[XML Viewer crash] Store the reference of pretty-print div as tree to be used during the processing

[XML Viewer crash] Store the reference of pretty-print div as tree to be used during the processing

XMLViewer creates a pretty-print div for rendering XML contents. Once the document onLoad is called upon, the processing of the tree takes place. In this if 
the source xml contains the id="tree", it will get conflicted. The resulting XML viewer
may appear blank if the id="tree" is present at the root node. Even worse, if the conflicting
id="tree" appears within any of the elements after the root, it causes the renderer to crash.

The fix is to store the pretty-print div as tree reference to be used during the internal processing.

R=abarth, pfeldman

BUG=284399
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=157201

[vivekg__, 2013-09-03 11:31:15]

Please take a look, thanks!

[abarth-chromium, 2013-09-03 20:57:35]

Won't this break extensions that hook prepareWebKitXMLViewer?

[vivekg_samsung, 2013-09-03 21:22:27]

On 2013/09/03 20:57:35, abarth wrote:
> Won't this break extensions that hook prepareWebKitXMLViewer?

Isn't "onAfterWebkitXMLViewerLoaded" the extension API to be used as the hook?
At least the test case
http://src.chromium.org/viewvc/blink/trunk/LayoutTests/xmlviewer/extensions-a...
makes use of this hook. And the source is attached back before this call is
made.

[abarth-chromium, 2013-09-04 00:40:21]

Looking around briefly, I'm pretty sure this CL will break
https://chrome.google.com/webstore/detail/efr-and-xbrl-addon/ghhlglecmgopkfda...

That extension assumes that the element with id webkit-xml-viewer-source-xml
will be in the document by DOMContentLoaded, which won't be the case after this
CL.

[vivekg__, 2013-09-04 01:47:54]

On 2013/09/04 00:40:21, abarth wrote:
> Looking around briefly, I'm pretty sure this CL will break
>
https://chrome.google.com/webstore/detail/efr-and-xbrl-addon/ghhlglecmgopkfda...
> 
> That extension assumes that the element with id webkit-xml-viewer-source-xml
> will be in the document by DOMContentLoaded, which won't be the case after
this
> CL.

hmm right! How about momentarily detaching the sourceXML from the DOM and
re-attaching it back once the processing is complete? Patch attached!

[abarth-chromium, 2013-09-04 02:02:40]

https://codereview.chromium.org/23604028/diff/6001/Source/core/xml/XMLViewer.js
File Source/core/xml/XMLViewer.js (right):

https://codereview.chromium.org/23604028/diff/6001/Source/core/xml/XMLViewer....
Source/core/xml/XMLViewer.js:80: var root = document.getElementById('tree');
Why not save off the |tree| reference instead of calling getElementById here?

[vivekg__, 2013-09-04 02:15:36]

On 2013/09/04 02:02:40, abarth wrote:
>
https://codereview.chromium.org/23604028/diff/6001/Source/core/xml/XMLViewer.js
> File Source/core/xml/XMLViewer.js (right):
> 
>
https://codereview.chromium.org/23604028/diff/6001/Source/core/xml/XMLViewer....
> Source/core/xml/XMLViewer.js:80: var root = document.getElementById('tree');
> Why not save off the |tree| reference instead of calling getElementById here?

Done. Thank you!

[abarth-chromium, 2013-09-04 02:45:45]

LGTM.  Thanks for being willing to iterate on this CL

[commit-bot: I haz the power, 2013-09-04 02:53:37]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/vivek.vg@samsung.com/23604028/1006

[commit-bot: I haz the power, 2013-09-04 05:57:19]

Change committed as 157201