Antwort: Re: Antwort: Re: [Mod_gzip] "mod_gzip_send_vary=Yes" disables caching on IE? ...

[mod_gzip@lists.over.net]

Hi Kevin,


> So in Opera's case... it really IS treating it's local cache
> as an 'inline' proxy

I am not sure it actually _is_. All I can say is that
the setting of the UserAgent seems to require it more
likely than with every other browser (although there
have been Mozilla versions in the past that allowed you
to change your "UserAgent:" header at runtime as well,
which now can be done only by editing the prefs.js file).

Anyway, the caching configuration dialog of Opera is
different from all other browsers I have ever used
... different caching strategies can be configured for
HTML, images and "other" ressources ...
Don't know how this interacts with all the "Expires:"
and "Cache-Control:" headers, as there it no such set-
ting like "Automatic" in both M$IE and Mozilla here.
They only know to validate "always", "never" and in some
time interval that is user configurable (and this even
separately for these three media type classe ... looks
like the Opera user is considered to _exactly_ know what
he wants).

> Geez... then I hope to heck that the "Opera" cache has full
> support for "Vary:" AND "ETag" or something is going to get
> screwed up (locally) sometimes.

I remember Opera to send "ETag:" headers together with
some conditional GET requests (must have been of the
If-Modified-Since type) but I cannot easily reproduce
the scenario at the moment. I have not yet seen any
other browser sending "ETag:" headers.

> How does Opera behave, then, if it has received a compressed
> page for MSIE User-Agent "imitation" because Origin Server
> allows MSIE to receive compressed data but "Opera" User-Agent
> string is being excluded?

I have no scenario where I can check that at the moment.

Regards, Michael