=== mandel_afaik is now known as mandel === jussi01_ is now known as jussi [07:21] morning all! [07:22] mandel: morning! [08:16] Good morning! [08:36] nhaines, morning :) [09:05] hi everyone [09:13] gatox, buenas1 [09:13] mandel, como va eso? :P [09:26] gatox, bien? no se que me pasa en la cabeza ultimamente... [09:26] debe de ser la temeratura [09:26] mandel, esta haciendo calor? [10:14] gatox, can you take a look at the following: https://code.launchpad.net/~mandel/ubuntuone-client/add-method-name/+merge/73400 [10:14] mandel, yep, on it [10:22] gatox, thx! [10:22] mandel, i've left you a comment with a Need Information [10:22] gatox, how do you fill doing a review on a fs related branch? is a really easy one? [10:22] gatox, ok, looking [10:22] mandel, shoot! :P [10:23] mandel, the last comment i think that maybe i'm wrong... so is just a need information :P [10:24] mandel, i saw the asserts, but i don't know how that is being made in ubuntuone-client [10:24] gatox, they are being called by the decorator, look at line 96 of the diff :) [10:25] mandel, okkk.... just wanted to know :P [10:25] gatox, I think yours make some sense? I have no problems to add it like that, let me update it :) [10:25] mandel, what about the other comment?? it makes sense? [10:26] gatox, tests? well that is hard.. the assert is tested but I'm not testing that the message of the assert is the one with the method... [10:28] mandel, is it possible to test that? (just asking... i don't want nessita to hit me later :P) [10:32] gatox, yes.. I suppose I can get the AssertError exception and test that the method name is indeed in the message? sigh I'll do it... [10:32] puto [10:32] gatox, as a punishment, take a look at https://code.launchpad.net/~mandel/ubuntuone-client/donot-add-watches-parents/+merge/73648 [10:32] :P [10:33] mandel, jejeje sorry... i'm afraid for my life :P [10:34] * gatox looking.... [10:42] gatox, that one those have tests :) [10:43] mandel, yes, i'm running the test on ubuntuone-client... a +1 is coming! jeje [10:45] gatox, the idea of the branch is very simple (I initially I had a harder implementation) and I blame the headaches for that hehe [10:45] jje [10:45] mandel, i used to have migraines.... but nothing like yours [10:52] gatox, miines are due to a bad circulation system.. I blame my parents ;) [10:53] mandel, +1 on https://code.launchpad.net/~mandel/ubuntuone-client/donot-add-watches-parents/+merge/73648 [10:53] gatox, superb! thx! [10:53] mandel, let me know when i can re-review the other one [10:53] gatox, I'm done adding the tests :) [10:53] gatox, superb! [10:54] :D [10:54] gatox, are you being this nice because I'm going to get your mibile phone? [10:54] hehe [10:54] mandel, crap........you figure it out! [10:54] jejejeje [10:56] morning all [10:56] duanedesign, hi [10:56] mandel, btw, did you receive the phone yet?? feel free to play with it... you are not going to see many devices with meego :P [10:57] * gatox write without thinking... :P [10:58] mandel, have you receive the phone? [10:59] gatox, no, not yet [10:59] gatox, at least no one has called [11:00] mandel, it might take a while... i wonder if meego will be cool [11:01] gatox, I've heard really good things? but who knows... [11:01] mandel, yep... and intel dismiss the project :S [11:05] yes? I recall all the drama hehe [11:06] mandel, yes, a few weeks ago i had a dinner with some people that work at intel and they told me that... [11:06] Holaaaa [11:06] facundobatista, buenas [11:07] facundobatista, morning! [11:08] facundobatista, can I ask you to look at an MP for ubuntuone-client? [11:08] facundobatista, it would be this one: https://code.launchpad.net/~mandel/ubuntuone-client/donot-add-watches-parents/+merge/73648 [11:12] mandel, sure! I'll exchange it with one of my two MP waiting for review... [11:13] facundobatista, sure :) [11:13] facundobatista, shoot! [11:18] mandel, did you update add-method-names branch? so i can take a look at that.... [11:18] mandel, i see it is updated...... reviewing.... [11:20] mandel, +1 [11:20] nessita, hi! [11:20] good morning! [11:20] nessita, holaaaaa [11:20] nessita, are you free for a review? (or 3? :P) [11:21] gatox: I will in 5 minutes! the same 3 from yesterday? :-) [11:21] nessita, yes! those ones [11:22] nessita, ping [11:25] mandel: pong [11:26] nessita, buenos dias! [11:27] nessita, two things, the branhc for the names in the sasserts was pushed and gatox did a review. I added some tests from his review can you take a look? [11:27] mandel: sure! [11:28] nessita, second things, and this is a quick one, looking at https://code.launchpad.net/~mandel/ubuntuone-client/donot-add-watches-parents/+merge/73648 facundobatista and I were wondering if adding a new module in platform was a good idea [11:29] mandel: looking [11:29] nessita, other approaches we though about were simply adding the var in the __init__ of platform (linux/__init__.py and windows/__init__.py) or creating a contants.py with all the diff contants [11:30] nessita, by the way, that is the simplest implementation I could think of.. I did a diff one yesterday but I looked at it and though it was too complicated for what we needed. [11:30] mandel, facundobatista: so, I think this issue needs a different solution, not a constant. I think that the filesystem_monitor should be the one avoiding putting watches in the parents of UDFs [11:31] facundobatista: does fs monitor knows about UDFs? [11:32] nessita, hm.. but the only one adding the watches is the local_rescan, that means we would have to move that logic out of the local_rescan module... [11:32] mandel: no, I mean the following: [11:32] maybe adding it to file_system_monitor but is that wise atm? [11:33] mandel: local rescan implementation does not change, it calls self._add_watches_to_udf_ancestors(volume) as always [11:33] nessita, yes it does, in linux handles it very different than normal watches, in windows it doesn't [11:34] * nessita looks at some code [11:35] ok, I remember onw [11:35] now* [11:36] mandel: so, I think the windows implementation of filesystem_notifications.py needs to be in charge of putting/not putting the real watch [11:37] nessita, uh! and how do you want it that to be done, just look at the watch to be added and asserta that is not added in '~' ? [11:37] facundobatista, mandel: does that make sense to you? [11:37] mandel: no, that layer knows about udfs [11:37] mandel: you need to check you're not putting a watch in a udf's parent [11:37] mandel: you can have a UDF not in home and still you can't put a watch in the parent [11:38] mandel: can you look at _AncestorsINotifyProcessor, in the linux implementation of filesystem_notifications.py? [11:38] nessita, that is done in line 282 of local_rescan_path, at that point the eq does not longer know that we are dealing with a path to a parent of a UDF [11:38] * mandel looks [11:39] mandel: what is local_rescan_path? [11:39] nessita, why you want filesystem_notification to deal with it? [11:39] facundobatista: because is the one that, I think, needs to ensure a given condition regarding watches [11:40] nessita, that is the NotifyProcessor, that reacts to the events, but here we want to add them before [11:40] nessita, so, you want it as a safeguard so nobody will add watches to the parents by mistake? that's the added value in your solution? [11:40] facundobatista: what if later we add another piece of code that adds a watch to a dir? in windows we can't do that blindy, since it can have sever consequences [11:40] I mean, we do not want to add them, which is before that [11:41] facundobatista: the added value is like a plus, I think the "barrier" should be in that layer... [11:41] mandel: right, what I mean is that that module knows about UDFs [11:42] nessita, ok [11:42] facundobatista: you think that having this in local_rescan is conceptually correct? [11:42] facundobatista: I was thinking in all the tests in eq_inotify that tests this as well [11:42] mandel: look at 450 def add_watch(self, dirpath): [11:43] nessita, facundobatista we can tell the FileSystemMonitor.add_watch to not add a watch to ~ [11:43] nessita, which file? [11:43] mandel: the issue is not only with '~' [11:43] mandel: the same one [11:43] (the same file as before) [11:44] mandel: suppose we're creating the UDF C:\User\Test\1\2\3\4\5 [11:44] mandel: the udf is name is '5' [11:44] nessita, ok, we can get the _is_udf_ancestor out of there so that is shared by both platforms and then do not add the watch in the FileSystemMonitor of Windows [11:44] mandel: we should not put a watch in C:\User\Test\1\2\3\4, nor in C:\User\Test\1\2\3, nor int C:\User\Test\1\2, etc [11:45] nessita, ok, got it? I'll make the changes as soon as I walk the dog [11:45] nessita, LR was alway the responsible of adding or not adding the watches [11:45] mandel: wait, let's finish thinking this idea [11:45] facundobatista, sure [11:46] mandel: I m not 100% sure is correct [11:46] nessita, that ^ was for you :P [11:46] but the proposed solution does not convince me either [11:46] nessita, but I understand (and I like) the needing of having a safeguard if adding a bad watch could bring other problems [11:46] facundobatista: you'd do both? (moving the constant somewhere else) [11:47] nessita, which "both"? [11:47] facundobatista: not adding watches in FS monitor and also checking 'watch_ancestors' in local rescan [11:48] I suppose implementing both is not a huge overhead? is a matter of adding some extra tests [11:48] nessita, it's a little redundant... [11:49] nessita, it would be *better*, but there're more urgent stuff to do [11:49] nessita, mandel, in any case, I do not prefer any solution over the other, it's up to you [11:50] mandel: put this branch on hold and let's work on the other solution, if we can we also improve the current one [11:50] * facundobatista -> kid to kinder, bbl [11:50] nessita, so, I'll make a diff branch with the other solution and will propose it, then we can see which is better [11:51] nessita, on the currelty proposed one, shall I move the contstant to linux/__init__.py and windows/__init__.py instead? [11:51] mandel: yes, sounds good [11:51] nessita, ok, I walk the dog and get that done, if I'm faste enough we can talk about it on mumble [11:51] * mandel runs to walk the dog [11:53] nessita, let me know if you have a minute to mumble for the wizard sizes.... [11:53] gatox: I finish a commnet in a MP and I'm free [11:53] nessita, okas [12:01] nessita: mandel just went to walk the dog. you know some about the Windows client, right? I have a simple question for you. [12:01] karni: shoot, let's see [12:01] nessita: I want to allow to pic auto-upload target directory (the one we upload pics to). But the "~" home concept is not known on Windows [12:01] nessita: essentially, it'll either be ~/Ubuntu One/ [12:02] or ~// [12:02] nessita: I'm thinking if I should not show that leading "~/" to users, as it may confuse them [12:02] and show "Ubuntu One/" or "/" instead. [12:03] nessita: I think the latter is more sensible ↑, don't you think? [12:03] karni: so, a couple of things: [12:03] On the other hand, if one day we allow uploading to places other than home, it's limiting. [12:03] karni: wait for the answers!!! :-) [12:03] I'm waiting :) <3 [12:04] * '~' does work on windows, at least from python I can do os.path.expanuser('~') [12:04] * anyways, we should never build the Ubuntu One folder path using ~/Ubuntu One [12:04] not even in linux [12:04] nessita: This talks to files REST API [12:04] nessita: Uhm [12:04] ah, REST api [12:05] karni: I thought you were dealing locally [12:05] nessita: The question is more about representation to the user [12:05] nessita: Sorry, my question wasn't detailed enough, my bad. [12:05] karni: we don't use any REST from windows, so I missed the connection [12:06] gatox: I'm entering mumble [12:06] nessita: Right. The question is only, that Windows user (using the Android app) would have no idea what "~/" is :) [12:06] nessita, can you hear me? [12:06] gatox: yes! [12:07] karni: one sec [12:08] nessita: Take your call first! :) [12:12] nessita: I think it was silly to ask, I should make such decisions on my own :) [12:20] good morning! [12:22] * facundobatista is back [12:22] Hola ralsina [12:23] lisette: ping [12:23] ralsina: hi [12:23] hi lisette [12:24] ralsina: are you super quick or did you not get my email yet? [12:24] I pinged you, then read it ;-) [12:24] haha [12:24] So both! [12:24] The congratulations page with skip setp button: +1 [12:25] The tray icon, I am thinking of the easiest way t do it :-) [12:25] The emblems... we can't now :-( [12:25] ralsina: i thought we couldn´t [12:25] And I need to know where the artwork is for the very last page of the wizard, if you have it handy [12:26] ralsina: that same image again? [12:26] the one with the three icons? Ok [12:26] ralsina: that is all we have for now artwork wise [12:26] ralsina: you got that? [12:26] In the wireframes, it has a square at the left of "Sync in progress". Is that empty space or should something be there? [12:26] lisette: I can copy it from the other page :-) [12:27] ralsina: we could put the arrow icon there? [12:27] sure [12:27] the one pointing down? [12:27] ralsina: same as the status top right [12:27] ralsina: the round one [12:28] lisette ok [12:28] blajk: did you have any specific ideas on the above? [12:30] blajk, cparrino: see above, no win explorer emblem icons possible for this release [12:39] karni: I can follow the conversation now! [12:39] karni: so, you were asking? [12:39] mandel: needs fixing the add_method_name [12:41] nessita: We've decided with beun'o to expose the upload path in form of "Ubuntu One/foobar", "Pictures cloud folder/foo/bar" "another_pic_udf" (without any leading characters) [12:41] nessita: No worries, thanks!! :) [12:42] karni: :-) [12:43] hello #ubuntuone! [12:46] hi alecu! [12:49] hola alecu [12:49] alecu, nessita, gatox, mandel: standup in 10' [12:49] ralsina, ack [12:54] ok [13:00] me [13:00] me [13:00] me [13:00] me [13:00] gatox: where did you take the 300 minimum for lineedit sizes? [13:00] gatox: was that given by design? [13:01] ralsina: standup? [13:01] nessita, nop..... but the size of the labels use when it was design [13:01] me [13:01] alecu go [13:01] DONE: Found the root cause of the multiple ipc connections, started working on a branch to fix this [13:01] TODO: work on ipc fixing branch [13:01] BLOCKED: no [13:01] NEXT! nessita [13:01] DONE: landed fix for bug #834900, reviews, started branch for bug #814113, talked about proxy branch submitted by a contributor [13:01] TODO: finish sign url branch, tons of reviews, talk with gatox re: wizard size [13:01] BLOCKED: no [13:01] NEXT: gatox [13:01] 300 is the width of the recaptcha, I think? [13:01] DONE: [13:01] Fixed some issues in my MP. [13:01] TODO: [13:01] Keep working on Wizard Size. [13:01] BLOCKED: [13:01] No. [13:01] mandel, go [13:01] DONE: Added tests and fixe branch: https://code.launchpad.net/~mandel/ubuntuone-client/add-method-name/+merge/73400. Proposed an MP with a fix for 806532 but after talk with nessita will propose a second branch with a diff type of fix. We should be able to talk about it in the mumble call. [13:01] TODO: Provide second implementation to fix bug 806532 [13:01] BLOCKED: no [13:01] COMMENTS: Will be away at my 5 pm for some time due to people visiting the house :( [13:01] ralsina, go [13:01] DONE: release, some IRL testing, debugged/fixed the handoff bug cristian had, worked on syncdaemon autostart, had kid at home. TODO: fix last page of the wizard, fix congrats page, finish autostart, release, reviews. BLOCKED: nope [13:02] gatox: can you please run by lisette if 300 is a good width for this? (in the reset password screen) [13:02] nessita, ok [13:02] lisette, ping [13:03] Hey, official UDS list in the mail! [13:04] comments anyone? [13:04] ralsina, already gave mine [13:04] comments! [13:04] I'll be 15 mins of.. hopefully they don't like the house ;) [13:04] alecu: shoot [13:04] nessita, facundobatista: when you have some bit of time, please finish the review of https://code.launchpad.net/~alecu/ubuntuone-client/hash-bash-dash-rash/+merge/73540 [13:05] alecu: yessir! [13:05] I have one: people, don't flood nessita with reviews. She's not *required* to review everything, and I am sure she would prefer to do some coding of her own :-) [13:05] ouch [13:05] ralsina, oh.. but it is sooo much better to have the nessita stamp on things [13:05] alecu: not meant for this specific case :-) [13:05] is like a 'make in the USA' thins ;) [13:05] ralsina, I know :-) [13:05] ralsina: I have an idea for that, that we can talk in the mumble [13:05] mandel, "mike wasowski, you have not filled in your paperwork yesterday!" [13:06] ein? [13:06] alecu, I dont understand and food is getting cold, will me back in few mins [13:06] nessita: cool [13:06] ralsina: but bottom line is I really like doing reviews because I'm more confident that we're moving in the rigth direction given the time constraints we have [13:06] gatox: hi [13:06] alecu, can I exchange that for reviews to my client branches? [13:06] (one of them, of course) [13:06] nessita: if it's fine for you, it's fine for me! [13:07] mandel, http://www.pixar.com/featurefilms/inc/chars_pop7.html [13:07] facundobatista, sure, point me at it. [13:07] ralsina: is like we can't afford "meter la pata" [13:07] nessita: yep [13:08] nessita: then again, turning you into a reviews-only person feels like a waste. We should all raise our level at reviewing. But agreed that time constraints are a problem. [13:08] ralsina: agreed on both, let's use 5-10 minutes today to talk about this [13:08] nessita: cool [13:08] lisette, hi!! can you tell me if the width for the notification area besides the line edits its ok, or could that be smaller??...... i'm sending you the screenshot in pm [13:15] gatox: comment added to https://code.launchpad.net/~diegosarmentero/ubuntu-sso-client/reset-password-page/+merge/73522 [13:16] nessita, ok [13:22] gatox: in close-on-license, there is no need to import from ubuntuone_installer.tests import TestCase and use that, since BaseTestCase already inherits from it [13:23] nessita, i'll try again...... but when i was extending from BaseTestCase i had some problems using _set_called....... weird [13:23] gatox: let me confirm [13:24] _set_called is defined in TestCase, not BaseTestCase [13:25] ralsina: yes, but BaseTestCase inherits from TestCase [13:25] so _set_called is inherited [13:26] class TestCase(BaseTestCase): [13:26] at least in my copy, no it doesn't :-) [13:26] 52 class BaseTestCase(TestCase): [13:26] what are we playing? :-D [13:26] ok, let me confirm [13:26] nessita: there are two TestCase classes [13:27] from ubuntuone.devtools.testcase import TestCase as BaseTestCase [13:27] FUN! [13:27] _set_called is defined in ubuntuone_installer.tests.TestCase [13:27] and BaseTestCase in ubuntuone_installer/gui/qt/tests, inherits from that one [13:27] exactly [13:28] which is the one gatox is using? or not, that may be the problem, let's see! [13:28] from ubuntuone_installer.gui.qt.tests import BaseTestCase [13:28] seems ok to me [13:28] yep, to me too [13:28] nessita, ok, i'll use thatone [13:28] that one* [13:28] gatox: confirmed you don't need to change the import clause nor the testcase parent [13:29] (already is class UninstallerTestCase(BaseTestCase)) [13:30] nessita, so the branch is ok? [13:31] gatox: your latest changes regarding this (adding from ubuntuone_installer.tests import TestCase and inheritng from TestCase) should be reverted [13:31] nessita, ahhhhhhh ok [13:32] gatox: and I added 2 more fixes in the MP, since addCleanup is not being used correctly [13:32] nessita, ok..... i had issues with that too trying to use it as it was in the comments [13:32] gatox: which issues? [13:33] nessita, when i tried to use it like this: self.addCleanup(setattr, sys, 'frozen', frozen) it fails [13:33] nessita, or like this: self.addCleanup(del, sys.frozen) [13:34] gatox: the del may fail, but what was the error with the first one? [13:35] nessita, i don't remember right now...... let me do it again, i'll ask you if i have any problem.... [13:35] please [13:36] duanedesign: Could you please try creating a "~/piczzz/foo/bar baz!" cloud folder and see if it appears in your Ubuntu One control panel? [13:36] duanedesign: scratch that, I'm using REST API, you can't reproduce that without a client [13:38] gatox: FYI, this works: http://pastebin.ubuntu.com/679739/ [13:39] nessita, ok, thanks! i'll try that [13:39] :-) [13:42] brb! [13:51] alecu: ping [13:51] alecu: unping [13:53] alecu: ping for real now [13:56] alecu: needs info added to the MP [13:56] * mandel back [13:56] Hello, is there anyway to see how many tcp sessions u1 sync daemon is throwing open? When it's running my router chucks me off and throws a "exceeds the max. number of session per host!" warning :) [13:56] nessita, ping [13:57] mandel: pong [13:57] nessita, the implementation you mentioned using the file_system_monitor, should I add it to the current branch or a new diff one? [13:57] mandel: a diff one [13:57] okis [13:57] nessita, pong [13:58] ha, those of you working on cmd, please press F7 :-) [13:58] nessita, thanks, looking. [13:58] ralsina, ay ay ay! [14:03] alecu: in windows, this test is failing for me (seems unrelated though( [14:03] ) [14:03] File "E:\u1\client\review_hash-bash-dash-rash\tests\syncdaemon\test_localrescan.py", line 625, in test_disc_symlink [14:03] "symlink", symlpath)) [14:03] twisted.trial.unittest.FailTest: None [14:03] seems like local rescan is not ignoring symlinks (???) [14:04] Chipaca, ralsina, mandel, gatox, alecu: shall we mumble-rumble-tumble? [14:05] nessita: in mumble [14:05] going! [14:05] nessita: ... talking with other people [14:05] nessita: ... people cooler than you [14:05] ok, let's wait for chipaca [14:05] Chipaca: madness! [14:05] * nessita waits [14:08] alecu, nessita, mandel, gatox: I would like a review for the simple branch https://code.launchpad.net/~ralsina/ubuntuone-windows-installer/put-in-autostart/+merge/73670 [14:08] simple, yet it took me 4 hours to figure out how to do it (WINDOWS IS FUN) [14:11] nessita: do we have a coding style guide, or should we just follow pep8? [14:12] elopio: we follow pep8 and pep257 for docstrings [14:12] elopio: we also run pep8 and pylint tools to check our code [14:15] nessita: cool. And you mentioned on a review that it was recommended to end every sentence with a period. Are that type of recommendations documented somewhere? [14:15] ralsina: the branch has some valid lint issues, I'm adding those to the MP [14:16] nessita: forgot to push! [14:16] nessita: there, pushed [14:16] ack [14:17] ralsina, 'Software\\Microsoft\\Windows\\CurrentVersion\\Run' should be a constant [14:17] alecu: it's used in exactly one place [14:17] ralsina, it's used on the test too! [14:17] alecu: and then in the test, but if I use a constant and the constant is wrong the test passes :-) [14:18] ralsina: constants can't be tested! we can make a typo in any place (test or live code) [14:18] :-) [14:18] nessita: but not in both (I hope ;-) [14:19] ralsina, also it would look better as a constant so both lines do not exceed 79 chars [14:19] ralsina: right, but constant duplication is very hard to maintain [14:19] alecu: that is a missing push, reload [14:19] ok then, constant it is [14:19] I see now. [14:21] ralsina: I still get this lint error, is it a false positive? [14:21] ubuntuone_installer/gui/qt/utils/windows.py: [14:21] 80: [E0602, add_syncdaemon_to_autostart] Undefined variable 'WindowsError' [14:21] well, you will get that on linux [14:21] I will disable that error [14:21] ralsina, I don't understand why you check for sys.platform == "win32" inside ubuntuone_installer/gui/qt/utils/windows.py [14:21] ralsina: but WindowError is not being imported from anywhere, is it a stdlib thing? [14:22] nessita: yes, it's stdlib but is not there on Linux [14:22] ah [14:22] alecu: forgot to remove that one, thanks [14:23] ralsina, in that case: would it make sense to add syncdaemon to autostart if SD is not run from the frozen .exe? [14:23] ralsina, my guess is "no" [14:23] * mandel hates circular imports! [14:24] alecu: hmmmm yes for the tests ;-) [14:24] ralsina, right. But for the tests we should be patching sys.frozen [14:24] that is not about SD being frozen is about the wizard being frozen. If it's not frozen but is in the path, it will work [14:24] let me rephrase that [14:25] "if the wizard is not frozen, but sd is in the path, it will work" [14:25] but it's a corner case [14:25] ralsina, besides the tests when (hasattr(sys, "frozen") == false) I think we should not be adding it to the registry. [14:26] hmmm [14:26]