[webkit-dev] Random Failures
eseidel at apple.com
Thu Jul 7 13:49:26 PDT 2005
Tracking this as:
On Jul 7, 2005, at 11:29 AM, Eric Seidel wrote:
> libgmalloc didn't seem to produce anything.
> This seems to be a timing issue.
> I was able to catch it failing in the debugger.
> Program received signal EXC_BAD_ACCESS, Could not access memory.
> Reason: KERN_PROTECTION_FAILURE at address: 0x00003ff3
> 0x01040c2c in QString::setLength ()
> (gdb) bt
> #0 0x01040c2c in QString::setLength ()
> #1 0x010414cc in QString::insert ()
> #2 0x010d75c8 in khtml::Decoder::decode ()
> #3 0x0105c7b8 in KHTMLPart::write ()
> #4 0x0022eda8 in -[WebDataSource(WebPrivate) _receivedData:]
> (self=0x4da3d20, _cmd=0x2ee0, data=0x4da5590) at /Volumes/Stuff/
> #5 0x0024d564 in -[WebMainResourceLoader addData:]
> (self=0x4da8450, _cmd=0x2ee0, data=0x4da5590) at /Volumes/Stuff/
> #6 0x0022b8e0 in -[WebLoader didReceiveData:lengthReceived:]
> (self=0x4da8450, _cmd=0x2ee0, data=0x4da5590, lengthReceived=16371)
> at /Volumes/Stuff/Projects/WKOpen/WebKit/WebView.subproj/
> #7 0x0024df14 in -[WebMainResourceLoader
> didReceiveData:lengthReceived:] (self=0x4da8450, _cmd=0x2ee0,
> data=0x4da5590, lengthReceived=0) at /Volumes/Stuff/Projects/WKOpen/
> #8 0x0022bdd8 in -[WebLoader
> connection:didReceiveData:lengthReceived:] (self=0x0, _cmd=0x2ee0,
> con=0x0, data=0x0, lengthReceived=0) at /Volumes/Stuff/Projects/
> #9 0x928af7d8 in -[NSURLConnection(NSURLConnectionInternal)
> _sendDidReceiveDataCallback] ()
> #10 0x928adc4c in -[NSURLConnection(NSURLConnectionInternal)
> _sendCallbacks] ()
> #11 0x928ada18 in _sendCallbacks ()
> #12 0x9074bd2c in __CFRunLoopDoSources0 ()
> #13 0x9074b25c in __CFRunLoopRun ()
> #14 0x9074acdc in CFRunLoopRunSpecific ()
> #15 0x9288c164 in -[NSRunLoop runMode:beforeDate:] ()
> #16 0x00002f74 in dumpRenderTree (filename=0x90a19534
> "runMode:beforeDate:") at /Volumes/Stuff/Projects/WKOpen2/
> #17 0x00003274 in main (argc=2, argv=0xbffff504) at /Volumes/Stuff/
> On Jul 7, 2005, at 8:45 AM, Eric Seidel wrote:
>> Thanks Mark.
>> One next step is to try running under libgmalloc. That should
>> catch whatever this is every time on the same test.
>> I'll try that when I get into work.
>> On Jul 7, 2005, at 5:59 AM, Mark Rowe wrote:
>>> On 8/07/2005, at 00:41 , Mark Rowe wrote:
>>>> I've spent the last two hours trying to track down which change
>>>> causes the test failures + DumpRenderTree crash. As was
>>>> mentioned, rolling out the patch for bugs #3882 and #3829 does
>>>> not resolve the issue. Rolling out #3826 and #3674 individually
>>>> didn't help either. It seemed that removing *both* stopped the
>>>> crashing but that may have just been luck as the crashes are
>>>> intermittent at best. I'm retesting this scenario now.
>>> It seems I was just lucky... As far as I can tell none of the
>>> four patches have any relation to the failing test + crash. I'm
>>> not sure which other changes have been checked in recently that I
>>> could check, so I'll leave that up to someone else.
>>> Mark Rowe
>> webkit-dev mailing list
>> webkit-dev at opendarwin.org
More information about the webkit-dev