Closed
Bug 851832
Opened 12 years ago
Closed 12 years ago
Outlook.com freezes in Firefox Nightly 22
Categories
(Core :: Layout, defect)
Core
Layout
Tracking
()
RESOLVED
FIXED
Tracking | Status | |
---|---|---|
firefox21 | --- | unaffected |
firefox22 | + | fixed |
People
(Reporter: kaixiluo, Assigned: jaws)
References
Details
(Keywords: regression, Whiteboard: [fixed by bug 851230])
Login and initial loading of Microsoft new webmail service http://outlook.com is extremely slow and freezes temporarily the whole browser in Firefox 22.
Comment 1•12 years ago
|
||
This can't be a firefox:General issue because it's about the content of a website.
IS this something new and does it work with older builds ?
Can you find the regression range with our tool ?
http://mozilla.github.com/mozregression/
Component: General → Untriaged
Reporter | ||
Comment 2•12 years ago
|
||
The website works fine in Firefox 20 beta and Firefox 19. I haven't tried Aurora or the regression range tool yet.
![]() |
||
Comment 3•12 years ago
|
||
Confirmed with Mozilla/5.0 (Windows NT 6.1; WOW64; rv:22.0) Gecko/20130317 Firefox/22.0 ID:20130317030923 CSet: 0b052daa913c.
Noticeable on first Load and also if you change to e.g. Calendar and back to Inbox.
SPS: http://people.mozilla.com/~bgirard/cleopatra/#report=12577b4f16e3cac38aadcc530cd492a69fbc695c
WFM with Mozilla/5.0 (Windows NT 6.1; WOW64; rv:21.0) Gecko/20130317 Firefox/21.0 ID:20130317042012 CSet: 8f067a1714bf
Component: Untriaged → Layout
OS: Mac OS X → All
Product: Firefox → Core
Version: 22 Branch → Trunk
![]() |
||
Updated•12 years ago
|
tracking-firefox22:
--- → ?
Keywords: regression,
regressionwindow-wanted
![]() |
||
Comment 4•12 years ago
|
||
Regression window(m-c)
Good:
http://hg.mozilla.org/mozilla-central/rev/ee4879719f78
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:22.0) Gecko/20130307 Firefox/22.0 ID:20130307030926
Bad:
http://hg.mozilla.org/mozilla-central/rev/71395a927025
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:22.0) Gecko/20130307 Firefox/22.0 ID:20130307091952
Pushlog:
http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=ee4879719f78&tochange=71395a927025
Regression window(m-i)
Good:
http://hg.mozilla.org/integration/mozilla-inbound/rev/5361e3f3590d
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:22.0) Gecko/20130306 Firefox/22.0 ID:20130306204411
Bad:
http://hg.mozilla.org/integration/mozilla-inbound/rev/ae59f51150c9
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:22.0) Gecko/20130306 Firefox/22.0 ID:20130306204611
Pushlog:
http://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=5361e3f3590d&tochange=ae59f51150c9
Triggered by:
ae59f51150c9 Jared Wein — Bug 253564 - Plain text (text/plain, text/javascript, text/css, etc) documents should word-wrap. r=bz
![]() |
||
Comment 5•12 years ago
|
||
And I confirm that setting plain_text.wrap_long_lines = false helps
![]() |
||
Comment 6•12 years ago
|
||
Alice0775, thanks!
I wonder whether the problem is the existence of the stylesheet per se, or the fact that we're now trying to wrap some long text (possibly in a subframe)...
Blocks: 253564
Keywords: regressionwindow-wanted
Updated•12 years ago
|
Comment 7•12 years ago
|
||
If at all possible, this would be great to fix before the 22 m-c->m-a merge
![]() |
||
Comment 8•12 years ago
|
||
I cannot reproduce any more.
This seemd to be fixed by Bug 851230.
http://hg.mozilla.org/mozilla-central/rev/0e9badd3cf39
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:22.0) Gecko/20130322 Firefox/22.0 ID:20130322031028
![]() |
||
Comment 9•12 years ago
|
||
Sounds like the answer to my question in comment 4 is "the latter".
It's maybe worth filing a separate bug to see why we end up hanging here, though.
Depends on: 851230
Assignee | ||
Comment 10•12 years ago
|
||
I looked in to the Outlook hang using the Gecko Profiler.
layout::DoReflow for these three scripts was the source of the hang:
https://a.gfx.ms/fullex_9GP2W-k-NPTasebleVep1Q2.js (432kb single line)
https://a.gfx.ms/i0a.mozilla_EQIKVE5BVqIlIKfazUROng2.js (440kb single line)
https://a.gfx.ms/i1a_DXrTuDkQu5MYcKflepX4cA2.js (85kb single line)
The Outlook webpage only referenced these files as <script src="..."></script>, not as an embedded iframe. We shouldn't be running layout on script source since it's not displayed. I created a test webpage that only had these three files included via <script src=""> and was unable to reproduce the hang, so my guess is that Outlook is doing something fishy with XHR + adding it to content in some way.
Here is a link to my profile: http://people.mozilla.com/~bgirard/cleopatra/#report=e101be78051fb20897c9579d816e454ce4773ec1
Loading the largest of the scripts directly does cause Firefox to hang as it is working on wrapping the text. We could either bail out on the text wrap for lines over a certain length (I'm not sure how hard that would be), or see if there are some optimizations we can make in our reflow code (possibly equally as hard).
Status: NEW → RESOLVED
Closed: 12 years ago
Hardware: x86_64 → All
Resolution: --- → FIXED
Whiteboard: [fixed by bug 851230]
Assignee | ||
Updated•12 years ago
|
Assignee | ||
Comment 11•12 years ago
|
||
(In reply to Jared Wein [:jaws] (Vacation 3/30 to 4/7) from comment #10)
> I looked in to the Outlook hang using the Gecko Profiler.
>
> layout::DoReflow for these three scripts was the source of the hang:
>
> https://a.gfx.ms/fullex_9GP2W-k-NPTasebleVep1Q2.js (432kb single line)
> https://a.gfx.ms/i0a.mozilla_EQIKVE5BVqIlIKfazUROng2.js (440kb single line)
> https://a.gfx.ms/i1a_DXrTuDkQu5MYcKflepX4cA2.js (85kb single line)
>
> The Outlook webpage only referenced these files as <script
> src="..."></script>, not as an embedded iframe. We shouldn't be running
> layout on script source since it's not displayed. I created a test webpage
> that only had these three files included via <script src=""> and was unable
> to reproduce the hang, so my guess is that Outlook is doing something fishy
> with XHR + adding it to content in some way.
>
> Here is a link to my profile:
> http://people.mozilla.com/~bgirard/cleopatra/
> #report=e101be78051fb20897c9579d816e454ce4773ec1
>
> Loading the largest of the scripts directly does cause Firefox to hang as it
> is working on wrapping the text. We could either bail out on the text wrap
> for lines over a certain length (I'm not sure how hard that would be), or
> see if there are some optimizations we can make in our reflow code (possibly
> equally as hard).
I filed bug 854425 for this.
You need to log in
before you can comment on or make changes to this bug.
Description
•