IRC Logs for #crux-devel Sunday, 2019-03-10

ryuowatch as 3.7 is the last python 3 version. they then turn their attention to python 4. =p00:59
pedjanot yet, there is a 3.8rc, afaik :)01:09
pedjathey still have a few numbers to go before 401:10
pedjaand, with the less then ideal transition from 2 to 3, I kind of doubt they'd want to inflict that pain on themselves again :)01:14
pedjaI don't think that python2 will really die off any time soon, considering there are already forks that plan to keep it updated01:17
jaegeryeah, I don't see it disappearing any time soon, for good or bad01:29
Romsternot every project has transitioned to python302:08
jaegerAnd if folks keep supporting it indefinitely some never will :/02:09
Romsterpedja, lets not forget the setuptools -> distribute -> setuptools mayhem they had and the breakage that caused.02:17
Romsterand then all the bootstrapping of pip setuptools ugh02:18
Romsterseriously i don't think they even know where they are heading towards.02:18
*** xor29ah has quit IRC02:21
*** xor29ah has joined #crux-devel02:25
Romsteropenssl is vulnerable in crux-3.506:11
ryuojue: i'm going to try a micro optimization on existing pkgutils. then maybe it's time to cut a new release for 3.506:37
ryuoi want to see how much an improvement in pkgadd's rule matching logic could improve its overall speed.06:38
ryuoOk. seems I have the improvement finished.07:03
ryuotime to see if caching the REs produces a noticeable benefit.07:03
ryuohm. code was causing double frees or so. i'll need to look into that. for now disabling the regfree stuff.07:17
ryuoone reason to hate C++... it's harder to get MM right in unfamiliar programs.07:19
ryuo... yep, a measurable improvement.07:20
ryuoshaved about 5s on an upgrade operation.07:24
ryuo1m 15s vs 1m 20s07:24
ryuo~7% faster.07:26
ryuono difference on install (shocker)07:29
ryuothere's no default install rules.07:29
ryuonot much, but it could make upgrades a little faster.07:31
ryuoi'll see if i can fix the double free issue.07:31
ryuohm. i see. so that's why...07:35
ryuotypical. it must be the default copy constructor.07:35
jueRomster: I'd say that we should raise the release of a port if the content of the package is different, which means a rebuild is required09:16
juewithout that a user has no sign that he has to rebuild the package09:16
jueI came across that with your latest footprint updates09:18
Romsteroh true it's just the footprint and hmm yeah if it's not on the iso it'll not get updated i didn't think of that.09:47
*** onodera has joined #crux-devel10:05
Romsterjue, fixed;a=commitdiff;h=bb330a2782f28296711ab361c2fc4983b8446f21;hp=97e47840d8b3a640cab4abb191f49edce6354a8b11:55
juethanks :)12:19
*** onodera has quit IRC12:59
*** onodera has joined #crux-devel17:42
*** jue has quit IRC18:01
*** jue has joined #crux-devel18:33
*** onodera has quit IRC21:32
*** xor29ah has quit IRC23:02
*** xor29ah has joined #crux-devel23:04
jaegerjue: sorry for the noise, the signature mismatch was from a misconfiguration of my local distfiles mirror at home23:11

Generated by 2.14.0 by Marius Gedminas - find it at!