[00:37] hola [00:37] trying to find some flows of the usb iniating stuff [00:38] seems i have a race condition for my device [00:38] sometimes it docks like a ttyusb** and sometimes it docks the driver to become an ethernet device [00:38] any pointers? [00:39] i think its related to the host. but i cannot debug it / disable the false action [00:39] host as in the usb part [00:41] is there any diagram of the usb support systems in ubuntu? === ben_r_ is now known as ben_r [01:15] is channel alive? [01:20] That is what I was wondering to about my ubuntu installer questions. [01:20] no one is here or just sitting [01:20] i dunno [01:20] i was redirected here [01:20] is there any arch documenations of ubuntu? [01:20] Do you know how to checkout with bzr the ubuntu installer? [01:20] nope [01:21] iam trying to figure out how the usb drivers flow looks like [01:21] i can hack a debugger [01:21] Are there development versions of ubuntu? [01:21] i dont know [01:21] iam merely a user [01:21] trying to understand how stuff work [01:21] ok me to. [01:22] i dont think they care tbh [01:22] not about irc that is [01:22] I think this is just a user channel not an officially supported channel. [01:22] agreed [01:23] It could be the fact it is the weekend now and you might have better luck on a week day. [01:23] could be [01:25] stefandxm, best chance in #ubuntu-devel is during officehours, uk, but most devs are on the mailinglist, not irc [01:25] so that might be it [01:27] That is understandable. [01:27] agreed === cpaelzer__ is now known as cpaelzer [12:15] rbasak: hi, I need some help with your chdist-if script again [12:15] rbasak: https://pastebin.ubuntu.com/p/rvx6NRMSN3/ is the autohinter section [12:16] rbasak: I also have this bit from you, last time you helped: https://pastebin.ubuntu.com/p/yj4nxccWSq/ [12:18] ahasenack: I don't see the issue immediately [12:18] :-/ [12:18] same, all combinations I try just work [12:18] Also looks like my method doesn't handle Conflicts well [12:27] W: Unable to locate package bind9-libs [12:27] ^ looks like that's new in focal-proposed [12:27] I wonder if that's related or a red herring. [12:29] bind9-libs is new in proposed only, yes [12:29] so there is src:bind9-libs in f-p, and bin:bind9-libs in f-p [12:30] not from the same source [12:31] I don't follow the rearrangement [12:32] Why doesn't the new bind9 in proposed depend on bind-libs or its dependants? [12:32] Uh [12:32] rbasak: src:bind9-libs is for 9.11.x, whereas src:bind9 (9.16.x) produces a binary package called bind9-libs that pulls together all the individual library packages that 9.11 used to produce [12:32] Why doesn't the new bind9 in proposed *build-depend* on bind-libs or its dependants? [12:33] isc-dhcp doesn't build with bind9-9.16.x, it needs 9.11.x [12:33] so we have two binds in main [12:33] one to provide only libraries for legacy apps, and the other package the server at its newest lts version [12:35] last I heard from vor-lon this was entangled in a debian-installer installability problem due to a new kernel, but that was last week, and seems to have resolved [12:36] Ah [12:36] I hadn't been keeping up-to-date with all the details of this [12:36] Should I catch up by asking a million questions, or leave it to you and vor-lon and others who have looked at it alreay? [12:36] this was discussed with foundations and security, I wouldn't do something crazy like this without peer checking :) [12:37] and in standups as well ;) [12:37] I had understood something complicated was going on [12:37] I'll keep looking, I was more wondering if I used your script correctly [12:37] I missed that the resolution was two source packages. I'm not complaining! [12:37] I think you are, but there might well be limitations in my script here [12:38] and my knowledge of udebs [12:38] where is this package stored, for example: libisccc-export161-udeb [12:38] I don't understand why britney says that autodns-dhcp would become uninstallable (starting at the beginning) [12:43] rbasak: ahasenack: Are you talking about the autohint "bind-dyndb-ldap/11.2-1build2 bind9-libs/1:9.11.16+dfsg-3~build1 isc-dhcp/4.4.1-2.1ubuntu2 [12:44] " [12:44] ? [12:44] yes [12:44] then it seems to come down to https://paste.ubuntu.com/p/9xXK6ySCxS/ [12:45] afaics [12:45] bind9's not in that hint [12:46] so bind9-host comes from src:bind9. There is src:bind9 9.11.x in release, and src:bind9 9.16.x in proposed [12:48] why isn't bind9-host 9.11.14 not upgrading to bind9-host 9.16? [12:48] I don't know why the auto-hinter didn't put it in there [12:48] I can ask the hint-tester and see what would happen [12:49] I don't know why my script can resolve that [12:49] But I agree src:bind9 isn't in that hint [12:49] libbind9-161 didn't change sonames between 9.11.14 and 9.11.16 [12:50] Laney: how does that work, ask the hint-tester? [12:52] I can run it with the --hint-tester flag on the archive machine [12:52] what does it do? [12:52] https://paste.ubuntu.com/p/jGwnq6JRMF/ [12:52] that [12:52] so, debian-installer-udebs [12:52] d-i was rebuilt a few times since bind9 was uploaded [12:54] I see its depends line [12:54] * ahasenack checks [12:54] the thing with vim in focal where middle click pastes the vim buffer not the x one is driving me bananas [12:55] mwhudson: +1 [12:56] ah https://bugs.launchpad.net/ubuntu/+source/vim/+bug/1864424 [12:56] Launchpad bug 1864424 in vim (Ubuntu) "new vim version in 20.04 sets mouse=a by default" [High,Fix committed] [12:58] in d-i-udebs, I see the diff in the libdns-export and libisc-export sonames, that looks correct [12:58] and I see errors installing kernel -di packages [12:58] so maybe it's still the issue vor-lon said [12:59] well https://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#debian-installer [13:09] ahasenack: but has the d-i been rebuild against the right kernel, and is that kernel out? =) as far as i can tell -16 kernel no longer exists, it's iether 14 (release pocket) or 17 (proposed pocket) [13:10] xnox: I don't have more details [13:10] ahasenack: i think we should rebuild d-i against 17 kernel, and wait for kernel to be released. [13:11] "debian-installer is all that's left, and it needs linux-meta to drop the snapdragon metapackages for a kernel flavor we're no longer building" [13:11] that's what vor-lon said last week, does that ring any bells? [13:11] whilst true, the statement is incomplete =) [13:11] well [13:11] rather things have changed since that was said [13:12] indeed vlan-modules-5.4.0-17-generic-di exists [13:13] but not vlan-modules-5.4.0-17-generic-di [13:13] er [13:13] but not vlan-modules-5.4.0-16-generic-di [13:13] yes, uploading d-i rebuild against -17 [13:13] thank you! [13:15] however [13:15] i think the snapdragon bug is still there [13:15] https://launchpad.net/~canonical-kernel-team/+archive/ubuntu/unstable/+build/18772755 [13:15] shows that meta for 17 still generates snapdragon packages, let's hope they are still installable.... [13:17] anyway [13:17] the rebuild for -17 was necessary anyway, right [13:17] ahasenack: let's wait for this https://launchpad.net/ubuntu/+source/debian-installer/20101020ubuntu603 to finish building and getting pickup by excuses. [13:17] then we will see what else is left [13:17] ok [13:21] OK so chdist-if-migrated has a bug I think [13:22] I was using a script I wrote a few years ago: https://code.launchpad.net/~laney/ubuntu-archive-tools/update-output-helper [13:23] "chdist apt-cache focal-proposed showsrc bind9-libs" outputs "Package: bind9" as well as "Package: bind9-libs" [13:23] I wasn't expecting that [13:24] Ah [13:24] I need --only-source apparently [13:24] Laney: the mp is pending since 2015? :) [13:25] That fixed it [13:25] rbasak: diff please ? :) [13:25] yeah... [13:25] I gave up on getting it merged and just use it locally :P [13:25] Laney: is that branch up-to-date? [13:26] https://code.launchpad.net/~laney/ubuntu-archive-tools/update-output-helper specifically [13:26] or, if it's standalone, could you just paste it? [13:28] ahasenack: http://paste.ubuntu.com/p/XHngKFM5K4/ [13:29] thx [13:31] it's not going to be up to date with respect to lp:ubuntu-archive-tools itself but you can grab the script out of it if you want to use it [13:31] kanashiro: did you prepare a ruby-defaults transition tracker? [13:32] sounds like chdist-if-migrated is doing the same thing though? [13:33] the later has some complicated preparation steps, that's why I wanted to look at yours too [13:33] latter [13:34] okey [13:35] well of course feel free to use, improve or steal ideas [13:35] :) [13:42] kanashiro: doko actually made a ruby2.7-only tracker now fwiw [14:18] mdeslaur: Do you care for usb-creator? I was wondering if it should still be in main. [14:18] bdmurray: what do you want to replace it with? [14:19] I don't particularly care about it [14:20] Time to stick this into git [14:21] ahasenack: I added chdist-if-migrated to https://git.launchpad.net/~racb/+git/tools/ and added a note on how to use it to detect d-i related migration issues also [14:22] Now it displays exactly what udebs need addressing: http://paste.ubuntu.com/p/SpxDtGKybr/ [14:23] mdeslaur: dd [14:23] nice! [14:23] bdmurray: lol [14:24] juliank, doko is this ruby2.7-only transition page live? I am not finding it [14:25] kanashiro: i think ben does not pull new configs very often or something [14:26] it does each time it runs [14:26] Laney: then it does not run very often :) [14:26] Laney: end result is the same :D [14:29] Looks to me like it's running right now [14:33] ack [15:04] Hi [15:05] i have a strange problem which is hard to explain well [15:05] we use ldap auth, and usernames are mail adresses [15:05] so we have some mail adresses > 32 chars [15:06] its not a problem it works, but in some cases, when i try to log in tty, it just don't asks for password [15:06] why dont you file a bug report? [15:06] where ? [15:06] I guess starting with whatever pam plugin for ldap you use [15:06] just to finish : problem is easy to reproduice [15:06] it's ubuntu package in launchpad [15:07] juliank: it isn't ldap related [15:07] then against the pam pacakge? [15:07] i precise that just to explain how i can have some > 32 usernames [15:07] so its maybe pam... [15:07] I mean, you wrote that all days ago already [15:07] yep [15:07] and this sounds like a clear bug [15:08] but i didn't know what component is involved [15:08] pam is the reasonable starting point for that [15:08] or login [15:08] ok, thx juliank [15:08] (login being in shadow) === hggdh is now known as hggdh-msft [15:16] ok didn't see but i get "checkname failed: Operation not permitted" when i reproduice the bug in /var/log/auth.log [16:19] xnox: the "linux" packae has a block proposed tag: https://launchpad.net/bugs/1865025 [16:19] Launchpad bug 1865025 in Kernel SRU Workflow regression-testing "focal/linux: 5.4.0-17.21 -proposed tracker" [Medium,In progress] [16:19] I wonder if that will be worked on this week [16:27] ahasenack: that's normal process for linux packages. [16:27] ahasenack: it will be removed, when linux is ready to migrate. [16:28] * ahasenack sits tight [16:28] for details you can go to the adt matrix [16:28] https://people.canonical.com/~kernel/status/adt-matrix/ [16:28] from there you can click on the focal-meta one [16:28] https://people.canonical.com/~kernel/status/adt-matrix/focal-linux-meta.html [16:29] from there you can see that there a few "MISS" => missing tests (because deps changed) [16:29] or regressions, i.e. on the linux package itself [16:30] ahasenack: but kernel team handles that [16:30] thanks [16:31] it is unfortunate like that that most udebs get entangled together. [16:46] hello. i performed a kernel bisection at the request of an ubuntu developer andnow i have about 30 broken unofficial kernel packages installed [16:47] how do i remove them and go back to the official kernel? [16:47] do i have to dpkg remove each one individually? [16:48] if so how do i tell which ones are not from the repos? [20:04] Hi, If I already have bzr installed and the git clone of the debian-installer how without making a branch on launchpad would I check out the ubuntu installer what would I type? [23:01] hey, for retriggering autopkgtests, I have a question on how to automate the sso/2fa process. [23:01] I know the trick to edit the URL, e.g. https://autopkgtest.ubuntu.com/request.cgi?release=focal&arch=armhf&package=asciidoctor&trigger=ruby-defaults%2F1%3A2.7%7E0 [23:01] and then load that in firefox [23:02] however if we need to requeue, say, all the arch's, then that's N urls to load up. [23:02] I'd like to load the URL in curl or wget or some such, however this requires SSO [23:03] I've been tinkering with retrieving the cookies doing the series of request/response stuff, but haven't gotten it to work end to end yet [23:04] I'm wondering if this is already a figured out process, and/or if anyone knows an existing example of CLI operation against login.ubuntu.com with SSO and 2fa? [23:07] I've been experimenting with usso-login, which seems like it should do it, but no luck. The docs for it are kind of sketchy, so if anyone's use that I'd be interested in tips/examples. [23:10] bryce: does this help? https://git.launchpad.net/ubuntu-security-tools/tree/README#n197 [23:11] bryce: it reads the cookie out of firefox, so you've got to be authenticated through that [23:12] sarnold, let me give it a shot [23:31] sarnold, hmm, seems the moz sqlite database tables have changed since this was written [23:31] bryce: hmm. I wonder which of our tools use this thing, and if they still work :) [23:31] heh [23:33] I found that usso-login is able to perform the 2fa transaction ok, but it stores the data as json rather than as cookie data.