etag should not be issued for mismatching content.

[GoogleCodeExporter]

When PageSpeed responds to a request for a .pagespeed. URL, and the optimized 
resource is cached, it sends
   Etag: W/"0"
This makes sense because the URL already contains the content hash.  However, 
when the optimized resource is not cached, PageSpeed may send back unoptimized 
content instead to avoid blocking clients waiting for the resource.  In that 
case it uses Cache-Control:max-age=300,private, so that proxies & CDNs don't 
cache the unoptimized response and serve it to other clients.  However, we 
still send:
   Etag: W/"0"
This does not seem appropriate.  The language in 
http://tools.ietf.org/html/draft-ietf-httpbis-p6-cache-25 indicates that it 
won't make the resource public:
   The "private" response directive indicates that the response message
   is intended for a single user and MUST NOT be stored by a shared
   cache. 
But it still doesn't make sense to provide a trivial validation for what is not 
the fully optimized resource content.

Original issue reported on code.google.com by jmara...@google.com on 27 Nov 2013 at 4:22