<?xml version="1.0" encoding="UTF-8" standalone="yes" ?>
<!DOCTYPE bugzilla SYSTEM "https://bugs.webkit.org/page.cgi?id=bugzilla.dtd">

<bugzilla version="5.0.4.1"
          urlbase="https://bugs.webkit.org/"
          
          maintainer="admin@webkit.org"
>

    <bug>
          <bug_id>15766</bug_id>
          
          <creation_ts>2007-10-30 12:46:26 -0700</creation_ts>
          <short_desc>[GTK] WebKit sometimes spews binary data as text/plain into iframes</short_desc>
          <delta_ts>2007-11-06 16:05:23 -0800</delta_ts>
          <reporter_accessible>1</reporter_accessible>
          <cclist_accessible>1</cclist_accessible>
          <classification_id>1</classification_id>
          <classification>Unclassified</classification>
          <product>WebKit</product>
          <component>WebKitGTK</component>
          <version>523.x (Safari 3)</version>
          <rep_platform>PC</rep_platform>
          <op_sys>Linux</op_sys>
          <bug_status>RESOLVED</bug_status>
          <resolution>FIXED</resolution>
          
          
          <bug_file_loc></bug_file_loc>
          <status_whiteboard></status_whiteboard>
          <keywords>Gtk</keywords>
          <priority>P2</priority>
          <bug_severity>Normal</bug_severity>
          <target_milestone>---</target_milestone>
          
          
          <everconfirmed>1</everconfirmed>
          <reporter name="Rodney Dawes">dobey</reporter>
          <assigned_to name="Nobody">webkit-unassigned</assigned_to>
          <cc>jhoneycutt</cc>
          

      

      

      

          <comment_sort_order>oldest_to_newest</comment_sort_order>  
          <long_desc isprivate="0" >
    <commentid>59862</commentid>
    <comment_count>0</comment_count>
    <who name="Rodney Dawes">dobey</who>
    <bug_when>2007-10-30 12:46:26 -0700</bug_when>
    <thetext>When a plug-in is not available for a type, webkit creates a new frame and appears to dump a textual representation of the file data into it, when plug-ins are enabled. The attached patch resolves this by checking the MIME type to see if we should be creating a frame or defaulting back to ObjectTypeNone.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>59863</commentid>
    <comment_count>1</comment_count>
      <attachid>16950</attachid>
    <who name="Rodney Dawes">dobey</who>
    <bug_when>2007-10-30 12:47:03 -0700</bug_when>
    <thetext>Created attachment 16950
Patch to validate mime type before dumping in an iframe</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>59933</commentid>
    <comment_count>2</comment_count>
      <attachid>16966</attachid>
    <who name="Rodney Dawes">dobey</who>
    <bug_when>2007-10-31 08:01:50 -0700</bug_when>
    <thetext>Created attachment 16966
Updated patch to match the win32 port

This is an updated version of the patch to match the win32 port, with a small change to return use ObjectContentNone if the type is empty, so that we don&apos;t assume we might be able to display the content, as we don&apos;t know what it is. It also lacks the check for an available plug-in, which will be added back when plug-ins are available and working.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>60004</commentid>
    <comment_count>3</comment_count>
      <attachid>16982</attachid>
    <who name="Rodney Dawes">dobey</who>
    <bug_when>2007-11-01 08:28:17 -0700</bug_when>
    <thetext>Created attachment 16982
Updated patch to fix one style issue and include a ChangeLog entry in header</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>60068</commentid>
    <comment_count>4</comment_count>
      <attachid>16982</attachid>
    <who name="Mark Rowe (bdash)">mrowe</who>
    <bug_when>2007-11-02 10:44:18 -0700</bug_when>
    <thetext>Comment on attachment 16982
Updated patch to fix one style issue and include a ChangeLog entry in header

For sake of compatibility I think you should match Windows and Mac in the return value when the MIME type is empty.  A minor coding style nit that should be fixed is the lack of whitespace around the + operator.  You should also include your ChangeLog entry, with spaces instead of tabs for indentation, as part of the diff.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>60081</commentid>
    <comment_count>5</comment_count>
      <attachid>16997</attachid>
    <who name="Rodney Dawes">dobey</who>
    <bug_when>2007-11-02 14:00:17 -0700</bug_when>
    <thetext>Created attachment 16997
Update patch to match win32/osx and include ChangeLog as patch</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>60122</commentid>
    <comment_count>6</comment_count>
      <attachid>16997</attachid>
    <who name="Alp Toker">alp</who>
    <bug_when>2007-11-02 18:07:51 -0700</bug_when>
    <thetext>Comment on attachment 16997
Update patch to match win32/osx and include ChangeLog as patch

&gt;Index: WebKit/gtk/ChangeLog
&gt;===================================================================
&gt;--- WebKit/gtk/ChangeLog	(revision 27376)
&gt;+++ WebKit/gtk/ChangeLog	(working copy)
&gt;@@ -1,3 +1,11 @@
&gt;+2007-11-02  Rodney Dawes  &lt;dobey@wayofthemonkey.com&gt;
&gt;+
&gt;+	Do some MIME type comparisons to avoid spewing flash/etc... into
&gt;+	an iframe as text/plain because the HTTP connection told us it was
&gt;+
&gt;+	* WebCoreSupport/FrameLoaderClientGtk.cpp:
&gt;+	(FrameLoaderClient::objectContentType):
&gt;+
&gt; 2007-10-29  Alp Toker  &lt;alp@atoker.com&gt;
&gt; 
&gt;         Reviewed by Maciej.
&gt;Index: WebKit/gtk/WebCoreSupport/FrameLoaderClientGtk.cpp
&gt;===================================================================
&gt;--- WebKit/gtk/WebCoreSupport/FrameLoaderClientGtk.cpp	(revision 27376)
&gt;+++ WebKit/gtk/WebCoreSupport/FrameLoaderClientGtk.cpp	(working copy)
&gt;@@ -226,14 +247,23 @@ Widget* FrameLoaderClient::createJavaApp
&gt; 
&gt; ObjectContentType FrameLoaderClient::objectContentType(const KURL&amp; url, const String&amp; mimeType)
&gt; {
&gt;-    if (url.isEmpty())
&gt;-        return ObjectContentType();
&gt;+    String type = mimeType;
&gt;+    if (type.isEmpty())
&gt;+        type = MIMETypeRegistry::getMIMETypeForExtension(url.path().mid(url.path().findRev(&apos;.&apos;)+1));
&gt;+
&gt;+    if (type.isEmpty())
&gt;+        return WebCore::ObjectContentFrame; // Match win32/OSX behavior for now

I guess we can&apos;t use MIMETypeRegistry::getMIMETypeForPath() because it returns &quot;application/octet-stream&quot; instead of an error status? That would explain why Win did the path parsing themselves.

&gt; 
&gt;-    // TODO: use more than just the extension to determine the content type?
&gt;-    String rtype = MIMETypeRegistry::getMIMETypeForPath(url.path());
&gt;-    if (!rtype.isEmpty())
&gt;-        return ObjectContentFrame;
&gt;-    return ObjectContentType();
&gt;+    if (MIMETypeRegistry::isSupportedImageMIMEType(type))
&gt;+        return WebCore::ObjectContentImage;
&gt;+
&gt;+    if (PluginDatabaseGtk::installedPlugins()-&gt;isMIMETypeRegistered(mimeType))
&gt;+        return WebCore::ObjectContentNetscapePlugin;

Careful. There is no PluginDatabaseGtk yet, so this will break the build. I appreciate that you&apos;re juggling a lot of changes. Perhaps it&apos;ll be OK to comment this out along with a TODO.

&gt;+
&gt;+    if (MIMETypeRegistry::isSupportedNonImageMIMEType(type))
&gt;+        return WebCore::ObjectContentFrame;
&gt;+
&gt;+    return WebCore::ObjectContentNone;
&gt; }
&gt; 
&gt; String FrameLoaderClient::overrideMediaType() const</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>60297</commentid>
    <comment_count>7</comment_count>
    <who name="Rodney Dawes">dobey</who>
    <bug_when>2007-11-05 09:12:33 -0800</bug_when>
    <thetext>(In reply to comment #6)
&gt; Careful. There is no PluginDatabaseGtk yet, so this will break the build. I
&gt; appreciate that you&apos;re juggling a lot of changes. Perhaps it&apos;ll be OK to
&gt; comment this out along with a TODO.

D&apos;oh. Forgot to remove that bit. This is what happens when you do too many things at once in the same tree. :) Attaching an updated patch now.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>60298</commentid>
    <comment_count>8</comment_count>
      <attachid>17043</attachid>
    <who name="Rodney Dawes">dobey</who>
    <bug_when>2007-11-05 09:13:44 -0800</bug_when>
    <thetext>Created attachment 17043
Updated patch to remove the plugin-specific bit for now</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>60350</commentid>
    <comment_count>9</comment_count>
    <who name="Alp Toker">alp</who>
    <bug_when>2007-11-05 16:20:54 -0800</bug_when>
    <thetext>My understanding is that under this condition:

  url=&quot;http://www.example.com/foo.unknownextension&quot;
  mimeType=&quot;&quot;

This code will return ObjectContentFrame.

And under this condition:

  url=&quot;http://www.example.com/foo&quot;
  mimeType=&quot;&quot;

It&apos;ll return ObjectContentNone.

I&apos;m not sure if this is good behaviour. Problem also seems to exist in the Win port. What do you think?</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>60352</commentid>
    <comment_count>10</comment_count>
    <who name="Rodney Dawes">dobey</who>
    <bug_when>2007-11-05 17:15:22 -0800</bug_when>
    <thetext>Yes. I think the behavior in this patch is the same for win32 and osx. It was suggested by bdash that we do the same as those for now, and then make a separate change proposal to have them all return ObjectContentNone, when we don&apos;t know what the type is. This seems to be how gecko behaves, by just ignoring the content, rather than dumping it in a frame.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>60405</commentid>
    <comment_count>11</comment_count>
      <attachid>17043</attachid>
    <who name="Darin Adler">darin</who>
    <bug_when>2007-11-06 07:51:50 -0800</bug_when>
    <thetext>Comment on attachment 17043
Updated patch to remove the plugin-specific bit for now

+        return WebCore::ObjectContentFrame; // Match win32/OSX behavior for now

I think that&apos;s a confusing comment. What&apos;s &quot;now&quot;? Will this need to change in the future? If so, why?</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>60406</commentid>
    <comment_count>12</comment_count>
    <who name="Darin Adler">darin</who>
    <bug_when>2007-11-06 07:53:31 -0800</bug_when>
    <thetext>(In reply to comment #9)
&gt; My understanding is that under this condition:
&gt; 
&gt;   url=&quot;http://www.example.com/foo.unknownextension&quot;
&gt;   mimeType=&quot;&quot;
&gt; 
&gt; This code will return ObjectContentFrame.
&gt; 
&gt; And under this condition:
&gt; 
&gt;   url=&quot;http://www.example.com/foo&quot;
&gt;   mimeType=&quot;&quot;
&gt; 
&gt; It&apos;ll return ObjectContentNone.
&gt; 
&gt; I&apos;m not sure if this is good behaviour. Problem also seems to exist in the Win
&gt; port. What do you think?

I&apos;d like to see more of the high level algorithm here in platform-independent code; that way we could correct this mistake (assuming it is one) once, rather than having multiple copies of it. We could consider changing the client function to do a smaller piece of the work.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>60410</commentid>
    <comment_count>13</comment_count>
    <who name="Rodney Dawes">dobey</who>
    <bug_when>2007-11-06 08:37:05 -0800</bug_when>
    <thetext>(In reply to comment #11)
&gt; (From update of attachment 17043 [edit])
&gt; +        return WebCore::ObjectContentFrame; // Match win32/OSX behavior for
&gt; now
&gt; 
&gt; I think that&apos;s a confusing comment. What&apos;s &quot;now&quot;? Will this need to change in
&gt; the future? If so, why?

&quot;Now&quot; would be &quot;the immediate future.&quot; As soon as this patch lands, the plan was to patch all the platforms to return ObjectContentNone in this case, to avoid dumping things into frames when we have no idea what they are, matching the behavior in Gecko.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>60411</commentid>
    <comment_count>14</comment_count>
    <who name="Rodney Dawes">dobey</who>
    <bug_when>2007-11-06 08:41:16 -0800</bug_when>
    <thetext>(In reply to comment #12)
&gt; I&apos;d like to see more of the high level algorithm here in platform-independent
&gt; code; that way we could correct this mistake (assuming it is one) once, rather
&gt; than having multiple copies of it. We could consider changing the client
&gt; function to do a smaller piece of the work.

Does it make sense to do this? The portion to check if a plug-in is available is platform-dependent. Would the method then have to check the parent method&apos;s return value, and then see if a plug-in is available, and return the parent method&apos;s return if not? We&apos;d have to duplicate the bits to get the MIME type String still though, in the child function, I think.

</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>60441</commentid>
    <comment_count>15</comment_count>
    <who name="Mark Rowe (bdash)">mrowe</who>
    <bug_when>2007-11-06 16:05:23 -0800</bug_when>
    <thetext>Landed in r27491.</thetext>
  </long_desc>
      
          <attachment
              isobsolete="1"
              ispatch="1"
              isprivate="0"
          >
            <attachid>16950</attachid>
            <date>2007-10-30 12:47:03 -0700</date>
            <delta_ts>2007-10-31 08:01:50 -0700</delta_ts>
            <desc>Patch to validate mime type before dumping in an iframe</desc>
            <filename>webkit-frameloader-fix.patch</filename>
            <type>text/plain</type>
            <size>1664</size>
            <attacher name="Rodney Dawes">dobey</attacher>
            
              <data encoding="base64">MjAwNy0xMC0zMCAgUm9kbmV5IERhd2VzICA8ZG9iZXlAd2F5b2Z0aGVtb25rZXkuY29tPgoKCURv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</data>

          </attachment>
          <attachment
              isobsolete="1"
              ispatch="1"
              isprivate="0"
          >
            <attachid>16966</attachid>
            <date>2007-10-31 08:01:50 -0700</date>
            <delta_ts>2007-11-01 08:28:17 -0700</delta_ts>
            <desc>Updated patch to match the win32 port</desc>
            <filename>webkit-frameloader-fix.patch</filename>
            <type>text/plain</type>
            <size>1310</size>
            <attacher name="Rodney Dawes">dobey</attacher>
            
              <data encoding="base64">SW5kZXg6IFdlYktpdC9ndGsvV2ViQ29yZVN1cHBvcnQvRnJhbWVMb2FkZXJDbGllbnRHdGsuY3Bw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=
</data>

          </attachment>
          <attachment
              isobsolete="1"
              ispatch="1"
              isprivate="0"
          >
            <attachid>16982</attachid>
            <date>2007-11-01 08:28:17 -0700</date>
            <delta_ts>2007-11-02 14:00:17 -0700</delta_ts>
            <desc>Updated patch to fix one style issue and include a ChangeLog entry in header</desc>
            <filename>webkit-frameloader-fix.patch</filename>
            <type>text/plain</type>
            <size>1587</size>
            <attacher name="Rodney Dawes">dobey</attacher>
            
              <data encoding="base64">MjAwNy0xMC0zMCAgUm9kbmV5IERhd2VzICA8ZG9iZXlAd2F5b2Z0aGVtb25rZXkuY29tPgoKCURv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</data>
<flag name="review"
          id="7206"
          type_id="1"
          status="-"
          setter="mrowe"
    />
          </attachment>
          <attachment
              isobsolete="1"
              ispatch="1"
              isprivate="0"
          >
            <attachid>16997</attachid>
            <date>2007-11-02 14:00:17 -0700</date>
            <delta_ts>2007-11-05 09:13:44 -0800</delta_ts>
            <desc>Update patch to match win32/osx and include ChangeLog as patch</desc>
            <filename>webkit-frameloader-fix.patch</filename>
            <type>text/plain</type>
            <size>2003</size>
            <attacher name="Rodney Dawes">dobey</attacher>
            
              <data encoding="base64">SW5kZXg6IFdlYktpdC9ndGsvQ2hhbmdlTG9nCj09PT09PT09PT09PT09PT09PT09PT09PT09PT09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</data>
<flag name="review"
          id="7212"
          type_id="1"
          status="-"
          setter="alp"
    />
          </attachment>
          <attachment
              isobsolete="0"
              ispatch="1"
              isprivate="0"
          >
            <attachid>17043</attachid>
            <date>2007-11-05 09:13:44 -0800</date>
            <delta_ts>2007-11-06 07:51:50 -0800</delta_ts>
            <desc>Updated patch to remove the plugin-specific bit for now</desc>
            <filename>webkit-frameloader-fix.patch</filename>
            <type>text/plain</type>
            <size>1884</size>
            <attacher name="Rodney Dawes">dobey</attacher>
            
              <data encoding="base64">SW5kZXg6IFdlYktpdC9ndGsvV2ViQ29yZVN1cHBvcnQvRnJhbWVMb2FkZXJDbGllbnRHdGsuY3Bw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</data>
<flag name="review"
          id="7247"
          type_id="1"
          status="+"
          setter="darin"
    />
          </attachment>
      

    </bug>

</bugzilla>