[02:23] <arraybolt3> cjwatson or whomever else works on debconf - Did you or anyone else happen to recently add conffile conflict handling to debconf? I remember seeing an old Debian bug report that asked for it to be added, but it seemed to have stalled. Asking because I just noticed a debconf prompt asking me to resolve a conffile conflict and thought it was neat that this was possible for at least the grub-pc
[02:23] <arraybolt3> package, but also wondered if there was any way to turn this off since it's happening in Lubuntu 24.04 and Lubuntu's updater has built-in conffile conflict handling that is quite a bit more efficient and prettier at least in our environment.
[02:23] <arraybolt3> *a debconf-kde prompt to be specific
[02:39] <sarnold> there's something like three different tools here: ucf, dpkg conffiles, maintainer script magic, and I think a fourth method of doing as little as possible
[10:05] <cjwatson> arraybolt3: nothing in this area has changed in debconf for quite some time.  I guess you must have encountered ucf
[10:45] <kpcyrd> hello! I'm new to this, I've uploaded a bugfix for a package to debian unstable and it made it into noble-proposed/universe. what would be the next steps to request/propose moving the update from noble-proposed -> noble? thanks in advance!
[10:50] <juliank> kpcyrd: you wait until it happens
[10:50] <juliank> kpcyrd: or you solve whatever the issues are
[10:50] <kpcyrd> juliank: where can I lookup the issues?
[10:50] <juliank> kpcyrd: so go to the britney output at https://ubuntu-archive-team.ubuntu.com/proposed-migration/update_excuses.html and see what it's blocked on
[10:51] <juliank> kpcyrd: note that with the t64 transition lots of stuff is blocked on that
[10:51] <juliank> kpcyrd: hint: it's best to search for 'packagename (' such that you find the section for your package
[10:52] <kpcyrd> ok, so it seems if I want to make sure the bugfix is going to make it I'd try to avoid the dependency to nettle
[10:52] <juliank> kpcyrd: for more questions I probably am going to need the package name
[10:52] <kpcyrd> ah yes sorry: repro-env
[10:53] <juliank> if that's the only issue, I do think it's fine the nettle will migrate eventually and all will be well
[10:53] <kpcyrd> ok awesome, thank you!
[10:54] <juliank> you're welcome
[11:00] <adrien> in the update_excuses.html page, you can also append that to the url: #your-package, like #openssl or #xz-utils
[16:52] <mkukri> i cant build bpfcc against llvm-18 https://bugs.launchpad.net/ubuntu/+source/llvm-toolchain-18/+bug/2056466
[16:52] -ubottu:#ubuntu-devel- Launchpad bug 2056466 in llvm-toolchain-18 (Ubuntu) "bpfcc fails to build against llvm-18-dev" [Undecided, New]
[16:54] <mkukri> is there suppossed to be a libLLVM.so.18.1?
[17:21] <ginggs> LocutusOfBorg: ^
[18:53] <LocutusOfBorg> mkukri, ginggs this is already fixed in Debian git
[18:53] <LocutusOfBorg> and sid
[18:53] <LocutusOfBorg> but until we finish the build, I don't think doko will sync :)
[18:55] <doko> already building llvm-18 in a ppa
[19:50] <tsimonq2> mwhudson, juliank, vorlon: Hello, I bring gifts: https://bugs.launchpad.net/ubuntu/+source/germinate/+bug/2056483
[19:50] -ubottu:#ubuntu-devel- Launchpad bug 2056483 in germinate (Ubuntu) "Provides and NBS" [Undecided, New]
[19:51] <tsimonq2> I'm happy to write the code, would just like advice on the approach, if that makes sense.
[19:53] <vorlon> tsimonq2: I don't think you're going to have success trying to fix this in germinate.  It is not universally correct to prefer a virtual package over a real one. The bug here is that seed traversal hit a dependency on libefivar1 first, marked it for inclusion; then hit a dependency on libefivar1t64, and couldn't resolve the conflict
[20:00] <tsimonq2> vorlon: I'm not advocating for a universal solution to prefer virtual packages, I'm advocating for Germinate to become smarter based on how apt's dependency resolution works.
[20:01] <tsimonq2> The piece that still puzzles me is, should livecd-rootfs's apt configuration be a hair more permissive? Or is this something that *can* be done in germinate?
[20:03] <tsimonq2> Alternatively... germinate *could* just exclude NBS packages entirely
[21:19] <tsimonq2> https://code.launchpad.net/~tsimonq2/germinate/+git/germinate/+merge/462006 :)
[21:51] <vorlon> tsimonq2: ok, I see your implementation specifically seeks to identify nbs packages, which is better than trying to interpret things via Provides:.  I will defer to juliank on this; there are still certainly trade-offs here where refusing to install an NBS package can break a build
[22:18] <tsimonq2> thanks :)
[22:18]  * tsimonq2 is curious to hear those tradeoffs
[23:43] <mwhudson> tsimonq2: there is a wider point here which is that we do conceptually bunk in livecd-rootfs
[23:43] <mwhudson> +something
[23:43] <mwhudson> tsimonq2: germinate is an apt simulator, to some extend. so we take a list of packages from the seed and simulate which packages would be installed if you install all of them
[23:44] <mwhudson> then we take that list that germinate made, and ask apt to install all of *them* in the chroot
[23:44] <mwhudson> i don't know why we do this! we could just feed in the top level packages from the seed
[23:45] <mwhudson> which would at least give apt the chance to avoid issues like this
[23:45] <mwhudson> ugh typos galore but you get the idea