[SCM] WebKit Debian packaging branch, debian/unstable, updated. debian/1.1.15-1-40151-g37bb677
trey
trey at 268f45cc-cd09-0410-ab3c-d52691b4dbfc
Sat Sep 26 07:22:10 UTC 2009
The following commit has been merged in the debian/unstable branch:
commit f925d4fe8b6b21a914755d65595984691b5a4201
Author: trey <trey at 268f45cc-cd09-0410-ab3c-d52691b4dbfc>
Date: Mon Jan 27 23:33:33 2003 +0000
3157104 - reproducible assert in _continueFragmentScrollAfterNavigationPolicy
Remove assertion, and it turns out the existing code will handle the case.
* WebView.subproj/WebFramePrivate.m:
(-[WebFrame _continueFragmentScrollAfterNavigationPolicy:request:]):
git-svn-id: http://svn.webkit.org/repository/webkit/trunk@3468 268f45cc-cd09-0410-ab3c-d52691b4dbfc
diff --git a/WebKit/ChangeLog b/WebKit/ChangeLog
index 4664140..cae6607 100644
--- a/WebKit/ChangeLog
+++ b/WebKit/ChangeLog
@@ -1,3 +1,12 @@
+2003-01-27 Trey Matteson <trey at apple.com>
+
+ 3157104 - reproducible assert in _continueFragmentScrollAfterNavigationPolicy
+
+ Remove assertion, and it turns out the existing code will handle the case.
+
+ * WebView.subproj/WebFramePrivate.m:
+ (-[WebFrame _continueFragmentScrollAfterNavigationPolicy:request:]):
+
2003-01-27 Richard Williamson <rjw at apple.com>
Make emptying the page cache synchronous when "Empty Cache"
diff --git a/WebKit/WebView.subproj/WebFramePrivate.m b/WebKit/WebView.subproj/WebFramePrivate.m
index 4755c33..d1d4d70 100644
--- a/WebKit/WebView.subproj/WebFramePrivate.m
+++ b/WebKit/WebView.subproj/WebFramePrivate.m
@@ -1353,10 +1353,16 @@ static CFAbsoluteTime _timeOfLastCompletedLoad;
[dataSrc _setURL:URL];
if (!isRedirect && ![self _shouldTreatURLAsSameAsCurrent:URL]) {
- ASSERT(![_private previousItem]);
// NB: must happen after _setURL, since we add based on the current request.
// Must also happen before we openURL and displace the scroll position, since
// adding the BF item will save away scroll state.
+
+ // NB2: If we were loading a long, slow doc, and the user anchor nav'ed before
+ // it was done, currItem is now set the that slow doc, and prevItem is whatever was
+ // before it. Adding the b/f item will bump the slow doc down to prevItem, even
+ // though its load is not yet done. I think this all works out OK, for one because
+ // we have already saved away the scroll and doc state for the long slow load,
+ // but it's not an obvious case.
[self _addBackForwardItemClippedAtTarget:NO];
}
--
WebKit Debian packaging
More information about the Pkg-webkit-commits
mailing list