RESOLVED FIXED 70725
[Qt][WK2] Crash when running pixel tests
https://bugs.webkit.org/show_bug.cgi?id=70725
Summary [Qt][WK2] Crash when running pixel tests
Balazs Kelemen
Reported 2011-10-24 08:00:03 PDT
Linux, 64-bit, up-to-date qt5. #0 0x00007f14d526f509 in memcpy () from /lib64/libc.so.6 #1 0x00007f14d62b5849 in MD5Update (ctx=0x764188, buf=0xc3cff0 "", len=2304) at tools/../../3rdparty/md5/md5.cpp:95 #2 0x00007f14d62b8b27 in QCryptographicHash::addData (this=0x7fff5f3ec840, data= 0x9cc8f0 "\377\377\377\377... #3 0x00000000004096ea in WTR::TestInvocation::dumpPixelsAndCompareWithExpected(OpaqueWKImage const*, OpaqueWKArray const*) () #4 0x000000000040c140 in WTR::TestInvocation::didReceiveMessageFromInjectedBundle(OpaqueWKString const*, void const*) () #5 0x00007f14d986d71e in WebKit::WebContextInjectedBundleClient::didReceiveMessageFromInjectedBundle(WebKit::WebContext*, WTF::String const&, WebKit::APIObject*) () from /home/balazs/work/master_clean/wk2BUild/Release/bin/../lib/libQtWebKit.so.4
Attachments
Patch (1.52 KB, patch)
2011-10-24 10:13 PDT, Balazs Kelemen
no flags
Balazs Kelemen
Comment 1 2011-10-24 10:07:50 PDT
Oops, my fault in http://trac.webkit.org/changeset/94524: hash.addData(reinterpret_cast<const char*>(image.constScanLine(row)), image.width() * image.bytesPerLine()); |width * bytesperline| is a bit too much ...
Balazs Kelemen
Comment 2 2011-10-24 10:13:23 PDT
Chang Shu
Comment 3 2011-10-24 10:19:33 PDT
Comment on attachment 112207 [details] Patch Just curious how this worked before. In addition, bytesPerLine sounds like it has taken width into consideration.
Chang Shu
Comment 4 2011-10-24 10:21:17 PDT
Comment on attachment 112207 [details] Patch sorry, i thought the code was added but it was removed. r=me.
Balazs Kelemen
Comment 5 2011-10-24 10:32:03 PDT
Comment on attachment 112207 [details] Patch Clearing flags on attachment: 112207 Committed r98252: <http://trac.webkit.org/changeset/98252>
Balazs Kelemen
Comment 6 2011-10-24 10:32:10 PDT
All reviewed patches have been landed. Closing bug.
Balazs Kelemen
Comment 7 2011-10-24 10:33:28 PDT
(In reply to comment #4) > (From update of attachment 112207 [details]) > sorry, i thought the code was added but it was removed. r=me. Sure, the value we passed was way too high, but sometimes it was working well with memory garbage :) Thanks for the review.
Note You need to log in before you can comment on or make changes to this bug.