[00:20] I am not completely sure that LP forbids copying source+binaries from a PPA to the primary archive if you aren't an archive admin. I suspect it doesn't actually forbid this but should. [00:21] It should be possible to test on staging; the publisher won't run, but the copy job will (eventually). [00:21] (Tests that need the publisher to run need to be done on dogfood.paddev.net with assistance from an LP developer.) [00:21] cjwatson: As far as I remember, the only way to do that is through the CI Train. [00:21] Otherwise an archive admin is needed. [00:22] (And that sort of thing would be documented in the CI Train logs publicly anyways.) [00:22] tsimonq2: From code inspection I believe that you are mistaken, but feel free to prove me wrong. [00:23] (I think your understanding is how it *should* work, but ...) [00:23] Well, anybody with upload access can use the tool, but an archive admin has to approve it because it comes from a PPA. [00:24] (I remember this from when we had problems landing something with the CI Train so I asked Gianfranco to use copy-package and he needed to poke an archive admin.) [00:25] cjwatson: Maybe there's an edge case I'm not accounting for, but as far as I remember, that's how it works. [00:26] tsimonq2: I don't believe any of the above is true. [00:27] cjwatson: Which part of it? [00:27] tsimonq2: Any of the last three lines :) [00:27] Well, it's true that anyone with upload access can also copy. The rest is AFAIK false. [00:27] cjwatson: It would help if I could access staging.launchpad.net. ;P [00:27] There may have been some other reason that approval was required. [00:27] staging is always down at the weekend for restoration from a production dump. [00:28] It takes a couple of days. [00:28] Alright. [00:28] Will it be back tomorrow? [00:28] Hopefully. [00:28] Alright, I'll check back with you once it's back up and once I've tested this. [00:28] Any quirks I should be aware of when using staging for this sort of thing versus production? [00:29] (Publisher shouldn't be needed for this.) [00:30] LP staging uses SSO staging for auth, but otherwise not that I can immediately think of. Oh, you won't be able to build test packages to copy, so you'll have to pick something already built. [00:30] If it doesn't work you can always use dogfood instead. [00:31] I might actually just use dogfood so I have the ability to test rebuild a package in a PPA with Backports enabled. [00:31] cjwatson: How would I get copy-package to work with dogfood then? [00:32] -l dogfood [00:32] Alright. [00:32] As I say you may need dev assistance - a lot of the usual cron jobs don't always run. [00:32] Alright. [00:33] Upload processing runs frequently, as does the copier. The PPA publisher doesn't currently, but can. [00:33] https://paste.ubuntu.com/26205792/ (obviously substituting your username) works for uploads. [00:34] All details subject to change as dogfood is a developer playground. [00:34] Alright. [00:34] cjwatson: Is that for PPAs or archive uploads? [00:34] (Would it be the same with s/upload/ppa/ ?) [00:36] Keep the FQDN the same, but you can upload to any archive that you have access to that way - it takes an archive reference following dogfood-sftp: on the dput command line. [00:36] i.e. the same syntax that you see at the end of copy-package --help [00:36] Alright. [00:37] (Except possibly a leading "ppa:" might not work, I forget.) [00:37] I'll play with it. [00:42] cjwatson: Could you please let my packages build? :) [00:42] ? [00:43] That shouldn't block on us ... [00:43] Oh, conceivably all the builders are down [00:43] As far as I can tell, the Launchpad builders for several arches are on Manual. [00:44] We often use dogfood for testing new builder clouds and such. [00:44] Oh, I can see that. [00:45] I've turned on a couple. [00:45] Er, or not. [00:45] Now I have. [00:48] (You might find it less time-consuming to use smaller packages for testing things like copies, though ...) [00:48] (I did realize that after the fact, but oh well.) [00:49] s/realize/remember/ [01:01] cjwatson: Ftr, this is an example of a copy including binaries from a Bileto PPA: https://launchpad.net/ubuntu/+source/qtdeclarative-opensource-src/5.9.2-3ubuntu1/+publishinghistory [01:01] cjwatson: Looking back in logs, Gianfranco executed "./copy-package --from=ppa:ci-train-ppa-service/ubuntu/3020 --from-suite=bionic --to=ubuntu --to-suite=bionic-proposed --include-binaries qtdeclarative-opensource-src" [01:04] tsimonq2: That was held for approval simply because bionic wasn't fully open yet at that time. [01:04] Proves nothing much - all uploads were held for approval. [01:04] cjwatson: Oh, good point. [01:04] Er, wait... was it? [01:05] Ah, hm, I'm wrong. [01:05] Yep: https://lists.ubuntu.com/archives/ubuntu-devel-announce/2017-October/001231.html [01:10] tsimonq2: So ... was it in fact held in unapproved? I see no mention of it in #ubuntu-release IRC logs around that time. [01:11] cjwatson: If I remember correctly, yes. [01:13] [2017-11-02 15:41:15,521: INFO/Worker-2] Running (ID 39344156) in status Waiting [01:13] [2017-11-02 15:41:21,426: DEBUG/Worker-2] Packages copied to Primary Archive for Ubuntu: [01:13] And in fact more clearly: [01:13] [2017-11-02 15:41:21,163: DEBUG/Worker-2] Subject: [ubuntu/bionic-proposed] qtdeclarative-opensource-src 5.9.2-3ubuntu1 (Accepted) [01:13] tsimonq2: This upload was not held for approval. [01:14] cjwatson: Could it be that CI Train PPA uploads are whitelisted? [01:14] tsimonq2: It is possible (even probable) that Bileto required an archive admin to sign it off, but that's it deliberately going over and above Launchpad's checks. [01:15] cjwatson: But isn't this intentional for the function of Bileto? [01:16] tsimonq2: Sure, Bileto has to run its copies with somewhat elevated privileges and thus has to do some checks itself first, but that mechanism wasn't used here. [01:17] tsimonq2: It looks like Gianfranco did the copy manually and thus bypassed Bileto's checks. [01:17] cjwatson: Hm, so this makes me wonder. When doing `./copy-package -l dogfood -s zesty --from=ppa:tsimonq2/ubuntu/testing-backports --to=ubuntu --to-suite=zesty-proposed vlc` I get "The signer of this package has no upload rights to this distribution's primary archive. Did you mean to upload to a PPA?" which makes me think that a proper mechanism *is* in place. [01:18] Er... [01:18] This database is from before I joined MOTU. [01:18] So that makes sense. [01:18] tsimonq2: There are checks that you're permitted to upload, but those checks don't include anything that cares about whether you used the --include-binaries flag, AFAICS. [01:18] That's all I'm saying here. [01:19] cjwatson: Right, so could you add me to MOTU on dogfood so I can try this one more time? (Unless I'm misunderstanding you here?) [01:19] Your test was wrong anyway, since you didn't use --include-binaries. [01:20] Right, and I've now included that. [01:21] tsimonq2: Added you to ~motu on DF [01:22] cjwatson: Oh, look: https://dogfood.paddev.net/ubuntu/zesty/+queue?queue_state=1 [01:22] So it does indeed show up as a sync in unapproved. [01:23] tsimonq2: Oh, look: https://dogfood.paddev.net/ubuntu/+series [01:23] zesty is CURRENT [01:23] you need to test artful [01:23] Indeed. [01:24] (otherwise you're just testing the "is this an SRU?" logic) [01:24] Gotcha. [01:24] * cjwatson -> bed [01:24] o/ === Elimin8r is now known as Elimin8er === NCommander is now known as KD2JRT === KD2JRT is now known as NCommander [15:00] !dmb-ping [15:00] bdmurray, BenC, cyphermox, jbicha, micahg, rbasak, sil2100: DMB ping. === Mister_Q_ is now known as Mister_Q [17:53] jbicha: ah, yes, teela is short of disk space [19:10] ahasenack: Congrats! [19:11] thanks tsimonq2 [19:17] // [19:17] === kmously is now known as kmously-afk [20:11] jbicha: ok, it should update again today [20:16] thank you :) === ogasawara is now known as Guest76312 [21:41] An SRU of mine was promoted from xenial-proposed to -updates, but was pulled after 10% propagation because of an end-user error report. I got an email with a link to the error report, but don't have authorization to see it. Filled out the form requesting access, and it's been pending for a week. Little help here? [21:46] bdmurray: are you around? ^ [21:46] rlink: a lot of people are on holiday break [21:55] RAOF: hi, if you are around tomorrow (tues) could you take a look at neutron in the xenial upload queue? that one is hot for a customer.