Closed Bug 610119 Opened 15 years ago Closed 14 years ago

Crash [@ XPCWrappedNative::HasInterfaceNoQI ] [@ unmarkPurple]

Categories

(Core :: General, defect)

1.9.2 Branch
All
Linux
defect
Not set
critical

Tracking

()

RESOLVED WORKSFORME

People

(Reporter: mozilla, Unassigned)

References

()

Details

(Keywords: crash, reproducible)

Crash Data

Attachments

(5 files)

User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.2.12) Gecko/20101027 Fedora/3.6.12-1.fc13 Firefox/3.6.12 Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.2.12) Gecko/20101027 Fedora/3.6.12-1.fc13 Firefox/3.6.12 I was surfing the web and I experienced a crash when opening the links from http://gallir.wordpress.com at http://ciberderechos.barrapunto.com/ciberderechos/10/11/06/0918214.shtml in a background tab (backtrace attached at first crash). I use Adblock Plus and HTTPS Everywhere, among others. HTTPS Everywhere redirects wordpress.com from HTTP to HTTPS. Just in case it helps, crash happened after redirection, already rendering the page. The second crash happened when I tried to opening those links in the background in private mode (backtrace attached as second crash). I opened a bug report, using firefox in private mode, and both pages from https://gallir.wordpress.com were loaded fine, but when I clicked on the blog title to go to the blog main page, firefox crashed again when it was rendering the page (backtrace attached as third crash). And opening again a bug report and clicking on a link from barrapunto.com that pointed to http://ciberderechos.barrapunto.com/ciberderechos/10/11/06/0918214.shtml (if I'm not wrong) crashed firefox for the fourth time (backtrace attached as fourth crash). Well, would you be so kind to fix these crashes? (I guess the bug might be the same.) Many thanks for your help, Pablo Reproducible: Sometimes
Attached file First crash
Attached file Second crash
Attached file Third crash
Attached file Fourth crash
The "HTTPS Everywhere" extension has problems, see bug 548102. Does the problem go away if you disable it? (and restart Fx)
Thanks for your reply, Mats. I don't actually know if the problems goes away, since it doesn't happen all the time. If I have a new crash, I'll report it here. Thanks again, Pablo
I disabled HTTPS-Everywhere and visiting http://www.frkncngz.com/, I clicked on the Shuffle button on the top bar (that links to http://frkncngz.com/random) and firefox crashed. Backtrace attached. Would you be so kind to check this? Thanks, Pablo
please don't use thread apply all bt. rather, feel free to save such a log, but don't attach it unless asked. it wastes a lot of space and makes it much harder to find the part that actually matters which is here: #2 0x0241aa61 in nsProfileLock::FatalSignalHandler (signo=11, info= 0xbfa61e7c, context=0xbfa61efc) at nsProfileLock.cpp:213 unblock_sigs = {__val = {1024, 0 <repeats 31 times>}} oldact = <value optimized out> #3 <signal handler called> No symbol table info available. #4 0x02882491 in unmarkPurple (this=0x3758ec0, s=0x8b0e7b60) at ../../dist/include/nsISupportsImpl.h:214 refcount = Cannot access memory at address 0x4 It's totally unhelpful since you have only one useful frame (#4). Basically, at best it says "we were doing GC/CC work and ran across something which was probably a null pointer at an unexpected point", precisely how we got there is lost because there's no frame #5. It doesn't match your other random stacks, but we do have general bugs for GC/CC not being perfect...
I can reproduce the crash quite easily by clicking the "shuffle" link at the URL. It crashes after 5 clicks or so. This profile has a few extensions, Adblock Plus being one of them. bp-2f1a72b5-0088-489c-a3a3-831122101107 bp-61cf93da-c427-4f7c-874c-3352d2101107
Status: UNCONFIRMED → NEW
Ever confirmed: true
Keywords: crash, reproducible
Product: Firefox → Core
QA Contact: general → general
Hardware: x86 → All
Version: unspecified → 1.9.2 Branch
I can reproduce the crash with Firebug (1.5.4) only, but it's a lot harder to reproduce. Adding Adblock Plus (1.2.2) makes it easier to reproduce. I can't reproduce it with just Adblock Plus.
Possibly related to bug 606316?
Depends on: 606316
Summary: four crashes when opening pages → Crash [@ XPCWrappedNative::HasInterfaceNoQI ] [@ unmarkPurple]
(In reply to comment #8) > please don't use thread apply all bt. rather, feel free to save such a log, but > don't attach it unless asked. Thanks for your help and sorry for attaching all the backtrace. > it wastes a lot of space and makes it much harder to find the part that > actually matters which is here: I'm an average user, I don't code and I don't know what might be relevant. Should I only attach the first thread? (Consider that the log it's almost Greek to me.) Thanks again for your help, Pablo
the debugger will automatically select the crashed thread, it may or may not be the first thread. simply using 'bt' without talking about 'thread...' will do the right thing (tm) in almost all cases (excluding deadlock, aborts due to potential deadlock, races which you wouldn't be able to catch anyway).
Crash Signature: [@ XPCWrappedNative::HasInterfaceNoQI ] [@ unmarkPurple]
Both of these signatures are no longer seen in crash stats. Resolving this as WFM. Please reopen if crashes resurface.
Status: NEW → RESOLVED
Crash Signature: [@ XPCWrappedNative::HasInterfaceNoQI ] [@ unmarkPurple] → [@ XPCWrappedNative::HasInterfaceNoQI ] [@ unmarkPurple]
Closed: 14 years ago
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: