[00:00:32] I'm ok for now, actually. [00:00:47] but thanks for the offer. [00:01:12] *AaronSchulz wants to actually *have* python-fu...time for more book-learnin ;) [00:01:16] I'll keep you in mind; it's good to know who the python people are. [00:40:50] *jeremyb is back with his python sign on forehead :) [00:51:33] AaronSchulz: upload-style urls are working on swift for any that match /wikipedia/commons/thumb/ [00:51:38] \o/ [00:52:15] jeremyb: it's not a big change, but it is python... want to review https://www.mediawiki.org/wiki/Special:Code/MediaWiki/105499 for me? [00:52:40] maplebed: can a non-commiter do so? or just as an extra review? [00:52:48] I have no idea. [00:52:52] I don't interact with svn much... [00:52:55] :P [00:53:09] You can comment/sign off [00:53:17] need coder right to mark ok/fixme etc [00:53:18] i mean, i know i can comment (or at least i'm under the impression i can) [00:53:29] it's really not a very interesting diff though. [00:53:45] anyway, looking [00:55:28] maplebed: so, not part of a changed line but why do this: url = req.url[:] ? [00:55:45] no idea. [00:55:54] I don't really know what it's doing... [00:56:04] I'm not even sure what kind of object req is. [00:56:24] probably request? :) [00:56:42] what I do know is that url was turning into http://127.0.0.1/path, which didn't work (because the thing listening is running on port 8080) [00:57:00] I was hoping it had a 'port' attribute, but no dice. [00:57:12] ahh [00:57:13] (req.port = self.bind_port would have been much nicer) [00:58:35] *AaronSchulz either gets Regexp failed or a blank page [00:58:37] when I put in a self.logger.warn("req is %s" % req"), I got req is GET /v1/AUTH_a6eb7b54-dafc-4311-84a2-9ebf12a7d881/wikipedia-commons-thumb/3/3f/Karl_Mack_von_Leiberich.jpg/8px-Karl_Mack_von_Leiberich.jpg HTTP/1.0#015#012Accept: */*#015#012Content-Type: None#015#012Host: 127.0.0.1:8080#015#012User-Agent: plain#015#012X-Cf-Trans-Id: txabfbcb36-d2ba-41d2-85b2-34a7244294c7 [00:58:48] AaronSchulz: what's the URL you're trying? [01:00:02] should I be hitting http://copper.wikimedia.org:8080 ? [01:00:11] yeah. [01:00:21] like http://copper.wikimedia.org:8080/wikipedia/commons/thumb/0/0c/McDonnell_XF-85_Goblin_USAF.jpg/100px-McDonnell_XF-85_Goblin_USAF.jpg ? [01:00:33] *maplebed tries that [01:01:01] http://upload.wikimedia.org/wikipedia/commons/thumb/0/0c/McDonnell_XF-85_Goblin_USAF.jpg/100px-McDonnell_XF-85_Goblin_USAF.jpg [01:02:32] weird. changing the size makes it work for me. but 100px doesn't. [01:02:57] eg http://copper.wikimedia.org:8080/wikipedia/commons/thumb/0/0c/McDonnell_XF-85_Goblin_USAF.jpg/120px-McDonnell_XF-85_Goblin_USAF.jpg [01:04:34] wfm [01:04:48] I just deleted it and re-requested [01:04:50] it worked that time. [01:05:03] I had just rebooted copper; maybe you hit it while it was still partially booted or something. [01:06:37] maplebed: why aren't you using python-requests.org ? [01:07:40] I didn't write most of what's there; I'm just trying to make it work. [01:07:48] there might be some reason, but I don't know it. [01:07:54] maplebed: also, is this file part of some swift middleware? [01:07:59] yes, it is. [01:08:24] ok, http://localhost/MW_trunk/phase3/wmf_swift_test.php is working via cURL on my wiki [01:08:27] *jeremyb resumes reading :) [01:08:38] AaronSchulz: woohoo!!!! [01:08:41] that's fantastic. [01:08:45] I think it's time for pie. [01:08:46] brb [01:12:45] Uncaught exception 'AuthenticationException' with message 'Invalid username or access key.' [01:12:56] maplebed: now I can't do the other tests ;) [01:15:20] eh? [01:15:40] AaronSchulz: can you point me to the other test? [01:15:46] or do you need the auth tokens? [01:15:57] did you change any auth stuff? [01:16:15] I don't remember what I told you so probably. pming you the password. [01:17:12] (not that they're actually secret; they're in wikitech already...) [01:17:44] Uncaught exception 'InvalidResponseException' with message 'Invalid response: (curl error: 7) couldn't connect to host: Failed to obtain valid HTTP response [01:19:30] would you pastebin the entire exception for me? [01:20:16] http://pastebin.com/0a3hzvA6 [01:21:23] AaronSchulz: are you sure it's using port 8080 the entire time? [01:21:33] and not throwing in port 80 somewhere? [01:22:01] all 8080 afaik [01:22:02] would you try again? [01:22:12] i just threw up tcpdump on port 80 [01:22:17] so I'll see any attempts. [01:22:37] nope nothin. [01:22:48] could you tcpdump 8080 too? [01:22:55] so i can see the actual reqs? :) [01:23:30] sure. [01:24:16] to avoid too much chaff though, let's coordinate. i'll say as soon as it's running and you say as soon as you've made a request (or a few) then I'll kill it. [01:25:13] ready? [01:25:26] sure [01:25:39] set? [01:25:46] go [01:26:40] *AaronSchulz hopes something showed [01:26:50] *maplebed is still waiting for the 'I'm done"... [01:27:05] *AaronSchulz is done [01:27:12] jeremyb? [01:27:23] erm? [01:27:36] I thought you were going to be throwing a few requests in there too. [01:27:41] oh, no [01:27:42] ok, it's stopped. [01:27:47] idk if that's even a public box [01:27:53] and i don't have a swift client handy [01:28:05] you didn't click on aaron's URLs earlier? [01:28:11] http://copper.wikimedia.org:8080/wikipedia/commons/thumb/0/0c/McDonnell_XF-85_Goblin_USAF.jpg/100px-McDonnell_XF-85_Goblin_USAF.jpg [01:29:10] ok, one sec while I make sure the cap is clean. [01:39:30] AaronSchulz, jeremyb: http://copper.wikimedia.org:8080/wikipedia/commons/thumb/tmp/swift-noauth.cap is getting a file like the one pasted above. http://copper.wikimedia.org:8080/wikipedia/commons/thumb/tmp/swift.cap is what AaronSchulz just tried. [01:40:02] and yes, I know they have the password in plaintext in there. it's ok. [01:40:18] (/me is splitting time with the otrs channel atm btw) [01:41:11] AaronSchulz: it looks like whatever you're running is successfully authing but then not doing anything with the token it gets. [01:45:30] so, one of the dumps is just hitting /wikipedia/commons/thumb... and one is just hitting /auth/v1.0 [01:46:58] the thumb one is succesful; I'm not sure what's up with the other yet. [01:50:22] they are both 200... [01:50:31] yeah, the auth attempt is successful. [01:50:41] AaronSchulz was saying that something after that timed out [01:51:02] I didn't see that on either port 8080 or 80, so I'm not sure what it is. [01:51:03] *AaronSchulz wonders why get_container fails now [01:51:45] I wonder if this is it: [01:52:01] X-Storage-Url: http://127.0.0.1:8080/v1/AUTH_... etc. [01:52:07] lemme change something. [01:53:17] (there are two spots that specify 127.0.0.1 in the config; I'm going to see which one is populating that and change it to copper.) [01:54:14] maplebed: is it listening on cooper? [01:54:28] yeah, it's listening on 0.0.0.0 [01:55:11] maplebed: well "something after that" isn't in the dump then [01:55:32] yup. I think this is probably it. [01:56:15] are php and python the same box? [01:56:23] it fails about instantly, wherever it happens [01:56:41] instant doesn't sound right for timeout [01:56:53] jeremyb: the only PHP is on my own box [01:56:59] no - aaron's tests are off-host. all of mine so far have been on localhost. [02:00:51] AaronSchulz: try again. [02:03:17] *AaronSchulz makes a commit to something else [02:03:19] ok, back [02:03:30] hooray! [02:03:34] looks like that worked. [02:03:39] *maplebed is watching logs [02:04:07] http://pastebin.com/SNCstphv [02:04:45] yup, that's it! [02:04:55] ok. I think I'm gonna end on a good note. [02:05:04] wtf are the last three files? :) [02:05:18] the placeholder I just threw in there cuz I wanted something to play with, [02:05:29] ok, as long as there is an explanation, heh [02:05:31] and the two .cap files are the tcpdumps I just got for you. (and jeremy-b) [02:05:44] I guess you didn't look at them. [02:05:44] :P [02:05:48] ahh, write, I didn't expect them to be *there* [02:05:50] heh [02:06:00] I looked at one cap file [02:06:02] right* :P [02:06:05] on pastebin [02:06:22] heh... "Hmm, I need a web-accessible filesystem to store these cap files so folks can download them. Oh, I know, I'll use Swift!" [02:06:23] :P [02:06:32] hehe [02:06:55] ok, nice...I'm glad to see the etag being set [02:08:05] *AaronSchulz should do some on-cluster perf testing [02:09:55] AaronSchulz: here are the rest of the details about two random files, just to see what swift's storing: http://pastebin.com/6LTs7KtF [02:11:10] ok, I'm heading out for real. [02:11:13] see you tomorrow. [02:11:14] same here [02:11:20] AaronSchulz: are you going to be in the office tomorrow? [02:11:36] all week [13:46:08] bugzilla is DOWN [13:46:16] perhaps here someone is online :D [13:58:12] hello, anybody there?? [13:59:06] i hav a suggestion to make for wikipedia.....anybody there?? [14:00:23] hey santhosh and multichill.... [16:06:05] yuvipanda: hashar: Inez_: thedj: a student is asking me, what's a good book to help him learn PHP? [16:06:08] any recommendations? [16:06:42] sumanah: no, sadly [16:06:44] get in class? [16:06:58] I myself learned it by looking at www.php.net documentation [16:07:00] sumanah: I do know that most of the online ones are pretty bad (atleast the ones i've checked) [16:07:12] hashar: "get in class"? The reason this student is asking me is because his coursework does not include PHP. [16:07:20] I had a background as an ASM / pascal / basic developer though [16:07:23] I learnt by diving in, looking up surprises as I hit them [16:07:33] sumanah: do let me know if you find something good [16:07:56] yuvipanda: :( This is one of the people you passed to me, btw. Argh. OK, I'll do what I can [16:08:01] sumanah: for tech matters, I usually grab a book from O'Reilly [16:08:09] and start a little and simple project just to play with the language [16:08:47] http://rosettacode.org/wiki/Rosetta_Code is a funny site. They list various algorithms and volunteers are implementing them in differents languages [16:08:50] hashar++ [16:08:55] so it is a good way to look at some PHP code [16:09:01] also the php doc is "good" enough [16:10:11] <^demon|away> The docs are usually pretty good. [16:10:20] <^demon|away> The comments below the docs are usually awful and should be ignored. [16:12:16] ^demon|away: hashar why does the mw docs have 'call graphs' anyway? [16:12:24] do they serve any purpose other than bloating up the page? [16:12:33] <^demon|away> Someone thought they'd be useful :) [16:12:38] who? [16:12:41] I guess it is me [16:12:45] <^demon|away> Don't remember. [16:12:47] <^demon|away> Or hashar. [16:12:48] <^demon|away> :) [16:12:50] hashar: why? [16:12:53] the idea was to eventually find out which functions were the most called [16:13:07] but that is probably useless [16:13:27] hashar: plus the images are huuugeee [16:13:32] hashar: can we please get rid of them? [16:13:41] edit maintenance/Doxyfile [16:13:51] look for CALL_GRAPH or CALLER_GRAPH and set that to false [16:14:04] I already have a local copy without the images ;) [16:14:24] was talking about copy at http://svn.wikimedia.org/doc [16:14:39] do you have any example of a bloated page? [16:14:51] oh, every page. let me find one [16:16:07] augh [16:16:12] not very linkable, the default frame based doc [16:17:09] right click the frame and choose "open in new tab" or something like that :D [16:17:16] hashar: linker, for example [16:17:41] hashar: http://svn.wikimedia.org/doc/classLinker.html [16:18:29] more than two screens of horizontal scrolling [16:19:22] brb [16:20:24] maybe there is a way to make the graph morecompact [16:20:27] and to filter out some methods [16:20:59] <^demon|away> We already filter wfMsg & friends iirc. [16:21:19] wfWarn wfDebug could use some filtering too [16:21:33] <^demon|away> wfWarn() is used like 5 places. [16:21:38] <^demon|away> wfDebug() perhaps [16:22:13] not sure how to filter them out though [16:22:52] <^demon|away> hashar: I have a git repo of phase3 now :) [16:23:01] awesome [16:23:11] <^demon|away> You don't want it just yet, still 4.2G ;-) [16:23:18] git gzip [16:23:27] <^demon|away> I'm doing some repacking now. [16:23:29] and get ride of all the languages files /D [16:23:53] with raymond doing commits every day on thases, that must takes a lot of space [16:23:57] thases/thoses [16:23:57] <^demon|away> But yeah, with a bit of luck we'll have a test phase3 repo up today for people to start playing with. [16:24:13] <^demon|away> The i18n commits made exporting take a long time ;-) [16:24:25] *hashar hands brion a donut [16:24:27] *brion smells a git repo [16:24:35] om nom nom nom [16:24:38] <^demon|away> brion: A very very smelly 4.2G phase3 repo :) [16:24:46] <^demon|away> Trying to remove some of the smell now [16:24:53] \o/ [16:26:25] <^demon|away> brion: http://www.mediawiki.org/wiki/Git_conversion/Rules/phase3 [16:26:36] <^demon|away> Those were the rules I ended up using. [16:30:00] is your tool able to detect svn copy? [16:30:15] <^demon|away> Yes [16:30:33] still 4G is a lot [16:30:46] <^demon|away> That's totally uncompressed. [16:30:57] <^demon|away> A raw SVN dump is something on the order of 10G, so I'm not totally surprised. [16:31:08] yay a test repo [16:31:39] fun fun [16:32:19] hopefully it will be much much less mess in it than in the current svn one [16:32:43] <^demon|away> We have over 337k objects over the course of the phase3 path, so yeah, definitely needed some efficient repacking :) [16:32:52] the current one messes up various unrelated things [16:33:07] has weird nonintuitive names etc [16:33:26] <^demon|away> Danny_B|backup: Well we're breaking our huge svn repo up into smaller git repos :) [16:33:34] <^demon|away> And we're dropping the phase3 name ;-) [16:33:40] oh rilly!!!! [16:33:51] apergos: for phase4 [16:33:52] phase4 [16:33:52] <^demon|away> I've been planning on mediawiki/core.git [16:33:54] I knew it :-D [16:33:59] <^demon|away> Unless you guys really want phase4. [16:34:02] <^demon|away> We can do that instead. [16:34:02] nooooo. [16:34:04] no. [16:34:07] mediawiki/phase4.git [16:34:08] no no no no and NO. [16:34:11] plizZZzzzz [16:34:16] *apergos glares [16:34:23] <^demon|away> How about mediawiki/2.0.git? [16:34:24] <^demon|away> :) [16:34:37] hmmmm :-) [16:34:41] or we might as well rename MediaWiki to something people will not mistake with WikiMedia or WikiPedia [16:34:58] something like BrioTim [16:35:04] <^demon|away> hashar: LDCWiki? [16:35:06] there is no 2.0 [16:35:20] I think brion reserved 2.0 for the new parser [16:35:32] mediawiki/core.git is fine anyway [16:35:45] core sounds confusing [16:36:00] or mediawiki/AAA_main.git so it get listed at the top [16:36:04] <^demon|away> mediawiki/mediawiki.git just sounds redundant. [16:36:12] i just don't care about version numbers that much [16:36:14] meh :D [16:36:18] I'm ok with core [16:36:25] <^demon|away> core would sort before extensions too. [16:36:27] why is there mediawiki/...? [16:36:27] we use that language plenty for exactly that part ofmw [16:36:33] <^demon|away> So that works with my planned repo layout: [16:36:36] <^demon|away> mediawiki/core.git [16:36:37] you can always git clone .../core.git mediawiki ;) [16:36:40] <^demon|away> mediawiki/extensions/foo.git [16:36:43] like i svn co ???.phase3 trunk [16:36:53] uh huh [16:36:58] or use it in french : mediawiki/C??ur.git [16:37:20] no, use czech! mediawiki/j??dro.git :-P [16:38:09] <^demon|away> We could give it an Esperanto name ;-) [16:38:17] you won [16:38:31] ^demon|away: so where is your structure available to see? [16:38:48] <^demon|away> Danny_B|backup: If you can read my mind! [16:38:54] <^demon|away> But I basically said it a second ago. [16:38:59] <^demon|away> mediawiki/core.git [16:39:01] that's all? [16:39:04] <^demon|away> And mediawiki/extensions/foo.git [16:39:11] there is bunch of other things in svn [16:39:27] <^demon|away> We'll migrate them a bit at a time as we find proper homes :) [16:39:41] <^demon|away> For example, trunk/debs was already moved to git, but ops didn't want the svn history [16:39:44] why not start with small thingies? [16:39:55] <^demon|away> Because phase3 is the important one :) [16:39:58] :-P [16:40:08] it could wait until 2.0 :-P [16:40:53] mediawiki/koro.git :) [16:41:19] ^demon|away: what else will be in 'mediawiki' folder anyway? [16:41:33] <^demon|away> Nothing that comes to mind other than core and extensions. [16:41:39] tools? [16:41:50] where's the "random stuff that doesn't deserve to be a maintenance script but I need to be able to put it somewhere" folder? [16:42:02] <^demon|away> /trunk/tools [16:42:14] <^demon|away> Danny_B|backup: We can probably split the tools off to various locations. [16:42:16] in git? [16:42:21] <^demon|away> Ah, in git. [16:42:24] <^demon|away> Not sure. [16:42:30] ^demon|away: yeah, that was my plan [16:42:45] <^demon|away> Pretty trivial to do, we just need to write some rules. [16:43:05] <^demon|away> tools is a very misc. folder, so they can probably find different homes. [16:43:32] ^demon|away: another scheme would be : wmf/mediawiki.git wmf/extensions.git and have extensions not on live site in a different directory extension/socialprofile.git [16:44:03] <^demon|away> I want core and extensions in the same base mediawiki/ [16:44:13] ahhhh a big bag of stuff [16:44:38] <^demon|away> And I'm not 100% sure how we'll do wmf branch yet. [16:44:53] <^demon|away> Maybe a wmf/mediawiki that's a fork of mediawiki/core with the various extensions imported in. [16:44:57] <^demon|away> We'll have to figure that out [16:45:39] <^demon|away> Although presumably, if core's going to a gated trunk model we could go back to doing deployments from master rather than maintaining a different branch. [16:46:10] obviously it's definitely necessary to discuss the structure befor migrating and find some solution comfortable for everyone [16:46:33] <^demon|away> Well we can do the conversion before we put them up. [16:46:44] <^demon|away> This naming convention is largely just bikeshedding. [16:48:53] wmf branches could be in the same core.git [16:49:04] then use submodules to get the extensions we care about? [16:49:14] <^demon|away> Perhaps something like that. [16:49:25] git clone bikeshet.git [16:49:35] <^demon|away> bikeshet? [16:49:40] heh [16:49:43] git clone speling.git [16:49:46] that can go nowhere good [16:50:25] <^demon|away> "I can't do any work right now, my repo's repacking" [16:50:41] :) [16:50:46] :-D [16:51:27] <^demon|away> I mean heck, the repo has almost 9 years of history :p [16:51:47] well see you later. Dad on duty [16:51:57] 5 million images with orientation info in the exif [16:52:03] ugh. how many of them are problems... [16:52:49] apergos: this isn't new; the rotatebot guys brought this up a couple months ago [16:52:53] and started on fixing them :) [16:52:56] no [16:53:03] the problem is that I am purging thumbs [16:53:08] but yeah there's still some cleanup to go [16:53:11] heh [16:53:14] because we are gonna run out of space if we don't [16:53:19] like next month [16:53:52] oww [16:54:30] yeah. ow indeed. hurry up swift! [16:54:50] <^demon|away> Too bad we don't use MongoDB, MongoDB is web scale. [16:55:02] <^demon|away> NFS is not web scale. You cannot use NFS for web scale applications [16:55:11] <^demon|away> MongoDB is made for web applications, because it is web scale. [16:57:03] it's still funny. every time... [16:58:21] <^demon|away> Compressing objects: 89% (302711/337395) [16:58:24] *^demon|away twiddles his thumbs [17:00:30] I purge my thumbs, you twiddle yers... hmmm [17:01:01] <^demon|away> You have many more thumbs than I. [17:01:07] <^demon|away> Millions, even [17:06:51] hexmode: FYI, one reason i didn't see your mail about the funcom triage until i got awjr's was because it was in teh spambucket [17:27:05] <^demon> How about 142M for a phase3 repo? ;-) [17:27:13] <^demon> 164, even [17:27:15] <^demon> But still :D [18:13:30] DarTar: I'm on now. [18:13:37] hey Krinkle [18:13:39] getting workflow started [18:13:43] fabulous [18:13:57] it looks like we need to find a dev who can deploy [18:14:20] Roan being on vacation, alolita suggested we ask Sam [18:14:27] and review [18:14:35] but I don't see him around [18:14:47] Roan's on vacation ? I didn't knew that, I was wondering where he was [18:14:55] Reedy: round ? [18:15:26] I need to send an urgent email after which I'll be working on CentralNotice [18:17:11] I'm just about to go AFK for dinner [18:17:15] I shouldn't be too long [18:17:42] hey Reedy [18:18:28] Reedy: hope we can have your help for a time-critical deployment scheduled for today [18:19:46] *apergos peeks in [18:29:56] Krinkle: re [18:30:05] let me see if I find Jerome [18:30:24] Krinkle: shall we copy your todo list to an etherpad so we can track its progress? [18:30:30] ok [18:30:51] btw the announcement went live on the WMF blog last night [18:31:05] http://blog.wikimedia.org/2011/12/08/experiment-decision-making/ [18:31:31] Eating dinner now, will be keeping an eye in the background [18:31:55] sure, I'll go through the latest updates from Philippe in the meantime [18:39:43] DarTar: Hi Dario! [18:41:32] hey Jerome_ [18:41:45] sorry, preparing an etherpad with Krinkle's todo list for the deployment [18:42:06] Reedy should be able to help us with the actual deployment once he's back from dinner [18:42:24] Great. [18:42:55] So I received the urls of the new backgroud image and banner button from Philippe. [18:43:04] I'll be sending the new banner code to you shortly. [18:43:09] yeah I have just quickly scanned through it [18:43:10] thanks [18:43:43] How much time do you expect the deployment and testing to take before we can actually go live? [18:44:21] I am just thinkinh that I would be nice to still have Anne around if we want to check that everything is going fine indeed after the launch... [18:44:38] it depends on the dependencies (code review, finalization of the design) [18:44:45] definitely [18:44:58] and I won't be around between 11.30-ish and 2pm today [18:45:16] so I want to do as much as possible in the next 45 mins [18:45:39] Okay. I'll make myself available anytime today if needed. [18:46:38] We still have to send you the banner code, beyonf that, is there any technical point that we still have to work through from our side? [18:46:55] Per you yesterday message, I assume that the answer is no, but want to make sure [18:47:44] Jerome_ Krinkle Reedy - here's the todo list: http://etherpad.wikimedia.org/HarvardBannerDeployment [18:48:18] Jerome_: as long as we have the final images and CSS nothing else (other than availability for testing) is needed on your end [18:49:37] Ok. I expect this will be done within the next hour at the very last. [18:50:05] maplebed: btw, how did you change swift to make it work via coudfiles again? [18:50:08] Do we have a sense of when Timo is going to be available online? [18:50:13] you'll need to get someone with staff access to save the CSS if I am not around [18:50:19] he's online but having dinner [18:50:24] should be back shortly [18:50:28] Okay. Cool. [18:50:40] AaronSchulz: by 'via cloudfiles' do you mean authenticated or unauthenticated URLs? [18:50:44] I think authenticated, right? [18:50:51] yes [18:51:17] in swift, when you create an account, you assign an authentication URL. I had assigned that URL to 127.0.0.1. I should have assigned it to $proxy-address. [18:51:17] We, I guess I can ping Philippe about this if you're not around? Is is just about replacing the old banner code with the new one... What do you think? Is that possible? [18:51:35] yes [18:51:37] while it was set to 127.0.0.1, cloudfiles obeyed it and tried to authenticate against your localhost. [18:51:40] which obviously didn't work. [18:51:43] let me see if I can find him [18:51:46] heh [18:51:52] Okay. [18:52:34] AaronSchulz: I documented that (at least noted it) here: http://wikitech.wikimedia.org/view/Swift/Hackathon_Installation_Notes#using_swauth [18:53:38] Philippe is joining [18:53:45] he should be able to back me up if I am not around [18:54:08] Nice :) [18:54:57] So Anne is coming back online in a moment. I e-mail you guys as soon as we updated the banner code. [18:55:45] Right, so [18:57:41] great, I just sent around the link to the etherpad [19:00:28] Jerome_: regardless of the CSS, are the graphics I see here the final ones? [19:00:33] http://meta.wikimedia.org/w/index.php?title=Special:NoticeTemplate/view&template=Harvard_small_linebreak [19:00:44] it looks like the RCom tagline hasn't changed [19:01:37] I thought Anne wanted to add the text to the image itself? [19:02:09] No. We add the text ourselves to the image. [19:02:29] The button will also be different. [19:02:55] Smaller so that the layout is nice in small screens. Same thing with the Berkman and Sciences Po logos. [19:03:02] ach, I thought we had the final version of the images ready? [19:03:29] let me mark this as not done in the etherpad then [19:03:38] Yes. I sent the latest versions of the images to Philippe, but we now have to include their urls in the modified banner code. [19:04:13] So the images are ready, we just need to call them in the banner code! [19:04:26] right, ok, that's reassuring [19:05:16] DarTar: I thought we agreed on the no_linebreak version, the one you linked to has the line break issue [19:06:06] Yes. We will use the no_linebreak version. [19:06:35] Can you link me the latest version of what you want to be in the banner ? [19:06:42] then I can try it out first on test.wikipedia [19:06:51] I do have banner access there, not on meta [19:07:02] https://test.wikipedia.org/w/index.php?title=Special:NoticeTemplate/view&template=Harvard_prototype [19:09:19] Jerome_: No changes since yesterday ? [19:10:01] Wait a minute I am updating the code myself to send it. [19:10:11] ok [19:11:52] This will be done in 10 minutes. [19:12:58] (really sorry to put you on hold guys...) [19:15:33] Krinkle: sorry, my bad [19:16:06] I heard graphics are uploaded ? [19:16:16] yes [19:19:47] Krinkle: nice checkboxes :) [19:21:23] thx [19:22:33] Jerome_: Krinkle I expect to go offline in about 10-15' from now, is there anything I can immediately help with? [19:23:15] No, not right now. as soon as the final banner version is ready I'll update the prototype so that it;s ready for testing as soon as deployment happends [19:24:00] Krinkle: one general note, it looks like the name we chose for the settings is specific to the Harvard campaign, I am wondering whether we shouldn't use something more generic? [19:24:17] $wgNoticeBanner_Harvard2011['enable'] = true; [19:24:17] $wgNoticeBanner_Harvard2011['salt'] = "14"; [19:24:46] for future campaigns, that is [19:25:14] I know what you mean, but I think this way makes sense for now. Also because different campaigns will need different kind of settings. [19:25:49] ok, fine with me as long as we document them in the CN help page [19:25:53] and allows to prepate/develop multiple compaigns at the same time. [19:25:58] prepare* [19:26:03] fair enough [19:28:13] so Jerome_ can we mark the upload final graphics as done? There is not other image dependency still missing, correct? [19:29:00] Mail for you, Dario, Philippe and Timo. [19:29:43] Got it [19:31:04] test.wikipedia banner updated [19:31:04] excellent, this is what we are using, correct? https://meta.wikimedia.org/w/index.php?title=Special:NoticeTemplate/view&template=Harvard_small_no_linebreak [19:31:14] latest version is now here: https://test.wikipedia.org/w/index.php?title=Special:NoticeTemplate/view&template=Harvard_prototype [19:32:05] we'll update the meta-banner once everything is confirmed to work on test.wiki [19:32:24] I'd say between deployment to test-wiki and en-wiki [19:32:26] looks good to me, the text of the button is not correctly vertically aligned though [19:32:38] Jerome_: ^ [19:33:20] other than that, the design looks fine and avoids the overlap problem [19:33:48] Jerome_: Alright if I change line-heigtht for button text from 1.6em to 1.9em ? [19:33:51] that seems to make it centered [19:34:01] Yep. Okay. [19:34:06] jh #wikimedia-commons [19:34:07] maplebed: yeah, swift still works [19:34:14] whoops [19:34:31] Jerome_: DarTar updated: https://test.wikipedia.org/w/index.php?title=Special:NoticeTemplate/view&template=Harvard_prototype [19:34:32] better now ? [19:34:46] way better [19:34:51] Looks fine. [19:35:08] maybe subject to cross-browser/platform font differences [19:35:24] yeah, testing right now on a few diff browsers on Mac OS [19:35:28] note that there is no font enforced on MediaWiki, just "sans-serif" [19:35:37] Hey DarTar :) [19:35:44] hey Philippe [19:35:45] Jerome_: What brwoser/OS you got at hand ? [19:36:01] Philippe: we are looking at: https://test.wikipedia.org/w/index.php?title=Special:NoticeTemplate/view&template=Harvard_prototype [19:36:19] That's the same version I just uploaded to meta? [19:36:25] Krinkle: on FF the vertical alignment is still not perfect [19:36:33] I'm use Firefox 8.0 / Windows 7 [19:36:46]