[
https://issues.apache.org/jira/browse/WICKET-3500?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13002606#comment-13002606
]
Georg Juttner commented on WICKET-3500:
---------------------------------------
The causing html snippet is
<!--[if IE 6]>
<link rel="stylesheet" type="text/css" href="ie-buggy.css" />
<![endif]-->
The closing comment tag is in a separate line. I don't know why the pull parser
splits this to separate chunks. This happens even in case of
<!--[if IE 6]><link rel="stylesheet" type="text/css" href="ie-buggy.css"
/><![endif]-->
The first "rawMarkup" remains "<!--[if IE 6]>" leading to the endless loop.
> Method AbstractMarkupParser.removeComment() causes an endless loop
> ------------------------------------------------------------------
>
> Key: WICKET-3500
> URL: https://issues.apache.org/jira/browse/WICKET-3500
> Project: Wicket
> Issue Type: Bug
> Components: wicket-core
> Affects Versions: 1.4.16
> Reporter: Georg Juttner
> Attachments: WICKET-3500.patch
>
>
> The last change (see WICKET-3277) of method
> AbstractMarkupParser.removeComment() in Wicket 1.4.16 causes an endless loop
> if the markup contains conditional comments for IE.
> For example, the following string causes a loop in method removeComment()
> rawMarkup = "\r\n <!--[if IE 6]>\r\n "
> According to the following code snippet
> private String removeComment(String rawMarkup)
> {
> int pos1 = rawMarkup.indexOf("<!--");
> while (pos1 != -1)
> {
> int pos2 = rawMarkup.indexOf("-->", pos1 + 4);
> final StringBuilder buf = new
> StringBuilder(rawMarkup.length());
> if (pos2 != -1)
> {
> ...
> }
> pos1 = rawMarkup.indexOf("<!--", pos1);
> }
> return rawMarkup;
> }
> one can see that pos1 remains unchanged if pos2 == -1.
> The previous version in Wicket 1.4.15 doesn't have this defect. As a
> workaround set StripComments to false.
--
This message is automatically generated by JIRA.
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
