Bug 152128 - [AX] WebProcess from WebKitGtk+ 2.10.4 hits SIGSEV in WebCore::AccessibilityRenderObject::document()
Summary: [AX] WebProcess from WebKitGtk+ 2.10.4 hits SIGSEV in WebCore::AccessibilityR...
Status: NEW
Alias: None
Product: WebKit
Classification: Unclassified
Component: Accessibility (show other bugs)
Version: WebKit Local Build
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Nobody
URL:
Keywords: InRadar
Depends on:
Blocks:
 
Reported: 2015-12-10 07:16 PST by Andres Gomez Garcia
Modified: 2015-12-15 07:03 PST (History)
6 users (show)

See Also:


Attachments
BT from gdb (42.34 KB, text/plain)
2015-12-10 07:16 PST, Andres Gomez Garcia
no flags Details

Note You need to log in before you can comment on or make changes to this bug.
Description Andres Gomez Garcia 2015-12-10 07:16:18 PST
Created attachment 267106 [details]
BT from gdb

I'm using WebKitGtk+ with my own JHBuild setting:
https://github.com/tanty/jhbuild-epiphany/tree/master

Epiphany 3.18.0 and WebKit 2.10.4

I'm running Epiphany with the dconf key:

"process-model" = "shared-secondary-process"

The compilation was done with CMake args:

'-DPORT=GTK -DCMAKE_BUILD_TYPE=Release -DDEVELOPER_MODE=ON -DCMAKE_C_FLAGS_DEBUG="-O0 -g1" -DCMAKE_CXX_FLAGS_DEBUG="-O0 -g1"'

When visiting several pages, eventually, WebKitWebProcess hits a SIGSEV.

This bug is not reproducible in a predictable way.
Comment 1 Radar WebKit Bug Importer 2015-12-10 07:16:48 PST
<rdar://problem/23839719>
Comment 2 Mario Sanchez Prada 2015-12-11 03:44:56 PST
It would be good to see more of the backtrace, looks like you don't have enough debug information available. Any chance you can get a more complete bt?
Comment 3 Andres Gomez Garcia 2015-12-14 03:12:01 PST
(In reply to comment #2)
> It would be good to see more of the backtrace, looks like you don't have
> enough debug information available. Any chance you can get a more complete
> bt?

I may eventually compile in fully Debug mode to try to get more information but don't expect it to happen soon. In addition:
* I've yet not been able to reproduce.
* Debug mode has proved to be difficult to use due to several ASSERT hits. It most probably will be even harder to reproduce this bug there.
Comment 4 Michael Catanzaro 2015-12-15 04:25:32 PST
You can do a release build with debug info. This is what all Linux distributions do. Building WebKit yourself, you can use -DCMAKE_BUILD_TYPE=RelWithDebInfo. There's not much point in reporting crashes without debug info.
Comment 5 Michael Catanzaro 2015-12-15 04:26:11 PST
Also, in the future, the backtrace should be produced with 'bt full' so we can see local variables, please.
Comment 6 Andres Gomez Garcia 2015-12-15 07:01:48 PST
(In reply to comment #5)
> Also, in the future, the backtrace should be produced with 'bt full' so we
> can see local variables, please.

I already answered similarly in another bug.

All my BT include the following commands, in order:

(gdb) backtrace full
(gdb) info registers
(gdb) x/16i $pc
(gdb) thread apply all backtrace
Comment 7 Andres Gomez Garcia 2015-12-15 07:03:41 PST
(In reply to comment #4)
> You can do a release build with debug info. This is what all Linux
> distributions do. Building WebKit yourself, you can use
> -DCMAKE_BUILD_TYPE=RelWithDebInfo. There's not much point in reporting
> crashes without debug info.

I suppose the point is showing that there is a crash in a specific point.

I may try to build with that but, please, have into account that I also need to be able to use my computer and not just have all my RAM used in the web browser.