| bgstack15 | Good news! It appears that building python3-ipalib without systemd is possible with very minor adjustments. I'm sending it up to jenkins now. | 15:37 |
|---|---|---|
| bgstack15 | Team, the `systemctl` replacement script exists in the `devuan-sanity` source package but maybe we should change the package name so we avoid appearance of insulting systemd. Everyone knows we don't like systemd, so we should be polite when we offer alternatives. | 17:25 |
| LeePen | I agree. And I have pushed bgstack15 into this (for which, apologies). | 17:27 |
| LeePen | My personal view is that we should avoid language which embodies judgements | 17:28 |
| LeePen | The systemctl replacement script is not Devuan specific. | 17:30 |
| LeePen | And everybody in Devuan is already sane | 17:32 |
| onefang | Only one of the voices in my head is sane. | 17:32 |
| LeePen | Listen to that one then ;) | 17:33 |
| bgstack15 | onefang: How do we know we are chatting with the right one? | 17:36 |
| onefang | Coz the one fang is the one on the right. | 17:36 |
| fsmithred | is there anything else in that package? vimrc? | 18:15 |
| LeePen | fsmithred: No. There is a devuan-sanity package in experimental | 18:17 |
| LeePen | with .vimrc | 18:18 |
| LeePen | https://pkginfo.devuan.org/cgi-bin/package-query.html?c=package&q=devuan-sanity=0.0.1 | 18:18 |
| fsmithred | so what's the new package called? | 18:18 |
| LeePen | It appears to come from https://git.devuan.org/devuan/sanity | 18:18 |
| LeePen | ATM devuan-sanity-systemctl, which I would like to change. | 18:20 |
| LeePen | And bgstack15 is tolerating my pickiness! | 18:20 |
| fsmithred | yeah, I'd be happy with a different name | 18:21 |
| fsmithred | if anyone wants a logical argument for that, "devuan-sanity" is already taken, and it's for replacing config files with better defaults. | 18:21 |
| fsmithred | eloginctl? | 18:22 |
| fsmithred | greybeardctl | 18:22 |
| LeePen | systemctl-service-shim | 18:22 |
| adam_free2air | systemctl-mock | 18:23 |
| LeePen | adam_free2air: SGTM | 18:23 |
| LeePen | bgstack15: Like any of these ^^^ | 18:24 |
| LeePen | ? | 18:24 |
| fsmithred | is it sysv-specific, or will it work with other inits? | 18:25 |
| fsmithred | sysvctl? | 18:25 |
| fsmithred | if that doesn't already exist | 18:25 |
| LeePen | bgstack15 can be more specific, but it translated systemctl to service and update-rc.d commands | 18:25 |
| fsmithred | servicectl/ | 18:26 |
| fsmithred | ? | 18:26 |
| fsmithred | that looks ugly, even without the slash | 18:26 |
| LeePen | So, I now wonder if we should add it to init-system-helpers package (which is already forked). | 18:27 |
| LeePen | Hmmmm | 18:27 |
| fsmithred | how does it compare to deb-systemd-helper? | 18:31 |
| fsmithred | nm, deb-systemd-helper is not to be used by users | 18:33 |
| adam_free2air | mocking systemctl is more polite than inferring it insane ;) | 18:34 |
| LeePen | fsmithred: Yes, it is just for maintscripts. | 18:34 |
| fsmithred | that package seems like a good place to put the new script | 18:35 |
| LeePen | Yes. Sorry to have been slow and not thought about it before. | 18:35 |
| fsmithred | you're not the slowest one around here | 18:36 |
| LeePen | bgstack15: What do you think? | 18:36 |
| bgstack15 | maybe systemctl-service-shim | 19:38 |
| LeePen | bgstack15: OK. | 19:39 |
| LeePen | Are you happy with it going in src:init-system-helpers? | 19:39 |
| bgstack15 | So what the script does is transliterates systemctl start/stop/restart/enable into the real "service [foo] start" and "update-rc.d enable" commands | 19:39 |
| bgstack15 | init-system-helpers; not directly familiar with that yet. Let me inspect that. | 19:40 |
| bgstack15 | Is that a required package for Devuan? | 19:40 |
| bgstack15 | I'm not sure this is ready for 100% saturation of the Devuan community; just those who like to tinker. | 19:40 |
| bgstack15 | This script isn't ready for warranties yet. | 19:40 |
| LeePen | That is the source package. It builds init and init-system-helpers (which contains service, invoke-rc.d, update-rc.d) | 19:41 |
| LeePen | I think we would put it in a separate binary package within that src. | 19:42 |
| bgstack15 | I would prefer that we have easy access to update this script when we learn its weaknesses outside of just my own environment. | 19:43 |
| bgstack15 | If it's in a larger source package it could be difficult. | 19:44 |
| bgstack15 | It's not actually a part of that init thing. | 19:44 |
| bgstack15 | Oh snap, yeah, there is no way I can recommend we place my systemctl translator in that package! | 19:45 |
| LeePen | OK, then just do a systemctl-service-shim source package that produces a single binary. | 19:47 |
| bgstack15 | OK | 19:48 |
| LeePen | I notice you have /usr/sbin/systemctl but a man1 manpage. Should be /usr/bin. | 19:48 |
| LeePen | I think that is where systemd systemctl goes. | 19:48 |
| bgstack15 | Oh the location of my systemctl is actually extremely important. | 19:54 |
| bgstack15 | Some things look for /usr/bin/systemctl and get stuck in a permanent loop calling /usr/bin/systemctl which calls the service script which calls /usr/bin/systemctl which calls... | 19:55 |
| bgstack15 | Would you find it acceptable if I just move it to a man8 page? | 19:55 |
| LeePen | From a consistency perspective it is fine. | 19:57 |
| LeePen | So other packages that want /usr/bin/systemctl can't use your script? | 19:58 |
| bgstack15 | Yes, things that will look for /usr/bin/systemctl will fail. In my limited experience I've only ever seen "systemctl" be hardcoded into stuff, not /usr/bin/systemctl. | 20:18 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!