[Webkit-unassigned] [Bug 11689] WebKit nightly crashes on launch

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Sat Nov 25 09:03:17 PST 2006


http://bugs.webkit.org/show_bug.cgi?id=11689





------- Comment #5 from bdash at webkit.org  2006-11-25 09:03 PDT -------
WebKit is a piece of code used by many applications.  I believe what you are
referring to as "WebKit" is WebKit.app in the nightly builds -- this simply
launches Safari with an updated copy of the WebKit framework.  The way Safari
extensions work is typically by modifying (patching) code at runtime.  If the
code that they hook into changes, the extension will not behave as expected. 
In most cases extensions are poorly coded so not behaving as expected becomes
equivalent ends up being a crash.  There are several well-designed Safari
extensions that do things properly, and thus will _not_ crash when WebKit is
updated.  They typically do this by performing version checks on both the
application and any frameworks that they interact with, and if they detect a
newer version than they were developed against they simply refuse to work
rather than risking a crash.

I can think of two reasons why you'd be seeing a crash with a nightly build of
WebKit but not with released Safari 2.0, or a Safari 3.0 seed:
1) A piece of code that the extension hooks has been reworked very recently, or
2) The extension causing you problems version checks only the Safari
application, and not the frameworks it hooks.  In this case it cannot tell the
difference between Safari 2.0 with system WebKit and Safari 2.0 with a nightly
build of WebKit, while it can tell the difference between Safari 2.0 and 3.0. 
Its very likely that it would be silently disabling itself in the case of
Safari 3.0.

There is nothing we can do _but_ have extension developers fix their
extensions.  We can't simply stop working on WebKit completely out of fear of
breaking badly coded extensions, all of which are third-party components that
are (ab)using the input management functions of Mac OS X to run arbitrary code
inside another application.  We have absolutely no control over what these
extensions do, so there is nothing that we can do to stop extensions from
causing crashes.  Yes, crashes are bad.  No, blaming WebKit isn't going to make
it stop happening as its not something we can control.


-- 
Configure bugmail: http://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.



More information about the webkit-unassigned mailing list