Re: Re: [PEPr] Comment on HTTP::HTTP_SessionServer

From: Date: Sun, 17 Oct 2004 14:18:20 +0000
Subject: Re: Re: [PEPr] Comment on HTTP::HTTP_SessionServer
References: 1 2 3  Groups: php.pear.dev 
Request: Send a blank email to pear-dev+get-33906@lists.php.net to get a copy of this message
Hi, Olivier Guilyardi wrote:
My point is : wouldn't it be more efficient to support an already existing protocol (as msession) so that one could make use of all of the existing API, wether they be C, Java, etc... ?
I could not find anything specific on the msession protocol, so I implemented my own protocol with all commands that I needed. And I did not like that you have to pass the session id with every get and set. There are a couple of other things I did not like about msession, like the lock/unlock, I prefer open in different modes and commit like I implemented it in HTTP_SessionServer. Also the plugin architecture is quite strange IMHO, as this is some kind of web service embedded in a session... But that's just my opinion, others may like it...
What do you think about this ? Have you been thinking of such HTTP-encapsulation ? I would not recommend using this, although something like this can be implemented in less than 100 lines of code. XML-RPC and HTTP will cause a lot of overhead, this will probably slow your site down.
And there's no need for the SessionServer, a simple web service with file_get_contents(), serialize/unserialize will enable you to do this. Of course you could integrate this with the storage container used by HTTP_SessionServer. The best solution probably would be buying one not-so-cheap server, where you may start your own daemon. The cheap servers may still access the SessionServer using fsockopen(). Stephan

« previous php.pear.dev (#33906) next »