Bug 14604 – crash if ASV3 installed

• Component: Plug-ins
• Status: RESOLVED WORKSFORME

[jay]

for builds post r24002 where ASV3.1 is installed in /Library/Internet PlugIns/ 
opening any SVG cause webkit to crash.

[jay]

ASV means Adobe SVG Viewer

[Alexey Proskuryakov]

I cannot reproduce this issue with r24241 (local debug build), ASV 3.0x76. Please provide more detailed steps to reproduce, and a full crash log.

According to <http://www.adobe.com/svg/viewer/install/>, the latest ASV version is 3.0 - are you sure that you have 3.1?

[jay]

Created attachment 15506 [details]
crash report

[jay]

#2

my mistake ASV: Version 3.0 beta 3 (Build 76)

[Alexey Proskuryakov]

Thread 0 Crashed:
0   com.apple.WebCore              	0x01365ae4 WebCore::PluginTokenizer::writeRawData(char const*, int) + 116
1   com.apple.WebCore              	0x0145e968 WebCore::FrameLoader::write(char const*, int, bool) + 216
2   com.apple.WebCore              	0x0110c488 -[WebCoreFrameBridge receivedData:textEncodingName:] + 408
3   com.apple.WebKit               	0x0032b970 -[WebHTMLRepresentation receivedData:withDataSource:] + 160
4   com.apple.WebKit               	0x0032670c -[WebDataSource(WebInternal) _receivedData:] + 92
5   com.apple.WebKit               	0x0037b848 WebFrameLoaderClient::committedLoad(WebCore::DocumentLoader*, char const*, int) + 120
6   com.apple.WebCore              	0x0147343c WebCore::DocumentLoader::commitLoad(char const*, int) + 92
7   com.apple.WebCore              	0x0147bf30 WebCore::ResourceLoader::didReceiveData(char const*, int, long long, bool) + 80
8   com.apple.WebCore              	0x01478664 WebCore::MainResourceLoader::didReceiveData(char const*, int, long long, bool) + 52
9   com.apple.WebCore              	0x0144d048 -[WebCoreResourceHandleAsDelegate connection:didReceiveData:lengthReceived:] + 152
10  com.apple.Foundation           	0x92c14624 -[NSURLConnection(NSURLConnectionInternal) _sendDidReceiveDataCallback] + 564

[David Kilzer (:ddkilzer)]

<rdar://problem/5335841>

[David Kilzer (:ddkilzer)]

Confirmed via Comment #5.

[Alexey Proskuryakov]

Actually, I still couldn't reproduce - I just copied the top of the crash log to a comment for easier searching.

[Alexey Proskuryakov]

One crasher in WebCore::PluginTokenizer::writeRawData() was fixed by Anders in r24631.

[Eric Seidel (no email)]

AFAIK this has nothing to do with SVG.  I also suspect it's been fixed by Anders' fix.

[Alexey Proskuryakov]

Now closing.

[jay]

Created attachment 24012 [details]
crash report

[jay]

#11 please do not close bug wfm without contacting reporter first.

are you using ppc? just one possible difference...

[Alexey Proskuryakov]

This is a new and different crash, we have at least a couple of bugs about it (see bug 20958, bug 21276).

> #11 please do not close bug wfm without contacting reporter first.

Why are you saying this? That's definitely a correct thing to do for a bug that hasn't been confirmed in a year, and that we even have a guess at when exactly it was fixed (r24631). A reporter can always re-open the bug if it still occurs.

In this particular case, re-opening was arguably the correct thing for you to do, because a crash still happens with the same steps to reproduce, but now I must close it again, because it is a different issue.

[jay]

#14,

you will also have to file the new bug, if you want a record, as I am secverly reducing my commitment to bug filing.

not least because of lack of progress on outstanding bugs.

regards