🦋 Welcome to the IRC channel of the core developers of the Raku Programming Language (raku.org #rakulang). This channel is logged for the purpose of history keeping about its development | evalbot usage: 'm: say 3;' or /msg camelia m: ... | Logs available at irclogs.raku.org/raku-dev/live.html | For MoarVM see #moarvm Set by lizmat on 8 June 2022. |
|||
01:03
guifa joined
01:37
guifa left
03:11
guifa joined
03:16
guifa left
04:19
Xliff joined
|
|||
Xliff | \o | 04:19 | |
04:35
guifa joined
04:53
guifa left
|
|||
patrickb | That script wrapper fix up PR is showing a CI test failure where `raku -V` is erring out. I can't repro it locally. Has anyone seen this before? | 07:29 | |
lizmat | works fine here | 09:05 | |
Geth | UK/main: 9d37049675 | (Elizabeth Mattijsen)++ | 16 files Initial commit from L10N.fresh-distribution |
09:27 | |
.github/main: 455c0f4f0d | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | profile/README.md Update README.md |
09:31 | ||
UK/main: cd93d67c0b | (Elizabeth Mattijsen)++ | 2 files Update README and META6.json |
09:34 | ||
RO/main: 716cbb783c | (Elizabeth Mattijsen)++ | 16 files Initial commit from L10N.fresh-distribution |
09:37 | ||
L10N/main: 50ee107ee0 | (Elizabeth Mattijsen)++ | 3 files 0.1.4 |
09:39 | ||
CZ/main: 92a023aa98 | (Elizabeth Mattijsen)++ | 16 files Initial commit from L10N.fresh-distribution |
09:43 | ||
SK/main: 52bd4c9dca | (Elizabeth Mattijsen)++ | 16 files Initial commit from L10N.fresh-distribution |
09:45 | ||
SV/main: feb3f61bd5 | (Elizabeth Mattijsen)++ | 16 files Initial commit from L10N.fresh-distribution |
09:47 | ||
NO/main: 31066f7ce1 | (Elizabeth Mattijsen)++ | 16 files Initial commit from L10N.fresh-distribution |
09:49 | ||
DA/main: 6849ee3f7b | (Elizabeth Mattijsen)++ | 16 files Initial commit from L10N.fresh-distribution |
09:50 | ||
FI/main: 2409e8ca8a | (Elizabeth Mattijsen)++ | 16 files Initial commit from L10N.fresh-distribution |
09:52 | ||
ET/main: 797566b742 | (Elizabeth Mattijsen)++ | 16 files Initial commit from L10N.fresh-distribution |
09:53 | ||
LV/main: 3d712ed582 | (Elizabeth Mattijsen)++ | 16 files Initial commit from L10N.fresh-distribution |
09:55 | ||
LT/main: 78978d5622 | (Elizabeth Mattijsen)++ | 16 files Initial commit from L10N.fresh-distribution |
09:56 | ||
PL/main: fc7d71eb0b | (Elizabeth Mattijsen)++ | 16 files Initial commit from L10N.fresh-distribution |
09:58 | ||
SL/main: ef6f2431f9 | (Elizabeth Mattijsen)++ | 16 files Initial commit from L10N.fresh-distribution |
10:00 | ||
HR/main: da643176b0 | (Elizabeth Mattijsen)++ | 16 files Initial commit from L10N.fresh-distribution |
10:01 | ||
BG/main: 747ed13461 | (Elizabeth Mattijsen)++ | 16 files Initial commit from L10N.fresh-distribution |
10:04 | ||
EL/main: d6893eb49a | (Elizabeth Mattijsen)++ | 16 files Initial commit from L10N.fresh-distribution |
10:05 | ||
TR/main: a7ab6cb5f7 | (Elizabeth Mattijsen)++ | 16 files Initial commit from L10N.fresh-distribution |
10:06 | ||
SQ/main: f0b55fdc42 | (Elizabeth Mattijsen)++ | 16 files Initial commit from L10N.fresh-distribution |
10:08 | ||
GA/main: eac9f144c9 | (Elizabeth Mattijsen)++ | 16 files Initial commit from L10N.fresh-distribution |
10:11 | ||
lizmat | ok, enough noise here for now | 10:17 | |
Geth | L10N/main: 8a3b9e9242 | (Elizabeth Mattijsen)++ | 7 files Add informational methods |
12:11 | |
12:48
guifa joined
|
|||
[Coke] | OK, putting a slide together for a lightning talk on "what's in 6.e". | 13:34 | |
... dammit, I just lost my WIP | 13:39 | ||
patrickb: can you give me a one liner on why 'script wrapper" is good in relation to 6.e? | 13:59 | ||
any thoughts on mentioning an EU TRF? | 14:00 | ||
patrickb | [Coke]: The new script wrappers fix argument processing on Windows. Without it, all use of CLI applications is limited. A prominent case: with zef it's impossible to install specific versions of modules, as it's impossible to pass < and > chars. (i.e. :version<1.0.0>) | 14:05 | |
[Coke] | p6c is the "just give us a github URL" eco list, yes? | 14:06 | |
14:08
guifa left
|
|||
patrickb | correct | 14:16 | |
14:16
guifa joined
|
|||
[Coke] is here with 0rir who just pinged me on github.com/0rir/touch/issues/2 | 14:37 | ||
Touch – MissingDependency – Dependency “v6” was not resolved (ok, but Touch doesn't depend on v6?) | 14:39 | ||
ah, HEAD != release. | 14:41 | ||
so I don't think I can use blin to test this easily. Can probably just use bisectable by hand, though, I guess. | 14:43 | ||
15:14
guifa left
|
|||
[Coke] | .seen finanalyst | 15:22 | |
tellable6 | [Coke], I saw finanalyst 2025-05-27T18:20:57Z in #raku-dev: <finanalyst> would `zef` need to change? or would simply adding in doc-provides in the Meta6 be suffcient? | ||
[Coke] | japhb: what's our current unicode version, which one are you working on now, and we'll get 17 right after it drops in 2025, yes? | 15:39 | |
latest p5 has U16 support | |||
for 6.e stuff: features: "it's a spec", AST, L10N, rakudoc v2, U17, language cleanup (maybe remove some stuff) | 15:41 | ||
non-features: web sites, docs (use rakudoc v2), docs (maybe do translations), blin (essentials), script wrappers | 15:42 | ||
AST is the big one. | 15:43 | ||
I only have 5 minutes, so it's just high level everything | |||
also we need to make cartouches work; karl isn't sure if that works in p5 | 15:45 | ||
:) | 15:46 | ||
lizmat, patrickb, timo, nine, etc.: any notes on the "6.e" list? | 15:50 | ||
japhb: karl from p5 also mentioned new line breaking algorithms that he had to implement based on specs. | 16:02 | ||
17:01
guifa joined
17:39
guifa left
|
|||
jdv | patrickb: first how did this happen and what guarantees are there that its fixed? | 17:40 | |
the release process isnt cheap and downstreams still have to do their bit agai. | 17:41 | ||
also, what is the impact of the issue? it needs to be pretty serious to trigger a point release. there will be another release in about a month... | 17:43 | ||
patrickb | jdv: don't make a release yet. There is still an open question I need to clarify. I'll get to this later in the evening (currently on the road) | 17:44 | |
17:55
guifa joined
|
|||
[Coke] | github.com/rakudo/rakudo/commit/cc...3d21011ffb - this should have been done behind a language version change. | 18:12 | |
m: ls | 18:22 | ||
camelia | ===SORRY!=== Error while compiling <tmp> Undeclared routine: ls used at line 1. Did you mean 'lc'? |
||
[Coke] | m: run('ls') | ||
camelia | Inline-Perl5 META6.json Perlito bin bla dalek-queue evalbot file.txt lib log mbox nqp-js p1 p2 precomp rakudo-j-1 rakudo-j-2 rakudo-j-inst rakudo-j-inst-1 rakudo-j-inst-2 rakudo-m-1 rakudo-m-2 rakudo-m-inst… |
||
[Coke] | hurm. what files exist where bisectable is run, I wonder. | 18:23 | |
bisectable6: run('ls') | |||
timo | we do have directory listing builtins you know :P | 18:24 | |
[Coke] | don't bash shame me! | ||
bisectable6: run('ls') | |||
bisectable6 | [Coke], Will bisect the whole range automagically because no endpoints were provided, hang tight | ||
[Coke], ¦6c (92 commits): «3rdpartyAkefileCONTRIBUTING.mdDockerfileLICENSEMETA6.jsonREADME.mdcompose.yamlconfig-default.jsondatalibmaintenancesandboxservicesxbinxt» | 18:25 | ||
[Coke], Nothing to bisect! | |||
[Coke] | (language version change) - hurm, maybe not. It's hidden inside the Instant class. | 18:30 | |
lizmat | yeah, I thought it was carefully hidden ? | 18:36 | |
[Coke] | There's a C interface here, so I'm guessing he needs to be careful with the result, but I can't see right now how he's creating it. | 18:46 | |
... and I should be writing my talk. :) | |||
but this is probably fixable with a patch to Touch. | |||
to warn you: it doesn't work on Mac, in case you got curious. | |||
Also: would love a simple sample showing how to get AST out of some code. | |||
lizmat | m: say Q|say "Foo"|.AST | 18:51 | |
camelia | RakuAST::StatementList.new( RakuAST::Statement::Expression.new( expression => RakuAST::Call::Name::WithoutParentheses.new( name => RakuAST::Name.from-identifier("say"), args => RakuAST::ArgList.new( RakuAST::QuotedS… |
||
[Coke] | why doesn't that need an env var or a use? | 18:54 | |
lizmat | because :-) | 18:55 | |
it doesn't interfere with anything | |||
nine | Just keep in mind that this is not an interface that is likely to stay forever | 18:56 | |
[Coke] | Yup, that's part of my 5m. | 18:57 | |
TBD: Review | |||
lizmat: I just installed L10N::JA but don't have "ryuuu" as an executable. | |||
lizmat | and ryuu ? | 18:58 | |
(only 2 u's | 19:05 | ||
) | |||
[Coke] | ah, there it is | 19:09 | |
in that case raku.land/zef:l10n/L10N::JA has a typo in the README | 19:10 | ||
also the second example doesn't work? | 19:12 | ||
Undeclared routine: 言う used at line 2 | |||
similar issue with L10N::NL's second example | 19:14 | ||
Undeclared routine: zeg used at line 2 | |||
FYI. Will simplify the example shown to the working ::NL one. :) | |||
also: dutku -e 'zeg "Hallo Wereld"' emits an extra newline before saying Hallo | 19:15 | ||
sorry for all the bug reports here. | |||
19:30
guifa left
|
|||
patrickb | jdv: The fallout is pretty serious. When using a relocatable rakudo build, it's not possible to install any module that contains any bin/ files. | 20:00 | |
This includes zef. This also means there won't be any compiled release. | 20:01 | ||
The fix is a pretty small change. Testing the fix is as simple as trying to install zef in a relocatable rakudo build. | 20:02 | ||
re How did it happen? The offending PR changes how our script executables are generated when installing a module. I've tested this on Windows, MacOS and Linux, but did not do so in different build variations. I.e. I always did non-relocatable builds. | 20:06 | ||
The issue only shows in relocatable builds. That's why it only surfaced after the source release when trying to do the binary release. | 20:07 | ||
Why is the PR that fixes this not merged? - There is a newly breaking test in the CI. I'm currently looking into that. But so far it seems it's the CI setup that has somehow changed. | 20:10 | ||
japhb returns from 2.5 days AFK ... | 20:34 | ||
[Coke]: I realize this is late for your lightning talk, but: | 20:35 | ||
We currently have Unicode v15.0 (This is available as `Unicode.version` directly in Raku) | 20:36 | ||
We need to upgrade across 15.1, 16.0, and (in September at the latest) 17.0 | 20:37 | ||
16.0 had some significant changes, including changing file formats, additional data files, and the new indic line breaking algorithm (probably what "karl from p5" was talking about). | 20:38 | ||
17.0 rearranges the directory tree similar to how 13.0 did (in a way that makes more sense, but also means that the tools need to adapt to it yet again). | 20:39 | ||
I have more to do before I can get the current data files to parse correctly and emit the correct C files; I am also working on maintainer documentation of how to do all this next year. | 20:40 | ||
patrickb | I have the MacOS CI fallout tracked down: github.com/rakudo/rakudo/issues/5921 | 20:41 | |
japhb++ Writing documentation for this is very kind to the future people. | 20:42 | ||
japhb | There is more documentation needed for the C data structures that are produced (and presumably the MoarVM, nqp, and/or Rakudo algorithms that operate on them), but I haven't even made it to the point I can do a good job of that yet. | 20:43 | |
Thanks patrickb. Yeah, it's clear that this whole area needs to be improved from a maintainability point of view, where the "maintainability" in question doesn't just consist of "can fix a bug in interpretation of *existing* Unicode rules", since the Unicode Consortium -- despite having lots of stability rules -- appears to be excellent and punishing implementors. | 20:45 | ||
*excellent at | |||
Geth | JA/main: 3bbd624ddf | (Elizabeth Mattijsen)++ | 3 files Fix documentation of the ryuu executable, Will Coleda++ |
21:38 | |
rakudo: patrickbkr++ created pull request #5922: Fix `raku -V` on strange non-standard MacOS |
21:50 | ||
rakudo/main: 2223dc4c70 | (Patrick Böker)++ | src/core.c/CompUnit/Repository/Installation.rakumod Fix wrapper installation in relocatable installations `<abs-slash>` is not a valid command for program arg in our wrapper generator. Doesn't make much sense either. The program path needs to be understood by the OS, so it requires platform separators. |
21:53 | ||
rakudo/main: 2d60b20179 | (Patrick Böker)++ | src/core.c/Main.rakumod Fix program name in USAGE With the new script wrappers the script is not executable anymore. With the new wrappers we always generate two files, a wrapper program that is executable and a .raku script that is not executable. Thus the path shortening logic isn't triggered anymore. Fix by additionally detecting such wrapper + script combinations and shortening them respectively as ... (6 more lines) |
|||
rakudo/main: 51fb0954aa | (Patrick Böker)++ (committed using GitHub Web editor) | 2 files Merge pull request #5920 from patrickbkr/script-wrapper-fixes Script wrapper fixes |
|||
patrickb | jdv: The fix for the script installation issue in relocatable builds is now merged. | 21:55 | |
lizmat | [Coke]: cannot reproduce the extra line | 22:08 | |
*can* reproduce the slangification issue | |||
:-( | |||
will look at that tomorrow, thanks for the nudge! | 22:11 | ||
[Coke] | japhb: thanks for the info, appreciate it! | 22:14 | |
lizmat: I'll file a ticket on the whitespace issue with sw_vers, etc. | |||
lizmat: the extra whitespace is *in the sample code in the readme* | 22:17 | ||
if I cut and paste it from the readme? whitespace. type it by hand? nope | 22:18 | ||
... and only if I do it weird, I guess. | 22:19 | ||
I did open github.com/Raku-L10N/NL/issues/1, though. :) | |||
so ignore the whitespace thing, was just freaking out before the talk. | |||
What units are Instant:1751239187.226795467 | 23:20 | ||
is that a fractional number of epoch seconds? |