New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 646087 link

Starred by 3 users

Issue metadata

Status: Fixed
Owner:
Last visit > 30 days ago
Closed: Sep 2016
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Android , Mac
Pri: 0
Type: Bug-Regression

Blocking:
issue 646152
issue 637051



Sign in to add a comment

[FATAL:url_request.cc(121)] Check failed: !connect_timing->connect_end.is_null

Project Member Reported by dalecur...@chromium.org, Sep 12 2016

Issue description

Sadly don't have any backtrace information since the Galaxy Nexus requires static builds with symbol_level=0|1.

Seen on vimeo.com just loading the page.
 
(ToT build from this morning (wasn't happening Friday), Jellybean, Galaxy Nexus).
Commenting out that line causes the subsequent DCHECK to also fail:

DCHECK(!connect_timing->ssl_end.is_null());
Owner: xunji...@chromium.org
Thanks for the report, dalecurtis. Helen, could you PTAL?

Your change (https://codereview.chromium.org/2324183002) only affects quic afaict but it's worth double checking (this might be for a Google property).


dalecurtis, was this before or after 417969?
Status: Assigned (was: Unconfirmed)
 Issue 646111  has been merged into this issue.
Built at #417986

Comment 7 by hubbe@chromium.org, Sep 12 2016

For me, this happend on a standard linux build.
The request was:

https://accounts.google.com/ListAccounts?gpsia=1&source=ChromiumBrowser&json=standard

I have it running in a debugger right now, anything else I should find out?

Comment 8 by mef@chromium.org, Sep 12 2016

Helen is OOO, should should we revert the https://codereview.chromium.org/2324183002 for now?

Comment 9 by mef@chromium.org, Sep 12 2016

I've reverted https://codereview.chromium.org/2333863003, could you check that it has fixed the problem?
Cc: kbr@chromium.org rch@chromium.org rsleevi@chromium.org
 Issue 646131  has been merged into this issue.

Comment 11 by kbr@chromium.org, Sep 12 2016

Blocking: 637051
Labels: -Pri-3 OS-Mac Pri-0 Type-Bug-Regression
This was not a P3; it was a P0. Chromium builds on multiple platforms were broken. The reporter should have been confident in labeling it as such. Adjusting the priority and blocking the feature bug on this one.

Thanks kbr@, it wasn't clear to me the scope of this issue until  issue 646131  but I agree this was a major problem.

Can you confirm that the revert has fixed things for you?
The reporter in this case (me) only saw it on one page on a crappy old device so I just left the defaults :)

Comment 14 by zmo@chromium.org, Sep 12 2016

Yeah I just sync and rebuild, now the crash is gone

Comment 15 by zmo@chromium.org, Sep 12 2016

It seems to me that the problem is really testing. Something like this landed with no tests catching it, this is just concerning.
Status: Fixed (was: Assigned)
Thanks a lot for verifying the fix. Agreed we should definitely add some QUIC tests to cover this case. I filed  issue 646087  for this.

Comment 17 by kbr@chromium.org, Sep 12 2016

Blocking: 646152
Cc: yolandyan@chromium.org

Sign in to add a comment