00:37 leont left 02:25 camelCaser left, camelCaser joined 07:09 Altai-man joined 07:19 |Tux| joined 07:25 sena_kun joined 07:26 Altai-man left 08:03 maggotbrain left 08:04 maggotbrain joined
|Tux| Working from $work shows that working form $hime *is* a load: 08:19
Rakudo version 2020.08.2-64-g1e552d21c - MoarVM version 2020.08-91-g590bac47e
csv-ip5xs0.888 - 0.910
csv-ip5xs-208.757 - 8.867
csv-parser24.231 - 24.578
csv-test-xs-200.389 - 0.397
test7.432 - 7.814
test-t1.796 - 1.798
test-t --race0.825 - 0.872
test-t-2030.712 - 30.888
test-t-20 --race8.795 - 8.904
08:38 leont joined 09:46 tobs left, tobs joined 11:24 Altai-man joined 11:27 sena_kun left 11:44 patrickb joined 12:05 finsternis left 12:07 finsternis joined 12:16 domidumont joined 12:26 patrickb left 12:30 patrickb joined
Geth rakudo: patrickbkr++ created pull request #3913:
Fix `--rakudo-home` Configure.pl parameter
patrickb o/ 12:40
Can someone with more NQP knowledge than me have a look at github.com/rakudo/rakudo/pull/3913...r487878887 ? 12:41
lizmat Interesting train of thought, that might be applicable to Raku as well? twitter.com/briandfoy_perl/status/...4184496137 13:11
[Coke] I would love it since I seem to always trip over this on mac os x. :) 13:14
Altai-man First, we need people to polish just a module to be rock solid everywhere, no? 13:15
vrurg Altai-man: I was about to ask you today as if you can run Blin against PR#3891. 13:20
Altai-man vrurg, sure, give me a bit... 13:21
vrurg Altai-man: whenever you can. Thank you!
Geth rakudo/master: 4 commits pushed by Altai-man++ 13:22
[Coke] .seen sergot
tellable6 [Coke], I saw sergot 2019-01-24T14:32:08Z in #perl6: <sergot> oh, Mar and May...
Altai-man vrurg, oops, I am afraid Blin/whateverable cannot work with forks. You want to push your changes to a rakudo branch, not fork branch. 13:23
lizmat Altai-man: remind me what Akefile is about ?
Altai-man .oO ( should've tried with spoons instead )
lizmat, it's a helper script for doing releases. 13:24
moritz at knife point! :D
vrurg Altai-man: that's what I didn't want to do. I find working in a fork less stressful. Ok, will think about it. 13:25
Altai-man vrurg, well, despite there are a lot of branches (200+?) and most of them are stale, I still don't see serious issues. I mean, there will be some notifications here, but development notifications make people cheerful (or should make), not annoyed. And if it does not go well, just delete the branch. Also, branches can not be tied to PRs. 13:27
lizmat Altai-man: you mean forks ? 13:28
Altai-man lizmat, where did I mean them?
lizmat "Also, branches can not be tied to PRs."
vrurg Altai-man: it's just a long time prejudice of mine. 13:29
Altai-man lizmat, no, I did not. You can do development in a branch if you don't want people to accidentally merge something and when it's ready, do a PR. Or when it's bad, well, just delete. If you want to merge - yes, PR, but for the development process it is not necessary to have a PR open/draft unless you want some attention. 13:30
vrurg, fair enough, up to you. Just noting the blin requirement. :)
Maybe "tied" is not the best word, because PR _is_ tied to source/target branches. What I wanted to say is that to try out various ideas in a branch it is not required to have a PR all the time. 13:32
vrurg Altai-man: BTW, I hoped that it could be sufficient to change repository URL in blin to make it work with a fork.
Altai-man vrurg, I think builds downloaded from whateverable are built automatically for rakudo repo, so blin does not build something itself by an url, but instead downloads compressed ready builds, as it also has to bisect commits. 13:33
[Coke] lizmat: you're on mac still? 13:34
lizmat yes, but on Mojave, *NOT* Catalina
vrurg To have a PR is always a good idea for anything significant. It then serves as a binding point for communications. Commenting on commits isn't as good when a general kind of discussion is needed.
[Coke] (or anyone?)
lizmat: can you test github.com/sergot/io-socket-ssl with a checkout, then NETWORK_TESTING=1 zef test --verbose . ?
vrurg Altai-man: Ah, right. I forgot about it downloading builds. Perhaps because I don't need a bisection, only to see if there is a fallout would be caused by the branch. 13:35
[Coke] I have openssl installed via macports, and it just immediately abort traps. Guessing it's related to 10.15 13:36
lizmat [Coke]: gist.github.com/lizmat/0d7b81e24fd...dc896b2b86 13:37
yes, I have heard reports of it failing on Catalina
one of the many reasons I hope Apple will get its act together on the next OS
vrurg I'm expecting my new home server to arrive this Wed. I'll be possible to have own Blin setup eventually.
[Coke] lizmat: are you using openssl via macports?
lizmat [Coke] 13:38
:good question: how can I tell ?
[Coke] 'sudo port installed | grep -i ssl' ?
(there's probably a better way
vrurg [Coke]: It's useless if the binary is outside of /opt/local tree. 13:39
lizmat sudo: port: command not found # I guess that answers the question :-)
[Coke] vrurg: what is useless?
vrurg [Coke]: checking against installed ports. For example, my rakudo is installed in $HOME and I have to take special measures to be able to use dynamic libs from /opt/local. 13:40
Altai-man vrurg, that would be pretty cool, as I can imagine. Still, running Blin is not too heavy on resources, just make sure you have enough threads. On desktop with 24 it takes just 2-3 hours to do a whole check depending on number of bisections to do. 13:41
lizmat notable6: weekly 13:42
notable6 lizmat, 4 notes: gist.github.com/1558e9fb9465b3a86f...48b4fd965a
vrurg Altai-man: 28 cores. :)
[Coke] vrurg: does IO::Socket::SSL work for you? 13:43
Altai-man vrurg, plenty. ;)
[Coke] vrurg: that command would tell her if she had it installed, though. (not whether or not anything was configured correctly to use it, of course)
vrurg [Coke]: I think so. Installs with errors, but otherwise – ok. Any specific test to try?
[Coke] ok. it dies here immediately. what version of os x are you on? 13:44
vrurg [Coke]: Catalania.
vrurg is away for a phone call.
[Coke] hates named releases.
10.15 - ok. how did you install openssl? 13:45
(Catalina?) 13:46
If I can duplicate what vrurg is doing, maybe we can get a doc update to IO::Socket::SSL for others. 13:51
also, IO::Socket::SSL hasn't been updated since 2015 ??
timotimo io socket async ssl has potentially been replacing io socket ssl in recently created modules 13:57
just a total guess
vrurg [Coke]: my openssl is from macports. I don't use homebrew. 13:58
14:00 |Tux| left
[Coke] vrurg: ok. I am also using macports - You mentioned you have some steps to recognize the library properly? 14:03
vrurg [Coke]: Simple: link /opt/local/lib to ~/lib. That's all. Makes sense for user-local installations only, not for when binary is installed into /opt/local/bin. 14:04
[Coke] I didn't do anything specific to macports, so I'm pretty sure it's installed into /opt/local/bin 14:07
s/bin// 14:08
14:08 |Tux| joined
lizmat notable6: weekly reset 14:08
notable6 lizmat, Moved existing notes to “weekly_2020-09-14T14:08:45Z”
vrurg [Coke]: Most likely. But with rakubrew the trick is something of "must have". 14:10
[Coke] when you say "user-local" installations - do you mean 'of macports'? (that is, if I'm the only one using macports on the box?) 14:12
vrurg [Coke]: no, pure user-local, under $HOME. Manual installation or with rakubrew. 14:17
[Coke] ... so not of macports, of rakudo? 14:28
ok, with the ln -s of ~/lib, all my "Abort Trap: 6" errors are gone. ;) 14:31
where the hell do we document that?
vrurg [Coke]: I was discussing the possibility of auto-linking for rakubrew with patrickb. Though rakubrew specific makes it better to link into ~/.rakubrew/<version>/lib, not into ~/lib. 14:35
[Coke] vrurg++
vrurg Otherwise, perhaps a note in INSTALL.md?
It'd be first platform-specific note in there. 14:36
Ah, wrong, windows-specifics are mentioned already. 14:37
[Coke] vrurg++ that works
still having issues with cro, but at least it's not the weird library error. 14:50
14:54 MasterDuke left
lizmat and another Rakudo Weekly News hits the Net: rakudoweekly.blog/2020/09/14/2020-...ally-raku/ 15:00
15:07 MasterDuke joined
vrurg lizmat++ 15:11
15:25 sena_kun joined 15:27 Altai-man left
Geth nqp: patrickbkr++ created pull request #666:
Move build logic to separate method so it can be called in subclasses
patrickb lizmat++ 15:46
I'm still in need of some NQP wisdom. I need to do subclassing and I can't figure out how the `BUILD()` methods should look. 15:49
The (rather simple) code is in this PR: github.com/rakudo/rakudo/pull/3913
I'm afk for a bit, but I will backlog! 15:50
16:39 |Tux| left 16:49 TreyHarris left 17:07 sena_kun left 17:18 domidumont left 17:19 sena_kun joined 17:31 vrurg left 18:16 TreyHarris joined 18:51 vrurg joined 18:57 vrurg left, vrurg joined
releasable6 Next release in ≈4 days and ≈23 hours. 3 blockers. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft 19:00
19:10 vrurg left 19:14 vrurg joined 19:24 Altai-man joined 19:26 sena_kun left 19:40 vrurg left 20:52 patrickb left 20:58 Altai-man left 21:09 sena_kun joined 21:50 sena_kun left 21:52 vrurg joined 21:57 vrurg left 22:06 Xliff joined