[00:33] wgrant: you around? [00:54] wgrant: I'll just drop Q here and if you have a chance I'd appreciate an opinion. [00:57] I think https://code.launchpad.net/conn-check is the tool OLS/snap store were using for external connectivity diagnosis. [00:57] Is that still your preferred direction for that function in headless environments where you have no ssh access? [01:27] blahdeblah: Hmm, I think that'd be my preference, yes. [01:42] wgrant: cool - thanks for the quick response. Hope things are going well with you. [02:09] blahdeblah: Likewise :) [09:56] jamespage, coreycb: python-pbr autopkg test fails [10:44] doko, what is the plan for linux/5.3.0-16.17/i386 autopkgtest in eoan? [10:44] doko, i can add a hint to let systemd migrate, but there are other options === ricab is now known as ricab|lunch [12:23] good morning o/ [12:24] dannf: tks for testing sg3-utils, ill ask someone to review the mr! [13:20] hi, is there a way (dch or similar) to change version and distro in the last entry of a d/changelog, without adding entries? [13:24] I tried something like dch -v -D "" but that does create a new entry === ricab|lunch is now known as ricab [13:50] ackk: How about doing it manually? [13:50] GunnarHj, yeah that's what I have right now, was just wondering if there was nicer way [13:51] ackk: Ok. [14:51] ackk: dch -e ? [14:58] why does systemd autopkgtest pull in linux from -proposed? https://people.canonical.com/~ubuntu-archive/proposed-migration/eoan/update_excuses.html#systemd [15:01] i mean why does linux's autopkgtest do that when triggered by systemd :-\ [15:09] vorlon, I just tried on a uefi machine, booted in legacy mode, install on entire drive, the ESP is not created. [15:09] jibel: ok. It's possible we missed the boat for 19.04, and only managed to land this in 19.10dev + 18.04.3 [15:09] 2 days old image [15:09] wait [15:09] you mean it doesn't create it even with eoan? [15:09] yes [15:10] then that sounds like a regression [15:10] eoan 20191001.2 [15:10] and if I reboot it'll obviously won't boot [15:10] ok, I will go digging and snarling [15:10] thanks [15:10] thx :) [15:11] uefi machine booted in legacy mode though [15:12] I don't think partman-auto ever accounted for creating the ESP [15:12] (it's marked $efi IIRC) [15:13] ergo, if it worked in previous releases, I wonder why; or somebody already addresses the recipes and I'm operating from incorrect data [15:14] cyphermox: there was a whole thing about this where I was insisting that we always create the ESP on amd64 and that we install both grub-pc and shim-signed [15:14] where did that go? [15:15] *shrugs* I don't know. I agree we should try to always create an ESP, but I don't recall who did the work or if it was finished [15:15] I just pulled partman-auto to have a look [15:17] cyphermox: https://bugs.launchpad.net/ubuntu/+source/grub-installer/+bug/1803031/comments/5 "system booted in UEFI mode; full disk automatically partitioned; no error prompts shown and system is ootable after install. system booted with UEFI ut with CSM enabled; idem." You say *you* tested this for bionic [15:17] Launchpad bug 1803031 in grub-installer (Ubuntu Disco) "error: cannot find EFI directory." [Undecided,Confirmed] [15:24] when booting in legacy mode, you'll install in legacy mode, and grub-installer shouldn't try to find an EFI directory [15:25] or at least, at the time of that bug, I think it would not [15:25] this would have changed since, now that we try to install both [15:26] vorlon: when I tested this, I did test all the test cases listed and all was working, whether or not you were booting in legacy mode or not [15:31] cyphermox: ok, my /intent/ with specifying "install in CSM mode" was that we verify that when legacy booting the install media, you end up with a target system with an ESP that's bootable [15:32] sorry if that wasn't clear [15:33] "install in CSM mode" is either installing in UEFI mode, or installing in legacy [15:34] you can't tell that it's "UEFI but booted with CSM", because CSM just does legacy [15:34] but again; I think we're looking at things around the point where plans changed [15:35] ie. we were fixing the missing ESP bug, separately from making the changes to have both bootloader flavors installed on disk [15:36] I mean, otherwise the bug is meaningless; you always want to have an ESP on disk, even in legacy (not UEFI w/ CSM) [15:36] I posit that's true; but this bug wasn't addressing that at all [15:37] jibel: did what you see lead to a crash at install time or not booting? or are we "just" checking for sanity? [15:37] cyphermox, the system doesn't boot [15:38] because it boots in uefi mode but there is no esp [15:38] ok; because it's EFI but booted the install media in legacy [15:38] right [15:38] cyphermox, on this machine, on the boot menu legacy is the first entry when booted in uefi mode [15:39] the probability is high that a user will boot the iso in legacy mode [15:39] yes [16:31] jbicha: ... have we considered just *disabling* that user from sending to the devel-discuss list until they stop sending the announcements? Not sure how active that one user is though [16:43] Unit193: ^^ xedniv doesn't appear able to address his constant disconnects.. any chance you've got Magic Powers that could help? [16:44] sarnold: i poked -ops btw [16:44] for here and -server and elsewhere [16:44] nobdoy's alive :P [16:45] and lo a magician appears :) [16:45] indeed. [16:46] I think xedniv is following me around, in at least four of my channels .. [16:46] dax: #ubuntu-server too if you can [16:47] i can't [16:47] dax: excellent, thanks :) [18:15] how can one select another component as the 'active' one on a lp bug report? if there's an external project as the active, it's not possible to propose it for an ubuntu series without changing the url and reloading the bug page, right? [18:15] tjaalton: I believe that's the only place in Launchpad that requires URL-hacking for a functional purpose. [18:16] right, ok [18:16] (And yes, it's gross.) [18:17] :) [19:22] marcustomlinson: Could you have a poke at LP: #1846555 ? [19:22] Launchpad bug 1846555 in libreoffice (Ubuntu) "libreoffice-style-hicontrast not installable due to libreoffice-style-sifr relationships" [Undecided,New] https://launchpad.net/bugs/1846555 [20:14] infinity: hey, sure, thanks for pointing that out! I'll prep an upload tomorrow morning [20:18] marcustomlinson: Ta. === nnm|2 is now known as e2 [21:15] vorlon: first step of looking into this ESP issue; partman-efi changes seem to do good; but I notice we'll need a partman-auto change as well, along with the probably grub-installer change you mentioned too. [21:19] * vorlon nods [21:41] sarnold: I'd say "better late than never", buut. [21:48] Unit193: oh don't let a few bans deter you from a kline if it's still going on :) [22:55] teward: I think it's appropriate to ask first (re: devel-discuss) :)