timo i might already be above 10k unread notifications on github, i have zero chance of noticing anything 00:14
guifa timo: o.O 00:42
tonyo yea, they're next to impossible to filter for relevance 00:46
antononcube I guess, they should be clustered first. 00:55
timo well, they are. it's not enough, i just end up with very few clusters of very many notifications that i can in theory probably all just dismiss immediately, but then it's hundreds of rather small clusters that are an amount of work each to dismiss 01:02
antononcube Are those automatically generated messages or human-made / sent/ 01:41
tonyo kind of a combination, work has a couple of hundred repos. i do too (not all of them are public) 01:50
if i could only get them when mentioned or when i'm the code owner rather than every ci/cd/issue comment on the repo/etc that'd be nice. i'm sure there's a way to do that at this point but i haven't looked in years. i also don't read my email for a similar reason 01:51
antononcube I am looking for motivation to implement (another) recommender system. 02:30
That probably, has to be anyway if I want LLM-RAG to be more seriously supported by Raku. 02:31
sjn El_Che: are you going to FOSDEM? 04:38
tbrowder [Coke]: not 11:59
*thanks, i was thinking about it 'cause of lizmat and 'stem'
tbrowder hm, File::Temp, which i use a fair amount, has an issue opened in 2020. rpt is one who commented. some raku and os expert i bet could close that (hint). 12:13
the issue doesn't bother me because i don't do stuff the issue addresses. 12:14
lizmat are you talking about the raku community modules version ? 12:15
El_Che sjn: I don't know yet. Are you? 13:23
lizmat I think sjn said he's going yesterday :-) 13:27
El_Che missed it :) 13:35
lizmat: are you?
lizmat no 13:36
weekly: dev.to/lizmat/finding-the-coverables-3l13
notable6 lizmat, Noted! (weekly)
sjn El_Che: yeah, I'll be there helping a little with the booth, and running between policy and sbom devrooms, talking about metadata, security and the CRA :-o 15:24
El_Che: We (Brian, Stuart and I) have been looking for someone local to help with logistics. I know you always are busy, but do you by chance have time this year? :-) 15:26
El_Che I am at the moment changing teams and looking what to do with my on the side academic endeavours (writing a presentation as we speak for a meeting with some profs from abroad) 16:02
so, not very likely, but I'll ping you if time opens up 16:03
sjn El_Che: nah, we'd need a committment early; we're thinking about renting a screen (large, with floor stand), and need someone to pick up/store/return. This isn't something we can do "at the last minute", and without knowing if someone can be there with a car, we can't even proceed with budgeting or reservations, sadly :-( 17:35
so no worries; if you don't know you can now, then you can't at all. 17:36
El_Che: do you know of someone else local that might be able to help?
I've pinged Dirk and ecocode, but that hasn't lead anywhere (no commitment/no reply) 17:37
El_Che sjn: have you tried serge and emmanuel (both are from Brussels) 18:04
[Coke] there's no way to declare modules required for author tests, correct? (test-depends are for items that would be needed for just t/ tests) 18:31
sjn El_Che: no, I don't have their info 18:36
lizmat [Coke]: that's my understanding 19:06
perhaps we need an "author-test-depends" tag in META6.json
lucs [Coke]: What do you mean by "author tests"? 19:13
lizmat tests run in the xt dir
lucs Oh. Um, I never used that, don't really know about it. 19:14
[Coke] tests that you don't want the user to run, like "how is my code coverage" - you'd run these before pushing a release, but no need for a user to run them to test functionality 19:15
lucs Ah, I see.
[Coke] or, like for raku/docs - things that break CI should be in t/, things that are less strict can go in xt/ 19:43
tbrowder lizmat: yes, the comm version 19:45
*community version 19:47
[Coke]: mi6 may let you get away with not dec 19:48
*declaring a module used only in xt
lizmat it does, and that's the whole point 19:50
tbrowder i guess i missed the point. 19:53
lizmat test in xt could need any number of modules, but those wouldn't need to be installed for users of that module 19:55
tbrowder yes, but if they're not delared in META6.json, which complains: zef, raku, or mi6? 20:21
ugexe there are going to be some dependencies that just dont belong in a META6.json. For instance if someone has an examples/ folder that demonstrates using the module but it also has a bunch of dependencies, should those be listed somewhere? those dependencies aren't really tied to the distribution per-se, they are tied to a specific use of the code 20:23
tbrowder [Coke]: what is yr specific use case? 20:24
for odd stuff i sometimes use a Makefile, or a special test dir like "xt-extra"... 20:26
ugexe I think if I wanted to immediately have a way of having my xt/ dependencies listed somewhere I'd probably put them into a text file under xt/ with instructions to do something like `echo xt/dependencies.txt | zef install -`
tbrowder hm, i thought we wanted to avoid having to formally declare dependencies? 20:29
lizmat I think ugexe meant only the top level ones 20:30
tbrowder ok 21:05
tbrowder m: use Test; my $f = True; ok $f 21:58
camelia ok 1 -
tbrowder m: use Test; sub f(--> Bool) { False}; ok f 22:01
camelia not ok 1 -
# Failed test at <tmp> line 1
patrickb ugexe: The fruit of my recent ecosystem generation wishes: sr.ht/~patrickb/EcosystemMakerFaker/ (should show up on raku.land momentarily) 23:01
It's pretty epic zef enables one to do such a thing. Thank you!
antononcube @patrickb Fake Zef news sounds good. Maybe, fake RakodoWeekly is next. 23:21