/srv/irclogs.ubuntu.com/2018/04/22/#ubuntu-devel.txt

tsimonq2In case nobody saw it yet: https://lists.ubuntu.com/archives/ubuntu-release/2018-April/004438.html00:01
Guest31513I'm downloading the image now to give it a try. Would like to investigate the hyper-v enhanced session functionality00:03
=== Guest31513 is now known as evancox10
=== sergiusens_ is now known as sergiusens
LaibschCan I still make an upload to a package with a targetted bugfix backported from upstream to a package in universe, a package I maintain in Debian where I have upload rights?  I have upload rights as PPU.  Should I target bionic or bionic-proposed?07:05
LaibschI did read https://wiki.ubuntu.com/FinalFreeze and it is not entirely clear to me what I should do now.07:06
LaibschI am trying to fix bug 134560507:07
ubottubug 1345605 in fslint (Ubuntu) "error when searching bad symlinks" [Medium,In progress] https://launchpad.net/bugs/134560507:07
LaibschBTW, the topic of the channel needs an update, doesn't it?07:07
=== tsimonq2 changed the topic of #ubuntu-devel to: Archive: Final Freeze | 18.04 final beta released | Devel of Ubuntu (not support or app devel) | build failures: http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of Trusty-Artful | #ubuntu-app-devel for app development on Ubuntu
=== tsimonq2 changed the topic of #ubuntu-devel to: Archive: Final Freeze | 18.04 Final Beta released | Devel of Ubuntu (not support or app devel) | build failures: http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of Trusty-Artful | #ubuntu-app-devel for app development on Ubuntu
=== tsimonq2 changed the topic of #ubuntu-devel to: Archive: Final Freeze | 18.04 Final Beta released | Devel of Ubuntu (not support or app devel) | Build failures: http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of Trusty-Artful | #ubuntu-app-devel for app development on Ubuntu
tsimonq2There.07:10
Laibschtsimonq2: Hi! In case you responded to me, I missed everything except the "There" due to internet outage.07:19
tsimonq2I didn't.07:21
tsimonq2:)07:21
Laibschgood, so I assume you just wanted to announce that you're back.07:24
LaibschWelcome back!07:24
LaibschI had posted a question that I will then assume that you missed and that I believe you might be able to answer.07:24
tsimonq2Ah, looking.07:24
LaibschCan I still make an upload to a package with a targetted bugfix backported from upstream to a package in universe, a package I maintain in Debian where I have upload rights?  I have upload rights as PPU.  Should I target bionic or bionic-proposed?07:24
LaibschI did read https://wiki.ubuntu.com/FinalFreeze and it is not entirely clear to me what I should do now. I am trying to fix bug 1345605. BTW, the topic of the channel needs an update, doesn't it?07:24
ubottubug 1345605 in fslint (Ubuntu) "error when searching bad symlinks" [High,In progress] https://launchpad.net/bugs/134560507:24
tsimonq2Is it seeded? (hint: run `seeded-in-ubuntu $package`)07:25
LaibschOh, I see you just did the update of the channel tagline.07:25
tsimonq2Yep. :)07:25
LaibschI doubt it would be seeded, it is in universe07:25
Laibschlet me check07:25
Laibschindeed, it is NOT seeded, so I can make the upload?07:27
tsimonq2Yep, if it isn't seeded, Feature Freeze applies.07:27
tsimonq2(Which is less strict.)07:27
LaibschYes, this only fixes a broken but current feature.07:28
tsimonq2OK.07:28
tsimonq2So, go ahead. :)07:28
LaibschThanks07:28
LaibschMuch appreciated07:28
tsimonq2No problem.07:29
tsimonq2And as to which you should target, target bionic.07:29
Laibschgood07:29
tsimonq2The only time you include a pocket in the changelog is for -backports or -security.07:29
tsimonq2It doesn't /hurt/ to target -proposed but it's just a tad bit weird. :)07:30
Laibschwell, the wiki says a bit otherwise.  apparently, you're supposed to target -proposed for main now07:49
Laibschhttps://wiki.ubuntu.com/FinalFreeze07:49
tsimonq2That hasn't been edited since 2012.07:49
LaibschWell, maybe it hasn't changed since then07:50
tsimonq2Notably, before we implemented https://wiki.ubuntu.com/ProposedMigration07:50
LaibschI wouldn't know the processes well enough07:50
tsimonq2I know for a fact it has changed. ;)07:50
Laibschgood07:50
LaibschI'm happy I can rely on your knowledge07:50
tsimonq2Believe me when I say I've had to cut it close with some uploads. :P07:51
LaibschOne more question, am I correct to assume that binaries are thrown away after upload?07:51
LaibschI'm used to Debian where I used to upload and then sync07:51
tsimonq2You can only do source-only uploads to Ubuntu.07:52
tsimonq2So syncs throw away binaries.07:52
LaibschBut what happens if you do include debs in your changes file and upload?07:53
Laibschthings break or the binaries simply get thrown away and rebuilt?07:54
tsimonq2The upload will get rejected.07:54
LaibschOK07:54
tsimonq2cjwatson might have more specifics; I'm only an Ubuntu developer and don't have upload access to Debian (yet). :)07:54
darkxsttsimonq2, hi07:55
tsimonq2darkxst: Hello, how are you?07:56
darkxstI am good07:56
tsimonq2Good to hear, what's up?07:56
darkxsttsimonq2, just wondering if grasp the value in archive freezes, doesnt seem you do from your reply07:57
tsimonq2I do understand the importance of them, but I disagree that they're still needed.07:58
tsimonq2If flavors can just upload whatever they want and respin, you won't ever get a release, because flavors can just keep respinning.07:58
tsimonq2Why even turn off the dailies? :)07:58
tsimonq2The archive freezes are to create a stable image, true, but the day after, the fixes people have been holding off on get uploaded.07:59
tsimonq2Thus, a stale image over that weekend.07:59
darkxstwell flavours could just choose to leave the dailies on if they want08:00
darkxstbut really having a semi-frozen archive where the flavours choose what to upload but universe etc is blocked out in general would allow you to make the most of your QA week08:00
valoriedarkxst: can you say more about what you value in the freezes?08:00
valoriebecause I've not heard the release team argue for keeping them08:01
tsimonq2darkxst: So, an inverted freeze?08:01
tsimonq2That's an interesting idea.08:01
darkxstvalorie, I am not release team, I was Ubuntu GNOME tech lead08:01
valoriethat blocks the release team08:01
valorieyes, read your email just now08:02
* valorie is the kub release manager and not so techy08:02
tsimonq2darkxst: Thinking about it though, the inverted freeze would just create a headache for non-flavor Ubuntu developers.08:02
tsimonq2The goal here is to not require any intervention from ~ubuntu-release.08:03
valorieI get why we need deadlines for new features, string freezes, etc.08:03
darkxstvalorie, assuming its on a similar schedule to previous milestones is that a problem08:03
valoriethat's obvious08:03
valorieright, but I fail to see the value in freezing08:03
tsimonq2Me too.08:04
valorieif you are testing and finding bugs, then the devels are fixing the bugs08:04
valoriethen you can test those fixes right away08:04
valorienot wait until freeze is over08:04
darkxstvalorie, if you don't freeze, other bugs may well creep in08:05
valoriebut I want to hear the argument for what freezes bring08:05
valoriesuch as?08:05
tsimonq2Right, but that's where testing comes in.08:05
darkxstvalorie, you have a moving target, who knows what will change08:05
valoriewell08:05
tsimonq2darkxst: But that's the development release...08:06
valorieyou are testing AN iso08:06
darkxstfreezes give a point in time image, testers identify issues, the dev's can focus on fixing those issues08:06
valoriethat's one moment in time08:06
tsimonq2Right.08:06
valoriefor this RC somehow kub. managed to test all the test cases08:06
valorieso respinning tomorrow would be cool08:07
darkxstthe point the milestones become stale is after the release, not when they are released08:07
valoriebecause we could at least partially re-test08:07
tsimonq2And often times, people report bugs after the release.08:07
tsimonq2With the release ISO.08:07
tsimonq2And they don't update.08:07
tsimonq2That's a problem.08:07
tsimonq2Much less so once you've entered a point where the archive is more stable.08:07
darkxsttsimonq2, even if you don't release milestone images, I think there is value in freezeing for your QA week (well days really)08:08
valorieI'm waiting to hear the advantages08:08
valoriecan you give an example?08:08
darkxstvalorie, I can;t give a real example, because its never been done before (unfrozen test weeks)08:09
valorieI meant a case where freezing was a benefit08:10
valorieall the people who test the dailies don't have freezes08:10
valoriealthough I don't have many DO that08:10
valorieI mean, I don't know how many do that08:10
darkxstand how much QA is done on the dailies?08:10
alkisgFreeze is a good signal for "everyone, please test now". I develop some packages and I want to test them 1-2 times for the new LTS Ubuntu version, but I can't test on a daily basis.08:11
tsimonq2alkisg: But is that a technical argument or a social one?08:11
valoriewe do have a few people who test a daily maybe one a week08:11
tsimonq2The point of this is to not have the technical downsides while keeping the social benefits.08:11
valorieor when a devel asks08:11
valoriebecause they did a fix on the installer for instance08:12
alkisgI don't think it matter how I would categorize it; that's how much time I'm able to spend in helping/testing the newer Ubuntu; it's up to the release team to somehow tell me when they want me to do those tests08:12
valorieor something else easier to test on an ISO than in a ppa08:12
valoriealkisg: I think part of the point of this is to take that pressure off the release team, and have us flavors helping one another out08:13
tsimonq2alkisg: Exactly. And if the Release Team said "test at this time" but didn't actually freeze anything, would that change things?08:13
valorieall getting together to do testing weeks08:13
tsimonq2valorie: Exactly.08:13
alkisgtsimonq2: a new package upload might break other packages, making my tests useless08:13
alkisgSo it would make sense for me to test at a point when the new uploads were minimal08:14
alkisgNow, minimal or frozen, no, it's not a big deal, but they should be minimal08:14
valorieI'm still not seeing your point08:14
valoriedoes it matter when this new package upload is made, if it is going to break your stuff?08:14
valorieI mean if you test before and everything's cool08:15
valorieand then the upload is made later, your stuff is broken.....08:15
darkxstvalorie, often things break on upload, only to be fixed a few days later08:15
darkxstnot you just wasted time debugging that breakage08:15
darkxsts/not/now/g08:15
valorieright, and that's why we shouldn't freeze!08:15
valorielol08:15
darkxstwhich would have been fixed anyway08:16
darkxstrather than focusing on your flavours QA08:16
alkisgvalorie: that's exactly my point, that that new upload that breaks stuff shouldn't be made. That's the point of the freeze, that only minor changes should be made after that, so that the big bulk of tests don't need to be done again08:16
valoriewell, I think that is the point of doing the new auto-testing *first*08:16
tsimonq2alkisg: Only seeded packages for some flavors are frozen.08:16
valorieso there are *never* uploads that break stuff08:17
Laibschvalorie: that's never going to happen08:17
alkisgYou can't test how your packages affects other packages with unit testing; they may not even be installed at all when you're testing08:17
Laibschautotests are nice, but "no bugs" is utopia08:17
tsimonq2alkisg: I'm not saying we should get rid of the Final Beta freeze. I'm saying we should get rid of the freezes for just some flavors.08:17
valorieno, no08:17
tsimonq2That's the point here.08:17
darkxstvalorie, what I was suggesting was more let flavours upload bug fixes as required, but otherwise freeze the archive for the few days08:17
valoriesoftware has bugs08:17
valorieah08:18
valorieso imo that is probably what flavors would be doing anyway that week, right?08:18
valorietesting week is fixing bugs08:18
alkisgtsimonq2: I don't mind at all if it's complete freeze or partial freeze or just "advice for fewer uploads"; but as a developer/user that wants to help make things better, I do like having a specific point in time where "my tests have additional value, because from that point on, little changes are going to be made"08:18
valorieI dunno if we need a formal freeze08:19
tsimonq2alkisg: The Alpha and Beta 1 freezes just aren't that.08:19
tsimonq2alkisg: It's putting a name on a daily, really.08:19
valoriebut I do see a value on Fixing Bugs08:19
valoriefor a week08:19
tsimonq2alkisg: You could grab any ISO from any point in the month and say "I'm going off of this"08:19
tsimonq2alkisg: And with those freezes, changes can and will be made in the underlying stack, but just those flavors hold off.08:20
tsimonq2alkisg: So this new procedure does just that.08:20
darkxstif there is no formal freeze, churn will continue08:21
tsimonq2Churn continues regardless.08:21
valoriepersonally I would rather not block the release team08:21
tsimonq2People just hold their breaths and then blast it at the archive once the freeze is done.08:21
tsimonq2So why even freeze?08:21
valoriebecause it is hell to have major stuff still happening the last week of the cycle08:21
darkxsttsimonq2, because it you a specific point to focus on08:22
darkxstyou can fix the bugs addressed, before more are raised08:22
valorieI think we are all stating positions here08:22
valorierather than thinking together08:22
darkxstwithout having to worry about what else churns08:22
valoriewe all want more quality08:22
valoriewe all want fewer bugs08:22
valorieand we all want the shiny new stuff too08:23
valorieso what is the best way to get there?08:23
valoriewe have only 6 months to make each batch of sausage08:23
* alkisg sees little value in the non-LTS releases, but that's another story :D08:24
valoriereally?08:24
valorieI do not like to be on stale anything08:24
tsimonq2darkxst: The only bugs fixed during that time are release-critical ones. Not just plain old bugfixes. So I still think it's pointless.08:24
alkisgYeah, I would much prefer having only LTS releases, and some method for package maintainers to more easily push new versions of their packages to LTS releases08:24
valorieI ran artful from the day the toolchain was uploaded08:24
valoriealkisg: that removes the value of LTS08:25
valorieimo08:25
alkisgI think it adds additional value08:25
alkisgI'm hearing a lot of users that would prefer newer versions of *some* packages in LTS releases08:25
darkxsttsimonq2, which is way I was saying to relax it a bit08:25
valoriewell, there is always backports08:25
valoriethat is what kub provides08:26
tsimonq2Regardless of the talk we're having right now, I've received ACKs from everyone with skin in the game. So I think it's worth trying, at least for a cycle. We can tweak it as we go, and as we see how it turns out.08:26
alkisgIndeed; but backports aren't used as much as they should...08:26
valorieI dunno if our users use them enough, but I often urge them to try them out08:26
valoriebecause if it doesn't work out, ppa-purge is your friend08:27
alkisgBackports aren't in ppas08:27
alkisgThey're in the official archive08:27
valoriewe test those too08:27
darkxsttsimonq2, valorie, I was only only offering my thoughts based on 5 years of releasing Ubuntu GNOME milestones, I have no vested interest in the outcome of these discussion, but if you are all going to be so defensive I will just get on with the others thing I need to get done08:27
valoriein Kubuntu, we have a backport PPA08:27
alkisgvalorie: that's usually a bad idea, mate does that too and I hate it08:28
valorieno, I wanted to hear your reasons08:28
alkisgIt would make a lot more sense to use backports there, rather than a PPA08:28
valorieI just didn't hear reasons08:28
valorieor examples08:28
tsimonq2darkxst: I'm not being defensive, I just see the idea being shot down before there's a chance to try it, that's all.08:28
tsimonq2s/being/trying to be/08:29
valorieI know that Simon didn't develop the idea alone; he discusses pros and cons with quite a few people (not me) before making the proposal08:30
valorieI spoke only after the release team said that they supported it, because to me that is the most important opinion08:31
darkxsttsimonq2, shot down by who?08:33
tsimonq2darkxst: Maybe I'm using the term liberally, but this discussion is happening as a result of making it official.08:35
tsimonq2(It seems like it, anyway.)08:35
tsimonq2I gave a good three weeks between the discussion and doing so.08:35
darkxsttsimonq2, no this has nothing to do with that, just I didnt get a chance to comment earlier08:36
valorieI'm glad to hear differing opinions08:36
darkxstagain I have no vested interest in the result08:36
valoriejust wish they had been stated earlier08:36
valoriegroupthink can be an issue when people don't push back when they have reservations about a new idea08:37
darkxstvalorie, unfortunately I have full-time job unrelated to ubuntu, which makes me kind of busy08:37
valorieyeah08:37
valorieanyway, creeping carrot will be a non-LTS08:38
valorieso if it doesn't work out well, we have time to revert to milestones and freezes, or modify slightly08:39
tsimonq2*Calculating Camel08:39
tsimonq2And, right.08:39
darkxstfreezes are useful even if there is not an official milestone image to come out of it!08:43
valoriewell, the lack of them might prove you right, darkxst08:44
LaibschWhat do you guys use to prepare the changes file to upload?  For uploads to Debian (that requires binaries to be included) I used to use pbuilder and now cowbuilder.  But now I need a source-only changes file and haven't quite figured out yet how to do that.09:30
LaibschI only build in a chroot like pbuilder, my main machine has very few if any dev-packages installed.09:31
Laibschanswering for myself, looks like it might be as easy as "dpkg-genchanges" even if the manpage says the source needs to have been built which outside of the chroot it hasn't been.09:40
darkxstLaibsch, sbuild09:43
Laibschdarkxst: thanks, I will look into. Never used it so far and never had to.09:44
Laibsch"debuild -S -d" or even simply "dpkg-buildpackage -S -d" (AFAIU, debuild calls dpkg-buildpackage)09:45
darkxstLaibsch, see sbuild-launchpad-chroot it produces almost identical environment to the Ubunutu archive and PPA builders09:45
darkxstLaibsch, and in chroot, so your not mutating your package sources09:46
Laibschdpkg-buildpackage seems to do it09:46
darkxstyou can also bootstrap debian into sbuild quite easily09:46
Laibschwell, same is true for pbuilder and I'm more familiar with that09:46
Laibschbut in the end, I think it seems to be necessary for Debian since you need to actually build the code and ship the binaries, the requirements for Ubuntu infrastructure apparently are already fulfilled by a lower-level tool like dpkg-buildpackage09:47
darkxstLaibsch,  are you uploading to ubuntu?09:48
Laibschyes09:48
darkxstthen use sbuild-launchpad-chroot09:48
darkxstto test09:49
Laibschtest what?09:49
cjwatsondebuild -S or any near-enough equivalent is fine, indeed.  Just building a source package doesn't usually need much in the way of build-deps.09:49
darkxstpackage builds, before uploading09:49
Laibschthanks for confirming, cjwatson09:49
Laibschdarkxst: I already built and started the package09:49
cjwatsonAnd indeed, tsimonq2 is correct, if you attempt to include .debs with an upload it will be summarily rejected.09:49
Laibschcjwatson: I tried ;-)09:49
Laibschor actually, what I tried was to fiddle with the changes file, remove the lines for the binary files so they would not be uploaded but it was still rejected (I assume because the changes file name had "amd64" in it, indicating an arch-dependent build)09:51
cjwatsonEditing .changes manually is fiddly.  Not recommended.09:51
LaibschI know09:51
cjwatsonAnd for Debian nowadays you should really do source-only uploads too, except in unusual circumstances.09:51
LaibschFirst attempt09:51
Laibschreally?09:51
Laibschdidn't know09:51
cjwatsonReally.09:51
LaibschWell, no more uploads to debian for me anyhow.09:52
tsimonq2cjwatson: When can we have source-only NEW in Debian? :P09:52
* tsimonq2 wishes for that.09:53
Laibschlooks like the upload was successful and accepted09:53
cjwatsontsimonq2: Do I look like a Debian ftpmaster?10:03
acheronukkirkland: byobu no longer shows a ubuntu logo in bionic?10:04
tsimonq2cjwatson: I was joking there. :P10:06
tsimonq2Right, time for bed.10:06
Laibschtsimonq2: are you in Hawaii or something?10:07
tsimonq2No, Wisconsin.10:09
Laibschouch10:30
Laibschinteresting time to go to sleep ;-)10:30
Laibschhappy Zzz10:30
=== sergiusens_ is now known as sergiusens
=== tomreyn_ is now known as tomreyn
=== realies- is now known as realies
realiesthe novnc package is missing its launch script, any clues why?16:28
realieshttps://github.com/novnc/noVNC/blob/master/utils/launch.sh16:31
realies^16:31
realiesalso is very old, the newer releases have great improvements16:46

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