IRC Logs for #crux-devel Saturday, 2020-01-18

*** [EaX] has joined #crux-devel00:18
*** xor29ah has quit IRC00:22
*** [EaX] has quit IRC01:56
*** xor29ah has joined #crux-devel01:58
*** pedja has joined #crux-devel08:48
TimB_jue: no idea why, but i just rebooted and lxc rc scripts do not complain about anything anymore10:40
TimB_rebooted shortly after it reports an error once more :B10:50
TimB_actually, two times in a row error now. Frankly no idea what's up with that10:51
jueerror message?13:29
TimB_i haven't tried to isolate the process yet, sorry13:29
TimB_when it's turned on via rc.conf it seems to just run, even though it reports a failure without any further output. When i try to manually restart it afterwards it says it's already running - but this can't be the real problem13:31
ryuojaeger: does CRUX have a solution for source file URLs that have weird filenames? I was thinking of writing a patch to allow pkgmk to accept other file names if you need to use such a thing.13:48
TimB_^ that would be awesome..13:48
ryuoIOW, you specify the local file name instead of accepting the implicit one.13:49
ryuoI'll do a quick test. I don't think I can use '#' for this but it would be perfect since it can't occur in regular URLs.13:49
ryuoit's a special type of hyper link.13:50
ryuoIt makes no sense for it to be in a URL.13:50
ryuointeresting. seems it is ignore if it's in the middle of text.13:51
ryuoat least by dash13:51
ryuoseems doable. a bit more work and i should be done.14:21
ryuojaeger, jue, frinnst:
ryuopatch for this feature which uses '#' to give an explicit file name for download URLs.14:50
ryuofor get_filename, you'd need to strip off the largest prefix to get the filename14:51
ryuothis is designed to remove everything before/after the last # in the URL14:52
ryuodepending on context14:52
ryuofilename? everything before. for downloading? everything after.14:53
ryuoi also removed get_basename since its output is the same as get_filename for URLs and this was the last thing using the custom function14:54
ryuosimple change overall14:55
ryuothough the logic in pkgmk could use an overhaul14:55
ryuolike using local more often for stuff that doesn't need to be global14:55
ryuoi fully expect this, if accepted, would not see use until at least the next major release where it is included15:04
*** jue has quit IRC15:24
*** jue has joined #crux-devel16:39
*** pedja has quit IRC19:14
frinnstjue: do you remember why we ship gmp.h with the port? testing 6.2.0 now..19:39
frinnsttalking about core/libgmp19:40
juefrinnst: it's related to -32, see here ->;a=commit;h=8645b10c8e68e29f8c19cd725f22ded25406bf8d20:48
jaegerryuo: would definitely be useful if implemented well, it's come up many times in the past but always with some kind of drawback22:21
jaegerfor example, does it work with prtwash, etc.22:23
*** xor29ah has quit IRC22:58
*** xor29ah has joined #crux-devel23:00
ryuojaeger: prtwash?23:13
ryuouh... no23:14
ryuoit wouldn't.23:14
ryuobecause these things use their own hacks -_-23:15
ryuoany solution would "break" non-pkgmk scripts23:15
ryuojaeger: i can't see this working w/o breaking them since it involves mapping file names. anything that reads the sources array will have the same problem.23:16
ryuomaybe this is a chance to do some modernizing of the old scripts...23:17
ryuovia some mechanic23:17
ryuoa parallel array might work23:19
jaegeryeah. It's all doable, of course, but more work to tie them all together23:55
*** junland has quit IRC23:59

Generated by 2.14.0 by Marius Gedminas - find it at!