Could you just trim out the "script" part of it, and leave behind the
Javascript example?

On Tue, Apr 14, 2009 at 3:28 AM, Chris Chabot <[email protected]> wrote:
> Any comments on this? I would like to commit this to break the currently
> broken trunk
>
> On Mon, Apr 13, 2009 at 11:30 AM, Chris Chabot <[email protected]> wrote:
>
>> Hey All,
>>
>> php-shindig's new gadget parser, which uses a dom parser, gets confused
>> when assigning a node value that contains a /* comment including a
>> &gt;script&lt; */ .. as in it translates it back to <script>foo</script&lt;,
>> which then causes the browser to trip because of the unterminated script tag
>> (odd btw that it doesn't translate the last &lt; but that's irrelevant to
>> the proposed solution)
>>
>> Would it be ok to remove that comment from core/prefs.js ? The patch would
>> be:
>>
>> Index: src/main/javascript/features/core/prefs.js
>> ===================================================================
>> --- src/main/javascript/features/core/prefs.js    (revision 764189)
>> +++ src/main/javascript/features/core/prefs.js    (working copy)
>> @@ -28,16 +28,6 @@
>>   *   var prefs = new gadgets.Prefs();
>>   *   var name = prefs.getString("name");
>>   *   var lang = prefs.getLang();
>> - *
>> - * Modules with type=url can also use this library to parse arguments
>> passed
>> - * by URL, but this is not the common case:
>> - *
>> - *   &lt;script src="http://apache.org/shindig/prefs.js
>> "&gt;&lt;/script&gt;
>> - *   &lt;script&gt;
>> - *   gadgets.Prefs.parseUrl();
>> - *   var prefs = new gadgets.Prefs();
>> - *   var name = prefs.getString("name");
>> - *   &lt;/script&lg;
>>   */
>>
>>
>> We've had trouble with url's and script tags in comments before (as in
>> hundreds of thousands of 404's a day on apache's servers due to them), and
>> now this issue with the xml parser i'm using, possibly it would be safer not
>> to include any form of script tags in the comments? :) (this is the only
>> occurrence of a script tag left in the features).
>>
>> Would like to have a second opinion before committing this 'patch'.
>>
>>    -- Chris
>>
>

Reply via email to