[00:22] <mwhudson> mwhudson@glamdring:go-test-cases$ go get -v qastaging.launchpad.net/pydoctor
[00:22] <mwhudson> Fetching https://qastaging.launchpad.net/pydoctor?go-get=1
[00:22] <mwhudson> Parsing meta tags from https://qastaging.launchpad.net/pydoctor?go-get=1 (status code 200)
[00:22] <mwhudson> import "qastaging.launchpad.net/pydoctor": parse https://qastaging.launchpad.net/pydoctor?go-get=1: no go-import meta tags
[00:22] <mwhudson> package qastaging.launchpad.net/pydoctor: unrecognized import path "qastaging.launchpad.net/pydoctor"
[00:22] <mwhudson> what?
[00:24] <wgrant> That seems unlikely.
[00:25] <wgrant> mwhudson: Oh, it's not updated yet.
[00:26] <wgrant> blr: Did you actually land the branch?
[00:28] <mwhudson> oh
[00:28] <mwhudson> doh
[00:34] <blr> wgrant: hmm I did
[00:34] <blr> but no response from pqm...
[00:34] <wgrant> Try again, I guess.
[00:35] <blr> resubmitting
[00:37] <blr> wgrant: nothing in the pqm queue afaict, yet I'm not getting an error on lp-land
[00:42] <blr> mwhudson: sorry, it appears I've killed pqm.
[00:42] <mwhudson> blr: congrats
[00:42] <wgrant> blr: Missing QA tags
[00:42] <wgrant> Jul 03 00:36:01 pqm [140255997114112] INFO: subject: [r=cjwatson,
[00:42] <wgrant>  wgrant] Fix first part of golang import path for the bzr case,
[00:42] <wgrant>  rendering root/product.name rather than root/branch.unique_name.
[00:43] <blr> oh, doesn't it usually complain about that?
[00:43] <wgrant> I'm surprised lp-land didn't complain, yeah.
[00:46] <blr> wgrant: is the issue that the linked bug is already tagged 'qa-needstesting'?
[00:48] <wgrant> blr: Or Fix Committed, probably, indeed.
[00:51] <blr> hmm, if a second branch is linked against a fix-committed bug, should we change the status?
[00:51] <blr> potentially that makes things more confusing, I don't know.
[00:52] <blr> mwhudson: at any rate, that's with buildbot now, sorry for the delay.
[00:52] <mwhudson> heh no worries
[00:52] <mwhudson> the delay meant i had no choice but to submit an expense claim, so thanks :)
[02:19] <blr> wgrant: hmm spurious buildbot failure?
[02:20] <wgrant> blr: Yep, haven't seen that one in a few days.
[02:20] <blr> wgrant: how do I restart the build?
[02:22] <wgrant> blr: http://lpbuildbot.canonical.com/force
[02:23] <blr> heh "carry on"
[02:24] <blr> ah, there's a link on the index...
[02:24] <wgrant> Yeah
[02:25] <wgrant> The index that nobody ever goes to, because it's just an index.
[02:25] <blr> I will say one thing for the UI, it actually renders ok on my phone.
[02:28]  * mwhudson tries to remember if the "carry on" message was him, it sounds kinda familiar
[02:31] <wgrant> 18.3.13: Michael Hudson 2009-09-03 less lame success html
[02:32] <blr> hahah
[02:34] <mwhudson> >:)
[02:38] <blr> hmm, failure in code.model.tests.test_branchjob.TestBranchScanJob.test_branch_deleted - that seems... improbable.
[02:43] <blr> passes locally
[02:43] <blr> wgrant: is that one you're familiar with?
[03:44] <maozhou> Package debian-install failed to build on my launchpad, a lot of package has no installation candidate.  http://chuantu.biz/t2/10/1435894813x-954497676.png
[03:46] <maozhou> The Base URL is set to "http://archive.ubuntu.com/" , is it incorrect?
[04:05] <blr> mwhudson: buildbot decided to behave finally, should be on qastaging soon.
[04:05] <blr> let me know how you get on.
[04:25] <mwhudson> will do
[04:26] <wgrant> maozhou: Those are udebs.
[04:26] <wgrant> debian-installer is a complicated package. You will need to debug it.
[04:29] <maozhou> wgrant: But this package is succeed  to build on launchpad.net , I download the same package
[04:30] <wgrant> maozhou: Yes, but your archive is not identical to the one that Launchpad has.
[04:31] <wgrant> maozhou: debian-installer pulls in udebs from the archive during its build, so it's by no means as simple as other packages.
[04:32] <maozhou> wgrant: Can I access the archive which Launchpad.net used?
[04:32] <wgrant> maozhou: That's archive.ubuntu.com
[04:33] <wgrant> Your debian-installer build is apparently not using it.
[04:33] <wgrant> I do not know how debian-installer constructs its sources.list, but it's not a Launchpad problem.
[04:33] <wgrant> For builds in Ubuntu itself, the entire archive is known to Launchpad.
[04:34] <wgrant> I assume you don't have all of it imported into your Launchpad instance, so you are using the OEM hack to add extra sources.list lines.
[04:37] <maozhou> wgrant: But in build.log which on Launchpad.net ,the archive is set to "http://ftpmaster.internal" , this it a internal archive which other people can't access?
[04:38] <wgrant> maozhou: archive.ubuntu.com is a mirror of that.
[04:44] <maozhou> wgrant: If set Base URL to "http://archive.ubuntu.com/" , the same error,  I don't understand why "launchpad.net" builded it succedd
[04:45] <wgrant> maozhou: Neither do I. You'll need to look at the package to see what it's doing to find why it might be failing.
[04:46] <maozhou> wgrant: ok , thank you.
[04:47] <wgrant> mwhudson, blr: Can you look at that rev on qastaging?
[04:48] <blr> wgrant: yep
[04:51] <blr> wgrant: lgtm, mwhudson might want to test with go-get however.
[04:52] <wgrant> Looks fine to me.
[04:57] <maozhou> wgrant: I guess the source.list is not intact , need I uncompress the chroot file ,  modify it ,  and  rerun "./manage-chroot ..... " ?
[05:01] <blr> wgrant: deployment unblocked now
[05:11] <wgrant> maozhou: launchpad-buildd overwrites sources.list, so the one in the chroot tarball does not matter.
[05:14] <maozhou> wgrant:  If modify "/usr/share/launchpad-buildd/slavebin/override-sources-list" on builder ,  the builder won't overwrites source.list.
[05:22] <wgrant> maozhou: It is the case that if you disable the sources.list overwriting, the sources.list will not be overwritten, yes.
[05:22] <wgrant> But I'm not sure why you'd want to do that.
[05:25] <maozhou> wgrant: The archive in source.list is set to "deb archive.ubuntu.com utopic  main" as default, is it not intact ?
[05:26] <wgrant> maozhou: That's not a valid sources.list line. I would expect "deb http://archive.ubuntu.com/ubuntu utopic main"
[05:27] <wgrant> maozhou: But you've never explained what your actual purpose is.
[05:27] <wgrant> For weeks you've been layering hacks upon hacks.
[05:27] <wgrant> If you can explain what you're actually trying to do with your Launchpad instance, we may be able to advise the lowest-friction way to go about that.
[05:32] <maozhou> wgrant: I want set up a local launchpad server to build packages.
[05:33] <wgrant> That's not an end goal.
[05:33] <wgrant> You want to build packages for a purpose.
[05:33] <wgrant> Are you creating a derived distribution? Do you just want to have a PPA that is hosted on your premises?
[05:38] <maozhou> wgrant: Yes, I want to build packages and creating a derived distribution.
[05:39] <wgrant> maozhou: And you understand the limitations that doing that as an overlay entails?
[05:40] <maozhou> wgrant: Yes
[05:42] <wgrant> maozhou: OK
[05:43] <wgrant> maozhou: So it sounds like you've set the archive base URL on /ubuntu/+pubconf to "archive.ubuntu.com"
[05:43] <wgrant> maozhou: That's not going to work very well if you're uploading packages to ubuntu, as they won't be able to see each other as build-deps.
[05:44] <wgrant> What you may want to do is set /ubuntu/+pubconf's base URL to http://archive.ubuntu.com/, then upload to an overlay derivative distribution that isn't ubuntu.
[05:44] <wgrant> That distribution will reference itself in its sources.list, followed by archive.ubuntu.com.
[06:04] <maozhou> wgrant: why they won't be able to see each other as build-deps?
[06:05] <wgrant> maozhou: Because if the archive base URL is set to http://archive.ubuntu.com/ rather than, for example, http://archive.launchpad.dev/, your archive will not be in sources.list. Only Ubuntu will be.
[06:11] <wgrant> maozhou: You can see the entries that will be in the final sources.list near the top of the build log.
[06:20] <maozhou> wgrant: I konw this, I don't understand why the package debian-installer failed to build , the same archive , the same package.
[06:22] <wgrant> maozhou: Examine the sources.list that debian-installer generates.
[06:22] <wgrant> maozhou: IIRC it logs it to the build log
[06:22] <wgrant> sources.list.udeb or something
[06:24] <maozhou> wgrant: Oh, you are right , may be that's the answer.
[06:27] <maozhou> but, if the source.list  is generated by debian-installer , the same package debian-installer, the  souce.list.udeb should be same.
[06:28] <wgrant> It is not generated in a vacuum.
[06:28] <wgrant> The sources.list in the environment is used as input.
[06:33] <maozhou> Ok, I konw.
[07:18] <vila> blr: re "another bzr release", I offered my help for that to the GNU maintainer but haven't received a reply.
[07:19] <vila> bzr, wgrant, cjwatson : and thanks for the less verbose MP comment emails ;-)
[07:52] <mwhudson> blr, wgrant: more hilarity!
[07:52] <mwhudson> mwhudson@glamdring:shared-lib-stuff$ go get -v qastaging.launchpad.net/pydoctor
[07:52] <mwhudson> Fetching https://qastaging.launchpad.net/pydoctor?go-get=1
[07:52] <mwhudson> Parsing meta tags from https://qastaging.launchpad.net/pydoctor?go-get=1 (status code 200)
[07:52] <mwhudson> get "qastaging.launchpad.net/pydoctor": found meta tag main.metaImport{Prefix:"qastaging.launchpad.net/pydoctor", VCS:"bzr", RepoRoot:"http://bazaar.qastaging.launchpad.net/~mwhudson/pydoctor/github"} at https://qastaging.launchpad.net/pydoctor?go-get=1
[07:52] <mwhudson> package qastaging.launchpad.net/pydoctor: cannot download, http://bazaar.qastaging.launchpad.net/~mwhudson/pydoctor/github uses insecure protocol
[07:52] <mwhudson> mwhudson@glamdring:shared-lib-stuff$
[07:52] <mwhudson> blr, wgrant: this check is new in 1.5
[07:52] <mwhudson> is there any anonymous way of getting the code over a secure protocol currently?
[07:54] <mwhudson> well bzr branch https://launchpad.net/product i guess but that is a bit hysterical raisins
[08:14] <wgrant> mwhudson: No, no anonymous secure Bazaar access today.
[08:14] <wgrant> mwhudson: I guess using the webapp branch reference might confuse it enough...
[08:34] <vila> wgrant, blr , cjwatson : MP comment emails... <3
[08:48] <blr> mwhudson: oh fun :(
[08:49] <blr> well, fair enough I suppose.
[08:50] <blr> wgrant: any suggestions?
[08:50] <wgrant> blr: You might be able to test it locally if you can disable its certificate validation.
[08:50] <wgrant> I assume it just passes the URL down to bzr once it checks it.
[08:51] <wgrant> So we can work around the security check by using https://launchpad.net/$unique_name as the path. BranchRefNavigation allows a bzr client to use that.
[08:52] <wgrant> (bzr looks up https://launchpad.net/~launchpad-pqm/launchpad/devel/.bzr/branch/location and follows that URL. which go get won't see, so can't complain about)
[08:54] <blr> mwhudson: is your fork on github?
[08:55] <blr> https://github.com/mwhudson/go>
[08:55] <blr> presumably?
[08:58] <blr> wgrant: shall we add supermirror_root_https to config.codehosting?
[08:59] <blr> oh I nvm, I see
[09:00] <blr> so, {vhost.mainsite.hostname}/{branch.uniquename}
[09:03] <wgrant> blr: It needs a protocol too.
[09:03] <wgrant> eg. allvhosts.configs['mainsite'].rooturl
[09:06] <blr> wgrant: thanks
[09:06] <mwhudson> blr: go get only special cases launchpad.net exactly
[09:07] <mwhudson> not qastaging.launchpad.net
[09:07] <mwhudson> blr: so just build go tip
[09:07] <wgrant> Or apt install golang-go
[09:08] <wgrant> Oh, that's too old.
[09:08] <mwhudson> yeah
[09:09] <mwhudson> this secure insistence is pretty new
[09:09] <mwhudson> <1 month i think