false warning: end of document without finding body

[GoogleCodeExporter]

This occurs on many web pages.  The cause appears to be various resources -- 
images and javascript is what I've seen -- that are served by mod_pagespeed, 
but which come into mod_pagespeed in the filter chain with a content_type of 
text/html.  These are likely being fixed up by mod_headers, but mod_pagespeed 
is trying to process them as if they were HTML.

One candidate solution has been found where the first bytes of any HTML 
streaming through mod_pagespeed are scanned and, if not looking like HTML, then 
mod_pagespeed just passes them through directly without processing.  This 
eliminates those messages.

Another possible solution is to put mod_pagespeed *after* mod_headers which 
might be a better way of achieving the same objective.  We'd need to check that 
we left the headers in the correct state (content-length in particular comes to 
mind).  But I'm not sure how that would play with mod_deflate, which we 
definitely want to run after mod_pagespeed.

Original issue reported on code.google.com by [email protected] on 12 Nov 2010 at 1:47