http://build.webkit.org/results/Leopard%20Intel%20Release%20(Tests)/r50717%20(7067)/plugins/document-open-pretty-diff.html
I think it's just flakey.
--- layout-test-results/plugins/document-open-expected.txt 2009-11-09 21:31:25.000000000 -0800
+++ layout-test-results/plugins/document-open-actual.txt 2009-11-09 21:31:25.000000000 -0800
@@ -1 +1,2 @@
-CONSOLE MESSAGE: line 0: PLUGIN: DOCUMENT OPEN SUCCESS
+
+This tests that document.open invoked by a plugin via NPN_Invoke without a javascript context succeeds.
CCing author and reviewer.
So far I've only seen it fail once on the bots, so this bug is far from urgent. I've filed this so that we can track if it fails repeatedly. I suspect that it's possible for the bots to be slow enough to cause the timeout to fire before the plugin does the document.open.
Comment on attachment 42901[details]
Updated patch with incorrect indentations fixed.
Note that the first ChangeLog is in the wrong folder (it's in the root instead of the WebCore folder), so this patch cannot be committed as-is.
I'll leave reviewing the actual code to plug-in experts.
Comment on attachment 42903[details]
Updated patch with the WebCore Changelog now in the right place
Braces on new lines:
38 static void pluginLogWithWindowObject(NPObject* windowObject, NPP instance, const char* message) {
check-webkit-style might help find some of these. It's just webkit's version of cpp_lint.py
I wonder if m_client shouldn't manage its own lifetime. Seems a little strange for stop() to clear m_client. But I really know next-to-nothing about this area of code.
The stream is created by PluginView when it receives a response. The PluginView instance pointer is passed in to the create call which is saved as m_client.
When the plugin view is going away it calls stop on existing plugin stream instances and releases them.
I think setting m_client to NULL in stop is safe. We could look into getting m_client to manage its own lifetime in a subsequent patch.
Comment on attachment 42977[details]
Fixed line endings in the layout tests.
Rejecting patch 42977 from commit-queue.
Failed to run "['git', 'svn', 'dcommit']" exit_code: 1
Last 500 characters of output:
ebCore/plugins/PluginStream.cpp
M WebKitTools/ChangeLog
M WebKitTools/DumpRenderTree/TestNetscapePlugIn.subproj/PluginObject.cpp
A repository hook failed: MERGE request failed on '/repository/webkit/trunk': Commit blocked by pre-commit hook (exit code 1) with output:
svnlook: Can't write to stream: Broken pipe
The following ChangeLog files contain OOPS:
trunk/WebCore/ChangeLog
Please don't ever say "OOPS" in a ChangeLog file.
at /usr/local/libexec/git-core//git-svn line 469
Comment on attachment 43132[details]
The reviewed by comment in the webcore changelog was moved incorrectly below the comment. New patch fixes that.
Rejecting patch 43132 from commit-queue.
Failed to run "['git', 'svn', 'dcommit']" exit_code: 1
Last 500 characters of output:
Tree/TestNetscapePlugIn.subproj/PluginObject.cpp
A repository hook failed: MERGE request failed on '/repository/webkit/trunk': Commit blocked by pre-commit hook (exit code 1) with output:
svnlook: Can't write to stream: Broken pipe
The following files contain tab characters:
trunk/LayoutTests/ChangeLog
Please use spaces instead to indent.
If you must commit a file with tabs, use svn propset to set the "allow-tabs" property.
at /usr/local/libexec/git-core//git-svn line 469
2009-11-10 15:02 PST, Ananta Iyengar
2009-11-10 15:04 PST, Ananta Iyengar
2009-11-10 15:17 PST, Ananta Iyengar
2009-11-10 16:49 PST, Ananta Iyengar
2009-11-11 09:21 PST, Ananta Iyengar
2009-11-11 10:41 PST, Ananta Iyengar
commit-queue: commit-queue-
2009-11-12 21:47 PST, Ananta Iyengar
2009-11-12 21:51 PST, Ananta Iyengar
commit-queue: commit-queue-
2009-11-12 22:21 PST, Ananta Iyengar