[10:34:32] <_joe_>	 I'm trying to untangle the myriad of ways one can make an http request from mediawiki
[10:34:40] <_joe_>	 some help would be appreciated
[10:37:19] <_joe_>	 as of now I've found uses of the Http class (deprecated), direct calls to MultiHttpClient, direct calls of MWHttpRequest, then via HttpRequestFactory
[10:37:29] <_joe_>	 which ones did I miss?
[10:37:42] <_joe_>	 I'm trying to reconstruct how timeouts are set in MediaWiki
[11:28:28] <kostajh>	 _joe_: includes/http I think has the complete list
[11:28:47] <_joe_>	 kostajh: yeah, that's pretty long :D
[19:44:37] <tgr>	 _joe_: there are basically two ways, MultiHttpClient (parallel curl) and MWHttpRequest (single request, has several implementations but in practice will also be curl, via the 3rd party library Guzzle)
[19:45:05] <tgr>	 Http and HttpRequestFactory are just helpers to get a MWHttpRequest object
[19:45:06] <_joe_>	 tgr: yeah I figured that everything boils down to those
[19:45:29] <_joe_>	 Http can also return a MultiHttpClient instance though IIRC
[19:47:20] <tgr>	 MwHttpRequest has three implementations, PHP stream handler, curl and Guzzle (which itself has a large number of implementations, but again in practice it will be curl)
[19:47:53] <tgr>	 basically we are in the process of moving from rolling our own request handling to Guzzle
[19:48:29] <tgr>	 not sure if MultiHttp is also planned to be moved