Created attachment 695654 [details] [diff] [review] testcase patch Steps to reproduce: (1) Apply the testcase patch. (2) In your objdir with tests enabled: make -C netwerk/test/ SOLO_FILE="test_streamcopier.js" libs check-one ###!!! ASSERTION: index out of bounds: 'indx < segCount', file c:\repos\trunk\mozilla\xpcom\io\nsSegmentedBuffer.h, line 58
Attachment #695654 - Attachment is patch: true
OK, I found something very interesting: the storage input stream isn't reading from the beginning of the storage stream.
Apparently it matters when I initialize the storage input stream: if I do so before there's any data in the output stream, then the crash happens.
When do we think this regressed? Are we certain this test represents a normal user scenario? If this is a longstanding issue it's unlikely we'd track for a specific release.
I'd be very surprised if this was a regression of any kind.
Crash Signature: xul!nsStorageInputStream::ReadSegments+0x00000000000000E3 (c:\repos\trunk\mozilla\xpcom\io\nsstoragestream.cpp, line 424) xul!nsStorageInputStream::Read+0x000000000000001D (c:\repos\trunk\mozilla\xpcom\io\nsstoragestream.cpp line 397) xul!nsScrip… → [@ nsStorageInputStream::ReadSegments]
(In reply to Alex Vincent [:WeirdAl] from comment #4) > I'd be very surprised if this was a regression of any kind. Given that, if you think this should still be pursued asap, please nominate with more context as to why you think this should track for a specific release.
tracking-firefox20: ? → -
Does this still reproduce?
No, it looks like it's working now. :-)
Status: NEW → RESOLVED
Last Resolved: 3 years ago
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.