| rwp | Devuan is eventually going to need to decide what to do about UsrMerge. I don't think Devuan can remain undecided because Unstable and Testing are currently broken without UsrMerge being applied. | 20:53 |
|---|---|---|
| rwp | I think Devuan needs to either decided to 1) Roll out UsrMerge to Devuan or 2) Fork packages that don't work without it. | 20:54 |
| rwp | Neither option is particularly likable options. Rolling out UsrMerge following Debian is probably the choice of least resistance. Even if it means accepting dealing with all of the breakage that accepting UsrMerge brings along with it just the same as in Debian. | 20:55 |
| LeePen | rwp: I agree. | 20:56 |
| LeePen | Unfortunately, we don't have an agreed mechanism for making such decisions. | 20:57 |
| LeePen | To my mind, 1) is the only practical and scalable option, but I think there are some who disagree. | 20:59 |
| rwp | I am of two minds on this problem. | 21:01 |
| rwp | Part of me wants to stay with Debian. It's been a comfortable home. | 21:02 |
| rwp | But part of me thinks that Debian will continue with hostility and instability. And committing to do whatever Debian does just keeps us committed to all of that negative baggage too. | 21:02 |
| rwp | Today's Unstable breakage for me is that /etc/init.d/hostname.sh sets PATH=/sbin:/bin which of course does not work after having moved /bin/hostname to /usr/bin/hostname. | 21:03 |
| rwp | It's clearly a bug. But any report will have several well known names attacking the idea of fixing it. | 21:04 |
| rwp | Two days ago nano and cpio both moved binaries (which is fine) but forgot to change their postinst,prerm scripts to change the update-alternatives location for their binaries to the new location. | 21:05 |
| rwp | bind9 is still broken due to Bug#1055429 from November which is still unfixed due to /etc/init.d/named hard coding /bin/ip but of course iproute2 moved it to /usr/bin/ip. (It should never have had a hard coded path anyway.) | 21:07 |
| rwp | JFTR but my forced non-UsrMerge Debian system I use for reference has these same problems too. | 21:09 |
| golinux | LeePen: A clear statement about the consequences of moving to Excalibut/testing or ceres at this time might be helpful. Also a recommendation to stay put until the dust settles. Perhaps we could start drafting a statement on a pad? | 21:15 |
| golinux | All I know is that an unacceptable number of users have run into issues which is having a negative effect on user confidence. | 21:17 |
| golinux | Such a message should be prominently displayed on www and all support channels imo. | 21:18 |
| golinux | Then users can make an informed decision. Official silence is no longer acceptable OMO. | 21:20 |
| golinux | Thoughts? | 21:20 |
| golinux | OMO . . . have no idea where that came from. Try OPTION | 21:21 |
| golinux | Or maybe IMO | 21:22 |
| golinux | LOL! | 21:22 |
| golinux | LeePen: I know you are busy but wish you could find time to engage more. It would be a good idea to do some damage control ASAP | 21:24 |
| gnu_srs | I have a small set of patches to avoid installing usrmerge. Will be published tomorrow :) | 21:40 |
| golinux | IIUC . . . unless we have a COMPLETE set of patches we cannot avoid usemerge. Cherry-picking what to patch is not a viable path for a default install | 21:59 |
| golinux | Please correct me if that is incorrect . . . | 22:03 |
| DelTomix | gnu_srs nice! looking forward to checking that out | 22:14 |
| DelTomix | I really thought this was kind-of decided to go along with usrmerge, however if its not - rwp makes excellent points, and ideas are worth looking at! Also if its not decided we should really consider it properly and collectively to understand what the direction of Devuan is to be. | 22:16 |
| rrq | imo, given Devuan's ratio of "talkers vs doers", going along with it is the only option. | 22:39 |
| rrq | even though it's an incredibly stupid and ugly notion | 22:39 |
| golinux | Would be good to have a statement about our policy and warning of the pitfalls if not done correctly. | 22:49 |
| gnu_srs | Again: I volunteer to fork the needed packages. Up to now there are around 10, maybe the total number will be 20. Please help identifying those :) | 22:52 |
| gnu_srs | Some are just minor problems, others we don't even have to revert! | 22:53 |
| Xenguy | Your efforts are admirable gnu_srs , I'm just wondering about the 'sustainability quotient' of the approach you've taken... Thoughts on that? | 22:54 |
| gnu_srs | And I'm talking about excalibur/ceres, not daledalus (will not be infected by usrmerge)! | 22:55 |
| Xenguy | How does that approach fare in the future? | 22:55 |
| gnu_srs | The patches are mainly reverting some of Debians move of files to /usr. Nothing else, no upstream changes :) | 22:56 |
| Xenguy | Okay, but going forward, does this mean more or less work of this kind, and does this project have the resources to manage that? | 22:58 |
| gnu_srs | I think we do have resources enough to do that yes! | 22:59 |
| Xenguy | Any sense of the relative workload for this, going forward? | 22:59 |
| gnu_srs | I think the work is not more than keeping systemd out of the distribution. | 23:00 |
| Xenguy | Okay, you seem to be suggesting generally that this approach is sustainable going forward... | 23:01 |
| Xenguy | I'm not qualified to comment on that assessment, but perhaps others here are | 23:01 |
| DelTomix | That has to be demonstrated - maybe it will become clear with the patches what is the thinking | 23:01 |
| Xenguy | Makes sense, let's look at the patches first | 23:02 |
| golinux | Whatever the strategy will ultimately be, we need a statement that there are currently issues with usrmerge and suggestions of how to avoid or deal with them | 23:05 |
| rrq | gnu_srs: have you added the forked projects to the git store? | 23:07 |
| rrq | once there is a forked proect, the "devuan/packages" team will need to keep it up to date with updates | 23:09 |
| rrq | j | 23:10 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!