Re: [pear-webmaster] I think your xmlrpc methods are psychic.

From: Date: Wed, 22 Dec 2004 19:45:32 +0000
Subject: Re: [pear-webmaster] I think your xmlrpc methods are psychic.
References: 1 2 3 4 5  Groups: php.pear.webmaster 
Request: Send a blank email to pear-webmaster+get-3629@lists.php.net to get a copy of this message
> It's not the cache. There is a separate cache for xml-rpc signatures, > it is located in PEAR_TMP_DIR / pear-database.php.index. Delete this > file and it will work again. Stupid piece of crap file :). Am I allowed to agree with that or will it cause political problems? ;) > There must be some kind of race condition when generating the signature > file. I will add in some locking, and that should fix it. Thanks. > FYI, I plan to implement a rest-based API for PEAR. Then we can ditch > the stupid xml-rpc signature crap, and xml-rpc altogether. HTTP caching > is much more efficient and robust than all this custom junk. Well, if you're going to take xml-rpc out please do me the honour of letting me know well in advance. This app of mine will be live tomorrow, and although I have to admit it's now armed to its back-teeth against pear.php.net's xml-rpc issues, it's not so well armed that it can psychically adjust to an entirely new API. > Implementing who's-who would be really, really easy. Yeah right. I thought that too, once :) The only part that's gone as planned is actually the chunk I spent last week writing, which relies completely on CVS logging. PEAR contributors change too quickly (ie you get too many new people/change package maintainers etc too often) for this to be a good solution in the PEAR area though. However, I wonder > if there is a secure cert for pear.php.net? This would be a good use > for donated funds, so that https://pear.php.net can be used > with HTTP > auth to do secure stuff. This will even allow remote upload of > releases, and remote validation, and so on. Hm - am I missing something here or are you talking about pearweb internals? - Steph

« previous php.pear.webmaster (#3629) next »