|
00:07
ShimmerFairy left
00:36
ShimmerFairy joined
00:37
wayland76 left,
wayland76 joined
01:18
wayland joined,
wayland76 left
01:33
hulk joined
01:34
kylese left
02:15
hulk left,
kylese joined
04:09
sibl joined
04:13
kst left
04:18
kst joined,
kst left
04:20
johnjay left,
johnjay joined
04:29
Sgeo left,
Sgeo joined
04:31
Sgeo left
04:32
Sgeo joined
05:24
ShimmerFairy left,
ShimmerFairy joined
06:19
abraxxa joined
06:24
hurufu joined
06:28
abraxxa left
06:29
abraxxa joined
06:53
Sgeo left
06:55
hwj joined
|
|||
| simon_sibl | while rakudo/moar arent supper easy to package, I just made a pr to add rakupp as a guix channel and allow to easily build it | 06:59 | |
| ab5tract | nice one | 07:15 | |
| antononcube: can you share which module this was? | |||
|
07:21
Evil-Bob joined
|
|||
| ab5tract | ah, Math::Matrix, as per your github issue | 07:29 | |
|
07:39
sibl left
07:57
dakkar joined
08:36
hurufu left,
hurufu joined
08:56
sivoais left
08:57
sivoais joined
09:15
kurahaupo joined
09:29
donaldh joined
|
|||
| El_Che | simon_sibl; being a compiler is static tar will do, probably (like GoLang). It's different for a runtime where you need to target everything | 09:36 | |
| tellable6 | 2026-07-25T18:25:52Z #raku-dev <[Coke]> el_che "source release is done for 2026.07" | ||
| simon_sibl | Guix has other advantages over simply giving a tarball to users, but I get what you mean | 09:52 | |
| its mostly reproducibility, the ease of modification (although here the code has no dependencies) | 09:53 | ||
| its like Nix but more focused on hackability and bootstrapping with a minimal binary blob to be easily audited | 09:54 | ||
| (I added the package to my own channel anyway just in case: github.com/4zv4l/sibl-channel/blob...akupp.scm) just it would be nicer if the repo in itself was the channel for its own package | 09:55 | ||
| ab5tract | antononcube: at a glance, I don't see how Math::Matrix actually compiles right now | 09:59 | |
| Math::Matrix::Util requires !clone-rows be defined, but Math::Matrix doesn't define this | 10:00 | ||
| but regarding the parsing / highlighting bug, it *should* be fixed in the next beta | |||
| ah, of course, priavte stubbed methods aren't required in the role consumer | 10:06 | ||
| antononcuber: please give this one a try when you get a chance: github.com/ab5tract/raku-intellij-...plugin.zip | 10:23 | ||
| melezhik. | o/ I wonder if anyone here is hosting a forgejo , and maybe someone would like to try dsci out, just let me know . Kinda resuming the project after having some positive feedback from forgejo community | 10:41 | |
| tellable6 | 2026-07-17T23:17:28Z #raku <wayland> melezhik the second last paragraph of dev.to/melezhik/hardening-rocky-li...tocol-58pe (beginning "If you like...") has some mistakes; I think it should be s/start oh/star on/ | ||
| timo | melezhik, I'm going to move my stuff from github to codeberg; does this also apply when I don't host the forgejo myself? | 10:44 | |
| lizmat | fwiw, I've started moving my repos to Codeberg as well, and all of the Raku Foundation stuff is on Codeberg as well | 10:49 | |
| melezhik. | codeberg is a public instance of forgejo ( with some custom skin I guess ) | 10:59 | |
| To use codeberg with dsci ci runner you need to host the runner on your vm with public ip cause this is required to send web hooks to | 11:00 | ||
| From codeberg to runner | 11:01 | ||
| So if you’d like to tinker / welcome ) I’d be glad to help | |||
| timo | i don't have a VPS of my own that I could put such a runner onto | 11:06 | |
| melezhik. | Yeah . I see. Target users are people who self host forgejo | 11:07 | |
| timo | yes, that makes sense | 11:08 | |
| lizmat | fwiw, I'm still using Github's CI: the primary storage is on Codeberg, but is mirrored to Github | 11:11 | |
| that at least feels as a good compromise for now, for open source projects at least | |||
| melezhik. | Yep. Using free gh runner minutes for free , why not | 11:31 | |
| Another possible layout is you can spin up a vm , host forgejo and dsci runner there , mirror your gh repos to your forgejo and run dsci ci without make it available through public ip, forgejo would send web hooks to dsci over localhost , this what I do now on forgejo.sparrowhub.io/root | 11:34 | ||
| Just like an option , but still involves VM ) | |||
| antononcube | Works -- no warnings, red flags, etc. | 12:45 | |
| lizmat | And another Rakudo Weekl;y hits the Net: rakudoweekly.blog/2026/07/27/2026-...thank-you/ | 12:55 | |
| antononcube | Some Raku projects are hinted here: | 13:03 | |
| cdn.discordapp.com/attachments/633...18770& | |||
| lizmat | antononcube where is that coming from ? | 13:05 | |
| feels sorta Antonov inspired :-) | 13:06 | ||
|
13:17
hurufu left
|
|||
| antononcube | Sure -- I modified a certain X-post in order to tease on of the Wolfram communities. | 13:20 | |
| Here is the original X-post: x.com/headinthebox/status/2080407137113743625 | 13:22 | ||
| ab5tract | antononcube: that bug had been around forever. it was parsing [and] as a array composer with contents as a call to 'and'. funny enough, another related bug was uncovered right after that was fixed, by the same file | 13:28 | |
| I'd also appreciate any bug reports of any misbehabior you might encounter, as everything should be working a lot smoother now | 13:29 | ||
| lizmat | antononcube understood | 13:31 | |
| antononcube | @ab5tract Sure! | 13:42 | |
| @lizmat In the latest rakudonews blog post this link is wrong: > PDF::Font::Loader “Provides API access to the xAI console” by Anton Antonov. | 14:01 | ||
| It should be linked to "WWW::XAI", raku.land/zef:antononcube/WWW::XAI . | 14:02 | ||
| lizmat | oops, fixed, thanks! | 14:17 | |
|
14:54
hwj left
16:10
Mason14 joined,
Mason14 left,
burpshirt joined
16:33
dakkar left
16:58
hwj joined
|
|||
| tbrowder | ref pdf::: | 17:01 | |
| font::loader: that's not what raku.land says | |||
|
17:06
jjido joined
17:14
abraxxa left
|
|||
| [Coke] | if I have some async code that is setting a single (unique each time) hash key, that should be safe, yes? | 17:40 | |
| tbrowder: I believe the error is that it shouldn't have mentioned PDF::Font::Loader at all on that line | 17:41 | ||
| timo | only safe if you've set up all the keys beforehand | 17:45 | |
| the growing of the hash when more keys are added is dangerous | |||
| [Coke] | bah. thanks. | 17:46 | |
| Was trying to add some racing to this file because it's very slow, and I haven't got the muscle memory to do it safely. :| | 17:47 | ||
| (thankfully an explicit lock on this seems pretty cheap) | 17:49 | ||
| is there a limit to the number of simultaneous taps that can be invoked when getting values via the supply? No mention in the docs about the amount of ||-ization there. | 17:55 | ||
| [Coke] also wonders the most efficient way to get # of values from a supply as you go. (.Seq.elems?) | 18:01 | ||
| timo | "as you go" as an incrementing value while it adds more? | 18:09 | |
|
18:12
abraxxa joined
|
|||
| timo | I'd probably zip the supply to a supply that just emits all numbers? or .map({ $++ }) | 18:12 | |
| [Coke] | Trying to do a progress bar - so I can have the tap do the "current progress + 1", but need to know the # of things emitted. (guess I can tie that to calling done in the supplier) | 18:15 | |
| ... right, my progress bar logic isn't thread safe either... and this script is basically shooting up to 98% and then moving veeeeery slooowly to 100. Feels like it's not happening actually in parallel when tapping the supply. | 18:32 | ||
|
18:32
hwj left
|
|||
| [Coke] | (fixed progress bar) | 18:33 | |
| Found a 8 year old stackoverflow entry from jnthn that may help. (Need someone to go through all his old posts and extract some tutorials/etc for the docs) | 18:43 | ||
| timo | there's a module like Async::Progress or something btw :D | ||
| [Coke] | (short answer, yes, tap is one at a time) | 18:44 | |
| I'm using Terminal Spinners, each enough to lock the things I need in my wrapper class of that. | |||
| *easy | |||
|
18:49
hwj joined
18:57
burpshirt left
|
|||
| [Coke] | ok, instead of a supply, going with multiple races over a worklist. (first race generates the second work list). hoping this is more || without more brain. | 19:01 | |
|
19:03
hwj left
|
|||
| timo | liz has an ecosystem module that builds a more sophisticated version of race sequences that handles things like automatically getting a sensible number for batch and the other one | 19:17 | |
| ParaSeq | |||
|
19:18
hwj joined
19:22
hwj left
19:32
jjido left
20:20
hwj joined
20:50
hwj left
21:04
human-blip left
21:06
human-blip joined
21:08
nine left
21:17
wayland76 joined
21:18
wayland left
21:27
jjido82 joined
21:37
dmvrtx left
21:38
dmvrtx joined
21:44
abraxxa left
|
|||
| tbrowder | El_Che: is there some delay between a new raku release and the generation of an updated rakudo-pkg? i'm guessing there is some delay on the debian side... | 22:08 | |
|
22:40
Sgeo joined
|
|||
| [Coke] | There's no guarantee the packages/binaries happen right away | 22:44 | |
| I think the last release, it took several days for El_che to get to it. | 22:47 | ||
| as I've mentioned recently, we could use more volunteers for the source and binary releases. | |||
|
22:48
acidsys left
22:49
wayland76 left
22:50
acidsys joined
22:59
jjido82 left
23:23
sjn left
23:33
sjn joined
|
|||
