Saturday, 2009-03-14

*** tpb has joined #melange00:00
*** bhy has joined #melange02:17
*** haoyu has quit IRC02:24
*** serphen`` has quit IRC03:27
*** dbentley-laptop has joined #melange04:48
dbentley-laptopHi #melange!04:48
*** solydzajs has joined #melange04:53
*** ChanServ sets mode: +o solydzajs04:53
*** solydzajs has quit IRC05:29
*** solydzajs has joined #melange05:53
*** ChanServ sets mode: +o solydzajs05:53
*** solydzajs_ has joined #melange06:23
*** solydzajs has quit IRC06:23
*** solydzajs_ has quit IRC06:28
*** Alex_UA has joined #melange07:32
*** WinterMute has joined #melange07:43
*** SRabbelier has joined #melange07:50
*** ChanServ sets mode: +o SRabbelier07:50
*** dhaun has joined #melange08:02
*** Merio has joined #melange10:08
*** Alex_UA has quit IRC11:08
dandersonSRabbelier: yo!11:29
SRabbelierdanderson: heya!11:29
dandersonSRabbelier: so, are we going with an in-repos google release branch?11:30
dandersonI'm getting to the point in my release script where I'm going to unbreak the svn code11:30
dandersonbut I need to know if I still need the broken O(n) copy11:30
SRabbelierdanderson: that's Pawel's call I think11:31
SRabbelierdanderson: I'm all in favor11:31
dandersontrue.11:31
SRabbelierlh: ping?11:31
dandersonOkay, how does one invoke Pawel? :P11:31
SRabbelierlh: you have any preferrences wrt soc-google?11:31
SRabbelierit's like this11:32
* SRabbelier invokes the polish11:32
dandersonhmm...11:33
SRabbelierwait for it11:33
dandersondid you forget the runic symbols? :P11:33
SRabbelierdanderson: nono, give it a second! :P11:33
SRabbelierthese things take time11:33
* SRabbelier kicks the amplifier11:34
*** solydzajs has joined #melange11:34
SRabbelierthere!11:34
solydzajsyo11:34
* SRabbelier grins11:34
dandersonwow, impressive11:34
dandersonhi solydzajs :)11:34
* SRabbelier grins11:34
solydzajshi11:34
solydzajs;)11:34
solydzajsim on g111:34
SRabbelierInvocation: success!11:34
solydzajsso might type slower11:34
dandersonwoot11:34
dandersonokay, I have a quick question for you11:34
solydzajswhats going on ?;)11:35
dandersonI'm refactoring the release script and stuff, and I now need to know if we're going to cut google SoC releases from the main melange repos11:35
dandersoninstead of keeping the soc-google repository11:35
dandersonthat would enable *much* faster releasing, less insane history11:36
*** ChanServ sets mode: +o SRabbelier11:36
SRabbeliermhhh11:36
SRabbelierweird11:36
SRabbeliermust be my irc client11:36
solydzajsi want to cut next release still using socgoogle later today and next release on soc11:36
dandersonsold!11:36
solydzajsi need to review release document from sverre11:36
dandersonthe soc-google release script is still up, if you want to use that11:36
solydzajsand add sections about official releases11:36
dandersonand the one in the soc repos will be ready for next release :)11:37
solydzajsgoogle official11:37
SRabbelierdanderson: awesome11:37
SRabbeliersolydzajs: sounds like a plan to me11:37
solydzajsand melange community releases11:37
solydzajsbut the release today will be still with socgoogle11:37
dandersonyes, that makes sense11:38
dandersonbut I'll start working on the revised script to do releases from soc11:38
solydzajsonce the proper wiki doc is in place we will switch11:38
solydzajsok great11:38
solydzajswe should have subfolders in branches and tags11:39
dandersonbranches/gsoc and tags/gsoc ?11:39
solydzajsgoogle and community or google and melange11:39
solydzajsi was thinking11:40
SRabbeliersolydzajs: melange :)11:40
SRabbeliersolydzajs: not 'community' :P11:40
solydzajssince its not only about gsoc11:40
solydzajsok11:40
SRabbelier(melange is shorter :P)11:40
dandersonyeah, melange; "community" gives the impression the google branch is not community11:40
danderson... which is wrong :P11:40
solydzajsok ;)11:40
solydzajsjust my first thoughts11:40
solydzajs;)11:40
dandersonokay, shall I move tags/branches into melange/ now?11:41
dandersonor will that break the release process for today?11:41
solydzajsi think we should do all of that in one go11:41
solydzajsnot in steps11:41
solydzajsi mean not part today and the other day11:42
dandersonI'm not sure I understand11:42
dandersonyou mean do it just before the first soc-based release?11:42
SRabbelierdanderson: no, like, do the release today11:42
solydzajsyes11:43
SRabbelierdanderson: and then after that, do everything needed to change the procedure11:43
dandersonah, okay11:44
danderson+111:44
danderson(anyway I'm just the release monkey, but it sounds sane anyway :)11:44
SRabbelierdanderson: :D11:44
solydzajsyes correct11:44
solydzajs;)11:45
dandersonthanks solydzajs!11:45
solydzajsno problem mobie internet ftw11:45
solydzajs;)11:45
dandersonI guess you want to get back to doing sth else now, so I'll leave you alone :P11:45
dandersonand see you later, I guess11:45
solydzajsok i should be online soon11:46
solydzajsi need to do code review11:46
solydzajsim in train now ;)11:46
SRabbeliersolydzajs: ok :)11:46
dandersonheh11:46
solydzajsyep funny ;)11:46
SRabbeliersolydzajs: should the 'list org apps' also show when the app has moved into 'pre-accepted' or 'pre-rejected'?11:47
SRabbeliersolydzajs: or even better, should we put the org in the 'needs review' list even if they are in 'pre-accepted' or 'pre-rejected'?11:47
solydzajsnope11:47
solydzajsone sec11:47
solydzajsthe status of org app should not be visible for users (org admins)11:48
solydzajsonly for hosts and developers11:48
solydzajsor are you talking about list org apps developer view ?11:48
SRabbeliersolydzajs: I'm talking about org_app/list_self/google/gsoc200911:49
SRabbeliersolydzajs: which shows you three lists11:49
SRabbeliersolydzajs: look at http://socghop.appspot.com/org_app/list_self/google/gsoc200911:49
tpb<http://ln-s.net/30Bv> (at socghop.appspot.com)11:49
*** dmitrig01|afk has joined #melange11:50
*** ChanServ sets mode: +v dmitrig01|afk11:50
solydzajsoh ok got iy11:50
solydzajsit11:50
solydzajsso it should show it in needs review11:51
solydzajsif its not accepted or rejected11:51
solydzajsok ?11:51
SRabbeliersolydzajs: that's whta I said11:51
SRabbelier....11:51
SRabbelier-_-"11:51
SRabbeliersolydzajs: or even better, should we put the org in the 'needs review' list even if they are in 'pre-accepted' or 'pre-rejected'?11:51
solydzajsthis is critical too11:51
solydzajssince the org would know which one is most propably accepted11:52
solydzajsyes correct11:52
solydzajsawesome11:52
solydzajscan you fix it before release ?11:52
SRabbeliersolydzajs: all the org can tell now is whether it has been reviewed11:54
SRabbeliersolydzajs: not whether it's in pre-accepted or in pre-rejected11:54
solydzajsok so your suggestion is ?11:55
SRabbeliersolydzajs: I will put all apps that are 'pre-accepted' or 'pre-rejected' under 'needs review'11:56
solydzajsyes and the list title should say that this is the list of orgs currently under review11:58
solydzajsnot with needs review status11:58
SRabbeliersolydzajs: Overview of Organization Applications whose status is "needs review"11:58
solydzajsas it is now ok ?11:58
SRabbeliersolydzajs: I can't change that11:58
SRabbeliersolydzajs: Lennie made it use the status name11:58
*** dmitrig01|afk is now known as dmitrig0111:58
solydzajsok then lennie will fix it later11:59
solydzajsno worries11:59
SRabbeliersolydzajs: ok :)11:59
solydzajsthanks ;)11:59
solydzajsok i will try to connect internet to my mac12:01
solydzajsand finish code review12:01
solydzajsso i will be offline for few minutes ;)12:01
solydzajsim not sure if i will survive ;)12:01
solydzajsdrama ;)12:01
solydzajsheh12:03
solydzajsbbs12:03
*** solydzajs has quit IRC12:04
dandersonSRabbelier: added comments to the release doc12:12
*** solydzajs has joined #melange12:13
*** ChanServ sets mode: +o solydzajs12:13
SRabbelierdanderson: awesome!12:15
dandersonSRabbelier: actually, I'm less and less sure I agree with the versionning format12:16
dandersonit doesn't make sense in various ways12:16
dandersonusually, a release branch is for the whole major-minor cycle12:16
dandersoneg. you have a 0-5-x branch12:16
dandersonand from that you roll 0-5-0, 0-5-1, 0-5-2...12:17
SRabbelierdanderson: one sec :)12:17
dandersonthe 'pX' at the end just looks like a weird way to specify the patch release, instead of using a known format12:17
SRabbelierdanderson: not my idea ;)12:17
SRabbelierdanderson: I'm open to suggestiosn12:17
dandersonmkay, should I post something to the mailing list?12:19
dandersonwith my reasoning12:19
dandersonI'm a little afraid of turning it into a bikeshed :)12:20
SRabbelierdanderson: Well, I think the point of the pX stuff is so that you can easily see which version of a branch we're running12:21
SRabbelierdanderson: say we have branch 0-5-.... and we cut a release12:21
dandersonyeah, but we're not running a branch12:21
dandersonwe're running a tag12:21
SRabbelierdanderson: that's why the pX is only for tags right?12:21
dandersonwell, no, according to the current doc, it seems it's for branches too12:22
SRabbelierdanderson: I fixed that :D12:22
dandersonah :)12:22
SRabbelierdanderson: yesterday12:22
dandersonokay, so here's what I think about what should change in the branch names12:22
dandersonbranch names should be for a whole major-minor cycle. So, just call them branches/0-5, branches/0.6 ...12:23
dandersonno date (svn can tell you), no patch level12:23
dandersonthen, when you tag, you add a patchlevel number and the svn revision number at which you tagged the branch12:23
dandersoneg. tags/0-5-0-r249512:24
dandersonthat gives you all the information to exactly define the release12:24
danderson - which branch it came from12:24
danderson - which revision it comes from12:24
danderson - is it the first release in this major/minor cycle, or a patch release?12:24
SRabbelierdanderson: nice, sounds sane12:26
SRabbelierdanderson: what's the patch number for though?12:26
dandersonhmm12:26
dandersongood point :)12:26
danderson0-5-r2495 exactly qualifies the release12:26
dandersonhmm12:27
dandersonthinking about it some more12:27
dandersonjust make the tag name 0-5-0, 0-5-112:27
dandersonthe revision number can be embedded in app.yaml, and svn can tell you when a release was tagged12:27
dandersonso no need to add the revision number12:27
SRabbelierdanderson: I'm a fan of using revision nr as branch name instead :P12:27
*** mib_3d0j4i has joined #melange12:28
dandersonand reporting an issue against 0-5-2 is nicer than reporting a bug against 3939 or 0-5-r292312:28
*** mib_3d0j4i has left #melange12:28
*** madrazr has joined #melange12:28
dandersonSRabbelier: I don't understand12:28
dandersonthe problem with using a revision number in general is that it doesn't define a namespace12:29
SRabbelierdanderson: Ok, fair enough12:29
dandersonif I report a bug for r2943, do I mean the r2943 branch, [email protected], ... ?12:29
dandersonif I report a bug against 0-5-4, it's clear that I mean "The 0-5-4 tag, which comes from the 0-5-x branch"12:30
SRabbelierdanderson: yeah, you're right, I forgot about svn's ambiguatiy in that aspect12:30
dandersonand we can embed the branch revision number in app.yaml12:30
dandersonso that people can see on the live site "This is 0-5-4, cut from 0-5-x at r394212:30
dandersonand if we're going to change shit anyway, how about s/-/./ ?12:31
dandersonso we have "classic" open source release numbers, 1.2.5, 0.3.4...12:31
*** solydzajs has quit IRC12:33
dandersonSRabbelier: am I making any sense?12:33
*** solydzajs has joined #melange12:33
*** ChanServ sets mode: +o solydzajs12:33
SRabbelierdanderson: we can't12:33
danderson?12:33
SRabbelierdanderson: we changed to - since we can't sue dots in app.yaml version12:33
SRabbelierdanderson: (since it is used as url)12:34
dandersonum...12:34
dandersonbut dots are allowed in urls...12:34
SRabbelierdanderson: *kicks* :P12:34
SRabbelierdanderson: don't be a smart-ass12:34
SRabbelierdanderson: yes they're allowed, but AppEngine doesn't allow us to use dots in app.yaml's version field12:34
solydzajsdanderson: can you post your suggestions on melange-soc-dev ?12:34
solydzajsdanderson: also do you use gtalk ?12:34
dandersonno, seriously, I don't understand; it's just that app engine refuses it?12:34
dandersonokay :)12:34
dandersonsolydzajs: yes, I can, and yes, I have gtalk12:34
SRabbelierdanderson: oh, ok, in that case I didn't kick you :P12:34
danderson[email protected] is my jabber-id12:35
solydzajsdanderson: ok adding now :-)12:35
solydzajsnice to have mobile internet on my mbpro in train :D12:35
danderson(also [email protected], but I prefer to keep only internal contacts on there, so that I don't accidentally leak information :)12:35
SRabbeliersolydzajs: are you using tethering?12:35
SRabbelierdanderson: understandable :P12:36
dandersontetherbot probably12:36
solydzajsSRabbelier: nope , didn't have time to configure it on g1 yet12:36
dandersonah12:36
solydzajsSRabbelier: i'm using BT on other phone12:36
dandersonwell, no :P12:36
SRabbeliersolydzajs: ah, how nice :D12:36
solydzajsbut tethering support on g1 is good12:36
solydzajsiptables :-)12:36
dandersonwell, not that good12:36
dandersoninside the dalvik sandbox you don't have enough network control to do tethering12:36
dandersonso you need a devphone12:37
dandersontetherbot just emulates it with a SOCKS proxy12:37
solydzajstetherbot is lame :-)12:38
solydzajsI have iptables on my dev phone12:38
solydzajs:-)12:38
dandersonokay, I'll go rant on the list about release numbers12:38
SRabbelierdanderson: please do :)12:38
dandersonand we an decide what color to pain the bikeshed12:38
SRabbelierI'm voting green12:38
dandersonbut first, time for a shower12:39
SRabbelierheheh, ok :)12:40
*** solydzajs has quit IRC12:46
*** Lennie has joined #melange12:56
*** ChanServ sets mode: +o Lennie12:56
*** dmitrig01 is now known as dmitrig01|afk13:00
*** MatthewWilkes has joined #melange13:04
*** James--Crook has joined #melange13:38
dandersonokay, posted14:09
dandersonhowever, it is now the time for sushi14:09
dandersonso, afk :P14:09
lhenjoy danderson14:09
SRabbelierdanderson: enjoy the sushi :)14:09
Lenniedanderson: douzo meshiagare14:10
scorchehrm...sushi sounds tasty...14:10
Lennieit should translate to something like "enjoy your meal" ^_^14:10
dandersonLennie: heh, thanks, I was wondering :P14:10
SRabbelierLennie: lol14:10
dandersonscorche: sushi is tasty, but not quite as tasty as in Sushi Tomi in mountain view14:10
dandersonbut, I should be spending a couple of weeks in mountain view in april, so \o/ sushi tomi14:11
LennieSRabbelier: don't ask me how I know, coincidences happen14:11
scorchei will have to try that place next time i am out there...i do, however, have a very tasty very cheap place near me =)14:11
SRabbelierLennie: heheh, right :P14:11
dandersonlh: speaking of which, are you in CA in late april/early may?14:13
dandersonor travelling?14:13
lhdanderson: both of the above, though brief stint in canada in early may14:13
lhbsdcan14:13
lhyou going to visit? :)14:13
dandersonyup14:13
dandersonbetween 10 and 14d on the mtv campus14:14
dandersonto help some folks out with our new shiny system14:14
dandersonwell, that's the excuse14:14
dandersonI'm just going for sunshine, sushi and friends :P14:14
dandersonlh: so, when I get there I'll drop by14:15
SRabbelierdanderson: where'd you mail it to?14:15
lhdanderson: excellent. i have a new place now with plenty of room if you want to couch surf.14:15
dandersonlh: works for me, though I suspect I'm probably getting a corp appt whether I want it or not :)14:15
dandersonSRabbelier: melange-soc-dev ?14:16
lhdanderson: fair enough14:16
SRabbelierdanderson: don't see it in my inbox nor on http://groups.google.com/group/melange-soc-dev14:16
tpbTitle: melange-soc-dev | Google Groups (at groups.google.com)14:16
dandersongrmbl?14:16
dandersonsent 23 minutes ago14:16
SRabbelierweird14:17
SRabbelierdanderson: where'd you mail from?14:17
SRabbelierdanderson: is that address registered as member?14:17
dandersonhrm14:18
dandersonlooks like this email is not :(14:18
dandersonbut I should have received a bounce from google groups14:18
SRabbelierdanderson: do I need to invite you on another address?14:18
SRabbelierdanderson: your @gmail.com is a member14:18
dandersons'okay, I subscribed myself14:18
SRabbelierkkay :)14:19
SRabbelierdanderson: got it14:20
*** dmitrig01|afk is now known as dmitrig0114:21
SRabbelierdanderson: Do we need the -x suffix?14:22
dandersonnot really14:22
dandersonI just lifted that from the Subversion release policy14:22
dandersonwe fork 1.6.x14:22
dandersonand from there 1.6.0, 1.6.1, 1.6.2 ...14:22
dandersonbut we can remove it if you prefer your bikeshed short :)14:23
SRabbelierhehe :D14:24
SRabbelierdanderson: I'm ok either way :)14:24
SRabbelierdanderson: if that's the way Subversion is doing it, I'm cool with using that14:24
dandersonlike I said, I don't care either way14:24
dandersonI just defaulted to a project I know :)14:24
SRabbelierhehe :)14:25
* danderson creates canned responses to politely tell someone to RTFM14:26
lhif it's the i have another job one, just sent14:26
dandersonno, that one was a semi-legitimate question afaict14:26
dandersonI meant the other ones14:26
dandersonclearly clueless, no idea how SoC works, etc.14:26
SRabbelierdanderson: good idea14:27
dandersonthe responses I sent were still manual. Now I have a polite but firm one recorded :)14:27
lhthat's awesome babe14:27
SRabbelierdanderson: set up some filters to automagically send a canned response :P14:27
SRabbelierdanderson: that'd be nice14:27
dandersonSRabbelier: I think I'd need NLP and an expert system to do that14:27
dandersonto assign a ClueRank score to each message14:27
SRabbelierLOL!14:28
dandersonand send an RTFM to posts with low ClueRank14:28
SRabbelier:D:D14:28
* SRabbelier rolls14:28
SRabbelierdanderson: sounds like another 20% project right there :p14:28
dandersonit's an old idea14:28
dandersonsome of us mused on the development of a Whinerank algorithm to filter some of our internal mailing lists14:29
dandersonso that you could auto-filter people who don't whine more than you can tolerate14:29
SRabbeliers/don't (whine more)/\1/14:29
dandersonum, yeah14:30
dandersonanyway14:30
dandersonI was saying something about sushi14:30
dandersonand that was long ago14:30
dandersonTo the sushi bar!14:30
SRabbelierdanderson: off youg o!14:31
SRabbelier**you go14:31
*** madrazr has left #melange14:40
*** Lennie is now known as Lennie|Gone14:45
*** Lennie|Gone is now known as Lennie14:55
dbentley-laptopHello all.15:03
*** dmitrig01 is now known as dmitrig01|afk15:03
SRabbelierdbentley-laptop: heya :)15:03
dbentley-laptopI just sent out a code review.15:04
SRabbelierdbentley-laptop: yup, replying :)15:04
dbentley-laptopSweet; thanks.15:06
SRabbelierdone15:06
lhdbentley-laptop: pawel will also take a look in about 2 hrs, he's on a train atmn15:07
lhdbentley-laptop: and good day to you. :)15:07
dbentley-laptopGreat.15:07
*** dmitrig01|afk is now known as dmitrig0115:08
*** Merio has quit IRC15:09
dbentley-laptopAnyone know of the top of their head (save me a google) :How do I test if two strings are equal in bash?15:09
dbentley-laptopOh, =.  That makes sense.15:11
SRabbelierdbentley-laptop: 'test foo = bar'15:11
SRabbelierright :)15:11
dbentley-laptopand [ foo ] is a synonym for test foo, right?15:12
*** Lennie is now known as Lennie|Gone15:12
SRabbelierdbentley-laptop: mhhh, I think so yes15:16
SRabbelierdbentley-laptop: if [foo = bar]15:16
*** MatthewWilkes has quit IRC15:18
*** dmitrig01 is now known as dmitrig01|afk15:27
dandersondbentley-laptop: if you're trying to follow the bash style guide, use the [[ ]] builtin15:50
danderson(the google bash style guide that is, melange doesn't have one)15:51
*** dmitrig01|afk is now known as dmitrig0116:06
*** James--Crook has left #melange16:06
*** dmitrig01 is now known as dmitrig01|afk16:20
*** dmitrig01|afk is now known as dmitrig0116:38
*** Lennie|Gone is now known as Lennie17:07
*** solydzajs has joined #melange17:10
*** ChanServ sets mode: +o solydzajs17:10
*** MatthewWilkes has joined #melange17:22
*** MatthewWilkes has quit IRC17:26
*** dhaun has quit IRC19:26
*** serphen`` has joined #melange19:47
serphen``hey19:47
dandersonhi19:50
dandersonsolydzajs: you said a release is happening today still?19:50
dandersonwell, today... before the next sleep period :P19:50
solydzajsdanderson: I'm still doing code review19:50
dandersonok19:51
solydzajsdanderson: if I won't make it today it's gonna happen in the morning :(19:51
solydzajsdanderson: I will test your script :-)19:51
dandersonif you want a googler to do the google pushing, I'm happy to help19:51
dandersoncool! :)19:51
solydzajsdanderson: I want to test your script at least once :-) since we are moving to /p/soc repo :-)19:51
dandersonyeah, but the /p/soc repo is going to be pretty different19:52
dandersonthe current script only supports soc-google stuff19:52
solydzajsdanderson: yep we will fix that :-)19:53
solydzajsdanderson: for now we are still doing releases on soc-google19:53
Lennieso no more p/soc-google?19:53
dandersonLennie: soon19:53
dandersonbut not yet19:53
Lenniedo not forget we need to have a place for soc-google specific issues :)19:53
dandersononce we have a release policy and a release script I guess19:54
Lennielike the login screen :p19:54
dandersonLennie: meh, unlikely to be any problems19:54
Lenniek fine, just saying :D19:54
dandersonwe'll be running a 100% vanilla melange afaik19:54
LennieI'm off19:54
Lennienn19:54
dandersonnight19:54
solydzajsgnighty19:54
Lennie(yeah but people will post OMG THERE IS A TYPO IN THE MENTOR AGREEMENT)19:55
Lenniein our issue tracker19:55
Lenniejust my 2 cents :D19:55
Lennieanyhow19:55
Lennielater19:55
dandersonI don't think we can expect users to do triage between melange engine problems and SoC problems :/19:55
dandersonit'd be nice, but to most of them, it's just one site19:55
solydzajsyep19:55
solydzajsand we still have google summer of code project on code.google.com19:56
solydzajswith issue tracker19:56
dandersonhm?19:56
*** Lennie has quit IRC19:56
solydzajshttp://code.google.com/p/google-summer-of-code/19:57
tpbTitle: google-summer-of-code - Google Code (at code.google.com)19:57
dandersonooh19:58
dandersonwell, yeah, but isn't it the same problem?19:58
dandersonWe can't expect users to do triage for us19:58
solydzajsyep19:58
solydzajsbut right now we have 3 issue trackers19:58
solydzajssoc soc-google and google-summer-of-code19:58
solydzajsnot good :-)19:58
dandersonindeed19:59
dandersonso, let's kill soc-google to start with :)19:59
solydzajsyep soon that's what will happen :-)20:01
solydzajsdanderson: ping20:18
solydzajsdanderson: would you like to mentor a student for Melange ?20:19
dandersonsolydzajs: I don't think I'll have the time to be a good mentor20:19
dandersonbut if you need mentors, I'm willing to give it a try20:19
solydzajsdanderson: ok awesome :-) we will see if we get in and how many proposal we get but just in case I will put you on the list20:20
danderson"if we get in"20:20
dandersonseriously now :P20:20
*** solydzajs has quit IRC20:26
SRabbelierdanderson: You never now! :P20:28
SRabbelierdanderson: maybe Leslie has a bad hair day, gets grumpy, and denies all orgs! :P20:28
dandersonSRabbelier: omg those tests scream for Selenium20:35
SRabbelierdanderson: no, they scream for automated test suite :P20:35
SRabbelierdanderson: we already have a stub-out for Django :)20:35
dandersonthat's what I said20:35
dandersonSelenium is an in-browser testing framework20:36
SRabbelierdanderson: I don't like having to start my browser to run the tests20:36
SRabbelieralso, isn't it hard to have a buildbot run those tests?20:36
dandersonwhy?20:36
dandersonwell, first, you'd plug in Selenium Grid as I understand it20:36
SRabbelierbecause a buildbot probably doesn't have X11 running?20:36
dandersonwell, it's the only sane way to test actual UI20:37
dandersonwhich matters a *lot* in a website :)20:37
dandersonhmmm, otherwise20:37
dandersonthere was another framework I think20:37
dandersonthat fired up a gecko renderer in memory20:38
dandersonheadless20:38
SRabbelierdanderson: This is true, but by stubbing out Django we don't need to look at HTML stuff20:38
SRabbelierdanderson: we can inspect just the python context object20:38
dandersonI don't understand what you mean by "stubbing out django"20:38
SRabbelierthat's what I figured20:38
dandersonif you match on the returned HTML, expect your tests to be uber brittle20:38
SRabbeliernono20:38
SRabbelierdanderson: we wont' do that20:38
dandersonand if you don't, then you can't detect UI breakage :)20:38
SRabbelierdanderson: We leave Django to render the HTML for us20:38
SRabbelierdanderson: we pass it a context object, and the path to a template20:39
danderson(note that I'm not saying selenium should be the only testing done, just that it tests an important aspect of the code that cannot be tested otherwise)20:39
SRabbelierdanderson: by stubbing out Django, we can test that the returned context is good etc20:39
SRabbelierdanderson: ok, then I agree with you :)20:39
SRabbelierdanderson: we should at least have some tests that verify that basic stuff like, submitting a form etc. works20:39
dandersonalso, I can solve the X11 issue20:39
dandersonI'll run a buildbot headless virtualbox20:40
SRabbelierdanderson: but Selenium can't really test the state of objects, etc.20:40
dandersonit'll happily have X11 running somewhere in RAM where I can't see it :)20:40
SRabbelierdanderson: lol, nice :)20:40
dandersonSRabbelier: as I said, not the only tests.20:40
dandersonbut what good is writing the right state if you forget to display the result?20:40
dandersonor if it displays wrong in IE? :P20:40
SRabbelierdanderson: *grins*20:40
SRabbelierdanderson: I think we should display as ugly as possible in IE20:40
SRabbelierdanderson: preferribly use <MARQUEE> a lot20:41
SRabbelierdanderson: I'm thinking #FF0080 as background colotr20:41
SRabbelierdanderson: added note on Selenium :)20:45
dandersonSRabbelier: I just mention selenium because I want to try it out20:47
dandersonhmm...20:47
SRabbelierdanderson: Same here ;)20:47
SRabbelierAnyway, off to bed I go :)21:02
SRabbelierg'night all21:02
*** SRabbelier has quit IRC21:07
*** dmitrig01 is now known as dmitrig01|afk21:07
*** dmitrig01|afk has quit IRC21:09
*** dmitrig01|afk has joined #melange21:09
*** ChanServ sets mode: +v dmitrig01|afk21:09
*** dmitrig01|afk is now known as dmitrg0121:39

Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!