[SCM] WebKit Debian packaging branch, debian/experimental, updated. upstream/1.3.3-9427-gc2be6fc

[bweinstein at apple.com]

The following commit has been merged in the debian/experimental branch:
commit e017fdb9fc8bd6fb4e4cb88a9dd757621d9df8b7
Author: bweinstein at apple.com <bweinstein at apple.com@268f45cc-cd09-0410-ab3c-d52691b4dbfc>
Date:   Thu Oct 28 23:43:10 2010 +0000

    Part of <https://bugs.webkit.org/show_bug.cgi?id=45980>
    userscripts/user-script-*-document.html tests are flaky on Windows
    
    Reviewed by Brady Eidson.
    
    WebCore:
    
    In WebFrameLoaderClient::committedLoad on Windows, we call through to PluginStream::didRecieveData
    if our manual loader is a plugin stream, and we can call didRecieveData with a 0 length.
    
    We shouldn't ASSERT that length is 0, because code paths exists where we call it with a 0 length.
    
    This is tested by re-enabling user-script-plugin-document.html on Windows.
    
    * plugins/PluginStream.cpp:
    (WebCore::PluginStream::didReceiveData):
    
    LayoutTests:
    
    Unskip the user-script-plugin-document test because its issue is fixed.
    
    * platform/win/Skipped:
    
    
    
    git-svn-id: http://svn.webkit.org/repository/webkit/trunk@70825 268f45cc-cd09-0410-ab3c-d52691b4dbfc

diff --git a/LayoutTests/ChangeLog b/LayoutTests/ChangeLog
index 09b1638..505da4c 100644
--- a/LayoutTests/ChangeLog
+++ b/LayoutTests/ChangeLog
@@ -1,3 +1,14 @@
+2010-10-28  Brian Weinstein  <bweinstein at apple.com>
+
+        Reviewed by Brady Eidson.
+
+        Part of <https://bugs.webkit.org/show_bug.cgi?id=45980>
+        userscripts/user-script-*-document.html tests are flaky on Windows
+
+        Unskip the user-script-plugin-document test because its issue is fixed.
+
+        * platform/win/Skipped:
+
 2010-10-28  Mihai Parparita  <mihaip at chromium.org>
 
         Reviewed by Darin Adler.
diff --git a/LayoutTests/platform/win/Skipped b/LayoutTests/platform/win/Skipped
index d5a2d8a..f496aee 100644
--- a/LayoutTests/platform/win/Skipped
+++ b/LayoutTests/platform/win/Skipped
@@ -988,7 +988,6 @@ userscripts/user-style-top-frame-only.html
 # https://bugs.webkit.org/show_bug.cgi?id=45980
 userscripts/user-script-video-document.html
 userscripts/user-script-audio-document.html
-userscripts/user-script-plugin-document.html
 
 # Directory upload is not enabled.
 fast/forms/input-file-directory-upload.html
diff --git a/WebCore/ChangeLog b/WebCore/ChangeLog
index c182a73..2fa6cab 100644
--- a/WebCore/ChangeLog
+++ b/WebCore/ChangeLog
@@ -1,3 +1,20 @@
+2010-10-28  Brian Weinstein  <bweinstein at apple.com>
+
+        Reviewed by Brady Eidson.
+
+        Part of <https://bugs.webkit.org/show_bug.cgi?id=45980>
+        userscripts/user-script-*-document.html tests are flaky on Windows
+
+        In WebFrameLoaderClient::committedLoad on Windows, we call through to PluginStream::didRecieveData
+        if our manual loader is a plugin stream, and we can call didRecieveData with a 0 length.
+        
+        We shouldn't ASSERT that length is 0, because code paths exists where we call it with a 0 length.
+        
+        This is tested by re-enabling user-script-plugin-document.html on Windows.
+
+        * plugins/PluginStream.cpp:
+        (WebCore::PluginStream::didReceiveData):
+
 2010-10-28  Mihai Parparita  <mihaip at chromium.org>
 
         Reviewed by Darin Adler.
diff --git a/WebCore/plugins/PluginStream.cpp b/WebCore/plugins/PluginStream.cpp
index 3e04c79..3be1da5 100644
--- a/WebCore/plugins/PluginStream.cpp
+++ b/WebCore/plugins/PluginStream.cpp
@@ -413,7 +413,6 @@ void PluginStream::didReceiveResponse(NetscapePlugInStreamLoader* loader, const
 void PluginStream::didReceiveData(NetscapePlugInStreamLoader* loader, const char* data, int length)
 {
     ASSERT(loader == m_loader);
-    ASSERT(length > 0);
     ASSERT(m_streamState == StreamStarted);
 
     // If the plug-in cancels the stream in deliverData it could be deleted, 

-- 
WebKit Debian packaging