517842 - A javascript string containing </script> is misinterpreted when using XHTML 1.1 doctype

Status: RESOLVED INVALID

(Firefox :: General, defect)

Description

[Johannes Stickel]

User-Agent:       Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.3a1pre) Gecko/20090920 Minefield/3.7a1pre
Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.3a1pre) Gecko/20090920 Minefield/3.7a1pre

When a HTML page with XHTML 1.1 doctype contains a javascript string like
'<div><script type="data">{}</script></div>', FF will treat the </script> in the string as end of the script tag. It does also happen when the javascript code is embedded as html comment. However the bug disappears if the doctype line is removed from the page.

Reproducible: Always

Steps to Reproduce:
1. create a minimal HTML document with XHTML 1.1 doctype
2. add a script tag that contains a string with </script> inside
3. open the document in FF

(example page provided in URL above)
Actual Results:  
 * Javascript error: 'unterminated string literal'
 * FF shows '; -->  

Expected Results:  
 * No Javascript error
 * Empty page

Comment 1

[Johannes Stickel]

Attachment: Example page that causes the bug

Comment 2

[David Vo [:auscompgeek]]

I think I see an error. You are meant to use <[CDATA[ ... ]]> instead of <!-- ... --> in XHTML.

Comment 3

[Mardeg]

Attachment: Escaped end tag in javascript string

By serving the document as text/html the xml parser is not used, so the html standards that describe what to do at http://www.w3.org/TR/REC-html40/appendix/notes.html#h-B.3.2.1 apply here.

Comment 4

[Johannes Stickel]

@auscompgeek: Using  <[CDATA[ ... ]]> doesn't change the outcome.

@Mardeg: Escaping the '</script>' as '<\/script>' does work. However when Javascript is inside a html comment or CDATA section, i expect it to be ignored by the parser. Am i wrong here?

Comment 5

[David Vo [:auscompgeek]]

Johannes, follow the link above.

Comment 6

[David Vo [:auscompgeek]]

Oh, and also, the file isn't actually being served as application/xhtml+xml so the XML parser doesn't kick in.

Comment 7

[Mardeg]

Attachment: Unescaped javascript string in a real XHTML document

auscompgeek is correct that CDATA is how to comment when serving the document as application/xhtml+xml then there is no error because it's using the XML parser.
But then if you want to serve the document differently depending on the browser (as IE still doesn't reconise application/xhtml+xml) you run into the problems outlined at http://www.hixie.ch/advocacy/xhtml
I resolved this INVALID based on your attachment being text/html but I can change it to WORKSFORME based on this latest attachment if you like.

Comment 8

[Johannes Stickel]

No, that's fine. Thx a lot for your work and excuse my ignorance.
Have a nice day.

Comment 9

[David Vo [:auscompgeek]]

Just another note, in IE, the document still works. IE just looks at the DOCTYPE and sees that the XML parser should be used. But, Firefox only uses the XML parser if the MIME type has XML in it. So, the page can be served as text/html to IE while it can be served as application/xhtml+xml to other browsers.