/srv/irclogs.ubuntu.com/2012/05/22/#ubuntuone.txt

trubborAre any UbuntuOne admins on this chat?07:05
mandelralsina, can be done, I usually do reviews in the morning when I'm lazy, is a good way to start the day :)08:27
=== yofel_ is now known as yofel
gatoxgood morning!11:04
mandelgatox, morning!11:15
gatoxmandel, hi11:15
mandelgatox, how is everything going?11:19
gatoxmandel, fine..... a little bit tired (i don't know why).... but fine..... you?11:19
mandelgatox, ok, getting the ipc to domain sockets is showing some error in the way the tests were written, so fixing it11:20
mandelgatox, dirty reactors, dirty reactors everywhere!11:21
gatoxmandel, yep..... yesterday before my sick leave..... i was trying to fix the fsevents tests..... obviusly there are some problems inn the way i implemented (because of the way the macfsevents lib works)....... but now with the tests it's easier to see what should do and start changing the code11:22
mandelgatox, I'm glad we have all this tests.. but is a PITA to have to run them.. in my machine the u1-client ones are really slow11:24
gatoxmandel, ahhhh yes..... but for it's quick because in this moment i'm only running the tests for filesystem_notifications in mac11:24
gatoxs/for it's/for me it's11:24
mandelgatox, yeah, I'm running all, it seems that we are doing something bad in the tests because when we use tcp as the transport things work, if we change it to domain sockets it times out..11:25
gatox:S11:26
mandelgatox, and is not an easy thing to debug, but once done we will have the perpective-broker tests running over tcp and domain sockets on linux11:26
ryealecu: please ping me when you are back11:39
alecuhello, all!12:06
alecurye, pingback12:06
gatoxalecu, hi12:07
mandelalecu, hello!12:10
mandelgatox, alecu, can I have a very simple review for a bug I found in u1-dev-tools, you can find the MP here: https://code.launchpad.net/~mandel/ubuntuone-dev-tools/correct-cleanup/+merge/10679112:10
gatoxmandel, ack12:10
mandelalecu, and, was it you the one that took all those legos? :P12:10
mandelgatox, by the way, that fixes the problem I was having with domain sockets.. and is my fault, stupid mandel!!12:14
ryealecu: i tried running tests within the trunk branch on precise and they don't actually pass12:15
ryealecu: http://paste.ubuntu.com/1000703/12:15
aleculooking12:15
ryealecu: that's with the IN_CLOSE_WRITE filter12:16
ryere-runnign with real branch12:16
ryemandel: also, where do you get skipIfJenkins for trunk?12:17
mandelrye, from ubuntuone-dev-tools12:17
mandelrye, should be in trunk12:17
ryemandel: quantal or precise? no such thing in precise version12:17
ryeah12:17
mandelrye, trunk as in nightlies12:18
gatoxmandel, +112:20
ralsinagood morning!12:20
mandelgatox, sweet, thx!12:20
alecuhola ralsina!12:20
mandelralsina, morning! so, what do we do about my review days?12:20
gatoxralsina, hi12:20
ralsinamandel: nah, let's keep it as it is, since you *are* doing reviews on monday mornings12:21
ralsinamandel: I just had not seen it12:21
ralsinaalecu: we are getting urther comments on the proxy bug: a windows user claiming .pac isnot being parsed12:21
mandelralsina, well, mondays are quite calm, I can also do some on tuesday mornings to add some extra review eyes there12:21
ralsinamandel: awesome, since tuesdays it's mmcc which can use a hand, and I have way too many calls12:22
mandelralsina, sorted then, mondays and tuesdays for me, I'll update the calendar12:22
ralsinamandel: awesome12:23
alecuralsina, we don't support .pac, right.12:25
mandelalecu, can you take a look at https://code.launchpad.net/~mikemc/ubuntu-sso-client/no-darwin-in-linuxnwindows-tests/+merge/10592412:25
alecumandel, sure12:25
ralsinaalecu: I thought I recall that on windows we do? Or maybe we need Qt 4.8 for that?12:25
alecuralsina, the Qt docs claim it's supported, but I've never tested it myself12:26
mandelalecu, we also need your review for https://code.launchpad.net/~mikemc/ubuntu-sso-client/fix-networkstate-darwin/+merge/105926 so that michels branches land :)12:26
ralsinaalecu: ok12:26
alecuralsina, and I'm waiting to see if more people request that feature, before working on it.12:27
alecuralsina, also, it's very likely that QNetworkAccessManager supports it (when downloading urls for webservice calls) but the low level qsockets do not support them automatically (as used by the proxy tunnel)12:29
alecuralsina, plus we won't be able to support it on linux and mac (other than interpreting the .pac ourselves)12:30
alecuralsina, so.... I'm just tempted to say "we don't do .pacs"12:30
alecuralsina, the user can always download the .pac and set the proxy values manually. (unless the .pac is complex and has weird rules)12:31
ralsinaalecu: ok, so we may have to do a manual config. dialog instead12:31
alecu(one of our windows users already did this)12:31
alecuralsina, by "manually" I meant on the system control panel.12:32
alecuralsina, but yes, that means that we'll be "unautomatizing" other apps....12:32
ralsinaalecu: yes, but that may break things in the case of complex .pacs right12:32
ralsinaso, ok, let12:33
mandelok, lunch for me12:33
ralsinaoops12:33
* mandel lunch12:33
ralsinalet's wait and see12:33
alecuralsina, right. We knew from the start that we would not be able to support *every* weird proxy use case. And as supporting this in a sane way means including a js parser, I'm tempted to push this forward for some time.12:34
alecu*js interpreter12:34
ralsinaalecu: yeah12:34
* mandel back13:42
ralsinadobey: 1920x1080 on a 11" ultrabook sounds nice? http://www.engadget.com/2012/05/22/asus-zenbook-prime/13:43
ralsinadobey: I'm guessing that's close to 180 dpi13:44
dobeybah, that url redirected to /13:46
ralsinadobey: it's exactly 200dpi13:48
ralsinabut it's IPS so the real DPI is slightly lower because it's pentile13:49
dobeyit's 189 dpi; presuming it's actually 11.6"13:49
ralsinaif it's 11" that's 9.6*5.4 which is exactly 200 though13:50
ralsinabut yeah, it depends on it being really 11". If it is, I want one :-)13:51
dobeywould be much better if it was 8.9" though13:51
dobeybut that resolution at 11" is at least starting to be respectable13:51
ralsinadobey: and cheap, really, about 900 for a loaded notebook13:51
ryealecu: re: https://code.launchpad.net/~alecu/ubuntuone-client/ignore-in-close-write-in-dirs - have your test run completed successfully?13:55
briancurtinralsina: FYI i sent the binaries in for signing yesterday afternoon...hopefully they come back today13:59
ralsinabriancurtin: cool, thx13:59
alecurye, yes, my test run worked just fine. I'm updating precise to see if I missed anything.14:03
alecurye, can you also try running the tests on an up to date trunk?14:03
alecumandel, why the "return defer.succeed(None)" in your branch?14:07
mandelalecu, I want to be consistent and always return a deferred (cleanup does return one) rather than return None sometimes and a defer other14:08
alecumandel, addCleanup *optionally* takes a deferred. If one is returned, it waits for it to complete before going to the next text.14:09
alecumandel, so that's why I think that line is redundant.14:10
mandelalecu, I can happily remove it, but the branch was already merge, I can me a new one if you want, not too much work14:10
alecumandel, it's part of addCleanup interface for that deferred to be optional.14:10
alecumandel, oh, right: it was merged. I didn't see eric's review.14:11
alecumandel, don't worry then.14:11
alecumandel, I got the tab opened to review laters :-)14:11
mandelalecu, are you sure you don't want me to remove it, is a second and with a +1 from you we can merge it14:14
mandelalecu, less code == less bugs and if it is redundant.. I have no problem in doing so14:14
dobeybah14:14
dobeynobody has tried my fix for the installer bug :(14:15
ralsinaduanedesign: I am clueless about what may be happening to http://askubuntu.com/questions/123580/ubuntu-one-sign-in-with-windows-client-fails14:15
ralsinadobey: well, I tried it but it does not fail for me without it either14:16
alecumandel, you've got my +1 if you want to remove it.14:17
mandelalecu, ok, will do then :)14:18
dobeyralsina: yes, but i commented on the bug with a link to the nightlies deb, asking people who are getting the problem, to try it14:18
ralsinadobey: ack14:18
dobeyhow is it that so many people are having problems with dbus giving "connection refused" from ubuntu-sso-client, as well.14:20
dobeyand not say, all the other multitude of things using dbus14:20
mandelalecu, here is the MP without that line: https://code.launchpad.net/~mandel/ubuntuone-dev-tools/remove-redundancy/+merge/10681614:21
mandelwow.. I have not had to do a bzr brake-lock in ages.. weird14:22
dobeybreak14:22
dobeybrakes are the things in your car14:22
mandeldobey, sorry, I always confuse them..14:23
mandelstupid brain14:23
briancurtinmandel: which branch(es) do you have which fix IPC on windows?14:26
mandelbriancurtin, I'm about to propose them, I had to fight a little with the reactor being left dirty14:27
briancurtinmandel: everyone's enemy14:27
mandelbriancurtin, are you block by them?14:27
briancurtinmandel: well i was just looking into why SD isn't starting up yesterday afternoon, and i think it's what you're working on14:27
briancurtinmandel: i'll gladly review and test the branch14:28
briancurtinor branches14:28
mandelbriancurtin, ok, will ping you as soon as I propose them14:29
mmccmorning folks, catching up14:30
ralsinamandel: just read that you now have to ask for government permission to broadcast sound or video over the internet in cataluña. You guys are weird.14:30
* mmcc wouldn't mind regulation of broadcasting blog comments over the internet14:31
mandelralsina, the country is really going down to hell, seriously, so far 5 of my friends left already and I have been warned that I might get into trouble do to the contract with canonical (I'm looking at what they mean)14:31
ralsinamandel: yeah, because bringing money into spain is something they really can't stand, I guess14:32
mandelralsina, they don't like that you have a contract with a company as a freelancer for an unlimited amount of time (they want their cut)14:33
mandelanyways, not the place to talk about this things14:34
dobeymandel: time to move to BsAs! :P14:35
mandelthisfred, after talking with alecu I removed a line, can you take a look: https://code.launchpad.net/~mandel/ubuntuone-dev-tools/remove-redundancy/+merge/10681614:37
mandeldobey, not just yet.. but looks like a possible outcome14:37
thisfredmandel: yeah that makes sense, I guess. +114:38
mandelthisfred, thx!14:40
alecumandel, +114:47
mandelgatox, I've got a branch that gives us ipc on mac for ui1-client (also fixes the broken tests on windows) :)14:47
mandelalecu, superb, thx!14:47
mandelalecu, you still have to tell me if you are the SAP executive that stole all those legos.. silence will interpreted as a yes :)14:48
gatoxmandel, awesomeeeeeeee14:48
gatoxmandel, i'm still working in the fsevents branch.... some events seems no to be properly processed14:49
mandelgatox, no worries, is a very complicated part of the project I'd be surprise if you had no problems :)14:49
gatoxmandel, jeje thx14:50
* alecu hits mute on his irc client. // cc: mandel14:51
mandellol14:51
alecumandel, I'm not getting caught just like that!14:51
alecumandel, and in fact, that dude looks like "Benjamin Linus" from Lost14:51
* alecu ought to get a better chair14:52
mandelalecu, he does look like him a lot!14:52
alecumandel, so, I fear that the Legos are a facade for a much more darker plan involving islands and time travel14:52
mandelalecu, if you have to do a time travel machine I think legos are well suited for the first version14:53
ralsinadobey, alecu, mandel, gatox, thisfred, briancurtin, mmcc: standup in 6'14:54
gatoxralsina, ack14:54
alecumandel, well, not if you are travelling before 1949... you need wooden blocks in that case.14:54
thisfredack14:56
mandelgatox, can you run the tests for lp:~mandel/ubuntuone-client/domain-sockets in you evil always failing manuel hating windows vm?14:56
gatoxmandel, LOL ok!14:56
mandelalecu, gatox, mmcc, we have the fsevents chat after the standup, is that correct?14:57
gatoxmandel, yes14:57
mmccyes mandel14:57
alecumandel, try to limit the machines that hate you to "virtual". Remember what happened to Sara Connor!14:57
mandelalecu, hahahaha14:58
gatoxjejeje14:58
gatoxmandel, your branch has some conflicts with trunk14:59
* gatox always execute merging with trunk14:59
gatoxmandel, no problem..... the conflicts resolve automatically with bzr resolve15:00
ralsiname15:00
gatoxme15:00
alecume15:00
mandelgatox,  great, I'll merge before I propose the merge15:00
mandelme15:00
mmccme15:00
briancurtinme15:01
ralsiname go!15:01
ralsinaDONE: reviews, minor progress in a bug, mgmt call, tech leads call, helped around, bled on my notebook (no, I can't get over it yet), I think I have my sinuses filled with rattlesnakes while I was asleep, pain TODO: go to the rattlesnake doctor, try to do stuff BLOCKED: by the previously mentioned rattlesnakes NEXT: gatox15:01
gatoxDONE:15:01
gatoxSick leave, keep debugging fsevents on mac, some events seems to not be process or throw properly.15:01
gatoxTODO:15:01
gatoxKeep fixing fsevents on mac.15:01
gatoxBLOCKED:15:01
gatoxNo15:01
gatoxmandel, go15:01
alecuDONE: 99% finished with sec patches15:02
alecuTODO: os-ten fs notifications mumble, finish above, bug sorting in u1-client15:02
alecuBLOCKED: no15:02
alecuPREV: time traveler mandel15:02
mandelDONE: Fixed bug 1002208 bug 1001296 and bug 1002834.15:02
mandelTODO: Muble chat about fsevents. Propose branch for ipc over domain sockets on u1-client. Look at cp and provide domain socket support. Move back to fsevents.15:02
mandelBLOCKED: no15:02
mandelmmcc, please15:02
ubot5Launchpad bug 1002208 in Ubuntu One Client "Darwin is using wrong transport" [Medium,In progress] https://launchpad.net/bugs/100220815:02
ubot5Launchpad bug 1001296 in Ubuntu One Client "Tests are broken on Windows" [High,In progress] https://launchpad.net/bugs/100129615:02
ubot5Launchpad bug 1002834 in ubuntuone-dev-tools "Not creating a client in a ServerTestCase results in dirty reactor error" [Medium,Fix committed] https://launchpad.net/bugs/100283415:02
mmccDONE: UX & App Store discussion, review, looking into py2app15:02
mmccTODO: review day, more py2app & pyinstaller15:02
mmccBLCK: NO15:02
mmccNEXT: briancurtin15:02
briancurtinDONE: installer created from release-3_0_1, installer tested, binaries sent for signing, looked into why SD not working (but i think it's what mandel's branch fixes)15:02
briancurtinTODO: get installer signed if binaries come back today, get SD working if the branch doesn't fix it15:02
briancurtinBLOCKED: None15:02
briancurtinNEXT: i think we're done15:02
gatoxyap15:03
ralsinaok, comments?15:03
gatoxnop15:03
gatoxalecu, mandel mmcc i'm already in mumble15:04
mandelgoing15:04
joshuahooverbriancurtin, ralsina: do we need to "nudge" IS on signing the binaries?15:04
ralsinajoshuahoover: can't hurt15:04
dobeyoh15:04
dobeyme15:04
ralsinajoshuahoover: at least get the standard "top of our queue!" response15:04
dobeyλ DONE: finished bug #853060 (hopefully), SRU uploads15:04
dobeyλ TODO: finish 3.0.1 SRU uploads, find someone to test installer fix15:04
dobeyλ BLCK: none.15:04
ubot5Launchpad bug 853060 in ubuntuone-installer (Ubuntu Precise) "ubuntuone-installer crashed with GError in function(): Failed to execute child process "ubuntuone-control-panel-gtk" (No such file or directory)" [High,Triaged] https://launchpad.net/bugs/85306015:04
joshuahooverralsina: do you mind doing that?15:05
mandelgatox, in that conflict I'm removing the tests/platform/ipc/test_darwin.py because /tests/platform/ipc/test_perspective_broker.py15:05
ralsinajoshuahoover: not at all, will do15:05
joshuahooverthanks!15:05
mandelgatox, and I think we can do the same with tools because the only thing that should change in the transport15:05
mandelbriancurtin, I have proposed a branch that fixes the windows tests: https://code.launchpad.net/~mandel/ubuntuone-client/domain-sockets/+merge/10683315:25
briancurtinmandel: cool, i'll take a look15:25
dobeybbiab, lunch time15:33
mandelthisfred, if you fancy an interesting review: https://code.launchpad.net/~mandel/ubuntuone-client/domain-sockets/+merge/10683315:36
mandelralsina, the jenkins slave is misbehaving :(15:36
ralsinamandel: whip it!15:37
mandelralsina, we are on it15:37
ralsinamandel: misbehaving how?15:37
thisfredwhip it good!15:37
ralsinathisfred: hi5! ;-)15:37
thisfredoh and: me15:37
mandelralsina, https://jenkins.errormessaging.com/view/Windows/job/ubuntuone-dev-tools-windows-test/11/console15:37
thisfredDONE: found and fixed a few memory leaks in u1db TODO: change u1db API so document.content is a dictlike object BLOCKED: no NEXT:15:37
thisfredsry got drawn into u1db discussion15:38
gatoxalecu, https://code.launchpad.net/~diegosarmentero/ubuntuone-client/darwin-fsevents15:54
alecugatox: awesome!16:06
mandelgatox, I believe I have fixed bug 997663 by accident16:18
ubot5Launchpad bug 997663 in Ubuntu One Client "platform/ipc imports are failing in MAC OS" [High,Triaged] https://launchpad.net/bugs/99766316:18
mandelgatox, not accident, but side effect of another branch16:18
gatoxmandel, cool.... do you want to assign that yo you?16:18
mandelgatox, sure16:18
mandelgatox, we need to merge the tools.py for darwin and windows so that we have ipc for control panel, want me to do that?16:19
gatoxmandel, merge how?16:20
briancurtinmandel: in the description of the domain-sockets branch, it says "using at least revno 73"...of what project?16:21
mandelgatox, they both use perspective broker, the real diff is that the transport is different, this means that we can use the exact same code and we are done :)16:21
mandelbriancurtin, sorry, ubuntuone-dev-tools16:21
mandelbriancurtin, I'll update the description16:21
gatoxmandel, ahhhhh so yes :P16:21
mandelgatox, if you look at https://code.launchpad.net/~mandel/ubuntuone-client/domain-sockets/+merge/106833 you can see is the exact same code :)16:22
gatoxmandel, ack16:22
mandelgatox, so, 2 birds with one stone!16:22
gatoxawesom16:22
mandelgatox, I'll take care of bug 100299416:24
ubot5Launchpad bug 1002994 in Ubuntu One Client "Tools implemenation is missing on darwin" [High,Triaged] https://launchpad.net/bugs/100299416:24
mandelgatox, should be very easy and is a matter of changing the tests and also make them run on linux like the ipc ones do in my proposed branch16:25
gatoxmandel, great...... let me know for review time!16:25
gatoxmandel, so i can take a better look at it and see how it works16:25
mandelgatox, I need you to run the tests for the other first hehehe16:25
mandelgatox, super easy: https://code.launchpad.net/~mandel/ubuntuone-client/domain-sockets/+merge/10683316:26
thisfredmanuel: should we be using real paths in tests? Or does this not actually write to the filesystem?: TEST_DOMAIN_SOCKET = os.path.join(basedir.xdg_cache_home, 'ubuntuone', 'ipc')16:26
gatoxmandel, i forgot to tell you that the tests for that one are failing really bad on windows16:26
mandelgatox, really? may I see?16:26
gatoxmandel, saving the output to a file16:27
mandelthisfred, is does not write in the file system, basedir is patch to write under _trial_tests16:27
mandelthisfred, is some magic added by u1trial16:27
thisfredmandel: excellent16:27
dobeyit's not patched, but the env var is set; but yeah, you can use the stuff from dirspec directly16:31
mandeldobey, yes, could not care enough to explain it properly (on a phone call)16:33
mandelgatox, hey, I'm leaving give me the output! :P16:33
gatoxmandel, wait a sec........ f#$%ing vm16:34
thisfredmandel: +1 with two small remarks16:34
mandelgatox, I expected more unicode in that word coming form you!16:34
* alecu is starving too16:34
mandelthisfred, cool, let me check16:34
gatoxalecu, starving! always forget16:35
alecugatox ;-)16:35
gatoxmandel, http://paste.ubuntu.com/1001185/16:35
gatoxmandel, jajajaa16:35
gatoxok..... lunch for me!16:35
=== gatox is now known as gatox_lunch
mandelgatox_lunch, looks like it was broke when merging with trunk16:36
mandelgatox_lunch, pull and run the tests I did the merge manually and pushed it16:37
mandelgatox_lunch, for example: http://bazaar.launchpad.net/~mandel/ubuntuone-client/domain-sockets/view/head:/tests/platform/ipc/test_windows.py16:39
briancurtinmandel: domain-sockets tests pass, now for an actual review16:42
mandelok, EOD for me16:55
* mandel goes to get hurt at rugby16:56
* mmcc lunch17:20
=== gatox_lunch is now known as gatox
ralsinathisfred: I am asking you to answer bug #1003028 because my head hurts17:42
ubot5Launchpad bug 1003028 in desktopcouch (Ubuntu) "Frequent and sudden logouts after desktopcouch update" [Undecided,New] https://launchpad.net/bugs/100302817:42
ralsinaalso, I am mean17:42
thisfredralsina, can I answer "won't fix, project abandoned" ?17:42
ralsinathisfred: you could answer 'why the heck are you installing that?'17:42
thisfredhow did desktopcouch even receive an update?17:43
dobeyhuh?17:45
dobeyoh i fixed a bug in it17:45
thisfredralsina: I have trouble believing desktopcouch is causing logouts, but nothing is impossible I guess. There is some strange code in there17:45
dobeyso this guy is probably on oneiric17:45
thisfreddobey: yep, but he says  someone else is seeing it on precise17:46
dobeyhowever, the update simply added an apport file17:46
thisfredwhy are we backporting anything wrt desktopcouch?17:46
dobeyto the package, as it was missing though being installed by setup.py install17:46
dobeywe aren't17:46
thisfredok17:46
dobeyi doubt it's desktopcouch though17:46
thisfredso then there must be something else that caused this right? Unless it's apport + dc17:47
ralsinasounds like OOM killer17:47
ralsinaas in "the guy has bindwood with 8GB of data and desktopcouch is killing his session" or something17:47
thisfredralsina: either way, I have no idea, sounds like dobey may be better suited to answer17:47
dobeyit's say "INCOMPLETE"17:47
ralsinahe17:47
ralsinawe are supposed to *ask* something before marking incomplete17:48
ralsina"what's your zodiac sign" -> incomplete17:48
thisfred"come here often?"17:48
dobeyi wonder what all actually uses desktopcouch, in the archive17:53
dobeyralsina: replied/incompleted17:53
ralsinadobey: thanks17:53
dobeyoh right, holiday17:57
ralsinadobey: +1 on the holiday18:01
alecurye, I've ran the tests on a fully-up-to-date precise, and they all pass.18:02
dobeyhrmm18:03
dobeyhmm18:20
ralsinaduanedesign: if you are ever so kind... https://bugs.launchpad.net/ubuntu/+source/ubuntuone-client/+bug/100299018:22
ubot5Ubuntu bug 1002990 in ubuntuone-client (Ubuntu) "when backing up on U1 I get:Got status code 500" [Undecided,New]18:22
ralsinaduanedesign: because I have no idea what he's talking about18:22
dobeygrr, clouds18:33
alecugatox, ping18:33
gatoxalecu, pong18:34
alecugatox: I've taken a look at the filesystem notifications branch you are doing for darwin, and it's looking very good.18:34
alecugatox: one thing that I'm not sure I'm understanding is how the events are being put back into the reactor main thread.18:35
gatoxalecu, great..... i was a little worry about some things that i didn't know very well...... but i'm understanding them better with the tests18:35
gatoxalecu, that is what i'm testing right now..... because i think that the problem i'm having in this moment is exactly that!18:36
gatoxthe problem about the events not going into the processor18:36
alecugatox: so, "fsevents.Stream(self._process_events....)" says that _process_events will handle all the events, right?18:37
alecugatox: but what thread is calling that method?18:37
gatoxalecu, yes...... but now i see that this is being called inside the macfsevents thread....... which is not what i want18:37
alecugatox, right: it must be ran inside the thread started by self.observer.start()18:38
alecugatox: it should be called inside the twisted main thread instead.18:38
gatoxalecu, yes.... i'll try to tweak that a little...... but that EXACTLY the problem i'm having right now :P18:39
alecugatox: so: _process_events should be renamed _process_events_in_main_thread, and in _process_events there should be just a call to "reactor.callFromThread" to the full method.18:39
dobeymmcc: care to review https://code.launchpad.net/~dobey/poauth/test-runner/+merge/105874 ?18:40
alecugatox: does that make sense?18:40
gatoxalecu, and that will call a function inside the reactor loop from another thread?18:40
mmccback18:42
mmccdobey, sure18:42
alecugatox: if you call that from any thread, it will safely store your function and your args in a reactor queue, and when the reactor is idle it will call your function.18:42
alecugatox, so, yes :-)18:42
gatoxalecu, awesome! that's the twisted magic i was needing!18:43
alecugatox: reactor.callFromThread(yourfunction, *args, **kwargs)18:43
alecugatox: that's the only safe method in the whole of twisted that you can call from a different thread.18:43
gatoxalecu, great!! i think that maybe that can fix some of the tests that are broken!! THX alecu!18:44
dobeyhow the heck am i supposed to find someone to test this fix18:46
ralsinadobey: want me to lie?18:46
* ralsina woul lie for dobey18:46
dobeys/for/to/ ?18:47
ralsina"Not only does this ix it, my computer now knows how to make belgian waffles"18:47
ralsinafeel free to quote me!18:47
dobeybring me a waffle and a trippel karmeleit, then we can talk18:48
ralsinadobey: the waffle is in the mail!18:48
dobeytoo bad customs in argentina will eat it and forward an empty box18:50
briancurtinif alfajores would make it here, i'd buy some right now18:52
dobeyindeed18:52
ralsinaI am fairly sure there has to be a place in Chicago that sells alfajores18:52
dobeyi would fill an atlas f silo with them18:52
ralsinaalthough it seems not: http://articles.chicagotribune.com/2004-01-09/entertainment/0401090415_1_cookie-argentina-uruguay18:53
dobeythe alfajores must be protected from nuclear attack18:53
briancurtinralsina: there's "buenos aires deli" here that supposedly has good ones along with empanadas, but every time i remember it, it's already closed for the day18:53
briancurtinit was on TV a few times for being pretty authentic, i guess18:54
ralsinait seems their specialty is maizena... ok18:54
dobeyheh18:54
ralsinahttp://lateral.netmanagers.com.ar/weblog/posts/BB985.html18:54
ralsinabriancurtin: http://www.facebook.com/LucilasHomemade18:55
ralsinaNo excuses, dude, you live in a big city :-)18:55
briancurtinoh man i forgot about that place18:56
dobeyand not only a big city; but one renowned for illegal immigrants and their connections to illegal activities18:56
briancurtinive been too busy making treks around the city for italian beef sandwiches and hot dogs...need to get back on the trail of argentine stuff18:57
ralsinahaha18:57
dobeyargentine, italian, they're pretty well connected18:57
mmccbriancurtin: how far do you really have to trek in Chicago for italian beef sandwiches? Don't they come out of the faucets there?18:57
ralsinadobey: being the grandson of Enriquetta Doffo and Umberto Primo, I endorse that statement18:57
dobeyheh18:58
briancurtinmmcc: they're everywhere, but i'll drive around all day just finding new places and "have to try them". good beef places here are like starbucks in the suburbs - everywhere18:58
mmccbriancurtin: ah, understood :) it's like BBQ here. hole in the wall joints in strip malls can be amazing18:59
ralsinaAnd if you guys want the real thing: online havanna alfajores: http://www.amigofoods.com/haalmi12al.html18:59
ralsinabriancurtin: that happened to me in Istanbul, I tried a different kebab place every day for two months18:59
ralsinabriancurtin: always looking for the mythical perfect kebab19:00
mmccthis food talk is distracting, I had oatmeal for lunch, after oatmeal for breakfast.19:00
ralsinaOTOH, 40 dollars for a box... I am taking some with me if I ever go back to the US and selling them in the black market19:01
* mmcc found in Austin: http://www.buenosairescafe.com/19:01
ralsinammcc: a distinct upgrade from oatmeal, guaranteed19:02
* ralsina measures the suitcase, he can probably fill it with alfajores and bring back two ultrabooks from the profit19:02
mmccyeah, it's been oatmeal and rice since saturday. dying to get some grease19:03
briancurtinthat's too long. the pre-game meeting for my series over the weekend was at a hot dog stand each time #america19:03
* briancurtin will never use a hash tag again in IRC19:04
* mmcc dug it19:05
ralsinathisfred, dobey: desktopcouch found not guilty19:05
dobeyof course19:06
ralsinaoh, good, doctor's here. I'm gonna get me some prescription!19:12
* alecu runs to kinder19:12
=== gatox_ is now known as gatox_macos
* briancurtin brb, late lunch19:20
thisfredralsina: I saw19:26
thisfredwhew19:26
ralsinanot only do I have access to doctors that do house calls, but he gave me free medicine samples. Socialism works dudes.19:28
ralsinaok, not really socialism, I am paying for this, but still, better than paying U$S 900 to be told "your kid has a fever, give him ibuprofen"19:28
dobeyralsina, alecu: i think we can mark https://bugs.launchpad.net/ubuntu-sso-client/+bug/617353 as wontfix no?19:30
ubot5Ubuntu bug 617353 in Ubuntu Single Sign On Client "The network state is not checked with ConnMan" [Medium,Triaged]19:30
ralsinadobey: yep19:30
ralsinadobey: mention that now, if NM is not installed, we assume there is network19:31
ralsinaconnectivity19:31
dobeyeh, alecu filed the bug :)19:31
dobeyi think he knows19:31
ralsinahaha19:32
joshuahooverralsina: have a windows proxy user behind an http proxy with no auth who can't connect, i attached logs to bug #100308519:34
ubot5Error: Launchpad bug 1003085 could not be found19:34
joshuahooverright, because it's private... http://pad.lv/100308519:34
ralsinajoshuahoover: ooooook, alecu? ^19:35
ralsinajoshuahoover: I get a 404 there, too19:35
* alecu is back19:35
joshuahooverralsina: i had to subscribe u1...can you try again?19:35
ralsinajoshuahoover: found it searching for the bug #19:35
* joshuahoover needs to figure out what the new lp bug privacy settings mean/do19:36
ralsinajoshuahoover: private means 'post in facebook wall, set to private'19:36
joshuahooverahhhh...makes sense19:36
ralsinabut yeah, me too19:37
dobeywhat do you mean "new" lp bug privacy settings?19:38
ralsinaalecu: looks like this is the "computer has no DNS whatsoever" case19:38
joshuahooverdobey: try to create a new bug and you'll get options for who can see it...loads of fun19:38
ralsinaFailure: twisted.internet.error.DNSLookupError: DNS lookup failed: address 'm.root-servers.net' not found: [Errno 11004] getaddrinfo failed.19:38
alecuralsina, have we seen that before?19:39
ralsinaalecu: we imagined it may cause trouble ;-)19:39
ralsinaalecu: never tested it though, AFAIK19:39
dobeywhat the heck19:39
dobeythos options make no sense19:39
ralsinadobey: thus the need to learn them19:40
dobeyhah19:40
dobeyand the "this is a security issue" thing is gone19:40
joshuahooverdobey: i knew you'd appreciate them19:40
ralsinathings that make sense, we just guess19:40
alecuralsina, the weird thing I see is that the proxy tunnel is started, but the SD logs then say: "Connection started to host fs-1.one.ubuntu.com, port 443"19:40
alecuralsina, so it's royally ignoring the tunnel.19:40
dobeyralsina: i guess we all need to have laweyers to answer those questions19:40
alecuthe tunnel sees no activity at all (other than being started).19:40
ralsinaalecu: could be a race condition if the proxy tunnel starts VERY slowly19:41
ralsina?19:41
ralsinaor crashing very badly19:41
dobeyjoshuahoover: looking at those options, i think we probably just always pick "proprietary" for the cases where we need to manually file private bugs that aren't private by default19:42
alecuralsina, "VERY" slowly may trigger a timeout in SD.19:43
joshuahooverdobey: yeah, i think you're right19:43
alecuralsina, there's a timeout in the code that starts the tunnel, in fact.19:43
ralsinaalecu: yeah, but that would mean, like 30 seconds to start19:43
alecuralsina, and VERY slow seems likely when DNS is broken.19:43
ralsinaright19:43
joshuahooverralsina: not that you don't already know this, but win 3.0.1 can't come fast enough...drowning in requests about it19:43
ralsinajoshuahoover: pinged IS, no response yet19:44
joshuahooverralsina: i figured19:44
ralsinabriancurtin: can you ping them in #is ?19:44
briancurtinwill do right now19:44
ralsinabriancurtin: thanks19:44
joshuahooverthanks guys, sorry this process gets bogged down so much19:44
alecuralsina, scratch that. DNS failing is fast in that computer: all the getaddrinfo for every root dns servers took less than a second.19:45
ralsinajoshuahoover: you could give them the unsigned installer19:45
ralsinaalecu: so it has a DNS server configured that doesn't do recursive to the internet19:45
ralsinaalecu: probably an internal-only DNS server19:45
joshuahooverralsina: i could, i suppose...but i'd like to not start that trend19:46
ralsinajoshuahoover: agreed19:46
* alecu starts a windows vm19:46
dobeyoh nice19:49
dobeynow this bug i can test for19:50
ralsinagatox: can you take care of bug #987976 eventually? Consider it a low19:52
ubot5Launchpad bug 987976 in Ubuntu Single Sign On Client trunk "checkbox is partially obscured on signup screen by "By signing up..."" [Undecided,Triaged] https://launchpad.net/bugs/98797619:52
gatoxralsina, yap! i'll assign it to me19:53
dobeyhuh19:53
ralsinagatox: thanks19:53
* dobey wonders why this code was written in this way19:53
ralsinadobey: there is always a reason.19:53
mmcchey dobey, it looks like your poauth code will try to ascii-encode a unicode URL if given one. This isn't new to your patch but the diff made me look. Does this matter? see oauth.py:137, and client.py:12219:54
dobeymmcc: http requires headers be in ascii19:55
dobeymmcc: in python3 certain apis also require encoding to ascii from the normal unicode string, to be able to get a bytes object19:56
* dobey fixes this SSO bug19:57
briancurtinralsina: binaries being signed right now, should be ready shortly and i'll turn it around quickly and get them to sign the installer right away19:58
ralsinayay19:58
mmccdobey, ok I guess I've got a gap in my knowledge here - I knew headers are ASCII, but URLs can be unicode, right?19:58
ralsinajoshuahoover: ^19:58
joshuahooverthanks briancurtin and ralsina :)19:58
mmccis this library not used in situations where the URL is user input?19:59
dobeymmcc: actually, no, they can't19:59
dobeymmcc: at least, the hmac API requires ascii, re oauth.py:13720:01
dobeymmcc: and client.py:122 isn't encoding to ascii ?20:02
dobeyat least, not in my branch anyway20:02
mmccdobey, ok I get it - unicode urls are user agent niceties...20:02
mmccdobey, no it's not. I pointed to it because it wasn't encoding to utf8 like the parameters in get_normalized_parameters were20:03
ralsinammcc: no, unicode URLs should be encoded as %-encoding thus are ascii already20:03
mmccralsina: ack. a gap in my web knowledge :)20:04
dobeymmcc: ah, not entirely sure why it's doing that. but it was like that before20:04
dobeymmcc: if it's a bug, can fix it later :)20:04
ralsinaIRIs can contain unicode, which should be utf-8 encoded and THEN %-encoded20:05
mmccdobey: yeah that threw me a bit. I doubt it's a correctness bug - if they're coming in from ASCII headers, never hurts to utf8'em, right?20:05
dobeymmcc: well, i suspect we need more tests in poauth for unicode stuff; there's like none right now20:06
dobeybut i think that's a separate branch to deal with20:06
ralsinajoshuahoover: bug #1000950 may require manual attention beyond what we can provide20:06
ubot5Launchpad bug 1000950 in Ubuntu One Client "UbuntuOne - connection problem" [Undecided,New] https://launchpad.net/bugs/100095020:06
mmccdobey: sure. I'm fine with this mp now that I understand all the internet20:06
joshuahooverralsina: k20:07
mmccis anyone else reviewing mandel's u1-client domain-sockets branch? seems important so I thought I'd do that...20:09
briancurtinmmcc: i haven't finished the review part of it, but i ran the tests on windows and it works, so i'm part way there20:12
gatoxmmcc, i was trying to run the tests for that...... but it keeps failing on my machine20:13
mmccgatox: which machine20:13
gatoxmmcc, my evil windows vm20:13
mmccgatox: oh right I saw that conversation as mandel was leaving20:14
mmccok I'll find something else to review20:14
gatoxmmcc, yes..... i'm going to mark it as need fixing to be able to ask with him tomorrow about that20:14
gatoxok..... i'm off for today! see you tomorrow people!20:16
ralsinaschool run. May see some of you later, maybe!20:31
dobeyeek, bad code20:34
dobeyhrmm, i think i need to actually add poauth to tarmac20:46
dobeyalso need to add lptools20:46
dobeygrr20:47
dobeythis run-mac-tests thing needs to go away20:47
mmccdobey why?20:48
dobeybecause i keep running it on linux now20:49
dobeyas it's the first tab complete result :(20:49
mmccah :\20:50
mmccmaybe alias rt=run-tests ? same keystrokes...20:51
mmcchonestly they could be the same file...20:51
mmccsame *number* of keystrokes, I meant20:52
dobeyheh20:52
dobeythey should be the same file20:53
dobeyor rather i don't know what run-mac-tests does, but it shouldn't be needed20:53
mmccfor SSO it's the linux script without GTK and with different ignore paths passed to u1trial20:55
dobeyi see that20:55
dobeywe need to get rid of it, and fix it so things can just use @skipIf or whatever instead20:56
* alecu has just reproduced bug #1003085 on a windows vm with working proxies, but no DNS server configured.20:57
ubot5Error: Launchpad bug 1003085 could not be found20:57
joshuahooverdobey: any clues on what to tell this 12.04 user trying to setup u1 and getting this error? http://paste.ubuntu.com/1001788/20:58
dobeyno, but i'd like to know how that error managed to happen21:01
briancurtinjoshuahoover, ralsina: 3.0.1 is now available, it's all up on the site and ready to go21:02
joshuahooverexcellent!21:02
dobeyjoshuahoover: there are some other reports of that same problem though21:02
joshuahooverdobey: hmmm...anything i should ask in particular to help troubleshoot?21:03
dobeyhrmm, i bet it's a race condition21:05
dobeyjoshuahoover: i can't think of anything, no21:07
joshuahooverdobey: k, first time i've seen it so not sure what to have them try21:07
dobeyjoshuahoover: well, it's showing up on errors.ubuntu.com21:08
dobeyjoshuahoover: this, i guess: https://bugs.launchpad.net/ubuntu/+source/ubuntuone-client/+bug/71116221:12
ubot5Ubuntu bug 711162 in ubuntuone-client (Ubuntu) "ubuntuone-login crashed with ValueError in call_async(): Unable to guess signature from an empty dict" [High,Confirmed]21:12
dobeyjoshuahoover: i'll put it on my list to poke at in the morning21:14
joshuahooverthanks21:15
dobeyalright, need to go. have a good evening all!21:28
mmccwell color me not surprised that neither py2app nor pyinstaller 'just works' ... both seem to have issues relating to our use of homebrew for setting up the dev environment21:32
mmccis the windows-installer project the right place to put mac packaging stuff, or should there be an analogous 'mac-packaging' project? we use windows-installer on the mac now....21:35
briancurtinmmcc: should probably be in its own project, especially as windows-installer is going to be used on jenkins to create the windows installer21:36
briancurtinso we might want a separate project for mac installer, and i'm eventually hoping to move the buildout stuff from windows-installer to its own top-level project as well21:37
mmccbriancurtin: ah ok. so, no good place for me to file bugs on this stuff yet.21:38
briancurtinnot really21:39
mmccI'll float the separate mac-pack project tomorrow when folks are back21:39
mmccanyone still around remember why we use homebrew instead of macports for the osx dev setup?21:49
briancurtinmmcc: urbanape and (kind of, barely) myself started the port, and i think he was just a fan of homebrew so that's what we went with21:51
mmccbriancurtin: ah, ok. I figured. py2app is confused by homebrew and pyinstaller is kind-of confused, but I think it'll be easier to fix. pyinstaller expects either a direct install of pyqt or macports - go figure.21:52
urbanapeyeah, momentum. I gave up on macports and fink years ago.21:59
urbanapeer, inertia, rather.21:59
mmccurbanape: ack. I note also that macports' portfiles are tcl, while brew's charmingly beer-themed system is all ruby. executable config files are a nice match for ruby22:01
mmccor are they really tcl? i dunno - they're tcl-mode anyway22:01
* alecu has not touched tcl for at least 15 years22:07
alecunot that I miss it that much.22:07
teknicoalecu, you insensitive clod22:09
alecuteknico, beuno still has you all slaving at the sprint? Go out, enjoy this lovely Buenos Aires weather!22:11
teknicoalecu, ralsina got the wheather point clear enough already, thanks :-P22:16
ralsinateknico: should improve by sunday! ;-)22:16
ralsinammcc, alecu, briancurtin: how about renaming windows-installer packaging-tools or something?22:17
teknicoralsina, I'm so glad for you then! ;-P22:17
teknico(weather not much to do with wheat, btw)22:17
mmccralsina: does that make a lot of work in renaming e.g. automated test scripts, etc?22:23
ralsinammcc: I am expecting "no" but could be wrong since I never did it before :-)22:23
mmccralsina: aha. well, I have no other concern there. I mostly just want a place to file bugs so you all can see what I'm doing :)22:26
mmcchooray. I got pyinstaller to work on its own example pyqt4 app22:34
mmccon that medium-high note I will leave for the evening, might be able to come back for some more pyinstaller tonight22:35
* mmcc goes to watch the baby22:35

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!