[06:37] ping ralsina, do you know if bug 946626 will be fixed before precise? Or should I try to modify the translations to make them fit? [06:37] Launchpad bug 946626 in Ubuntu One Control Panel "ubuntuone-control-panel does not adapt to length of translated messages" [High,In progress] https://launchpad.net/bugs/946626 [08:05] Happy Tues^WWednesday, all! [08:23] morning! [11:19] good morning [11:22] gatox, morning! [11:22] gatox, can I ask you to do reviews and let me know the errors you get when running the ubuntuone-control-panel tests, I get all green.. [11:22] gatox, I suspect you will get issues :P [11:23] mandel, jeje ok [11:23] gatox, you don't have a static ip or something so that I could get an account in that machine, right? [11:24] gatox, would be more efficient than this hehe [11:24] mandel, nop [11:24] gatox, fuuuuuuu [11:24] :) [11:28] gatox, here are the MP for the reviews, each of them fixes a subset of the failing tests (so that the changes were not too big): [11:28] https://code.launchpad.net/~mandel/ubuntu-sso-client/fix-broken-tests/+merge/99770 [11:29] https://code.launchpad.net/~mandel/ubuntu-sso-client/fix-webclient-tests/+merge/101410 [11:29] https://code.launchpad.net/~mandel/ubuntu-sso-client/fix-activation-tests/+merge/101411 [11:44] ralsina, ping? === teknico_ is now known as teknico [12:46] good morning! [12:46] mandel: pong [12:47] ralsina, I think we are very close to start merging the windows tests fixes and get sso pasing tests on jenkins, can I have some reviews? [12:47] mandel: sure! [12:47] * mandel ralsina mas, echame una mano para que alecu y dobey miren lo de u1-dev-tools :) [12:47] mandel: let's do it now before I spend the rest of the day doing perf. reviews :-/ [12:48] mandel: sure, will do [12:49] ralsina, ok, so, ideally, you need to add to you path lp:~mandel/ubuntuone-dev-tools/mocked-webserver [12:49] ralsina, which fixes adds a number of test cases to clean the resource properly [12:49] ralsina, and the mp are: [12:49] https://code.launchpad.net/~mandel/ubuntu-sso-client/fix-broken-tests/+merge/99770 [12:49] https://code.launchpad.net/~mandel/ubuntu-sso-client/fix-webclient-tests/+merge/101410 [12:49] https://code.launchpad.net/~mandel/ubuntu-sso-client/fix-activation-tests/+merge/101411 [12:51] mandel: ok, this is going to take a while ;-) [12:52] ralsina, I know.. [12:53] I could do a code review and let gatox do the actual test runs since he has the evil vm [12:53] and I really don't want to boot into windows [12:53] ralsina, yep.... i'm doing that already [12:53] ralsina, sure! [12:53] ralsina, i'm mandel's bot [12:53] ralsina, just don't accept them cause we need the devtools landing first :) [12:54] gatoxit's better than being mandel-brot [12:54] ralsina, BUAH, chiste matematico.. [12:56] mandel: yes. But it's because Mandelbrot is DEAD [12:56] or mandel-broth [12:56] that would be bad [12:56] ahhhhhhhh now i understand [12:57] i thought that maybe was something nassty... and i was afraid to ask :P jeje [12:58] lol [13:02] mandel: this code is a bit out of my depth :-/ [13:02] ralsina, ok, no worries [13:02] mandel: I am very unfamiliar with the code, and I suspect I don't understand the real problem [13:11] ok, time for my lunch! [13:15] i need to do some important errands..... i'll be back in a while (i'll stay a little more today) === gatox is now known as gatox_brb [13:20] ping ralsina, do you know if bug 946626 will be fixed before precise? Or should I try to modify the translations to make them fit? [13:20] Launchpad bug 946626 in Ubuntu One Control Panel "ubuntuone-control-panel does not adapt to length of translated messages" [High,In progress] https://launchpad.net/bugs/946626 [13:20] trijntje: we didn't fix it in time, sorry :-( [13:21] trijntje: we'll try to make a SRU for it soon after release, though [13:21] ralsina: ok, thanks, good to know. I'm not sure if I'll be able to squeeze the translation even more, so maybe I'll just leave it as it is and wait for the fix [13:22] trijntje: great, thanks for trying [13:25] ralsina: sure, thank you for looking into it ;) [13:25] oh crap [13:25] perf reviews [13:27] hola mandel! === m_conley_away is now known as m_conley [13:29] the b^Hpeer reviews are due today === yofel_ is now known as yofel [13:51] briancurtin: pig [13:51] I mean ping [13:51] ralsina: pong [13:51] briancurtin: how about we do a windows release today? :-) [13:51] briancurtin: I want to go through all the steps with you so I don't have to do it anymore ;-) [13:52] ralsina: i like the sound of that [13:52] briancurtin: 1) we need to do a patch for this branch with the version numbers updated: https://code.launchpad.net/~dobey/ubuntuone-windows-installer/update-from-trunk/+merge/101466 [13:52] briancurtin: this release is 3.0.0 [13:54] ralsina: so should i start by building installers based off of trunk, plus the Cloud To Computer hack branch from yesterday, plus this branch, plus 3.0.0 version change? [13:54] briancurtin: no, we'll do it from the stable branch + that hack branch [13:54] ah [13:54] briancurtin: this patch is to make dobey merge it into stable with the correct version numbers [13:55] ralsina: *what* version number exactly? [13:55] dobey: the XML that builds the installer contains the version number (used for many things, including installer filename) [13:56] dobey: also, it appears in windows' "installed software" thingie [13:56] another thing we should automate [13:57] ralsina: i can change that xml file when i bump the version in the setup.py [13:58] ralsina: so that doesn't need to happen in this branch [13:59] dobey: on what setup.py? [13:59] dobey: the one on windows-installer? [13:59] ralsina: yes [14:00] for doing eg. ./setup.py sdist to make the tarball :) [14:00] dobey: but we have not you know, writtten the code to do that. If we don't do it on this branch, we will be building the release from something that is not tagged [14:00] dobey: that's not how this thing works [14:00] ralsina: huh? [14:00] dobey: that's not even remotely close to how this thing works ;-) [14:00] ralsina: i think you are confused [14:00] ralsina: this branch has no tag [14:00] dobey: probably [14:01] ralsina: when i make the tarball, i will commit the version bump to the branch by hand, and tag it [14:01] dobey: I want to use the tarball we are "releasing" of -windows-installer to build this release for windows [14:01] dobey: ah, ok then [14:01] dobey: all the places that need touching on version bump are described in the README [14:02] ok [14:02] hopefully ;-) [14:02] ralsina: unless i'm mistaken, should i wait for dobey to make the tarball and commit the version bump before proceeding? [14:02] briancurtin: yes, that is right [14:03] ralsina: and re: your comment on the ubuntuone-client branch; i thought we agreed to leave that as-is and rely on revnos.txt for now, and we'd automate/fix that issue after precise [14:03] briancurtin: i think you also need to wait for the releases of ubuntuone-control-panel and ubuntuone-client as well, which aren't done yet [14:03] dobey: just for this once, so the user gets the correct version on the logs, please? [14:03] dobey: then we automate it after P [14:03] ralsina: ok, but i'll also do that with the version bump of configure.ac, rather than in this branch [14:04] dobey: fine by me, +1ing both branches then [14:04] ok [14:04] and there's also https://code.launchpad.net/~dobey/ubuntuone-control-panel/update-from-trunk/+merge/101470 [14:04] if you hadn't noticed it yet :) [14:05] dobey: there, +2 on those, looking at u1cp now [14:10] dobey: +1 on u1cp [14:11] ok. as soon as they land, i'll get the tarballs done. then work on getting the uploads done [14:12] and hopefully actually get the peer reviews done today as well [14:12] dobey: if I missed one, let me know [14:12] dobey: I have to do 14 allhands tasks today. Gonna be a long day [14:13] * mandel back [14:14] alecu, buenos dias! [14:14] dobey, :) [14:14] hola mandel! [14:14] you'll love this one: [14:14] "Teacher: Billy, can you say the word ‘politics’ in a sentence?" [14:14] "Billy: Sure, Polly the Parrot swallowed a clock and now Polly ticks." [14:15] alecu, hahahahaha [14:15] hahah [14:15] alecu, yes I do hehe [14:21] hi mandel [14:21] dobey, hello! how is it going? [14:21] dobey, do you have a busy day, can I beg for reviews? [14:21] i have a very busy day, yes [14:22] and yes, your branches are on my list to review :) [14:22] dobey, cool, but if you have a busy day, we can wait 'til tom, I know there is work to be done for the linux review [14:22] mandel: but first, i will have to review *you* [14:23] dobey, sorry I ment linux release [14:23] dobey, just say I'm hairy :P [14:23] dobey, and I can't spell hehe [14:23] yep, need to finish releases, and do perf reviews [14:25] So, my mom uses Ubuntu One a lot. Has relied on it for quite a while for note taking, and has built up an extensive number of notes. Hundreds? [14:26] Just wanted to say I'm a little disappointed w/ you guys for pulling a Google and just killing off a service that people had become dependent on. You'd think you could at least just hide it for people who aren't using the notes sync :( [14:26] You know. do it a bit more slowly. Or. Maybe, and Google at least does this, give like a 6 month warning period so people can try to find an alternate service without something they rely on vanishing. [14:26] And, yeah, I know that Ubuntu One tomboy sync still works, but without the ability to access notes from work, she's crippled. [14:26] So now I'm reading the API trying to figure out how hard it would be to reimplement a subset of the functionality you removed :( [14:30] back === gatox_brb is now known as gatox [14:31] nemo, yeah, it was a really hard decision to make. We did broadcast it many months before, maybe we should of done a better job at communicating, sending out emails [14:31] nemo, essentially, it had a lot of problems and users were loosing data [14:31] fixing it was a huge amount of work, which we can't afford to invest [14:32] nemo, but you are right, it's really bad that we have to shut it down [14:33] beuno: correct place to do it is in the Notes interface [14:33] ideally a migration path to an alternately hosted service [14:33] nemo, indeed, and we're building a notification capability into the website [14:34] or sharing the code you used for the notes interface [14:34] would simplify my task now of replacing it [14:34] a rewrite will be tedious [14:34] well. how hard is it to build in a notification. edit the template for the website, add "Hey! We are shutting this down in 6 months!" [14:34] nemo, well, the problem is it's based on top of couchdb, couchdb really doesn't scale well [14:34] * nemo sighs [14:34] so we've shyed away from sharing the code [14:34] well. I only need it to function for one person :) [14:34] there's no need for it to scale, at all [14:35] ohhh [14:35] wait. what? [14:35] You still support notes sync! [14:35] so the db is still active! [14:35] yeah [14:35] If I trigger an API call, from my website, you'll still honour it [14:35] assuming she is authenticated... [14:35] yeap [14:35] so. how have you solved the load issues? [14:35] all the people still using tomboy notes are still syncing [14:36] also. I find it amusing that couchdb did not scale well, given one of the arguments of the anti-SQL crowd (pardon me, NoSQL) is that their databases scale better :) [14:37] alecu, just in case: http://www.mapofthedead.com/ [14:37] nemo, well, added more hardware and shut down a few services, shelved a few new ones we've had. As it stands, it works well most of the time, working hard to build U1DB to replace it [14:37] hm [14:37] nemo, it was amusing to us as well. Well, amusing may not be the right word :) [14:37] beuno: does this mean if I tell her to wait a few months, it might come back? [14:37] nemo, and this is not NoSQL in general, this was specifically couchdb :) [14:38] beuno: eh. I'm sticking w/ PostGreSQL :-p [14:38] alecu, dobey: https://code.launchpad.net/~mandel/ubuntuone-dev-tools/mocked-webserver can use reviews [14:38] nemo, well, we may not bring back a notes web ui, so I wouldn't count on that, no [14:38] beuno: hrm :-/ [14:38] alrighty, sooo. gonna have to write some kind of replacement [14:39] ralsina, looking [14:39] the problem is Yahoo! is probably less likely to survive than Ubuntu [14:39] nemo, but we will make it easy for others to build these applications :) [14:39] and Google shut down theirs [14:39] And the nice thing about ubuntu is it synced to a desktop app [14:39] was really convenient that way [14:39] kind of a unique service [14:39] indeed [14:39] it wasn't something we looked forward to doing [14:40] in the end, there's only so much we can do with a small team [14:40] well. migration path, and actually editing the website to put a notice in [14:40] would have done a world of good [14:40] yes, learn and improve :) [14:40] the latter one would have taken... 1 minute? [14:40] oh well :-/ [14:40] briancurtin, there is nothing like XVFB on windows, right? [14:40] beuno: oh. and super super nice. leaving it active for people who have a ton of notes in sync [14:40] and just hiding it for everyone else [14:40] mandel: i'll have to look up xvfb...no idea what it is [14:40] beuno: you could still do that one actually ;) [14:40] surely that would avoid load issues [14:40] nemo, the problem is, people were loosing data [14:40] mandel: ohh, no there isn't [14:40] nemo, so leaving it on was a terrible experience [14:40] I heard about that [14:40] but she wasn't :) [14:40] m'k [14:41] mandel: at least as far as im aware of [14:41] briancurtin, ins rendering on a buffer the ui, it might not exist.. [14:41] beuno: so. it'd be even better then if I synced to something other than UbuntuOne - rehosted it all on my machine [14:41] beuno: is the codebase for ubuntuone open sourced by any chance? [14:41] briancurtin, mandel: you could use a terminal server and rdp for it but it would be incredibly annoying [14:41] mandel: it might be because im not a UI guy, but i've never heard of anything like that [14:42] nemo, not the server, no [14:42] durn [14:42] * nemo ponders [14:42] ugh. guess I'll just have to reimplement the API on my local SQL db :-/ [14:42] nemo, and why not just use tomboy? [14:42] beuno: she *DOES* use Tomboy [14:42] she uses tomboy a lot [14:42] ralsina, briancurtin I wanted to remove the need to draw the ui in the tests on the screen, but nevermind [14:43] the problem is she counts on all the hundreds of items she inserted into tomboy to be accessible when she's not at her ubuntu machine [14:43] beuno: that means, when she's signed into a web browser at the hospital [14:43] nemo, I see, remote access [14:43] aquarius, any ideas here? ^ [14:43] beuno: I guess I'll start with a read-only interface to ubuntu one. a basic one might be as simple as a bookmarklet [14:43] I'll checkout your file upload. maybe I can insert a web page that when clicked on does local XHR [14:44] ralsina: yes, the branches are on my list; but priorities :) [14:44] beuno: then, look into replacing the read/write [14:44] nemo, beuno I though that he could simply install http://live.gnome.org/Snowy and the do a webui on top of that if needed [14:44] * aquarius reads the backscroll [14:44] mandel: I checked that out [14:44] I saw "alpha" [14:44] and "you will lose data" [14:44] and thought I'd look into alternate stuff [14:44] mandel, back again.... connection problems [14:44] nemo, beuno, since couchdb is an implementation detail for us and AFAIK we use the snowy protocol and store in couchdb [14:45] nemo, so, is it important that your mum can *edit* notes from a browser? [14:45] aquarius: well. read is most critical [14:45] if I could restore that to her, she'd be a lot better off [14:45] I don't know if she often edits from work, but I imagine it'd be nice [14:45] nemo, OK. Take a look at u1.to [14:45] gatox, can you pull from the fix-activation branch and run the tests again? [14:45] aquarius: does tomdroid sync notes with u1? [14:45] mandel, yep! [14:45] aquarius: oh. cool! [14:45] though i guess cell phone + hospital might not be acceptable [14:46] aquarius, is u1.to open source? === gatox_ is now known as DiegoSarmentero [14:46] nemo, it's a private project by Chipaca and I -- it's not an official Ubuntu One thing. However, it has (very rudimentary) read-only access to your notes. === DiegoSarmentero is now known as gatox_ [14:46] aquarius: sweet sweet sweet [14:46] thank you very much [14:46] I want to call her right now, but she's probably still asleep [14:46] timezones and all [14:47] nemo, the code for u1.to is at launchpad.net/youoneteeoh, so if you're a hacker and are interested in improving the notes support, I'd love to help you with that -- I want to myself (I liked the web UI for notes, too, for my mobile phone) but I just haven't had time [14:47] nemo, does your mum have a smartphone, or is she using an actual on-computer browser? [14:47] aquarius: actual computer. but I think she has a smart phone too [14:47] I just don't think it works in the hospital (shielded) [14:47] I believe her phone is running android [14:47] we don't meet that often :) [14:48] she actually installed ubuntu on her laptop herself [14:48] she loves it, apart from the "grey screen of death" as she terms it where occasionally her X session freezes up. I need to look into that [14:48] nemo, on Android, there are two apps which sync notes (sort of) with U1: tomdroid and Chalk. Neither are great yet (they're both under-resourced), and they're both read-only, but they may be alternatives [14:48] I'd recommend updating from 11.04 except, well, she despises the Unity interface. I got a panicked phonecall from her after her 11.04 upgrade [14:48] and XFCE4 is not nearly as friendly as gnome 2 [14:49] nemo, so, u1.to exists and is already there, but I need to stress that it's a private project run by chipaca and I, not something official, so it might crash or be unavailable or whatever :) [14:49] aquarius: hm. 'k. [14:49] aquarius: eh. I'll mention [14:49] you're a dev, so you are probably semi-official anyway [14:50] nemo, the notes code in u1.to uses the Snowy notes API, so if you wanted to write a separate notes app and maintain its uptime yourself, that might be a place to start; the notes API isn't hard, is documented at http://live.gnome.org/Tomboy/Synchronization/REST/1.0, and u1.to is an example of using it from a django view :) [14:50] dobey, mandel, alecu, gatox, briancurtin, thisfred_ (and Harry!) : standup in 10' [14:51] aquarius: yeah. the API seemed so easy, that that was why I thought a bookmarklet might be practical [14:51] aquarius: editing, seemed a bit more problematic, even if I didn't support the XML internal to the note for things like links [14:51] yo! [14:51] nemo, hopefully there's enough there that your mum can either use one of the existing things or that you're able to use our APIs to provide for her; if you're interested in building your own thing (or hacking on u1.to's code) then I'd be more than happy to give some pointers or advice on that === gatox_ is now known as gatox [14:52] nemo, yeah, one of the big reasons that the U1 web app for notes became unmaintainable is that it's very, very hard to translate Tomboy XML into editable HTML and then back into Tomboy XML again, losslessly. [14:53] well [14:53] she doesn't really need HTML [14:53] a text-editor only would be fine, apart from the links syntax [14:53] And the links syntax isn't that critical. [14:54] nemo, yeah; if you are prepared to limit her to a subset, then something which translates tomboy XML into, say, Markdown format (or just text) would be relatively easy to do, I agree [14:54] deciding whether I want to do that is part of why I haven't had time to continue with u1.to's notes view :) [14:54] tomboy XML can contain arbitrary things (Tomboy plugins can define their own XML tags to go in notes, for example) [14:54] yeah. I don't think she uses any plugins [14:55] she just wanted searchable notes on her machine that were available "in the cloud" [14:55] ok. she uses the sync plugin, but apart from that... [14:55] defining a limited syntax (that is: saying "if you use this weird thing in your notes, the web notes view will blow up") makes the problem quite a bit easier; that's not something we can define at an Ubuntu One level, obviously, but in a more limited case like u1.to or a standalone server that you write, it's potentially doable :) [14:55] is funny how many notes interfaces there are online that just overcomplicate this concept. like google thinking that google docs was a good replacement for google notes [14:55] nemo, I agree with you entirely on that point :) [14:56] aquarius: well. not blow up. but "strip the weird syntax leaving you with a plain text note" [14:56] that seems relatively easy to do [14:56] text is more critical than markup. losing text is bad [14:56] supporting links might not be too hard [14:56] aquarius: do you have the code to u1.to somewhere? [14:56] I'm a relatively extensive tomboy user, and in all these years and 100+ notes I've used lists and strikethrough and very occasionally bold and italic and that's it; I don't even use links, myself ;) [14:56] nemo, yep, the code is at launchpad.net/youoneteeoh [14:57] oh. right [14:57] thanks [14:57] aquarius: hm. since you use those things, have you looked into supporting those features? [14:57] nemo, so, as I say, if you're interested in building something or hacking on an existing something like u1.to, then I'd be happy to give pointers. Hopefully u1.to itself will at least partially alleviate the problems [14:57] one thing with online RTEs is they generate crap HTML [14:57] nemo, I have looked into supporting them -- the reason I have not is purely a matter of time. [14:58] so converting back to XML leads to exploding bad docs [14:58] aquarius: eh. first thing is to see how well read-only satisfies her needs :) [14:58] Hedgewars sucks up a ton of time. especially during GSoC [14:58] and of course work [14:58] nemo, yeah, what I'd probably want is somethnig which translates tomboy XML into markdown. All HTML RTEs are dreadful (see earlier point about reasons that the Ubuntu One notes web view was hard to maintain) [14:58] and family [14:59] aquarius: yeah. wiki syntax would be more than adequate [14:59] translating to markdown would be fine for *me*, certainly, but I'm not a typical person, I'm a developer :P [14:59] aquarius: naw. she'd have no trouble w/ that [14:59] ordinary people instinctively use * for bullets already [14:59] kudos to your mum, then :) [14:59] well. she did install ubuntu herself :) [14:59] she's in her late 50s sooo. \o/ [14:59] ok. I installed it the first time, but she's really gotten used to it [14:59] me [15:00] me [15:00] nemo, nice! [15:00] she even figured out how to rip a DVD herself :) (noooo, not illegally - some preacher she likes) [15:00] went to ubuntu software centre, installed dvd rip, put file on desktop [15:00] figured out that doubleclicking on it in gnome offered an option to play it... [15:00] nemo, so, if you run into problems or want to hack on stuff, you can find me here, or various other places (twitter as @sil and http://kryogenix.org/contact) [15:01] m'k. welp. thanks. been a big help already [15:01] my dad's grasping that sort of level of dealing with technology, to his credit [15:01] I told her about u1.to, and I'll take a look at your code [15:01] nemo, excellent! [15:01] nemo, you're a python hacker? [15:01] me [15:02] aquarius: naw. I despise python. sorry [15:02] mandel, alecu ralsina dobey ? [15:02] but I modify it as needed. [15:02] meh [15:02] me [15:02] aquarius: herm. python. so is this hosted in mercurial? :) [15:02] nemo, no worries -- the reason I ask is that u1.to is python/django. What's your tool of choice? (Implementing something which talks to the notes API will be fairly trivial in any language, and there are openid and oauth libraries for pretty much anything these days) [15:03] me [15:03] ok....... i'll go..... alecu last [15:03] aquarius: eh. I'm flexible. For web interface I'd probably pick PHP, but other options are Java and Perl. I *have* been considering doing more w/ node.js though. [15:03] DONE: [15:03] Fixed remove current device issue, Fixed opening syncdaemon only when it's necessary, Run A LOT of tests for mandel :P. Working on Bug #973702 [15:03] TODO: [15:03] Finish with Bug #973702 keep fixing some UI issues. [15:03] BLOCKED: [15:03] nemo, although if you wanted to build something *just* for your mum then you could hardcode one of her oauth tokens into it and then just slap basic auth on the front and put it on nemo.com or wherever ;) [15:03] No [15:03] Launchpad bug 973702 in Ubuntu One Control Panel trunk "The "Settings" panel should have its items set to wrap" [High,Triaged] https://launchpad.net/bugs/973702 [15:03] thisfred_, go [15:03] DONE: wrapped up lp:~thisfred/u1db/c-SPLIT_ALL_THE_WORDS / started on improving tests TODO: finish test improvements BLOCKED: no NEXT: briancurtin [15:03] DONE: mostly finished the first rev of installer automation, had a half day so that's it [15:03] TODO: finish up this installer script, prepare windows release once its ready [15:03] BLOCKED: none [15:03] NEXT: dobey [15:03] me [15:04] nemo, although if you wanted to build something *just* for your mum then you could hardcode one of her oauth tokens into it and then just slap basic auth on the front and put it on nemo.com or wherever ;) [15:04] yep. got it :-p [15:04] herb, looks like people from cordoba like standups :P [15:04] oops, double entry :) [15:04] λ DONE: releases [15:04] λ TODO: finish releases, uploads, peer reviews, reviews, SRUs, finish u1db packaging [15:04] λ BLCK: none. [15:04] mandel: go [15:04] DONE: Read a lot about ssl certs. Added an extra deferred to make tests cleaner. Proposed all branches to fix the sso tests. [15:04] TODO: look at the failures in control panel and u1-client. Talk with briancurtin and start automating things. [15:04] BLOCKED: no [15:04] COMMENTS: country is going down to hell, is fun to watch :) [15:04] ralsina, please [15:04] DONE: some allhands, reviews TODO: 14 more allhands tasks, reviews, guide brian through windows release process BLOCKED: allhaaaaaaaaaands! NEXT alecu [15:04] DONE: reviews, 1-1, resumed working on security bug [15:04] TODO: keep working on security bug, discuss some issues regarding this with mandel [15:04] BLOCKED: no [15:05] joshuahoover: could you add briancurtin as a CC in the windows release RT? I am guiding him to lead it this time. Also, windows release today! ;-) [15:05] right, also DONE: 1-1 with alecu [15:05] ralsina: yep...and great news! [15:05] ralsina, and with me 1-1 [15:05] ralsina, busy day :P [15:06] gatox: he, and crappy memory [15:06] * ralsina forgot to take notes yesterday [15:06] * ralsina takes note: remember to take notes [15:07] ralsina, if only there was a note taking application that you could use in the web.. /cc beuno [15:07] hehehe [15:07] jejejeje [15:07] mandel, good one [15:07] mandel: I have a notebook just for that. [15:08] mandel: it's portable and never runs out of batteries [15:08] aquarius: WRT hardcoding oauth token. I didn't realise oauth worked like that. [15:08] mandel: can probably store like 200kb of notes! [15:08] I honestly haven't looked into the specifics of how it works [15:08] another "been meaning to" [15:08] ralsina, I loose the pen, which is like running out of batteries [15:08] aquarius: I figured tokens expired [15:08] nemo, there are three steps to oauth: get a request token, ask the user to authorise that token so that it can be swapped for a real access token, and thereafter use that access token to sign requests [15:09] nemo, so, if you're building a site thatanyone can use (as u1.to is) then your site needs to handle getting tokens for an arbitrary user [15:09] nemo, but if you're building nemosmumsnotes.com and no-one else will ever use it, you can just do the first two steps once and then hardcode the token into the backend ;) [15:09] cool [15:10] anyway: EOM [15:10] nemo, not that I'd necessarily *recommend* this approach, but if you're not interested in building a public tool but instead just want to help your mum (which is reasonably noble in itself) then why not :) [15:10] heh [15:10] And people: remember your allhands deadlines! [15:10] aquarius: well. I'd *like* to help the public, but, resources are limited :( [15:11] nemo, I know the feeling; I have exactly the same problem, which is why u1.to's notes view is unformatted and incomplete and sometimes a bit broken and read-only. :) [15:11] aquarius: I like to think that my contributions to Hedgewars and Mozilla and Audacious Media Player and various debugging and patches on free software stuff [15:11] has built up sufficient karmic balance ;) [15:11] aquarius: heh. I kinda noticed ;) [15:11] nemo, you're already ahead of the game, then ;) [15:11] aquarius: the publish interface is kinda neat [15:11] shame the formatting gets blown up [15:12] nemo, I feel a bit guilty about its unfinished state, but entirely giving up going to the pub just so I can build *more* free software projects is a bit of a bridge too far ;) [15:12] aquarius: also odd, is that the first time I click on a note, nothing happens [15:12] I have to click on it twice [15:12] hm. that's not consistent [15:12] maybe it is about idling [15:12] nemo, yeah, the formatting thing is because I'm not sure *how* I want to do formatting; if you want to display the notes formatted then you have to at least have a one-way tomboyxml->html translation (or possibly tomboyxml->markdown->html, of course) [15:13] unless "displaying things in markdown/wikimarkup" *counts* as formatting :P [15:13] aquarius: huh. you'd think someone would have already written a tomboy->html xsd somewhere [15:14] most modern browsers could then just load the XML+XSD [15:14] probably wouldn't be too much work to write one [15:14] there is one, but it's a little out of date [15:14] eh. still [15:14] and I didn't look at it recently because I wanted a two-way translation; now, a one-way translation may be enough (store in tomboy xml, view in xsd'ed HTML, edit in markdown) [15:15] again, it's purely a matter of having time to look at it rather than it being immensely difficult :) [15:15] yeah. this weird clicking on links twice thing. really odd 'cause you're not using JS. [15:18] dunno what that's about. Chipaca did tell me that the way I'd done the caching stuff was stupid, which I agree with him on but haven't had time to fix it, so that may be part of the problem :) [15:18] huh. I wonder if it is a firefox nightly bug. not reproducing in my normal firefox w/ noscript and firebug running. [15:18] ehm. will keep an eye for it [15:18] welp. thanks. plenty to start with. makes me feel a lot better about this [15:19] sure the other ubuntu one users are screwed, but at least my mom is fine! [15:19] patches are naturally welcome ;-) [15:19] we're all about helping your mum. A waterfall begins with one drop of water. :) [15:23] aquarius: Ubuntu One is probably the main reason I really can't recommend Mint + Maté||Mint Gnome shell as a solution to the new candified tablet interfaces problem. She really has gotten used to it. [15:24] aquarius: I have no idea what I'll do right now once 11.04 is EOL'd :( :( [15:24] My main hope is that XFCE4 becomes sufficiently user friendly, really fast [15:25] nemo, we shall have to agree to differ on the Unity interface (although my usual advice is that it's worth taking a week and thinking "I will try and work how unity works, rather than trying to make unity work how my existing desktop works", and then if you still don't like it, no problem, use something else) [15:25] but I'm really pleased to hear that your mum likes U1 :) [15:28] ralsina: ok, so the scripts/README is kind of nondescript about life :) [15:29] ralsina: what all versions should i bump in windows-installer exactly? [15:30] ralsina: there's a "version_id = 203" in ubuntuone.xml; should i bump that to 1000 also? or should i bump it, and the autoupdate.xml to both have 1001 or something? [15:30] aquarius: well. it was more like. mom gives me panicked phone call after a week of trying to use it [15:30] aquarius: look. she's in her late 50s. she doesn't want to learn a new desktop [15:31] ralsina, do you know if sidnei is around today? [15:31] aquarius: she spent 5 years or so learning gnome2, which was luckily similar to Windows XP which she had spent years learning. which was similar to windows 95 which she'd spent years learning :) [15:31] * aquarius grins [15:31] I understand :) [15:31] aquarius: but. yeah. not going to argue with her on this. she wants a classic desktop. so. I gotta find her one. [15:31] that's fair [15:32] personally I switched to XFCE4 on 3 machines. I did spend quite a while trying to use unity. A 4th runs gnome shell + mint extensions PPA for ubuntu, but is really sluggish. My work machine is on XFCE4, 'cause gnome-shell repeatedly crashed due to fglrx suckiness. Unity did not, but, sorry, it was getting in the way of getting my work done. [15:34] My SO is using XFCE4 now, but kinda finding it annoying to find stuff since the config is far from unified, unlike Gnome 2. she's been suggesting she wants to try Gnome Shell again - She's spent several weeks on unity and gnome-shell [15:34] oh well. whatever. moving on. totally unrelated to helping mum w/ u1 [15:34] nemo, I feel your pain. However, Unity in 12.04 has been a massive change in polish and has generally solves my gripes with it [15:34] beuno: aight. I'll give it another try. it is still installed [15:34] ralsina, briancurtin, gatox just confirmed me that we have a branch with all control-panel tests passing \o/ [15:35] mandel: \o/ [15:37] ralsina: did you see my questions above? [15:37] dobey: looking... [15:38] dobey: give me 1' and I will clear it [15:39] dobey: bump build to 1000, and version to 3.0.0 [15:39] dobey: also version_id to 1000 [15:39] ok [15:40] dobey: version in ubuntuone_autoupdate.xml should be 1000 [15:40] dobey: because it hates versions with 3 numbers in it :-/ [15:46] bug 973498 [15:46] Launchpad bug 973498 in Ubuntu One Control Panel "Cloud-to-Computer screen does not work on Windows install" [Undecided,Triaged] https://launchpad.net/bugs/973498 [15:47] ralsina, briancurtin: ok, releases done. and all the stable-3-0 branches have the release-3_0_0 tag [15:47] and on that note, i'm going to get some lunch [15:47] bbiab [15:47] dobey: thanks! [15:56] briancurtin, do you have time in 10 mins for a mumble so that we can talk about jenkins and what was wrong with the windows tests? [15:57] briancurtin, mainly so that you know the reasons and can tackle the problem if it happens again :) [15:57] * mandel hides from dirty reactor errors! [15:58] mandel: i'm putting together this windows release right now and need to take my girlfriend to the doctor in 45 minutes. would you mind doing the call tomorrow morning since i know you have to EOD soon? [15:58] briancurtin, sure, no problem, gf first! [15:58] briancurtin, specially regarding a doctor appointment [16:03] * gatox lunch === gatox is now known as gatox_lunch [16:05] briancurtin: looks like we need to wait for a SSO fix, though (see u1-internal) [16:05] briancurtin: or rather, no we don't since that oly affects the gtk ui [16:06] ralsina: alright, i'll push on with the release then. got the right tags setup, building it out right now [16:06] briancurtin: cool [16:08] ralsina: set SHOW_CMD=False on this one? [16:09] briancurtin: yes please [16:20] ralsina, briancurtin unless gatox vm of death says the opposite, all tests of u1-client pass on windows which means that by merging the already proposed branch everything should be back to green [16:20] mandel: awesome! [16:21] ralsina, briancurtin, if we have time tom we can jenkins running tests and done some automation for the package, I really don't know how we exactly want to do that [16:21] mandel: i have a script almost done which takes care of automating the package [16:22] so that's kind of the last stage of jenkins, the final step if everything is green [16:22] mandel: briancurtin was working on automating the last few bits. Once we get that, you just have to tell jenkins how to find the exe and it will be stored in the artifact archvie [16:23] ralsina, briancurtin, awesome! so let do this tom and see if we can DOS attack QA with packages :P [16:25] ralsina, briancurtin FYI talking with sidnei on why we have ec2-windows is offline [16:34] ralsina: http://u1.to/~brian.curtin/g/3.0.0-windows-installer - i have to run in 12 minutes, should we pick up after i get back or is there a next step i can start doing? [16:35] briancurtin: I'll wait for you [16:35] briancurtin: now, starts the boring part ;-) [16:42] sigh; stupid lunch hour traffic [16:43] complaining about traffic == complaining about yourself :P [16:45] * briancurtin lunch+doc [16:45] ok, EOD for me, catch you all tom! [16:46] gatox_lunch, please add a +1 in the branches you tested, the intermidiate ones 'til activation are the ones with some tests failing, activation should have none [16:46] dobey, alecu if you can review today the ones in ubuntuone-dev-tools I'll really appreaciate it [16:47] mandel, I am reviewing and re-reviewing both of them. [16:47] mandel, tcp-testcases and mocked-webserver [16:47] alecu, sweet! thx a lot! [16:47] mandel: i'll get to them when i can [16:47] dobey, I now, I not trying to push you to do them :) [16:47] dobey, is the EOD bye message hehe [16:47] mandel: can you clarify the mocked-webserver bug though? [16:48] dobey, sure, I'll do that now [16:49] mandel, roger that! [16:51] dobey, updated, let me know if it makes more sense before I go [16:54] despite the spelling errors, it is better, yeah :) === gatox_lunch is now known as gatox [18:21] * briancurtin back [18:24] briancurtin: welcome back! [18:24] hola friholes! [18:24] briancurtin: ready to continue the adventure that is doing a windows release? ;-) [18:25] ralsina: yep, i'm just doing a few quick checks of the installer i built to make sure i built something that works, before we continue on with it [18:25] i need to write a magical script. which will do all this pbuilder stuff in-memory [18:25] also, setting up a caching transparent squid would help i guess [18:25] briancurtin: next step: get all exes and QtCore.dll zip them, gpg-sign them and attach them to RT #51590 at rt.admin.canonical.com [18:26] briancurtin: with a message saying "here are the binaries to be signed for this windows release" [18:26] ralsina: sounds good, will work on that [18:26] brb restarting [18:59] ralsina: i don't think we're ready for release :/ something with creds seems to be screwed up [19:00] briancurtin: boooo [19:00] briancurtin: what's the problem? [19:01] on one machine, i installed 3.0.0 over one of the installers created in the last few weeks. that worked. i went to remove the device so i could uninstall and try installing with no creds, it gave an "error has occurred" dialog, then i had to close, then i open up again and it's in a state where CP shows up but gets AUTH_FAILED, and i can't remove the device to clear creds [19:02] getting AttributeError 'module' object has no attribute 'delete_password' when removing [19:04] briancurtin: you are not using the patched keyring? [19:05] ralsina: i dont think i knew about a patched keyring [19:05] or if i did, i forgot...so that's probably it [19:05] briancurtin: it's the one buildout installs [19:05] ralsina: yeah i've always been using that keyring then [19:05] briancurtin: ok, then we need to debug that [19:06] briancurtin: and I have a deadline in 2 hours and 6 performance reviews to do... [19:06] ralsina: i'll go back to running everything as python scripts instead of as a packaged installer of exes, and see what happens [19:12] briancurtin: ack [19:26] briancurtin: that sounds a lot like https://bugs.launchpad.net/bugs/973830 which should have been fixed in 3.0.0, no? [19:26] Ubuntu bug 973830 in ubuntuone-control-panel (Ubuntu Precise) "Error after removing current device" [Undecided,New] [19:26] briancurtin: are you certain it pulled in the right version for the installer? [19:27] briancurtin: if killing the process and restarting it makes it go away,it may be that bug, yes [19:27] dobey: i manually branched everything from the stable 3.0 branch and the release-3_0_0 tag and am currently re-testing each branch [19:27] ok [19:28] ralsina: it never works after restarting. i currently have two broken machines [19:28] briancurtin: yikes [19:28] briancurtin: removing credentials from keyring manually? [19:29] ralsina: the creds no longer appear inside the registry, where i used to see them in a keyring key with a value ubuntu_sso [19:30] briancurtin: ok, that is just weird then [19:52] briancurtin, do you still get the """object has no attribute 'delete_password'""" error? [19:53] alecu: yep. now i get that upon clicking on the "devices" tab [19:54] briancurtin, afaik, "delete_password" is the method that mandel added to pykeyring [19:55] briancurtin, so it might be the case that the unpatched pykeyring was used by py2exe [19:56] alecu: i'll double check what keyring is going in. ive never used keyring until i started working on u1 and it comes from a buildout script which hasn't changed [19:56] briancurtin: maybe you did an easy_install keyring sometime and stepped over the patched one [19:58] ah, it looks like i have a keyring 0.7 from jan 20 and a keyring 0.8.1 from monday [19:59] so the buildout work i was doing earlier picked up a new one. that must be it [20:00] briancurtin: yes, that is it [20:00] mandel: if you are still around, you should update your patch to latest keyring, I am fairly sure we are missing sme fixes there [20:06] brb [20:08] yay. new furniture arrived [20:09] ralsina, ping [20:09] gatox: pong [20:10] ralsina, i've looking at setBuddy, and stuff.... but all the options involved a lot of changes adding frames or stuff like that.... i've just implement a independent function that receives a widget and do wordWrap over that widget.... even qcheckbox, using "\n"... when is properly..... does this sounds ok for you? [20:11] ralsina, also..... the other options presents other problems..... this function works with the code just the way it is..... we only need to call it after the window is shown [20:11] gatox: can you use \n inside a qcheckbox? weeeeeird [20:12] gatox: another idea I just had! You can examine the qcheckbox.children and see if there is a qlabel there? If there is... well, we can fix this easy ;-) [20:12] yep..... in every widget..... is kind of a manual wordWrap..... the function determines based on qfontmetrics and another stuff where is the best place to wrap.... and check if that position is a space if you are in wordWrap mode.... or if you are in wrapAnywhere just cut the word [20:12] gatox: no hyphenation? tsk tsk [20:13] hyphenation? [20:13] ;-) [20:13] separa con gui-ones [20:13] o sea gui- [20:13] oes [20:13] ralsina, ahhhhhh that's good too [20:13] gatox: it was a joke because it's crazy hard to implement ;-) [20:14] ralsina, i'll check..... but i'm 90% that you don't have a label inside the qcheckbox....... as you don't have a label inside the qpushbutton..... and so..... you have something that behave like that... but its not a qlabel really....... [20:14] and you have qlabel stuff missing there [20:14] oh debuild, why are you not applying this patch [20:15] gatox: I am sure you are right, but just checking is cheap [20:15] ralsina, of course..... i'll check....... the main reason i said is not a qlabel.... it's because you can set html text in those widgets :P [20:16] ah, it did [20:16] yay [20:16] ralsina, either way..... if that is not possible..... i think this is a clean and nice solution..... just wanted to check with you if it sound right [20:16] gatox: sure [20:16] gatox: you will need to overload resizeEvent [20:17] ralsina, yes [20:19] fml. text selection in evolution is so horribly broken now :( [20:22] yay, uploads done [20:23] dobey: broken the way gedit text selection seems to be intermittently broken for me (let's pretend we never got the mouse-up and just drag the end of the select any which way) [20:24] should have been a ? there [20:25] rmcbride: sounds similar [20:25] it likes to select a huge block of text that is nowhere near my mouse, but somehow ends where the pointer is [20:25] also, it selects a huge block of text, instead of opening links, which can be quite annoying [20:26] "You look like you're editing a block of text. Let me take another hit off my meth pipe and help you out there." [20:30] o/~ one toke over the line, sweet jesus, one toke over the line o/~ [20:30] heh [20:31] ralsina, nop..... children is an empty list [20:31] for qcheckbox [20:31] gatox: then let's go with your idea [20:31] usually those things are inside a private object inside the widget [20:31] in the c++ implementatoin [20:31] in the qt code [20:32] salgado: u1-deps revno 0.215 uploaded to hackers' PPA [20:32] wow, that was quick! :) [20:32] ok [20:32] * dobey does this peer reviws stuff [20:32] salgado: doesn't take too much time, I have it partially scripted on my end [20:33] rmcbride, cool. does tarmac take care of merging my branch as well? [20:34] gatox, I don't quite understand what you are trying to do with labels and newliness [20:34] salgado: yea tarmac did the merge [20:34] i really wish launchpad recipes could support private branches/ppas [20:35] alecu, the problem is that qcheckbox, radio and some other widgets don't have wordWrap..... so we are trying to add word wrap functionality to those widgets, but without reimplementing them and dealing with the issues of changing the internal structure [20:36] alecu, because in some translations those widgets exced the width of the window [20:36] gatox, ok so far [20:36] alecu, but, i already have it working [20:37] gatox, I just don't like on matter of principle doing manual wordwrap in the app instead of using something that the toolkit provides or should provide. [20:38] gatox, there are so many issues with i18n that can go wrong with doing things manually, like RtoL languages, and such. [20:38] alecu, in this case..... word wrap on those widgets is not provided [20:39] alecu, do you want to take a look at the function? [20:39] gatox, that I understand. But I think we should be using something provided by the toolkit, not half-assedly rewritting it ourselves. [20:39] gatox, I want, yes. [20:40] gatox, and by that I really mean: "we may write something that works for some cases, but will surely not work for every case" [20:40] alecu, we have search for something in the toolkit with ralsina, nessita and i... but the other solutions presents more problems [20:40] gatox, text layout is incredibly difficult to get *right*. [20:40] alecu: left-align-split-at-spaces is easy to get right [20:41] alecu: it's just ugly that's all ;-) [20:41] ralsina, this is not an exaustive list: "what about RTL languages?" [20:42] alecu: doesn't matter in this case [20:42] alecu, ralsina this is the implementatoin with a working example: v [20:42] http://paste.ubuntu.com/925469/ [20:42] alecu: what this does is split the string and insert \n, if it's RTL it still has to fit in the same width [20:42] alecu: so it doesn't matter [20:43] alecu, gatox: you can try LTR by adding -inverted to the CLI BTW [20:43] or something similar [20:43] gatox, what was the bug for this, again? [20:43] alecu, https://bugs.launchpad.net/ubuntuone-control-panel/+bug/973702 [20:43] Ubuntu bug 973702 in Ubuntu One Control Panel trunk "The "Settings" panel should have its items set to wrap" [High,Triaged] [20:45] gatox, ralsina: is each checkbox in those screenshots only one widget? Why can't we use a standard label for the right part of them? [20:46] alecu: focus issues, among others [20:46] gatox, ralsina: also, that screenshot is from a VM; do we care for screens with a width < 1024? [20:46] alecu: yes, our target is 800x600 [20:46] alecu: which is why we went to great pains to be 525px tall [20:47] alecu, focus issues with the stylesheet...... not reacting to the click on the text unless we extend the qlabel or add an event filter..... a lot of more code, and more issues [20:50] On the "Syncing your computer with the cloud" page i just got...AttributeError: "'NoneType' object has no attribute 'stop'" [20:50] alecu, i need to go now.... i'm not working tomorrow, but i can step here for a while so we can discuss this a little bit more [20:50] ok, gotta go be an uncle for a bit [20:50] ralsina, byeeee [20:51] I will be back tonight, so feel free to ask things by email [20:51] gatox, ok, I agree that changing this is a lot more complex. [20:51] gatox, so, let's do it like you propose, but let's open a bug to fix it in a more elegant way: [20:51] gatox, like it's proposed on the first screenshot here: http://stackoverflow.com/questions/1839194/qcheckbox-qradiobutton-line-wrap-qt4-6-0 [20:52] gatox, that is, having a label with a smaller font and the longer description, below the checkbox with a shorter text. [20:52] i don't think i'll have these done in the next 8 minutes :-/ [20:52] alecu, we analyze those options.... and one of them is wrong.... and the other one presents another issues in our case [20:52] gatox, which one is wrong? [20:53] alecu, the option you mention would be the proper one [20:53] gatox, and what are the issues that it presents? [20:53] alecu, the one about adding a buddy..... that one is wrong.... don't apply for our case [20:53] right. [20:53] because adding a buddy means that we get the same focus and hover issues [20:54] gatox, but the other option looks much better. I guess the problem would be with doing all the translations again. [20:54] alecu, among other problems....... the one you mention, that one will the best choice.... but we will need more height for all the cases [20:55] gatox, yes: in all cases we will need more height [20:55] alecu, ok..... gotta go..... do you want to have a quick mumble tomorrow for this?? or just implement my proposal and open the issue? [20:55] gatox, even if we manually add the \n, right? [20:55] alecu, yes [20:55] gatox, and for some languages we'll surely need scrollbars... [20:56] gatox, yes, go ahead, make your branch. we can still mumble tomorrow. [20:56] alecu, what i mean..... when we don't need the \n.... we are going to be fine with the height..... [20:56] alecu, also.... we could add everything inside a QScrollAarea for those cases..... i don't know which is the design opinion about this..... [20:56] ok..... let's talk tomorrow..... [20:56] see you alecu!! o/ [20:57] gatox, lets talk tomorrow, right. bybye [20:58] alecu: the idea of having a "title" doesn't work, because eventually someone is going to do a translation of the title that is too long [20:59] alecu: the texts may need to be made shorter in english so translations don't get ridiculous, too. Like, "sync all folders" [20:59] instead of "automatically sync all folders shared with me on this computer" [20:59] which is way too long even in english [21:00] Even just removing "automatically" everywhere would be a vast improvement [21:00] ralsina, yes, that's the kind of solution I like. [21:00] Or using "Kbps" instead of "Kilobits per second" [21:01] alecu: so, let's do this gatox is suggesting, and let's ask for better copy for the next version [21:01] ralsina, and not going full Knuth ahead and doing all the text layout. [21:01] which we can't do now :-) [21:01] ralsina, ok, awesome. [21:01] alecu: yes, no hyphenation, no stretchy interlettering.... sigh === yofel_ is now known as yofel [21:03] or maybe we use TeX markup in each Qt label, and add the page number to each tab of the dialog! [21:04] [Folders] [Devices] [Settings 1] [Settings 2] [Account information] [21:05] See Devices[1] [21:06] ok, really off now [21:07] * alecu can't wait for next year's rewrite of control panel in go+webkit. [21:11] i don't think so [21:16] ok. i'm generally off, but will be around at some point tonight, doing peer reviews [21:16] later all === m_conley is now known as m_conley_away === thisfred_ is now known as thisfred [21:46] anyone know the process to signing the windows binaries? "gpg --sign file" or "gpg --detach-sig file"? (gpg newbie) [21:49] alecu: have you ever had to do this? ^ [21:56] ralsina: when you get back, can you mail me how the signing should be done? i found a wiki page that explains some of it, but i'm still not sure about the usage of GPG. also, not sure if i should sign each file or just the zip file of EXEs+QtCore4.dll === salgado is now known as salgado-afk [22:32] briancurtin2: just the zip file so IS knows the files are not tampered. Use --detach-sig. [22:32] ralsina: ah cool, doing that now [22:41] ralsina: signed and uploaded to the ticket [22:41] briancurtin2: awesome