On Oct 15, 2012, at 12:46 PM, Victor Dubiniuk <[email protected]> wrote:

> Hi,
> 
> In case of the webdav access session is not used hardly so the single 
> session_start won't have a great influence on the performance.
> From the other hand the webdav clients that are not pass the session id while 
> syncing. So one will get a new session on each request.
> It might cause issues like this one 
> http://bugs.owncloud.org/thebuggenie/owncloud/issues/oc-1459

I would stongly advice using sessions for this very reason.

Sessions suck for a list of reasons. You're in a unique position when creating 
Web API's and WebDAV servers to avoid them completely; and this would be a very 
wise thing.

If you rely on sessions for some reason to make things faster, it's likely that 
your business logic is sub-optimal. Caching should really be handled on a 
different layer; functions should ideally not have a response that varies based 
on the contents of HTTP header, with the exception of the functions that are 
specifically built to handle this very thing.

So push your caching to a different layer, this will also benefit you when you 
need to pull in user-specific data for people who are not the currently 
logged-in user :)

Evert
_______________________________________________
Owncloud mailing list
[email protected]
https://mail.kde.org/mailman/listinfo/owncloud

Reply via email to