| === JanC is now known as Guest60369 | ||
| === JanC_ is now known as JanC | ||
| === chihchun_afk is now known as chihchun | ||
| === chihchun is now known as chihchun_afk | ||
| === chihchun_afk is now known as chihchun | ||
| === chihchun is now known as chihchun_afk | ||
| === chihchun_afk is now known as chihchun | ||
| === athairus is now known as afkthairus | ||
| === chihchun is now known as chihchun_afk | ||
| === chihchun_afk is now known as chihchun | ||
| popey | dpm / bzoltan who owns this tutorial documentation? https://developer.ubuntu.com/en/phone/apps/sdk/tutorials/building-cross-architecture-click-applications/ | 10:00 |
|---|---|---|
| bartbes | hmm, I stuck '"template": "undefined"' in my apparmor file, yet apparmor is still blocking my dbus calls | 10:06 |
| popey | unconfined | 10:06 |
| bartbes | sorry, that's what I've got there as well, I just can't copy from my vm | 10:07 |
| bartbes | and related, it seems you can ask for access to specific folders, but not to dbus destinations | 10:08 |
| popey | sounds right | 10:09 |
| bartbes | and here I was thinking I was much better off talking to a service over dbus than modifying its configuration directly | 10:10 |
| bzoltan | popey: Do you want me to own it? | 10:20 |
| dpm | popey, I can't remember who wrote it, but surely davidcalle knows ^ | 10:21 |
| dpm | oh, looking at the history, it seems I published it first, but IIRC someone else wrote it | 10:24 |
| popey | bzoltan: dpm had reports of it not working. Developer unable to follow that guide to build their C++ app. So whoever owns it, it needs updating | 10:28 |
| davidcalle | dpm: popey: not mine, but I'm happy to help bzoltan with an update (screenshots, styling, etc.) | 10:30 |
| bzoltan | popey: I do not own it as now. It mush have been somebody from the community team to contribute that content. | 10:30 |
| bzoltan | davidcalle: dpm: popey: same here... i am happy to fill in gaps and keep documentation up to date | 10:30 |
| popey | I think perhaps we need a review of these pages, as I'm seeing various people trying somethings here which don't work | 10:30 |
| popey | they get tips from knowledgeable people "use static chroots" or "use lxc" or "no, that doesn't work" | 10:30 |
| popey | I feel we should have working documentation :) | 10:31 |
| bzoltan | popey: what comes out from our team is not a "tip" :) | 10:45 |
| popey | what? | 10:46 |
| bzoltan | popey: that page is badly outdated... talking about saucy and trusty targets. The SDK we have out there is fully supported and should be working by now. | 10:46 |
| popey | Tell that to the guy who just wasted some time trying to get his app to build and gave up. | 10:47 |
| bzoltan | popey: the static chroot package is an official workaround for the very unpleasent but not too rare occasion when the Ovelay PPA breaks the SDK | 10:47 |
| popey | where's that documented? | 10:47 |
| bzoltan | popey: I have wrote up the story here - https://developer.ubuntu.com/en/blog/2015/09/01/next-generation-sdk/ | 10:48 |
| bzoltan | popey: since then those packages were promoted to the Release SDK | 10:49 |
| popey | That's a blog post, not documentation. | 10:49 |
| popey | People shouldn't have to read 6 month old blog posts to find out how to use our sdk should they? | 10:50 |
| bzoltan | popey: you are right, the situation is not good and yes these documentation need owners | 10:50 |
| bzoltan | popey: but to be precise ... our sdk is usable as it is relased. if there is a problem then we can help... the click chroots keep breaking, but I have fixed the last issue some time ago. I have not heard about new regression since. Is there a known problem again? | 10:52 |
| popey | bzoltan: a developer on the ubuntu apps telegram group was trying to compile their app for a different arch (previously armhf, adding i386) and had trouble installing and updating the click chroots. | 10:54 |
| popey | it was an hour or more of conversation | 10:55 |
| popey | they got frustrated at the broken-ness of it all, and gave up | 10:55 |
| bzoltan | popey: I am not on that group... i am here on IRC and on the ML | 10:55 |
| popey | Yes, that's clear. | 10:55 |
| bzoltan | popey: I am happy to help anybody who has problems with the chroots. They are sensitive beasts, i know. | 10:56 |
| popey | They aren't on the mailing list (too high volume) and not on irc (at work) | 10:56 |
| popey | so they use telegram, where many of us hang out | 10:56 |
| bzoltan | popey: my best offer is that of somebody is reporting an SDK problem on the Telegram group, please direct them to us. We are here almost 24/7 | 10:56 |
| bzoltan | popey: there is web interface for the IRC | 10:57 |
| popey | yes. but now the user is so frustrated (and busy with work). I am just being a messenger and getting frustrated that I can do nothing to help this guy | 10:58 |
| bzoltan | popey: and my mail box zoltan.balogh@canonical.com has no high traffic :) I will respond to anybody who drops me a line, | 10:58 |
| popey | tell you what, I'll screenshot the entire telegram conversation and attach it as bitmaps to your email | 10:59 |
| popey | that'll work | 10:59 |
| bzoltan | popey: you are not the only one... we had the same discussion with faenil too some time ago. Ubuntu folks are frustratd because app developers are frustrated... because they use the _WRONG_ communication channel. Telegram is not good for this. No highlights, no public logs, too much noise what you can not filter out. I will not spend hours every day reading Telegram backlogs. | 11:00 |
| bzoltan | popey: in my view it was a mistake to direct developers to Telegram | 11:01 |
| bzoltan | I tried telegram for some time... but it is impossible to follow up discussions. On IRC I can use highlights for "sdk" "qtcreator" "toolkit" "zoltan" ... easy. On Telegram one must read kilometers of backlogs. Not too productive. | 11:02 |
| popey | bzoltan: here is a plain text log of the conversation. missing screenshots, I can grab those if you need them http://paste.ubuntu.com/16917761/ - Popescu Sorin is the person with the issue | 11:03 |
| popey | Yes, I don't need to have the "which communication method" argument right now. | 11:03 |
| popey | I'm trying to focus on helping a developer who happens to use a communication method nobody on the sdk team uses | 11:03 |
| bzoltan | popey: the best way to help them is to direct them to the right channel. Telegram is the wrong channel. | 11:03 |
| faenil | bzoltan: I had a chat with QA to get them to build an automated daily task that tests chroot creation. It seems a very low hanging fruit with a lot to gain. You should have received an email about this today, have a look at it ;) | 11:04 |
| bzoltan | faenil: I have read it. Yes, it is a good solution.. I was asking for it for years :( | 11:04 |
| popey | bzoltan: http://imgur.com/a/5q9FM here are all the screenshots they posted | 11:05 |
| faenil | bzoltan: awesome! :) glad a quick chat was enough this time then :) | 11:05 |
| bzoltan | popey: duuuude :) you are spreading wrong information -"Alan Pope, [02.06.16 08:49] chroot creation has been broken for ages" | 11:06 |
| bzoltan | popey: the click chroots were fixed.. I personally have fixed them :) | 11:06 |
| popey | lies | 11:07 |
| popey | am I going to have to nuke a computer to prove this again? | 11:07 |
| faenil | he means he fixed that particular dependency problem, I think | 11:07 |
| faenil | not the chroot situation in general and forever | 11:07 |
| popey | i tried to create one within the last two weeks and it failed | 11:07 |
| davmor2 | popey: nah vm is enough | 11:08 |
| bzoltan | popey: as I said .. the Overlay PPA can break the chroots of _ALL_ developers without telling us. So it is not like that the click chroots are broken for ages... they get broken, then i fix them.. then they break again and i fix again... that is how it is going. the Overlay PPA is not under my control. | 11:09 |
| popey | right, so as a process, it's broken | 11:09 |
| popey | as a concept | 11:10 |
| bzoltan | popey: if there is a new breakage then please tell me which pakage | 11:10 |
| popey | "it's not broken on tuesdays between 3pm and 6:34pm" | 11:10 |
| bzoltan | popey: the concept is broken indeed... I am telling it since it was introduced. | 11:10 |
| popey | I would, but i avoid the sdk because I get so incredibly frustrated that it breaks whenever I try and use it | 11:10 |
| bzoltan | popey: :) It is not that dramatic... but yes, that is the picture | 11:10 |
| popey | I have a laptop which is frozen in time | 11:10 |
| popey | I never update it because when I *breathe* near it, it breaks | 11:11 |
| bzoltan | popey: remember that it is a public channel :) you are talking about our product... | 11:11 |
| popey | and the answer I hear is "read a blog post from last year" or "lxc is the new hotness" or whatever | 11:11 |
| popey | and I work for canonical | 11:11 |
| popey | I can't imagine what it's like for people who don't! | 11:11 |
| popey | I am well aware of where I am. | 11:11 |
| bzoltan | popey: https://en.wikipedia.org/wiki/Osborne_effect | 11:12 |
| davmor2 | bzoltan: why is the sdk not just using stable and no overlay? Developers in theory should only be releasing against the image that is released surely | 11:12 |
| popey | Ok, if that makes you feel better, I'll shut up. There, there's no more problems in the SDK. | 11:12 |
| popey | Job done. | 11:13 |
| bzoltan | popey: there are problems. I am very much aware of them | 11:13 |
| bzoltan | davmor2: is there a PPA with the stable packages of the last OTA? | 11:14 |
| bzoltan | davmor2: The click chroots should be static indeed. | 11:14 |
| faenil | davmor2: (btw, a developer could want to get read for the changes brought by the next OTA before that OTA is out) | 11:15 |
| faenil | ready* | 11:15 |
| faenil | s/could/might | 11:15 |
| davmor2 | faenil: no developer I know would, effectively what should happen is when there is an sdk platform change called the click chroots should be built of that image the only time that should change is when there is a new sdk platform bump | 11:17 |
| faenil | davmor2: a developer might want to use a feature that he knows will be released in the next OTA; so that as soon as the next OTA is out he can publish the app. Anyway, this was just a note, let's not shift the topic :) | 11:18 |
| zbenjamin | popey: we work really hard on fixing that thing. But as its just me and bzoltan sweating as hell on the SDK IDE , and ONLY us two, we are running between improving and fixing the breakage | 11:21 |
| faenil | if there's a way to prevent the chroot breakages, even better davmor2 :) | 11:21 |
| davmor2 | bzoltan: can the chroot instances not be built from the images? If so that would solve all the issues, you default to stable and in faenil case if a dev want to target the next ota he can run on rc-proposed with a warning that it is in flux and the chroot may break and done? | 11:23 |
| bzoltan | davmor2: ohhh... I would soooo wish | 11:23 |
| zbenjamin | davmor2: we are preparing a beta test of the new LXD based QtC, which will indeed use images instead of bootstrapping | 11:23 |
| davmor2 | zbenjamin: oh that would be so much better :) | 11:24 |
| zbenjamin | davmor2: i hope so | 11:24 |
| zbenjamin | davmor2: all images will be tested by us and then released. | 11:24 |
| bzoltan | davmor2: I was asking for this feature from the day zero.... boostraping the builder chroot from Overlay PPA and vivid (!!!) archive brings its own weakneses | 11:25 |
| popey | zbenjamin: I'm sorry to hear that. This is clearly not a viable long term solution. Meanwhile we have people running contests to develop apps, and developers who can't develop apps using our documentation. | 11:25 |
| davmor2 | popey: when you say sorry do you mean happy? | 11:26 |
| popey | behave | 11:27 |
| zbenjamin | popey: no its clearly not ... | 11:27 |
| bzoltan | popey: davmor2: faenil: zbenjamin: I have just set up an i386 click chroot with this command -> sudo click chroot -a i386 -f ubuntu-sdk-15.04 create | 12:06 |
| bzoltan | The logs are here - http://pastebin.ubuntu.com/16918331/ Note, that I am using the correct click package from the SDK Release PPA - http://pastebin.ubuntu.com/16918338/ | 12:06 |
| bzoltan | So the first thing whenever somebody claims that the SDK does not work, the best is to check if the user has the right versions installed from the right sources. | 12:07 |
| bzoltan | Way too many times I have seen that the developer is using an outdated SDK or does not even have the SDK PPA enabled. | 12:07 |
| bzoltan | So, the click chroot creation is good. The update can be broken, but since the update is an apt-gte update... it might require some extra knowledge. We all know that .deb package management is not trivial. That is why the phones and snappy has imag based updates. | 12:09 |
| bzoltan | Anyhow, I know from kalikiana that the developer Popescu Sorin was helped by kalikiana and he managed to create fat packages after he set up the right Kits. | 12:10 |
| === _salem is now known as salem_ | ||
| slvn | The "Stats" graph of downloads for my snap/click packages! It goes from current day to the past, makes think the time is going backward! | 14:17 |
| === chihchun is now known as chihchun_afk | ||
| === afkthairus is now known as athairus | ||
| popey | bzoltan: zbenjamin apologies for being grumpy with you earlier. I was over the top. Sorry. | 17:31 |
| bzoltan | popey: it is okey, no problem. All your points are valid 1) the documentation is outdated 2) click chroots tend to break 3) developer support is fragmented | 17:33 |
| bzoltan | 1) we need doing hands 2) click is sacked and the LXD solution is going to be bet test released tomorrow 3) I have just joined Telegram | 17:33 |
| zbenjamin | popey: accepted. But you were still right... i share the same frustration about the breaking tools ... | 19:37 |
| zbenjamin | popey: and you should join the beta testers once we have it released | 19:38 |
| === salem_ is now known as _salem | ||
| === _salem is now known as salem_ | ||
| === salem_ is now known as _salem | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!