=== salgado is now known as salgado-afk === Odd_Blok1 is now known as Odd_Bloke === mwhudson_ is now known as mwhudson === salgado-afk is now known as salgado === mrevell is now known as mrevell-lunch === mrevell-lunch is now known as mrevell === EdwinGrub is now known as EdwinGrubbs [15:00] #startmeeting [15:00] Meeting started at 09:01. The chair is barry. [15:00] Commands Available: [TOPIC], [IDEA], [ACTION], [AGREED], [LINK], [VOTE] [15:00] yay! [15:00] hi everyone and welcome to this week's ameu reviewer's meeting. who's here today? [15:00] me [15:00] me [15:00] me [15:00] me === sinzu1 is now known as sinzui [15:00] me [15:00] me [15:01] me [15:01] me [15:02] me [15:02] me [15:03] [TOPIC] agenda [15:03] New Topic: agenda [15:03] * Roll call [15:03] * Next meeting [15:03] * Action items [15:03] * Queue status [15:03] * Mentoring update [15:03] * cprov mentor, abentley mentor [15:03] * Review process [15:03] * Ensure that all outgoing http connections go through a proxy (intellectronica) [15:03] [TOPIC] next meeting [15:03] New Topic: next meeting [15:03] += week(1) ? [15:03] me [15:04] same time and place next week [15:05] anybody know they will not be there? [15:05] [TOPIC] action items [15:05] New Topic: action items [15:05] * intellectronica to file bug on lint issue regarding elementtree import [15:05] sorry, didn't yet [15:06] nor did i add any checlist items for js [15:06] intellectronica: cool, we'll carry it over [15:06] intellectronica: k [15:06] * gmb to update PythonStyleGuide for long argument lists (use The Salgado Way) [15:06] is gmb here today? [15:06] That was done [15:06] no, he's on leave [15:06] * sinzui should have removed it from the agenda [15:07] intellectronica: k, thanks. sinzui, thanks, np [15:07] * barry to ask lifeless to summarize what he knows about the PQM Mysteries (e.g. autopacking bug losing branches) [15:07] not done [15:07] [TOPIC] * Queue status [15:07] New Topic: * Queue status [15:08] i don't really have anything on this today. any comments from y'all on either the pqm queue or PR queue? [15:08] guess not :) [15:09] [TOPIC] * Mentoring update [15:09] New Topic: * Mentoring update [15:09] so, abentley definitely wants to start up soon as a reviewer, and i'm all for that. he needs a mentor. any volunteers? [15:10] okay, i'll do it [15:11] barry: well, so do I. [15:11] also, and i haven't talked to cprov yet about this, but your name came up too. [15:11] cprov: right, we need a mentor for cprov. can someone in his timezone volunteer? [15:11] maybe allenap and i can co-mentor? it's a bit easier, since we share the shift [15:12] I'm happy with that. [15:12] we're a bit earlier than him, so if there's an american that can do it maybe it's better [15:12] intellectronica: allenap: I'd more than happy to follow you both. [15:12] intellectronica, you're a bit later than him, in fact [15:12] of course, i forgot that :) [15:12] the Americans are harder? [15:13] salgado: I'll be back to BR in August. [15:13] is'nt cprov in +0200 these days [15:13] but only for 2 more weeks! [15:13] cprov: i'm sure you'll graduate by then ;) [15:13] :) [15:14] intellectronica: ehe, couldn't start any better ;) [15:14] intellectronica, allenap i'm okay with that if you and cprov are. [15:14] +1 [15:14] +1 [15:14] great, thanks! [15:14] +1 [15:14] two mentors, is this a first? :) [15:14] bigjools: it is! [15:14] two senators [15:14] bigjools: that just means intellectronica and allenap have to be doubly hard on cprov :) [15:15] so he should graduate in double quick time! [15:15] bigjools: or take twice as long [15:15] barry: why ? [15:15] cprov: i'm kidding [15:15] [TOPIC] * Review process [15:15] New Topic: * Review process [15:15] please don't leave him in a gibbering mess, I need him to fix Soyuz bugs :) [15:15] :-D [15:16] barry: uhm, I just thought you would come with a funny reason, nevermind ... [15:16] cprov: i'm limited to one joke per meeting, if i'm lucky [15:16] * Ensure that all outgoing http connections go through a proxy (intellectronica) [15:16] intellectronica: you have the floor [15:16] so, on production, we can only make http connections to the outside world via a proxy [15:17] but time and again we get code in that doesn'y handle proxies well [15:17] since we can't really test this, we must pay attention in reviews when we notice code that makes connections to the outside world and make sure that it works via proxies [15:18] what sort of problems arise? [15:18] ehm ... we can't connent? [15:18] bigjools: timeouts [15:18] ok - terminal problems then :) [15:18] intellectronica: do you remember any example ? [15:18] The most recent problem was using xmlrpclib. [15:19] It doesn't cope with proxies. [15:19] really? [15:19] cprov: i'm currently working on one [15:19] And just hits its shiny head on the inside of the firewall. [15:19] flacoste: yeah, xmlrpclib sux [15:19] intellectronica: i wonder if a better xmlrpclib could be built ontop of httplib2? [15:20] barry: there's a python bug, and a patch to fix it [15:20] so hopefully not before long [15:20] ok, it seems you need a custom TRansport for this [15:20] intellectronica: can you point us to an example of what is the right way to do it? [15:20] flacoste: yeah, that's exactly my fix [15:20] but i think this points to another problem [15:21] we are lacking tests for things that connects to the outside [15:21] "in end-to-end" kind of way [15:21] bac: there's no right way to do it. we simply have to scrutinise any new code that makes http connections to the outside world and ask questions [15:21] intellectronica: can you forward to me the bug #? [15:21] can we use pylint to find out callsites using xmlrpclib with bare HTTPTransport ? [15:21] flacoste: indeed. but it's quite difficult to test, since we need to block normal connections [15:21] and by "outside" you mean outside the data center. this doesn't affect connections to niobium [15:21] intellectronica: i might be able to at least convince the python RM to sneak it into py2.6 [15:21] well, it's not a xmlrpclib problem pe-se [15:22] right [15:22] cprov: i wouldn't worry too much about rpclib specifically. this is a general problem [15:22] barry: see https://bugs.edge.launchpad.net/malone/+bug/243634 [15:22] or in fact, monkey patch HTTPTransport in a way it couldn't be used. [15:22] Launchpad bug 243634 in malone "ExternalBugTrackers that use XML-RPC need an XML-RPC transport that can work across proxies" [Critical,In progress] [15:22] intellectronica: yes, it might be too much trouble. [15:23] intellectronica: thanks [15:23] and since we don't really open connections to the outside world in the test suite, i don't think there's any automated way to do that. we simply have to pay attention to that [15:23] welcome [15:24] intellectronica: is there some way we can catch and elevate these bugs when testing on staging? [15:24] barry: i don't understand your question [15:24] well [15:24] on staging, things won't work [15:25] we could have a proxy that has to be used wihtin the test suite [15:25] to show that connections to the "outside" world are proxied [15:25] flacoste: and what if we don't use it? how would the test fail? [15:25] well, the test has to show that it goes through a proxy [15:25] flacoste: yes [15:25] that's a nice idea, actually [15:25] but i guess this is just icing over your initial comment [15:26] which is reviewers: pay attention to the proxy issue [15:26] if we had a standard way to show that it was handled, it would be easier [15:26] but reviewers would still have to spot a missing test [15:28] yeah, i don't think there's much more to say about this issue other than that [15:28] intellectronica: please send a message to the mailing list on this issue and please add a note to TIpsForReviewers [15:28] barry: ok [15:28] [ACTION] intellectronica to communicate on ml and wiki about watching for outside connections [15:28] ACTION received: intellectronica to communicate on ml and wiki about watching for outside connections [15:28] thanks [15:29] that's everything on the agenda. does anybody have anything /not/ on the agenda? [15:30] that sounds like a no. okay everybody, thanks! we're done 15 minutes early today [15:30] #endmeeting [15:30] Meeting finished at 09:31. [15:30] thanks barry [15:31] thanks barry === mwhudson_ is now known as mwhudson === salgado is now known as salgado-afk