mason | fsmithred: I've been several hours longer than anticipated and I've got to go again for a bit, but: 1) I'd love to see us package a fix. 2) We should think about the current bandaid and decide if we want to try to anticipate or otherwise determine if we're able to release a device or not. | 00:50 |
---|---|---|
mason | For 2) we can either allow for manual flagging, look for what is currently using a LUKS device and see if it's root or otherwise unlikely to go away, or both. | 00:51 |
mason | The current fix - just try once and give up if this fails - is probably reasonable, but I'm certainly not tied to the method. | 00:51 |
fsmithred | I've cloned the mxlinux repo for cryptsetup-modified-functions | 00:52 |
mason | fsmithred: If you have thoughts and/or ideas I'd love to hear them - also some notion of whether Debian would take such a fix or if it should be packaged locally for Devuan. | 00:52 |
mason | Oh, hey there. | 00:52 |
fsmithred | I want to tweak it a bit and put it in devuan | 00:52 |
fsmithred | hi | 00:52 |
mason | How do they fix the issue? | 00:53 |
fsmithred | for i in 1 | 00:53 |
mason | Ah. | 00:53 |
fsmithred | they also change the logging of the return value of some other function | 00:53 |
mason | That feels sloppy. It's leaving the loop in place but explicitly not looping. | 00:53 |
fsmithred | from log $rv to log 0 | 00:53 |
mason | Which is to say, the code lies about what it's doing just to preserve i, which isn't actually necessary. I like my fix better honestly. | 00:54 |
fsmithred | yes, it makes sense | 00:54 |
mason | Why lose the return value if we can save it? | 00:54 |
fsmithred | I couldn't remember what else you changed | 00:54 |
mason | I don't think I like their changes. :P | 00:54 |
fsmithred | and I can't find a copy of it | 00:54 |
mason | That was it. | 00:54 |
mason | Oh, let me fix that. | 00:54 |
fsmithred | they also don't include the code for dealing with lvm | 00:55 |
mason | Here's the beowulf version: https://bpaste.net/QAAA | 00:55 |
fsmithred | thanks | 00:55 |
mason | ASCII version: https://bpaste.net/S3NA | 00:55 |
mason | I don't do anything with LVM myself, so that might be someone else's addition. | 00:55 |
fsmithred | it is | 00:56 |
mason | fsmithred: Do you have an email I can use? I'll mail you the patches so you don't have to worry about copy/paste whitespace. | 00:56 |
fsmithred | me at gmail | 00:56 |
fsmithred | I changed something else but haven't tested yet. Should be ok | 00:56 |
fsmithred | line 11 | 00:56 |
mason | I'd like to see what LVM changes needed to be done. | 00:56 |
fsmithred | if [ $rv -ne 0 ]; then log it | 00:57 |
fsmithred | you want link? | 00:57 |
mason | Sure, please. | 00:57 |
fsmithred | https://dev1galaxy.org/viewtopic.php?pid=14530#p14530 | 00:57 |
mason | So, I just sent the raw patches. Dot is ASCII, dash is Beowulf. | 00:58 |
mason | ty | 00:58 |
fsmithred | yeah | 00:58 |
mason | If we need such a thing, I'd be happy to explicitly declare the new content in my patches to be public domain or CC0 or whatever the heaviest release is we can use. :P | 00:58 |
mason | Argh, and now I'm in hot water for making folks wait for dinner. I'll be back! | 00:59 |
fsmithred | bye | 00:59 |
onefang | You need to put something more edible in that hot water. | 00:59 |
enyc | Hrrmph... | 09:48 |
enyc | having all sorts of trouble trying to reboot my devuan desktop still ;-( | 09:48 |
enyc | The "udev" errors with devices /dev/sda etc etc not populated after 10000000 microseconds thing happens a lot | 09:48 |
enyc | if there are testing images, very happy to test new-beowulf isntall somewhere.. | 09:49 |
rrq | enyc: there are a couple of pre-freeze trial i386 isos at https://borta.rrq.id.au | 09:56 |
enyc | hrrm, i'd prefer amd64...... | 09:57 |
enyc | Probably be bhelpful nonetheles i sohw that it doe/oesn't work on this machine ta currentyly-triggers-issues | 09:57 |
enyc | put in a different hdd and test-away | 09:57 |
enyc | amd64 testing images definitely good $plan, i'd like to test install MATE desktop etc | 10:32 |
LeePen | Evilham: Hi. | 18:00 |
LeePen | Been thinking about the britney output and avoiding it being propagated to mirrors. | 18:01 |
Evilham | does it have to be on pkgmaster? | 18:02 |
Evilham | because changing mirrors' rsync settings would be a PITA | 18:02 |
LeePen | Yes, that is where britney needs to work. | 18:02 |
LeePen | If we add location /britney with a root /home/britney/public, would that work? | 18:03 |
Evilham | I don't mean teh server btw, I mean pkgmaster.devuan.org | 18:04 |
Evilham | if we setup a different directory and alias it on the nginx level, mirrors won't mind | 18:04 |
Evilham | if it really has to be pkgmaster.devuan.org/britney | 18:04 |
LeePen | Mirrors just take /devuan and /merged not /? | 18:05 |
Evilham | they rsync the dir with everything in it directly | 18:06 |
Evilham | there is just one rsync job, not multiple | 18:06 |
LeePen | The root dir? | 18:07 |
Evilham | https://pkgmaster.devuan.org/devuan_mirror_walkthrough.txt | 18:08 |
Evilham | `rsync -avz vesta@pkgmaster.devuan.org:~/devuan ./` | 18:09 |
Evilham | that's pretty much the command that they have | 18:09 |
LeePen | So I think setting an nginx location /britney will not get propagated to the mirrors. | 18:12 |
Evilham | yes :-D maybe we meant the same: as long as the files are not in that dir, it's fine, we can use nginx with an alias directive or location or sth like that | 18:14 |
LeePen | OK. I will only put it on pkgmaster itself, doesn't need to be on the enabled sites. | 18:15 |
LeePen | britney output is now available on https://pkgmaster.devuan.org/britney/ | 18:46 |
LeePen | Updates every 6 hours. | 18:46 |
LeePen | Once we are happy with it we can feed it back to dak to manage unstable -> testing migrations. | 18:47 |
Evilham | nice :-) | 18:51 |
LeePen | I am all for reducing workload! | 18:51 |
Evilham | same | 19:00 |
Evilham | just checked your changes to pkgmaster: totally aprove of them :-) | 19:01 |
Evilham | wasn't doubting, just making sure :-p | 19:02 |
Evilham | since I was explaining while on the phone, wasn't sure I made sense :-D | 19:02 |
LeePen | Evilham: Excellent! I am always happy to be checked up on. Thanks. | 19:47 |
xrogaan | The project should consider opting for something else than SLiM as default DM (if it still is). The SLiM project being abandoned since 2013. | 23:57 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!