script tags missing from rewritten HTML output after inlined css

[GoogleCodeExporter]

What steps will reproduce the problem?
1. Load page.  Watch it work because all the javascript files are loading.
2. Turn on mod_pagespeed
3. Load the page.  See it break because not all the javascript files are linked 
from the HTML any more.

What is the expected output? What do you see instead?
I expect all the linked javascript to remain in the modified output.  The HTML 
rewriter is stripping out two <script src=""/> tags from the output.


What version of the product are you using (please check X-Mod-Pagespeed
header)?
X-Mod-Pagespeed: 0.9.16.9-576


On what operating system?
Ubuntu 10.04 LTS

Which version of Apache?
Server: Apache/2.2.14 (Ubuntu)

Which MPM?
Milwaukee Public Museum?

Please provide any additional information below, especially a URL or an
HTML file that exhibits the problem.

I'm attaching the relevant parts of the original HTML before mod_pagespeed 
runs, and the modified output, busted.html.  I know it's something of a mess, 
but this is real life building sites on a deadline.
The last two javascript files, dragface.js and make_postcard.js are completely 
missing from the mod_pagespeed version.  

I'm guessing it has something to do with the inlined CSS before it.  If I 
re-order the CSS that gets inlined to higher up in the file, the missing 
javascript files re-appear.

Original issue reported on code.google.com by leo.dirac@gmail.com on 30 Mar 2011 at 11:30

• Merged into: combine_css on XHTML file with unbalanced tags can strip elements #252

Attachments:

• original.html
• busted.html

[GoogleCodeExporter]

The problem is caused by

a) you claim that this is XHTML in the doctype at the top and
b) you missed a "/" line in the 'link' tag on line 52.

However, you are correct that this is a mod_pagespeed bug, although as it 
happens we've already found & fixed it.  It hasn't made it to a binary release 
yet. 

An XHTML validation site would probably also have found your missing "/".

Original comment by jmara...@google.com on 31 Mar 2011 at 1:20

[GoogleCodeExporter]

Just to be clear, mod_pagespeed should (a) be more conservative the face of 
broken markup and (b) use loose HTML semantics when parsing documents that 
claim to be XHTML.

Original comment by jmara...@google.com on 31 Mar 2011 at 1:23

• Changed state: Duplicate