mwhudson | ok unmkinitramfs is terrifying | 00:06 |
---|---|---|
Odd_Bloke | alias unmkinitramfs=sudo rm -f /boot/initrd* | 00:20 |
xnox | mwhudson: lets do a binwalk, then lets skip things, and do stuff. | 00:30 |
mwhudson | xnox: i love parsing binary formats in shell | 00:31 |
xnox | mwhudson: note to self, when grub loads multiple initrds it inserts padding between them for alignment. also note that intel-ucode tries to insert weird cpio members too for alignment. note that we add amd64-ucode too. i'm pretty sure is that end result is _not_ aligned. | 00:31 |
mwhudson | xnox: i have this side project to make customizing installer isos easier, one of the things i want to do is update .disk/info/casper-uuid-generic if you replace the initrd | 00:32 |
mwhudson | but right now i just want a beer | 00:32 |
xnox | mwhudson: i'm not sure if there is a better way. in bios days one didn't know where one booted from; but in UEFI days we do. | 00:40 |
xnox | mwhudson: sdboot does store in efi variable the drive one booted from, hence one should look up squashfs from the same place. | 00:40 |
xnox | mwhudson: i wonder if we can make casper leverage that. | 00:41 |
mwhudson | hm that would be nice | 00:42 |
xnox | mwhudson: i.e. not sure if we can inspect "BootCurrent" standard efi variable to figure out where abouts shim got loaded from this time around. | 00:50 |
xnox | and decide that that's the badger | 00:50 |
mwhudson | xnox: i suspect a distressing number of server installs are still bios too | 00:52 |
mwhudson | although maybe that's changing finally | 00:52 |
xnox | mwhudson: new servers not only don't have bios, they crutially do not have bios or efi ipxe, meaning efi-httpboot as the only option | 00:54 |
xnox | plus bmc which is the weird "virtual cdrom" option. | 00:55 |
mwhudson | nice | 00:55 |
=== DarkTrick_ is now known as DarkTrick | ||
=== shivaram is now known as slingamn | ||
DarkTrick | Does this mean, my actively used for debian unstable? https://imgur.com/a/SSK4YJi | 11:06 |
JackFrost | ...No, it means you set the homepage URL to that. | 11:08 |
DarkTrick | JackFrost, thank you. ... I wonder when I did that :D | 11:09 |
=== didrocks999 is now known as didrocks | ||
TJ- | possible problem with the hirsute aarch64 ISO images; the GPT/EFI-SP is failing to be recognised from a Samsung Galaxy Book2 (Qualcomm Snapdragon 850) - manually writing a GPT + EFI-SP and copying grubaa64.efi to /EFI/BOOT/BOOTAA64.EFI works fine | 12:36 |
JackFrost | ddstreet: Any idea why ubuntu-dev-tools prints keyring modules it can't load when firing up a lot of the commands? It's rather annoying. | 12:55 |
JackFrost | Eg, `syncpackage $anything` will print out a lot of 'Loading $foo' starting with kwallet and going down a line of a bunch of them. Granted, redirecting stderr from all programs un ubuntu-dev-tools would "fix" the problem, but this is far from ideal. | 12:56 |
rbasak | cjwatson: o/ just checking you still have my changes file proposal in your queue please? | 13:21 |
cjwatson | rbasak: I do, sorry | 13:24 |
cjwatson | It has been a Week | 13:24 |
rbasak | np, thanks | 13:28 |
rbasak | I don't have a particular deadline; I just wanted to make sure it wasn't lost. | 13:29 |
=== ricotz_ is now known as ricotz | ||
Odd_Bloke | rbasak: We maintain cloud-init's packaging in our upstream repo; we don't want to change that, but I was wondering (idly, mostly) if there's a good way of us getting that history into git-ubuntu somehow? | 14:35 |
ddstreet | JackFrost i don't see the prints you're talking about | 15:00 |
rbasak | Odd_Bloke: there is a method that will work, but I'm not sure if it's a good idea. So I don't suggest doing it immediately, but we can ponder it. If you were to merge git-ubuntu's history into your packaging branch in your upstream repo, then git-ubuntu would be able to adopt subsequent commits that you make into its own history. But that might make a complete mess of the git history in a not-useful | 15:20 |
rbasak | way, which is why I'm not sure it's a good idea. | 15:20 |
rbasak | Apart from that, I'm not sure of what an integration might look like. Ideas welcome! | 15:21 |
Laibsch | Can I interest someone in sponsoring bug 1916250? I used Conflicts/Replaces even if only Replaces by itself might be enough. | 16:38 |
ubottu | bug 1916250 in libsignon-glib (Ubuntu) "gir1.2-signon-2.0 needs to declare replace on older releases (Groovy2Hirsute)" [Undecided,New] https://launchpad.net/bugs/1916250 | 16:38 |
rbasak | Laibsch: sure, although Conflicts is a little too strong perhaps, and are best avoided to give apt the most leeway to resolve things. | 16:56 |
rbasak | Laibsch: can you pick from https://wiki.debian.org/PackageTransition please? | 16:56 |
rbasak | Unversioned is also unusual; I'd expect versioned relationships in this case I think. | 16:58 |
=== mnepton is now known as mneptok | ||
Laibsch | rbasak: Thank you for your comment. Frankly, I lack some history of the package. Looking through the changelog, it seems as if this package was neglected for a long time and recently a new Debian maintainer stepped in with frequent updates. | 17:22 |
Laibsch | There's been an API change from upstream apparently. Versions were also bumped from 1.x to 2.x upstream. | 17:23 |
=== jdstrand_ is now known as jdstrand | ||
Laibsch | Versions 1 and 2 don't coexist in the same Debian pocket FWIW. | 17:24 |
rbasak | Laibsch: sorry, I don't know anything about the package either. I appreciate your efforts in fixing bugs! | 17:24 |
rbasak | But to fix it we need to understand what is necessary unfortunately | 17:25 |
Laibsch | digging deeper, I find that hirsute provides both gir1.2-signon packages (which is most likely not desirable) | 17:37 |
Laibsch | rbasak: Will it drop automatically from the archive, given that the source is now missing? | 17:39 |
Laibsch | for hirsute | 17:39 |
Laibsch | https://packages.ubuntu.com/hirsute/gir1.2-signon-2.0 and https://packages.ubuntu.com/hirsute/gir1.2-signon-1.0 | 17:40 |
=== seb128_ is now known as seb128 | ||
rbasak | Laibsch: yes - it's in the NBS report, so the old source will be deleted when the archive admins clean up - usually only once per cycle I think. | 18:06 |
rbasak | https://people.canonical.com/~ubuntu-archive/NBS/libsignon-glib1 | 18:06 |
rbasak | But I think that shouldn't make a difference to what's needed for the binary packages? | 18:07 |
cjwatson | NBS clean up is quite a bit more than once a cycle | 18:07 |
rbasak | Ah, OK | 18:07 |
* cjwatson does that one now | 18:07 | |
cjwatson | Well, NBS doesn't involve removing *source* | 18:08 |
cjwatson | but rather binaries | 18:08 |
rbasak | Yes, sorry. | 18:10 |
rbasak | Looks like they're both from the same source. | 18:10 |
Laibsch | rbasak: After some looking into it, I believe that indeed the Conflicts line is too strict and not necessary. https://bugs.launchpad.net/ubuntu/+source/libsignon-glib/+bug/1916250/comments/3 | 18:40 |
ubottu | Launchpad bug 1916250 in libsignon-glib (Ubuntu) "gir1.2-signon-2.0 needs to declare replace on older releases (Groovy2Hirsute)" [Undecided,New] | 18:40 |
Laibsch | Concerning the link to the wiki page, I don't it is either of the options listed there with #5 being closest but I think that adding an empty transitional package would be overkill here. | 18:44 |
Laibsch | I don't think it is either of the options ... | 18:44 |
=== Laibsch is now known as Laibsch|bbl | ||
JawnSmith | Is any core dev available to restart the autopkgtest for iputils that was blocked by systemd? The new version of systemd should resolve the issue | 23:15 |
JackFrost | ddstreet: Hmm, have kwallet installed? | 23:40 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!