[09:17] <jamespage> if there are any archive admins around; I just added armhf to google-perftools in Packages-arch-specific after it had build in proposed with the associated packaging changes;
[09:18] <jamespage> its now building for armhf direct for the release pocket so it will look a bit odd
[09:20] <jamespage> Daviey, ^^ as you are in my tz
[09:21] <Daviey> jamespage: Hmm, sorry.. I'm not understanding
[09:21] <infinity> jamespage: That doesn't need an AA...
[09:21] <cjwatson> jamespage: not a problem
[09:21] <Laney> He made the change and is just informing AFAICT
[09:21] <jamespage> Laney, you got it
[09:21] <jamespage> normally the NEW packages would appear in -proposed
[09:22] <jamespage> but not in this case
[09:22] <infinity> Nothing new here.
[09:22] <cjwatson> Yeah, it won't count as NEW.  That's based on name.
[09:22] <Daviey> Ah, you mean it will show as binNEW for the release pocket.  I see.
[09:22] <infinity> And it won't get caught in proposed because britney finds no regression in armhf not being built yet.
[09:22] <infinity> No big deal.
[09:22] <infinity> Daviey: It won't.
[09:23] <cjwatson> Once a binary name has been NEWed for >=1 architecture and published, it won't thereafter count as NEW for other architectures.
[09:23] <jamespage> cjwatson, I did not know that
[09:23]  * jamespage learns something new again
[09:24] <Daviey> cjwatson: arch all to separate does show, right?
[09:24] <infinity> all to any, you mean?
[09:24] <Daviey> yes
[09:24] <infinity> Neither all->any or any->all should count as NEW.
[09:25] <Daviey> Hmm, i thought i had seen that. I guess not.
[09:25] <infinity> Note that I said "shouldn't", it's possible there's a bug where it does, but I don't recall ever seeing such behaviour. :P
[09:26] <cjwatson> Daviey: like infinity, I'm not aware that it does, no.
[09:26] <cjwatson> And it would irritate me if it did so I think I'd notice.
[09:28] <infinity> In other news, it's probably about time for us to retire P-a-s.
[09:29] <cjwatson> I had a brief look at importing it into the LP DB a while back in order to be able to fix some of the outstanding bugs where certain kinds of source operations don't honour P-a-s
[09:29] <cjwatson> But never got very far
[09:29] <infinity> I might go through it and see if I can find any valid arguments for us keeping it, push packaging changes back to Debian for things that are obviously arch-restricted but not in control, and call it done.
[09:29] <infinity> cjwatson: I meant retire it entirely, not remodel it.
[09:30] <infinity> cjwatson: It's nowhere near as useful as it used to be.
[09:30] <cjwatson> That would certainly be more fun
[09:30] <cjwatson> I suspect there are cases where architecture restrictions need to be synced across many packages due to some bit of foundational code not being ported, which tends to be non-fun to put in all the packages in question
[09:31] <cjwatson> I can see an argument for centralising such things
[09:31] <infinity> cjwatson: Yeah, but I also don't mind the inverse argument that, if you have the resources to take a few build failures, it's nice to just have the FTBFS logs there as a reminder to effin' port the code.
[09:32] <infinity> cjwatson: And for obviously arch-restricted stuff (things that depend on binary drivers, etc), it really should be arch-restricted down the chain, IMO.
[09:33] <infinity> I'm not sure I'm a fan of the "it's not ported, so put it in P-a-s" use case.  It just buries the list of things that need porting.
[09:34] <infinity> (I was a much bigger fan of that use-case when I was an m68k buildd admin, but...)
[09:34] <cjwatson> I don't think things that aren't ported themselves should be in P-a-s, but collateral damage is a bit different
[09:34] <cjwatson> Perhaps
[09:35] <infinity> Collateral damage often just ends up in perpetual dep-wait, which isn't so bad.
[09:35] <cjwatson> It at least makes no sense to me that google-perftools is in P-a-s at all
[09:35] <infinity> (Like all the rdeps of V8 on powerpc)
[09:35] <cjwatson> Actually the way that LP doesn't quite consistently honour P-a-s in all cases (e.g. copies) is one of the reasons people have ended up overdoing things there, I think
[09:36] <Daviey> it does urk me that there isn't a tool i can say $ what-ya-wanna-build-for *.changes
[09:36] <cjwatson> Hm, no, that makes no sense.  Maybe it failed to honour Architecture or something
[09:36] <infinity> Daviey: Misspelling irk irks me.
[09:36] <Daviey> infinity: urks me more.
[09:37] <infinity> *flinch*
[12:00] <cjwatson> ogra_: Given the legal debacle over the last thing called "utouch", I don't think it's a good idea to abbreviate Ubuntu Touch to utouch.
[12:01] <ogra_> oh
[12:01] <ogra_> there was a legal debacle ?
[12:01] <ogra_> i can indeed rename it to just -touch or -ubuntu-touch (the latter feels a bit longish)
[12:04] <cjwatson> We had to go through everything called utouch in the archive and rename it ...
[12:04] <cjwatson> Trademark dispute
[12:04] <cjwatson> So rejected, please call it something else, don't mind exactly what :)
[12:05] <cjwatson> This is why the touch-as-in-touchscreen stuff is called OIF nowadays
[12:05] <ogra_> ah
[15:26] <stgraber> cjwatson: is it just a local setup issue or are the cdimage unittests failing on saucy?
[15:27] <stgraber> I've been trying to figure out how I broke some random bits with my qa_product rework just to notice that the tests failed even without my changes ;)
[15:28] <stgraber> cjwatson: nevermind, it's just a missing test dependency (zsync)
[15:35] <cjwatson> stgraber: Hmm, not totally sure I intended that to be a dependency.  Let me look
[15:36] <stgraber> cjwatson: ah, ok. I already added it as a dependency in run-tests, so if you can somehow make it unnecessary for the tests, remove it from there
[15:40] <cjwatson> stgraber: done
[15:55] <stgraber> nice, finding a few problems in the current set of tests for qa_product, we were testing for products that don't exist in reality, so now that I check against a fixed list all of those are showing up as failures ;)
[16:40] <stgraber> cjwatson: hey, there's one bit I'm a bit unclear on, are the paths in the images parameter of post_qa supposed to include the .iso or not?
[16:42] <stgraber> cjwatson: hmm, looking at some of the other tests, it looks like the answer is no and test_post_qa_wrong_date is just buggy
[16:42] <stgraber> and testsuite passes!
[16:50] <cjwatson> I believe they should not contain .iso
[16:50] <cjwatson> Indeed that test is wrong, sorry
[16:52] <stgraber> alright, I'll do one more pass through the tracker to make sure my updated tests don't regress in test coverage, then push that and the new etc/qa-products file+parser
[17:54] <stgraber> cjwatson: is there a magic way to have a file copied from the real etc/ into the temporary etc/ used to run the tests? (I want etc/qa-products to be copied over)
[19:09] <stgraber> cjwatson: went with a rather ugly workaround in setup() for now, we can always move to something cleaner later on.
[19:10] <stgraber> cjwatson: and I just pushed the changes now to production including the new cdimage_project function (opposite of qa_product). All tests pass and I triple-checked the product list so I think we're good, I'll still do a bit of grepping in the logs to make sure we don't fail to publish something.
[19:12] <stgraber> cjwatson: qa_product also now returns a tuple instead of just a string, the second value is the name of the tracker instance (iso or localized-iso). I'll push another update to actually use that properly (only zh_CN should be affected).
[19:20] <cjwatson> stgraber: Copying it wouldn't be very unit-testy; better to write out temporary files with just the features you're trying to test
[19:20] <cjwatson> stgraber: Thanks; will have a look later :)
[19:23] <stgraber> cjwatson: well, our existing unit tests cover every single line of etc/qa-products, though I suppose we could change that to use a temporary input file with a smaller subset of tests that actually test the corner cases
[19:25] <cjwatson> The latter is my preferred approach for unit testing, indeed
[21:04] <phillw> cjwatson:  ubuntu server 12.04.2 has been moved from http://cdimage.ubuntu.com/ubuntu-server/precise/daily/20130214/precise-server-amd64.iso on the QA tracker... any idea where it is now?
[21:05] <infinity> phillw: Given that 12.04.2 was released quite a while ago, I'd expect it to be at http://releases.ubuntu.com/12.04.2/
[21:06] <infinity> (Not sure that the ISO tracker is meant to be all that useful post-release)
[21:07] <stgraber> it's useful to grab bug information and statistics, we indeed don't expect the URLs to be of much use though
[21:08] <phillw> infinity: point taken but maybe http://iso.qa.ubuntu.com/qatracker should be updated to a correct link. As a tester, I am used to heading there and grab the link for an install :)
[21:09] <infinity> Definitely not the write place to get "golden" install images, IMO.
[21:10] <infinity> Wow, fingers, did you really just type "write" instead of "right"?
[21:10]  * infinity smacks his fingers around a bit.
[21:10] <phillw> stgraber: In that case, what is the point of having the entries on that page? Just a thought... please do not shout at me.
[21:10] <stgraber> phillw: as a tester you're not supposed to test something that's released ;)
[21:11] <phillw> stgraber: I am supposed to check for regressions :D
[21:11] <phillw> 12.04 is an LTS :P
[21:11] <stgraber> however we're supposed to have a flag that tells you the links won't work but I guess the scanner doesn't run against release milestones (I can probably change that)
[21:13] <phillw> stgraber: thanks, it is the first time I've come across this. I wanted to zsync up a 12.04.1 to 12.04.2 and it failed with the "Oh, it's gone" (404) errors
[21:13] <phillw> -s
[21:14] <phillw> I know server is a pain to track! I'll just do a 'get' on it.
[21:15] <infinity> phillw: You can zsync from the above URL, where it's released...
[21:18] <phillw> infinity: okies :) I did rename it back from 12.04.1 to the expected name. Sadly, I only use the server install for people who request it. I did try to install an 10.04 one for xnox but I could not get it to talk to the world (They are on KVM system).
[21:30] <phillw> infinity: I have http://cdimage.ubuntu.com/ubuntu-server/precise/daily/current/ where should I be heading?
[21:39] <infinity> phillw: releases.ubuntu.com?
[21:39] <infinity> phillw: Expecting released images to be in a directory marked "daily" seems a bit of a disconnect. :)
[21:41] <phillw> infinity: have a look at http://releases.ubuntu.com/ Ubuntu server is not mentioned?
[21:42] <infinity> phillw: If you read scrollback, I pointed you at http://releases.ubuntu.com/12.04.2/
[21:42] <infinity> phillw: ubuntu-server is just another Ubuntu image.
[21:42] <infinity> (releases.ubuntu.com/precise also works, lands you in the same place)
[21:43] <phillw> infinity: and as some one looking for a system... http://releases.ubuntu.com/ it says to head to cdimage server?
[21:44] <infinity> phillw: It tells you to go to cdimage for other flavours (which it then lists).  Ubuntu isn't a flavor, and is linked at the top before that blurb.
[21:45] <phillw> Ah.. thanks, scrolling down I found it :)
[21:45] <cjwatson> It should qualify its mention of Ubuntu there.  I'll fix that.  Otherwise, most people have no trouble finding server images given they're linked from the website and other prominent places
[21:46] <cjwatson> (done)
[21:49] <phillw> sorry for annoying, I'm a tester and get asked away from here, to grab an iso / update an iso not and again. Having a simple link to zsync is what I ask for when people have a problem installing. I know that this is not a release team issue, thank-you for your understanding :)
[21:49] <phillw> s/not/now
[22:08] <phillw> cjwatson infinity: I have the link of http://releases.ubuntu.com/12.04.2/ubuntu-12.04.2-server-amd64.iso.zsync and I get the error :
[22:08] <phillw>  http://releases.ubuntu.com/12.04.2/ubuntu-12.04.2-server-amd64.iso.zsync
[22:08] <phillw> -bash: http://releases.ubuntu.com/12.04.2/ubuntu-12.04.2-server-amd64.iso.zsync: No such file or directory
[22:09] <cjwatson> You might like to try a command in front of that.
[22:10] <cjwatson> Like zsync ...
[22:13] <phillw> cjwatson: there are times I hate you :D As  I said, I'm used to qa-tracker which does include it when copying and pasting :D
[22:14] <phillw> 3:04 mins to go :)
[22:36] <phillw> cjwatson: not that people ever bother, but http://phillw.net/isos/ubuntu-server/precise/ has been updated.