Bug 39700 – Clean up MimeClassInfo and PluginInfo

• Status: RESOLVED
• Resolution: FIXED
• Component: New Bugs

[Anders Carlsson]

Clean up MimeClassInfo and PluginInfo.

[Anders Carlsson]

Created attachment 57055 [details]
Patch

[WebKit Review Bot]

Attachment 57055 [details] did not build on chromium:
Build output: http://webkit-commit-queue.appspot.com/results/2457005

[Anders Carlsson]

Created attachment 57110 [details]
Patch

[Anders Carlsson]

The new patch should fix the Chromium build.

[WebKit Review Bot]

Attachment 57110 [details] did not build on chromium:
Build output: http://webkit-commit-queue.appspot.com/results/2529033

[Anders Carlsson]

Created attachment 57120 [details]
Patch

[WebKit Review Bot]

Attachment 57120 [details] did not build on chromium:
Build output: http://webkit-commit-queue.appspot.com/results/2500036

[Anders Carlsson]

Created attachment 57129 [details]
Patch

[Darin Adler]

Comment on attachment 57129 [details]
Patch

>  PluginData::~PluginData()
>  {
> -    deleteAllValues(m_plugins);
> -    deleteAllValues(m_mimes);
>  }

It'd be nice to just let the compiler generate this and not define it explicitly at all.

[Anders Carlsson]

(In reply to comment #9)
> (From update of attachment 57129 [details])
> >  PluginData::~PluginData()
> >  {
> > -    deleteAllValues(m_plugins);
> > -    deleteAllValues(m_mimes);
> >  }
> 
> It'd be nice to just let the compiler generate this and not define it explicitly at all.

The destructor? Sure!

[Anders Carlsson]

Committed r60258: <http://trac.webkit.org/changeset/60258>

[WebKit Review Bot]

http://trac.webkit.org/changeset/60258 might have broken GTK Linux 32-bit Release and Qt Linux Release
The following changes are on the blame list:
http://trac.webkit.org/changeset/60257
http://trac.webkit.org/changeset/60258

[Julie Parent]

This looks to have introduced a crash in fast/dom/prototype-inheritance-2.html.

I see the crash on the following bots:
* Chromium (linux, win, mac)
* Leopard Intel Debug (Tests)
* Snow Leopard Intel Leaks

[Darin Adler]

Is there a URL for the backtrace from the crash?

[Anders Carlsson]

I think that the crash was caused by another change - looks like Geoff fixed it in http://trac.webkit.org/changeset/60261

[Julie Parent]

Here is the Chromium linux bot stack:
	StackTrace::StackTrace() [0x80eeacf]
	base::(anonymous namespace)::StackDumpSignalHandler() [0x80de9da]
	0x4001c420
	WebCore::collectionIndexedPropertyEnumerator<>() [0x89648a8]
	v8::internal::GetKeysForIndexedInterceptor() [0x81ba92e]

Geoff's change was fixing a different crash.

[Eric Seidel (no email)]

Looks like there may be new failures on Qt after this change?
http://build.webkit.org/results/Qt%20Linux%20Release/r60258%20(12428)/results.html

[Csaba Osztrogonác]

(In reply to comment #17)
> Looks like there may be new failures on Qt after this change?
> http://build.webkit.org/results/Qt%20Linux%20Release/r60258%20(12428)/results.html

It seems this patch broke tests. I try to roll-out it locally now to confirm it.

[Roland Steiner]

The crashing of fast/dom/prototype-inheritance-2.html also occurred during my WebKit gardening. I submitted and committed a quick fix in WebKit r60276.

See https://bugs.webkit.org/show_bug.cgi?id=39811