There's a new crash in Firefox 3.6b5 with the signature "nsJARInputStream::InitFile(nsZipHandle*, nsZipItem*)" that hasn't been seen in any of the versions 3.5,3.5.1,3.5.2,3.5.3,3.5.4,3.5.5,3.5.6. So far we've seen 23+ of these crashes in the wild in the first 2 days of 3.6b5 testing. Please see http://crash-stats.mozilla.com/query/query?product=Firefox&version=Firefox%3A3.6b5&range_value=1&range_unit=weeks&query_search=signature&query_type=exact&query=nsJARInputStream%3A%3AInitFile%28nsZipHandle%2A%2C%20nsZipItem%2A%29&do_query=1 for more crash info. there was some work going on in the source files on the stack earlier during 3.6 development. ad1b7a04fbba 2009-08-12 13:50 -0700 Taras Glek - Bug 504864: mmap io for JARs r=bsmedberg stack looks like Frame Module Signature [Expand] Source 0 XUL nsJARInputStream::InitFile modules/libjar/nsZipArchive.h:276 1 XUL nsJAR::GetInputStreamWithSpec modules/libjar/nsJAR.cpp:342 2 XUL nsJAR::GetInputStream modules/libjar/nsJAR.cpp:316 3 XUL nsJARInputThunk::EnsureJarStream modules/libjar/nsJARChannel.cpp:166 4 XUL nsJARInputThunk::Read modules/libjar/nsJARChannel.cpp:203 5 XUL nsInputStreamTransport::Read netwerk/base/src/nsStreamTransportService.cpp:233 6 XUL nsStreamCopierOB::FillOutputBuffer xpcom/io/nsStreamUtils.cpp:566 7 XUL nsPipeOutputStream::WriteSegments xpcom/io/nsPipe3.cpp:1137 8 XUL nsStreamCopierOB::DoCopy xpcom/io/nsStreamUtils.cpp:583 9 XUL nsAStreamCopier::Process xpcom/io/nsStreamUtils.cpp:323 10 XUL nsAStreamCopier::Run xpcom/io/nsStreamUtils.cpp:439 11 XUL nsThreadPool::Run xpcom/threads/nsThreadPool.cpp:219 12 XUL nsThread::ProcessNextEvent xpcom/threads/nsThread.cpp:527 13 XUL NS_ProcessNextEvent_P nsThreadUtils.cpp:250 14 XUL nsThread::ThreadFunc xpcom/threads/nsThread.cpp:254 15 libnspr4.dylib _pt_root nsprpub/pr/src/pthreads/ptthread.c:228 16 libSystem.B.dylib _pthread_start 17 libSystem.B.dylib thread_start Is there a better component for libjar?

0-1 crashes in sept-oct and up until we started getting more users at beta1 daily crash numbers are pretty spikey. maybe increase volume days around the releases of new betas. highest peak was 60 crashes per day right around the release of beta2. near beta 4 also had a spike of 50 crashes

Crash Address 0x8 tglek@31399 275 NS_ABORT_IF_FALSE (aHandle, "Argument must not be NULL"); tglek@31399 276 if (aOffset > aHandle->mLen) Taras: "it was null"

(In reply to comment #3) > > *** This bug has been marked as a duplicate of bug 510844 *** Josh, I don't understand why this bug was marked as a DUP of 510844.