Closed Bug 151163 Opened 22 years ago Closed 22 years ago

Sometimes pages aren't loaded M1BR crash [@ GetOffsetFromView]

Categories

(SeaMonkey :: Tabbed Browser, defect)

x86
All
defect
Not set
critical

Tracking

(Not tracked)

VERIFIED FIXED
mozilla1.2alpha

People

(Reporter: bugzilla, Assigned: jag+mozilla)

References

Details

(4 keywords)

Crash Data

Attachments

(1 file)

This has been happening to me since the day before yesterday or so, but I
assumed it was a local problem, until Matti told me that he had the same issue.

I have tabbrowser set to open links in new tabs in the background, using
middle-click. In rare cases, pages opened like that don't get rendered. I can't
reproduce it at the moment, but I seem to remember that the *titles* of the
pages always did get displayed, so I assume that the networking part worked
fine, but the rendering was broken.
The tab has the correct title but the content is empty.

I clear the cache, clobbered my build, deleted the XUL fastload file ..
Status: UNCONFIRMED → NEW
Ever confirmed: true
*** Bug 151719 has been marked as a duplicate of this bug. ***
from the dupes and my own builds : Mozilla sometimes crash if you reload the
page to get the content. 
Severity: normal → critical
Keywords: crash
See bug 129098. Maybe that is what you are seeing?
i've been noticing this as well, using branch builds on linux rh7.2.  when i see
it, it's when i've clicked to the tab before it has finished loading the page.
when it "finishes" the title on the tab is there, but the content area is blank.

i've even encountered a crash (talkback below) once, while trying to reload the
blank tab. not sure if/how it's related...

Soren and Matti, have you been seeing this on trunk or branch builds?

Incident ID 7633468
Stack Signature GetOffsetFromView() 18fbc18f
Product ID Gecko1.0
Build ID 2002062107
Trigger Time 2002-06-23 13:39:43
Platform LinuxIntel
Operating System Linux 2.4.7-10
Module libgklayout.so
URL visited reloading a[n empty] tab
User Comments
Trigger Reason SIGSEGV: Segmentation Fault: (signal 11)
Source File Name
Trigger Line No.
Stack Trace
GetOffsetFromView()
Invalidate()
nsImageFrame::FrameChanged()
nsImageListener::FrameChanged()
imgRequestProxy::FrameChanged()
imgRequest::FrameChanged()
imgContainer::Notify()
nsTimerImpl::Fire()
handleTimerEvent()
PL_HandleEvent()
PL_ProcessPendingEvents()
nsEventQueueImpl::ProcessPendingEvents()
event_processor_callback()
our_gdk_io_invoke()
libglib-1.2.so.0 + 0xff9e (0x4039bf9e)
libglib-1.2.so.0 + 0x11773 (0x4039d773)
libglib-1.2.so.0 + 0x11d39 (0x4039dd39)
libglib-1.2.so.0 + 0x11eec (0x4039deec)
libgtk-1.2.so.0 + 0x94333 (0x402b8333)
nsAppShell::Run()
nsAppShellService::Run()
netscape-bin + 0x8fe9 (0x08050fe9)
netscape-bin + 0x97d7 (0x080517d7)
libc.so.6 + 0x1c507 (0x404e3507)
Keywords: regression
OS: Windows XP → All
nominating...not sure if this a bug specific to tab browsing, or another area
(layout? networking?), so reassign as needed.

if this gets minused, here's a relnote: if you see this bug, try reloading.
Keywords: nsbeta1, relnote
yes, that looks like the stack i get, but i have no symbols..
(libimg2 as the crash point)

I use trunk builds and i see it near every day ~5-10 x 

> ------- Additional Comment #6 From sairuh (se)  2002-06-24 13:58 -------
> when i see it, it's when i've clicked to the tab before it has finished loading
> the page.

Yes, that might be it.

> Soren and Matti, have you been seeing this on trunk or branch builds?

I'm only using trunk builds.
Nav triage team: nsbeta1-
Keywords: nsbeta1nsbeta1-
Target Milestone: --- → mozilla1.2alpha
I've been seeing this on branch builds, too...
I think that I also have came across this bug, but for me the page background
color is rendered fine, and if I view source I get all the HTML.

using 2002062008 in win98
This ESPN page is loading blank at times, or if you reload:
http://worldcup.espnsoccernet.com/report?match=48868&lang=us

If you notice a MAJOR layout problem, that has been logged at:
http://bugzilla.mozilla.org/show_bug.cgi?id=154441
from the talkback data, looks like it's only happening on linux for mozilla branch.
Keywords: topcrash
OS: All → Linux
Summary: Sometimes pages aren't loaded → Sometimes pages aren't loaded M1BR crash [ @ GetOffsetFromView]
The stack data show this stack possible only on linux but i see also very often
a crash which ends in libimg2 if I reload the empty tab.
It's possible that the windows stack trace is a little bit different..

This bug is reported on WinXP and I see it also on win2k.
The crash seems to be related but the bug report deals with a balnk Tab where
the title is shown and view source shows the source but the content ois not
rendered.
OS: Linux → All
I experience the same on W2K Build 2002062308

To replicate:

- Switch on "Load links in Background" in options. 
- Middle click a link (preferably to a slow page) to open it in a new tab.
- Click the new tab before the page has finished loading.

The page is now empty or only showing the background image

- Clicking refresh results in a crash.
I see this behaviour (not sure about the crash) on linux trunk build 2002061808.
Jan:  Can you try to reproduce your crash again and post your Talkback
incidents?  Are you using Talkback enabled builds?  It would be helpful to find
out what stack signature this crash is being reported under for Windows.  Thanks.
I have just replicated the bug on W2k build 2002062308 with "Load links in
background on". Talkback information has been sent with my email addres janderk
AT digitaldutch DOT com. Yes I run a talkback enabled version and have send in 3
of these crashes.

I can replicate this behavior any time by following these steps:

"Load links in background on" must be enabled in the Tabbed Browsing Options.


1) Go to http://forum.rackshack.net/
2) Middle mouse click the top left Rackshack logo to load the main page in
another tab
3) Click the new tab, before the new page has finished loading.

The page now shows only the background for me. It just stops.

4) Hitting the Refresh button raises the exception causing Mozilla to die.

I tried to upload the crash information as an attachment to this bug, but
Bugzilla would not accept it because the file was too large.
The signature of Jan's call stacks is nsImageBoxListener::OnStopDecode, which
is bug	153815.
Looking at the stack posted in comment #6, it looks like this might be another
variant of bug 153815.  If others agree, maybe it would be best to mark this a
dup of that one.  If not, we should look into this some more.

Has anyone been able reproduce this crash (opening site in new window, switching
to tab before load is complete and then refreshing the blank tab) with recent
MozillaTrunk or Gecko1.0 Branch builds?
QA Contact: sairuh → pmac
Resolving fixed.  There are only 2 incidents in the database and they're both
from very old builds.
Status: NEW → RESOLVED
Closed: 22 years ago
Resolution: --- → FIXED
verified 
this is long gone (but "worksforme" instead of "fixed" should be better)
Status: RESOLVED → VERIFIED
Product: Core → SeaMonkey
Summary: Sometimes pages aren't loaded M1BR crash [ @ GetOffsetFromView] → Sometimes pages aren't loaded M1BR crash [@ GetOffsetFromView]
Crash Signature: [@ GetOffsetFromView]
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: