
#8 g_signal_emit
at gsignal.c line 3365
#9 ephy_web_extension_proxy_allow_tls_certificate
at ephy-web-extension-proxy.c line 254
#10 web_extension_allow_tls_certificate
at ephy-embed-shell.c line 292
#11 emit_signal_instance_in_idle_cb
at gdbusconnection.c line 3753
#12 g_main_dispatch
at gmain.c line 3064
#13 g_main_context_dispatch
at gmain.c line 3663
#14 g_main_context_iterate
at gmain.c line 3734
#15 g_main_context_iteration
at gmain.c line 3795
#16 g_application_run
at gapplication.c line 2219
#17 main
at ephy-main.c line 488

Comment 1 Michael Catanzaro 2014-08-16 13:23:33 UTC

Carlos, do I need to disconnect the signal handler in dispose? If so, why do some handlers need to be disconnected there but not others?

I have no idea what destroyed the web view, though. :/

Comment 2 Carlos Garcia Campos 2014-08-25 13:17:21 UTC

(In reply to comment #1)
> Carlos, do I need to disconnect the signal handler in dispose?

We should probably use g_signal_connect_object

> If so, why do
> some handlers need to be disconnected there but not others?

I guess it's because the form auth callback doesn't really use the web view, it only shows the confirmation info bar.
 
> I have no idea what destroyed the web view, though. :/

Yes, it would be better to know what destroyed the web view first.

Comment 3 Carlos Garcia Campos 2014-08-25 13:19:45 UTC

(In reply to comment #2)
> > If so, why do
> > some handlers need to be disconnected there but not others?
> 
> I guess it's because the form auth callback doesn't really use the web view, it
> only shows the confirmation info bar.
> 

I does use the web view, so I have no idea.

Comment 4 Michael Catanzaro 2014-08-25 15:11:16 UTC

Created attachment 284414 [details] [review]
ephy-web-view: fix crash using Load Anyway button

Not sure if there's much value in attaching this patch, since it's surely wrong (the commit message is very optimistic), but Fredric you could try it out to see what happens. I guess it should fix the crash, but the Load Anyway button will do nothing. But probably something else will go badly wrong if you continue to use that browser tab after the web view has been destroyed....

Comment 5 Frederic Peters 2014-08-25 15:26:23 UTC

Well; I just can't get it to crash at the moment (without the patch); once I'm sure I can crash it I'll apply the patch.

Comment 6 Michael Catanzaro 2014-08-25 15:59:51 UTC

Actually I don't think that patch will fix anything.

Comment 7 Michael Catanzaro 2014-08-30 16:02:59 UTC

This probably has the same cause as bug #735706

Comment 8 Michael Catanzaro 2014-09-10 22:36:39 UTC


*** This bug has been marked as a duplicate of bug 735706 ***