The default bug view has changed. See this FAQ.

Latest TinderBox (Azure Build) crashes upon opening Psychedelic Browsing (IE Test Drive Experiment)

VERIFIED FIXED in mozilla7

Status

()

Core
Graphics
P1
critical
VERIFIED FIXED
6 years ago
2 years ago

People

(Reporter: Optimizer, Assigned: bas)

Tracking

Trunk
mozilla7
Points:
---
Bug Flags:
in-testsuite +

Firefox Tracking Flags

(Not tracked)

Details

(crash signature, URL)

Attachments

(2 attachments, 1 obsolete attachment)

(Reporter)

Description

6 years ago
User-Agent:       Mozilla/5.0 (Windows NT 6.1; WOW64; rv:7.0a1) Gecko/20110625 Firefox/7.0a1
Build Identifier: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:7.0a1) Gecko/20110625 Firefox/7.0a1

The latest hourly azure builds are having problem with specifically this IE test : Psychedelic Browsing ( http://ie.microsoft.com/testdrive/performance/psychedelicbrowsing/Default.html )

Reproducible: Always

Steps to Reproduce:
1.Have the latest Hourly (at point of bug posting) having the Azure Rendering engine ( http://ftp.mozilla.org/pub/mozilla.org/firefox/tinderbox-builds/mozilla-central-win32/1309018896/ )
2.Open link ie.microsoft.com/testdrive/performance/psychedelicbrowsing/Default.html
3.Start the Psychedelic Browsing test by clicking Find Out


Actual Results:  
Firefox Crashes

Expected Results:  
The test should run properly as it runs on latest Nightly (7.0a1 (2011-06-25))
(Reporter)

Updated

6 years ago
Priority: -- → P1

Comment 1

6 years ago
Crash on mozilla-inbound nightly:
Crash Report [@ _invalid_parameter_noinfo ]

Crash Id
bp-a88dc130-2960-4259-9658-5a09e2110625
bp-be57e6f6-dccc-4c20-9b0c-eb6f42110625
Component: General → Graphics
Product: Firefox → Core
QA Contact: general → thebes
Version: unspecified → Trunk
If Alice sees it, this is confirmed. :)
Assignee: nobody → bas.schouten
Status: UNCONFIRMED → NEW
Ever confirmed: true

Comment 3

6 years ago
Could you add this a dependency of bug 651858?
Blocks: 651858

Comment 4

6 years ago
Confirmed on latest m-c hourly build

https://crash-stats.mozilla.com/report/index/bp-d8ccec74-293d-45c6-9ed0-24f5b2110625
(Assignee)

Comment 5

6 years ago
Created attachment 541992 [details] [diff] [review]
Silently ignore drawing glyph runs where no glyphs are to be drawn

This demo draws fill text call where the textrun reports no glyphs to be drawn. In this situation we should ignore the draw call and go on drawing rather than invalidly use the vector and cause a crash. The exact cause as to why the textrun reports no glyphs to be drawn should be examined further.
Attachment #541992 - Flags: review?(joe)
(Assignee)

Comment 6

6 years ago
Created attachment 541993 [details] [diff] [review]
Silently ignore drawing glyph runs where no glyphs are to be drawn v2

I was unable to create this situation but we probably should just continue to draw any runs in the textrun that -do- draw valid glyphs. I've adjusted the patch for this.
Attachment #541993 - Flags: review?(joe)
(Assignee)

Updated

6 years ago
Attachment #541992 - Attachment is obsolete: true
Attachment #541992 - Flags: review?(joe)
(Assignee)

Updated

6 years ago
Status: NEW → ASSIGNED
(Assignee)

Comment 7

6 years ago
Just to be clear, the same situation exists in traditional glyph drawing but it's silently ignored there as well, at gfxFont.cpp:1185. glyphCount there is found to be 0 similarly as for Azure canvas.
(Reporter)

Comment 8

6 years ago
Please push this to inbound as early as possible.
There might be several other places where a similar crash can occur once today's Nightly is out
(Assignee)

Comment 9

6 years ago
This is pending review, once reviewed it will be pushed. I'm not sure what causes the condition, once we've figured out we should probably add a test for it. Until then I doubt this is a common situation but I'll keep an eye out.

Updated

6 years ago
Crash Signature: _invalid_parameter_noinfo
Comment on attachment 541993 [details] [diff] [review]
Silently ignore drawing glyph runs where no glyphs are to be drawn v2

Make sure you add a test case before pushing this.
Attachment #541993 - Flags: review?(joe) → review+
Keywords: testcase-wanted
(Assignee)

Comment 11

6 years ago
Created attachment 542033 [details] [diff] [review]
Add a testcase for the bug

This adds a crashtest for this bug, it seems to occur when trying to draw 0-sized fonts.
Attachment #542033 - Flags: review?(jmuizelaar)
Comment on attachment 542033 [details] [diff] [review]
Add a testcase for the bug

Drop the xmlns stuff and give the test a better name like 0px-size-font-667225.html
Attachment #542033 - Flags: review?(jmuizelaar) → review+
(Assignee)

Comment 13

6 years ago
Landed on mozilla-inbound:

http://hg.mozilla.org/integration/mozilla-inbound/rev/9a79edc67610
http://hg.mozilla.org/integration/mozilla-inbound/rev/59374caa391e
Merged:
http://hg.mozilla.org/mozilla-central/rev/9a79edc67610
http://hg.mozilla.org/mozilla-central/rev/59374caa391e
Status: ASSIGNED → RESOLVED
Last Resolved: 6 years ago
Flags: in-testsuite+
OS: Windows 7 → All
Hardware: x86_64 → All
Resolution: --- → FIXED
Target Milestone: --- → mozilla7
(Reporter)

Comment 15

6 years ago
Did It miss today's Nightly ? I am not seeing it under the Nightly Changeset
http://hg.mozilla.org/mozilla-central/rev/5e072732cb45
(Assignee)

Comment 16

6 years ago
It did miss today's nightly.
(Assignee)

Updated

6 years ago
Duplicate of this bug: 667673
Duplicate of this bug: 667443

Comment 19

6 years ago
The issue is not reproducible anymore. Verified fixed on:

Mozilla/5.0 (Windows NT 6.1; WOW64; rv:7.0a2) Gecko/20110706 Firefox/7.0a2
Mozilla/5.0 (X11; Linux i686; rv:7.0a2) Gecko/20110706 Firefox/7.0a2
Mozilla/5.0 (Windows NT 6.1; rv:7.0a2) Gecko/20110706 Firefox/7.0a2
Mozilla/5.0 (Macintosh; Intel Mac OS X 10.6; rv:7.0a2) Gecko/20110706 Firefox/7.0a2
Status: RESOLVED → VERIFIED
Keywords: testcase-wanted
You need to log in before you can comment on or make changes to this bug.