webbrowser-app test failure on armhf with Qt 5.3

Bug #1321679 reported by Timo Jyrinki
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
webbrowser-app
Fix Released
High
Olivier Tilloy
webbrowser-app (Ubuntu)
Fix Released
High
Unassigned

Bug Description

With the FTBFS bug fixed, it seems there's still a test failure on armhf only when using Qt 5.3:

https://launchpadlibrarian.net/176026671/buildlog_ubuntu-utopic-armhf.webbrowser-app_1%3A0.23%2B14.10.20140515.1-0~537%2B201405201947~ubuntu14.10.1%2Bfixqt53~test1_FAILEDTOBUILD.txt.gz

---
 4/12 Test #1: tst_QmlTests ...................................***Failed 1.12 sec
libEGL warning: GLX/DRI2 is not supported
file:///build/buildd/webbrowser-app-0.23+14.10.20140515.1/obj-arm-linux-gnueabihf/tests/unittests/qml/undertest/AddressBar.qml:55:13: QML QQuickImage: Cannot open: file:///build/buildd/webbrowser-app-0.23+14.10.20140515.1/obj-arm-linux-gnueabihf/tests/unittests/qml/undertest/assets/reload.png
process 18699: D-Bus library appears to be incorrectly set up; failed to read machine uuid: Failed to open "/etc/machine-id": No such file or directory
See the manual page for dbus-uuidgen to correct this issue.
********* Start testing of QmlTests *********
Config: Using QtTest library 5.3.0, Qt 5.3.0
PASS : QmlTests::AddressBar::initTestCase()
Segmentation fault
---

--
Refer to kept-update info on the front page of https://launchpad.net/~canonical-qt5-edgers/+archive/qt5-beta2 regarding where/how to get Qt 5.3. Note that it's not necessarily that PPA itself.

Tags: qt5.3
description: updated
Revision history for this message
Olivier Tilloy (osomon) wrote :

The failure is actually a segfault, here is the backtrace I managed to get (running the tests on a chromebook):

********* Start testing of QmlTests *********
Config: Using QtTest library 5.3.0, Qt 5.3.0
PASS : QmlTests::AddressBar::initTestCase()

Program received signal SIGSEGV, Segmentation fault.
QV4::StringPrototype::method_split (ctx=<optimized out>) at jsruntime/qv4scopedvalue_p.h:313
313 jsruntime/qv4scopedvalue_p.h: No such file or directory.
(gdb) bt
#0 QV4::StringPrototype::method_split (ctx=<optimized out>) at jsruntime/qv4scopedvalue_p.h:313
#1 0xb613024e in QV4::BuiltinFunction::call (that=<optimized out>, callData=<optimized out>) at jsruntime/qv4functionobject.cpp:583
#2 0xb6164c68 in call (d=0xaff020e8, this=<optimized out>) at jsruntime/qv4object_p.h:270
#3 QV4::Runtime::callProperty (context=0xbeffe678, name=..., callData=...) at jsruntime/qv4runtime.cpp:938
#4 0xafacb0ba in ?? ()
Backtrace stopped: previous frame identical to this frame (corrupt stack?)

Changed in webbrowser-app:
status: New → Confirmed
Revision history for this message
Olivier Tilloy (osomon) wrote :

I’m seeing a different backtrace when commenting out a few tests (thus getting some tests to pass):

Program received signal SIGSEGV, Segmentation fault.
0xb6f0baa8 in __memcpy_neon () at ../ports/sysdeps/arm/armv7/multiarch/memcpy_impl.S:332
332 ../ports/sysdeps/arm/armv7/multiarch/memcpy_impl.S: No such file or directory.
(gdb) bt
#0 0xb6f0baa8 in __memcpy_neon () at ../ports/sysdeps/arm/armv7/multiarch/memcpy_impl.S:332
#1 0xb6564892 in memcpy (__len=<optimized out>, __src=<optimized out>, __dest=<optimized out>) at /usr/include/arm-linux-gnueabihf/bits/string3.h:51
#2 QString::append (this=this@entry=0xbeffe538, str=...) at tools/qstring.cpp:8695
#3 0xb6144122 in operator+= (s=..., this=0xbeffe538) at /usr/include/arm-linux-gnueabihf/qt5/QtCore/qstring.h:411
#4 QV4::StringPrototype::method_replace (ctx=<optimized out>) at jsruntime/qv4stringobject.cpp:593
#5 0xb613024e in QV4::BuiltinFunction::call (that=<optimized out>, callData=<optimized out>) at jsruntime/qv4functionobject.cpp:583
#6 0xb6164c68 in call (d=0xaff01eb0, this=<optimized out>) at jsruntime/qv4object_p.h:270
#7 QV4::Runtime::callProperty (context=0xbeffe7c0, name=..., callData=...) at jsruntime/qv4runtime.cpp:938
#8 0xaba068dc in ?? ()
Backtrace stopped: previous frame identical to this frame (corrupt stack?)

Revision history for this message
Olivier Tilloy (osomon) wrote :

When running the tests with QV4_FORCE_INTERPRETER=1, the crash goes away and the tests pass.

Olivier Tilloy (osomon)
Changed in webbrowser-app:
assignee: nobody → Olivier Tilloy (osomon)
status: Confirmed → In Progress
Revision history for this message
Timo Jyrinki (timo-jyrinki) wrote :

As discussed, this may be similar to the upstream split string problem that was workarounded successfully in UI Toolkit:
http://bazaar.launchpad.net/~ubuntu-sdk-team/ubuntu-ui-toolkit/staging/revision/1080

Another very big fix in many apps resulted from this UI Toolkit fix adding default keyword too:
http://bazaar.launchpad.net/~ubuntu-sdk-team/ubuntu-ui-toolkit/staging/revision/1085

The armhf crasher related to the split function is tracked at bug #1323997

Bill Filler (bfiller)
Changed in webbrowser-app:
importance: Undecided → High
Changed in webbrowser-app (Ubuntu):
importance: Undecided → High
status: New → In Progress
Revision history for this message
Timo Jyrinki (timo-jyrinki) wrote :

I'll mark this as Fix Committed to indicate that it has landed to landing PPA.

Changed in webbrowser-app:
status: In Progress → Fix Committed
Changed in webbrowser-app (Ubuntu):
status: In Progress → Fix Committed
Olivier Tilloy (osomon)
Changed in webbrowser-app:
status: Fix Committed → Fix Released
Changed in webbrowser-app (Ubuntu):
status: Fix Committed → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.