=== chris14_ is now known as chris14
[05:21] <luna_> morning
[07:24] <luna_> https://social.linux.pizza/@bittin/112608103026405646
[15:00] <ahasenack> mjt0k: hi
[15:00] <ahasenack> mjt0k: I'm about to enter a sequence of meetings and then lunch, so I'll be unresponsive for a while
[15:01] <mjt0k> ok
[15:06] <mjt0k> so, the talk was/is about samba package (and also about qemu package), and me being slightly sad about ubuntu and debian heading into different directions sometimes, despite all my efforts to keep things the same as much as possible
[15:14] <mjt0k> ahasenack: the prob with samba for me is because users still ask me to provide samba binaries for ubuntu (including noble), and at the very least I'd rather be compatible with ubuntu changes
[15:15] <mjt0k> (I see numerous installs of samba-4.20 on noble once I made that one available on my repository)
[15:18] <mjt0k> ahasenack: so, there's no particular question from me really, I'm just trying to figure out how to go from there without making things worse for one or another party.  It'd be really nice if I knew about at least some bigger changes in ubuntu (like binary package rearrangements) a bit before.  It's mostl likely true for you as well (wrt my changes)
[15:31] <luna_> almost done now, just gonna update some snap packages on the last machine
[17:03] <ahasenack> mjt0k: the distinction between main and universe has always been a source of delta between ubuntu and debian, there isn't much that can be done there specially when different packages are built because of that
[17:03] <ahasenack> the glusterfs case is one such example
[17:03] <ahasenack> in ubuntu I have had no problems in massaging our delta on top of the debian changes, and even less after you started to either incorporate them, or make the build consider ubuntu or debian packages and behave differently, thanks for that
[17:04] <ahasenack> I didn't think you would care about our package split, but what I didn't realize is that you also provide ubuntu packages
[17:05] <ahasenack> in this cycle we are starting to document our delta as part of the commit message that introduces it, and we are doing it retroactevely to all our delta. Quite the exploration
[17:05] <ahasenack> I'm sure we will find out some delta doesn't make sense anymore, and drop it. Or find other bits that were never sent to debian to begin with.
[17:05] <schopin> y
[17:06] <ahasenack> In this samba 4.20.1 merge that I started now, I'll revisit all the delta and try to reason why we still have it, or drop it
[17:07] <ahasenack> regarding glusterfs specifically, the bit that cannot be avoided is that we cannot have something in main depending on glusterfs anymore. Some time ago, our delta was to not build the glusterfs vfs modules
[17:08] <ahasenack> then we moved glusterfs to main, and that delta was dropped. But now we demoted it again, given upstream resourcing is lacking, but didn't want to leave users who got used to it, not have it all of a sudden. Hence, the universe package for samba-vfs-modules-extra and qemu-block-<something>-supplemental (IIRC on the name)
[19:25] <hamburgler> Perhaps an add question, is there a record of checksums for daily builds? trying to find one for 2023-08-19 for server-cloudimg-amd64.img looking for one built on 08-19-2023.
[19:25] <hamburgler> odd*
[19:30] <oerheks> https://cloud-images.ubuntu.com/releases/
[19:30] <oerheks> no, 15-8 or 28-8 https://cloud-images.ubuntu.com/releases/jammy/
[19:31] <oerheks> sorry
[19:31] <hamburgler> no worries appreciate the help :)
[20:16] <mybalzitch> I don't understand why /boot wants to be 2G and /boot/efi wants to be 1.05G
[20:18] <oerheks> i don' t get those numbers either, efi 512 mb is enough for multiple instances
[20:24] <mjt0k> ahasenack: thanks for the explanation.  I do care about things being easy for ubuntu, and I know about main vs universe thing, but I didn't know gluster were moved back to universe (though I knew about when it moved to main).  I really am into incorporating back ubuntu things, - besides universe-vs-main, there are some other changes which are useful to debian too, there's no reason to keep it in 
[20:24] <mjt0k> ubuntu delta
[20:25] <mjt0k> ahasenack: so, please keep me updated where possible.  We had a few less pleasant examples in the past which would've been much better if done in a coordinated way
[20:28] <mybalzitch> static network config on one interface, box has multiple interfaces, 24.04 fresh install, appanrelty networkd is waiting for dhcp to time out on all the other interfaces I didn't configure?
[20:28] <mybalzitch> https://i.imgur.com/lGiAbdW.png
[20:33] <mjt0k> I hate it when something depends on network-online.  Especially on a notebook with no known wifi nearby :)
[21:08] <ahasenack> mjt0k: ugh, network-online is such a mess
[21:08] <ahasenack> mjt0k: about samba, ack, and sorry for having made incorrect assumptions about what I thought debian would care about, that was a mistake. Better to always ask.
[21:08] <ahasenack> mjt0k: and I'm happy to try to find a way to make this glusterfs mess more palatable
[21:09] <mjt0k> I don't think it's messy, it just clashed a bit with my plans, nothing more than that :)
[21:10] <mjt0k> it's fine to have that in main vs universe
[21:10] <ahasenack> oh, the qemu changes were more complicated, due to that /run/<versioned-qemu-directory> which holds the previous copies of the .so files
[21:11] <ahasenack> the block-*.so ones I mean
[21:11] <mjt0k> it's not really complicated, but more "political".  The difference is trivial
[21:11] <ahasenack> true
[21:11] <mjt0k> I asked a few times to provide a mount unit in debian too. I still don't see why it can't be done.  And I think I solved it finally
[21:11] <mjt0k> don't remember anymore :)
[21:12] <mjt0k> yes, I moved whole thing to a postinst script
[21:12] <mjt0k> it eliminates the difference
[21:43] <ahasenack> I had to do some smarts because `cp *` became `cp all-except-gluster`
[21:43] <ahasenack> don't remember the details now
[21:44] <ahasenack> when copying the block modules