IRC Logs for #crux-devel Saturday, 2016-05-28

*** onodera has quit IRC00:49
*** erdic has quit IRC01:14
*** erdic has joined #crux-devel01:15
*** _________mavric6 has quit IRC02:31
*** _________mavric6 has joined #crux-devel02:32
*** Romster has joined #crux-devel03:04
*** Romster has quit IRC03:38
*** Romster has joined #crux-devel03:39
*** frinnst has joined #crux-devel08:03
*** frinnst has quit IRC08:03
*** frinnst has joined #crux-devel08:03
*** onodera has joined #crux-devel08:23
*** heroux has quit IRC09:15
*** heroux has joined #crux-devel09:31
*** onodera has quit IRC09:35
*** mobile_teK has joined #crux-devel10:12
mobile_teKHm fair enough. I have to attend by mobile tonight.10:13
mobile_teKSo behold of Major Sperling Mostar10:13
mobile_teKMajor spelling mistakes10:13
frinnstwtf, you suggested the time! :)10:23
*** mobile_teK has quit IRC10:27
*** mobile_teK has joined #crux-devel10:30
frinnstteK_: rtorrent/libtorrent could do with some cleanups. site url has changed to https://rakshasa.github.io/rtorrent/10:31
frinnstand proper tarballs exist (no github crap): http://rtorrent.net/downloads/libtorrent-0.13.6.tar.gz10:31
frinnsthttp://rtorrent.net/downloads/10:32
mobile_teKCool10:33
mobile_teKNo git here though :p10:33
*** mobile_teK has quit IRC10:34
*** mobile_teK has joined #crux-devel10:35
*** mobile_teK has quit IRC10:44
*** mobile_teK has joined #crux-devel11:16
*** AndChat|8064 has joined #crux-devel11:16
*** mobile_teK has quit IRC11:20
*** mobile_teK has joined #crux-devel13:44
*** mobile_teK has quit IRC15:37
*** mobile_teK has joined #crux-devel15:42
*** mobile_teK has quit IRC16:08
*** mobile_teK has joined #crux-devel16:37
*** sepen has joined #crux-devel18:02
sepenhi18:02
sepen10 min please, I just arrived at home18:02
*** onodera has joined #crux-devel18:15
sepen*re18:26
juehi sepen18:27
jaegerheyo18:28
juemobile_teK: are you around?18:33
sepenIt would be possible to talk later? I think I'll have to bathe my daughter18:36
sepenHow difficult it is to be a father :D18:36
juewell, I'll start and try to give a short report about my tests with signed ports18:36
sepenok18:36
jueI'm using the patched pkgmk for several weeks now and found zero incompatibility with 'normal' ports18:37
sepenI'll stay connected, but until she falls asleep I will not totally focused, sorry guys18:38
jueI've added signing to my private repo and I can report that it just works18:38
sepenyeah, but how about the next release flow?18:39
juethe only small effort is that you have run pkgmk -us after each edit of the Pkgfile or run pkgmk -is in the test-phase of a port18:40
jaegersounds pretty promising18:43
sepenand about the user experience?18:43
jueit's more than simple, the signing pkgmk works just well for ports without a .signature file so it would be possible to commit a new pkgutil tomorrow and add signinig to, let's say, core after that18:43
jaegerwhat do you think about rolling that into 3.3 since we've talked a tiny bit about a new release anyway?18:44
juesure, why not, I don't have a strong opinion about a timetable18:45
sepenagreed18:46
sepenhmm but the thing is that the repository is signed, then would not have the command 'ports -u' to check this?18:47
juebut the whole thing is a non-breaker IMO, so we can do it whenever we want18:47
jaegerfair enough, just a suggestion18:48
juesepen: to check what?18:48
sepenI refer not only to check each port individually but the whole repo18:49
sepenwhat about ports from third party repos?18:50
juewell, that's a feature we don't have currently18:50
jueto check the whole repo18:50
sepenbut I still do not see advantages to signing ports, I have always advocated that everyone is responsible and mature enough to keep your operating system18:53
sepensorry bath time, bbl18:54
frinnstwoi18:56
juewell, the only "problem" with third party repos is that the public key has to go into /etc/ports, but that's the same as with the .httpup or .rsync files18:56
jueand for 'shared' repos like our offical ones we have to find a way to distribute the secret keys18:57
juebut we are a small group, so a encryted email should work18:58
jueservus frinnst18:58
frinnstWhile I agree it might not be critical for us to set up signing as it would be for a larger distribution with more users; signify seems to simple I find it hard to say something negative about it18:59
frinnstI see your point about being responsible sepen. And if it would break users workflow or add a whole visible layer of complexity for users / maintainers I would probably suggest we should look for another solitution.19:00
frinnstbut signify seems to "tick all the boxes" for me19:00
jueyeah, indeed :)19:04
juehmm, looks like our friend teK_ is missing the meeting he invites to ;)19:16
jaegerI really don't have any strong feeling about it. I know that's not super helpful but it doesn't sound like it gets in the way very much19:17
frinnstjaeger: yeah i share your feelings19:18
frinnstmy main motivation for supporting it is to end the fucking debate over md5sums :)19:19
jaegerheh19:19
frinnstalso its nice when speaking with friends and explaining what crux is to be able to say we have signed updates19:19
juefrinnst: that's exactly my main motivation too :)19:22
sepenI hate md5sums too19:25
sepenI just considering that our ports are plain text which can be download in many ways and we do not distribute binaries, so I do not see an obligation to sign files19:27
sepenanyway, if everyone agrees then I do not get against19:27
juesepen: I don't see a real reason to hate md5 in our case, but it would be just silly to do a s/md5/sha.../19:27
sepensha is the same19:27
sepenI mean, I do not see the greatest of our problems is to sign pkgfiles, but for example in 'nice to have deps' or things like 'depupdate' feature19:29
mobile_teKI am SO sorry19:30
sepensecurity is important, but we have more shortcomings with the current system of ports19:30
frinnstyou suck tek! :D19:30
sepenhey teK!19:30
mobile_teKNo YOU do19:30
mobile_teKHi19:30
mobile_teKJust a sec. Reading up19:30
juesepen: yeah, I agree, but we have a working solution for one problem right and not for the others19:30
sepenso +119:31
jue*now19:31
mobile_teKY19:31
sepenjue: signed ports sounds like good replacement to md5 / sha19:31
jueyes, that's my feeling as well19:32
jueand a bit more, of course :)19:32
frinnstAnd everyone is ok with slowly branching off 3.3 ?19:32
sepenif most developers bridge the learning curve and do not have users against, I do not see why not start with it for 3.319:33
frinnstI have 3.5 weeks vacation starting on thursday and i have *NOTHING* planned so I'll need something to do19:33
frinnst:>19:33
sepenI think it has been very good job and is transparent enough, thanks teK ;D19:34
sepenfrinnst: lol19:34
sepenfrinnst: make a child and see :D19:34
frinnsthaha19:34
frinnstIm too selfish to be a dad :)19:34
sepenlol19:34
sepenI though the same19:35
jaegerI have 2 nephews, that's close enough for me currently19:35
frinnstI have one - thats enough19:36
sepenleave for a few minutes, sorry19:36
frinnstand I can give him back to his mom when im tired :p19:36
sepen:P19:36
sepenbbl19:36
juefrinnst: no objections to start a 3.3 cycle19:37
frinnstgoodie :>19:38
jueof course only if jaeger agrees too19:38
mobile_teKSo i Do  Not need to red alle the log?19:38
mobile_teKWill Do. ...19:38
jaegerno objections here19:39
frinnst\o/19:41
jueok, so the summary is easy to do: we will introduce port signing with 3.3 and we will start slowly but soon with the 3.3 devel cycle19:45
mobile_teKOk create i reading the log too19:45
mobile_teKI will try to Come up with some things for the iso too of you like19:47
jaegerlinux-firmware is one of them19:47
frinnstyeah. we should probably provide a few different firmware packages if its not too hard to maintain; linux-firmware-gpu, linux-firmware-block, linux-firmware-net and whatnot19:48
frinnstthat is, if its not hard to maintain19:48
mobile_teKI can split Them yes19:48
onoderateK_: can you please fix the man location of contrib/mpc, also the URL is dead19:49
jaegersemi-related, are we possibly able to upgrade crux.nu sometime?19:49
jaegerI know we talked about it in the past19:49
frinnstyeah but it just fizzled out19:49
onoderaif you are talking about the look, I really love how it currently looks tbh19:49
frinnstI should be able to be on location if we need hands on with the server19:50
frinnstif its ok with _________mavric619:50
frinnstlive just 140km away19:50
mobile_teKI am ok with Helping19:55
frinnstbut it needs to be replaced, no?19:55
frinnst32bit horrorshow19:55
mobile_teKI can invite for the Server Migration agaik :D19:55
jaegerwe either upgrade it with a 32-bit build or replace it with a VPS or another server, I suppose19:57
jaegerthe first two options are preferable to the third in my opinion19:57
frinnstyeah19:57
frinnstone issue that will come up if we move to a vps is the domain name19:58
jaegerthough I do have a 1U xeon box here I bet it would be expensive to ship to Kalmar :D19:58
frinnstcharlie controls that, no?19:58
frinnsthehe yeah it would be cheaper for me to buy one :)19:58
mobile_teKOr we can dpnate for a Tony New Server with the help of mav19:59
jaegerperhaps a good question would be this: is there any real advantage to having a physical server anymore?20:00
frinnstnot at all imo20:01
frinnstI could host it at work but that would make it very tied to me - if I get fired or whatnot :)20:03
frinnstbut even if I were i dont think that would be a problem for you guys20:04
sepenofftopic: teK! could you create /home/crux/git-to-rsync-working-copy/xfce/3.2? I created 3.2 branch a long time ago (some days after CRUX 3.2 released)20:04
mobile_teKWe should give mav a Chance20:05
mobile_teKWill Do tomorrow sepen20:05
sepenthanks20:06
mobile_teKWith hangover :D20:06
sepenlol20:06
sepentake my hangover too :P20:07
mobile_teKWill do20:20
*** mobile_teK has quit IRC21:26

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!