01:33 Manifest0 left 02:10 kwolcott joined
kwolcott Newbie here 02:11
Been learning/using Raku by exploring Rosetta Code.
I use MacPorts (obviously I'm on MacOS) 02:12
antononcube To Raku or programming in general?
kwolcott I've installed the Primesieve Library.
But I am unable to install the Raku Primesieve module even though I was able to install the Perl Primesieve module. 02:13
antononcube Consider using rakubrew: rakubrew.org
kwolcott I'm (re)learning Ada, learning Julia, (re)learning Perl, learning Raku, almost everything that I can do on my Mac that is open-source. 02:14
yes, I tried that once, perhaps I screwed it up :-) 02:15
I have a Raku from MacPorts and one that I installed from source.
antononcube Installed from source!!
kwolcott When using zef to install Raku packages, how does it know where to install them? I can specify with cpanm when using Perl. 02:16
antononcube I kind of know that for rakubrew… But I think this question has been asked before. 02:18
I.e, search IRC’s logs. 02:19
The creator / maintainer of zef @ugexe frequently is seen here, so might answer. 02:20
kwolcott my Mac is using arm, not Intel/AMD, and I do not want Intel/ARM (or hybrid/universal) on the Mac, so I'll try the manual install and try going for the shell approach rather than the shim approach. Learning as I go here :-)
antononcube Hmmm… do you think it is faster ? 02:21
I mean, Raku installed from sources on arm. 02:22
I am not sure why would anyone learn Ada. Julia is worth knowing, but if one knows or uses R or Mathematica, Julia is (much) less interesting or needed. 02:24
kwolcott Well, I started with BASIC, then Pascal, then C, then Ada (that was in college). Then I used Perl. I've always wanted to re-learn Ada and I'm getting better at it now. I'm also expermining with Maxima and Octave, as well as R. I don't have Mathematica. 02:27
Where is zef under rakubrew? 02:28
I tried ~/rakubrew build-zef which fails because it is trying to install zef where my raku is that I built from source; I'd like to have a zef based on the rakubrew. 02:36
I tried rakubrew available, how do I choose MoarVM? The most recent Raku version will be fine. I guess I can't build zef until I have a local Raku installed under rakubrew :-) 02:38
ok, building a Mac raku for arm using rakubrew now... 02:45
02:48 kylese left, kylese joined 02:55 leah2 left
kwolcott So, how do I tell zef where to look for the Primesieve Library? zef help spews out lots of info, not grokking it (yet) 02:55
ab5tract Why does anyone learn anything? Ada is a great language 03:05
kwolcott: so did you install `Math::Primesieve`? (raku.land/github:CurtTilmes/Math::Primesieve) 03:07
you can configure where zef installs to by using either `--install-to` or via the env var `ZEF_INSTALL_TO` 03:08
I recommend using `rakubrew` with defaults 03:10
03:15 kylese left, kylese joined
kwolcott not able to install it, zef complains about not being able to find the library; but cpanm had no problem finding it. 03:19
ab5tract ah, I see 03:20
did you install `primesieve` via brew?
kwolcott no, I use MacPorts, not brew 03:21
ab5tract ah, ok. I think it's the same issue either way
if it's what I think it is, it is related to mac security settings for loading shared libraries 03:22
which makes the successful installation of perl's version interesting to note
kwolcott MacPorts installs stuff under /opt/local
ab5tract does the perl module work as expected in addition to installing successfully? 03:23
kwolcott brief testing of the Perl Priemesieve module was successful, yes. 03:24
ab5tract ok thanks for confirming that 03:33
I have a feeling that the problem is related to what this module is designed to solve: raku.land/zef:lizmat/MacOS::NativeLib
however I'm not getting it to work that way either :/ 03:34
03:37 guifa joined
ab5tract ah! I found the problem 03:41
the version of primesieve installed by brew is 12
but `Math::Primesieve` is looking for versions v7, v8, and v9 03:42
kwolcott ah but what version of Primesieve do I have from MacPorts? Looking...
port installed | grep primes 03:43
primesieve @12.4_0 (active)
ab5tract it seems weird to require a hardcoded list of versions for module loading 03:45
but I've never used the `NativeLibs` module that this code relies on 03:46
so, the fastest option to install is `git clone github.com/CurtTilmes/raku-primesieve.git` and adding `v12` to the for loop on line 7
then `zef install .`
kwolcott ok, thanks for the debugging and hints; will try later tonight. 03:48
ab5tract this is a more comprehensive solution (especially in case there is still an issue with the the macOS library loading) gist.github.com/ab5tract/8aad1605a...594c1cf58b 03:52
kwolcott: it bothered me enough to clone and fix here : github.com/ab5tract/raku-primesieve 03:57
it's not in the ecosystem yet though, so you will need to clone and do a `zef install .`
happy hacking!
04:06 kwolcott left 04:16 Xliff left 04:54 japhb joined 05:08 japhb left 05:10 japhb joined 05:12 zara left 05:28 leah2 joined 06:01 Aedil joined 08:06 japhb left 08:07 zara joined 08:10 japhb joined 08:11 Sgeo left 08:17 zara left 08:25 leah2 left 08:41 leah2 joined 08:47 sena_kun joined 08:48 zara joined
SmokeMachine m: 42.say: "!!!" # LTA Error message? 08:55
camelia Cannot resolve caller print(Int:D: BOOTStr); none of these signatures matches:
(Mu $:: *%_)
in block <unit> at <tmp> line 1
SmokeMachine why is that giving a print error?
and shouldn't that work? `say 42, "!!!"`, shouldn't `42.say: "!!!"` also work? 08:56
m: say 42, "!!!" 08:57
camelia 42!!!
SmokeMachine must be related to this: github.com/rakudo/rakudo/blob/dc03...#L843-L845 09:05
m: augment class Cool { method say(**@args) is default {say self, |@args} }; 42.say: "!!!" # would something like this make sense? 09:19
camelia 42!!!
09:25 Manifest0 joined
timo SmokeMachine: i think 42.say: "!!!" should not work like say 42, "!!!" 09:42
SmokeMachine timo: that's ok, I think... but the error should be better, no? 09:46
timo you mean because your code calls say and it's complaining about print? 09:47
and/or because BOOTStr is an odd detail 09:48
SmokeMachine yes, because a call on "say" errs print... 09:58
BOOTStr also doesn't help... :)
timo right, i'm not sure if we have anything for when one call mostly defers to something else, where we could have put the original thing in the error message. i guess a stack trace is kinda that haha 10:01
10:12 sena_kun left
timo maybe there should be a method more like AT-POS / BIND-POS for the purpose of outputting, that you can manually override instead of having to know you must pick "print" if you don't want to override print, put, as well as say 10:30
but also i'm still not very happy about .say meaning "the invocant is a file handle and you want to output something into it" and "the invocant is an object that you want to output to $*OUT", but i can understand why it exists, and i don't have a good suggestion how to change it 10:31
SmokeMachine why can't classes where .say will use self as the "where" to implement their own .say? speed wise or something like that? 10:35
lizmat m: class A { method print(|c) { say "42" } }; A.say 11:15
camelia 42
lizmat m: class A { method print(|c) { say "42" } }; A.say("foo")
camelia 42
lizmat m: class A { method print(|c) { dd c; say "42" } }; A.say("foo") 11:16
camelia \("foo\n")
42
timo can you rephrase that sentence? i'm not parsing it correctly 11:28
lizmat Mu.say and Mu.put will call Mu.print 11:36
so that a class would only need to provide a .print method 11:37
m: class A { method print(|c) { dd c; say "42" } }; A.say("foo")
camelia \("foo\n")
42
lizmat ^^^
SmokeMachine: re "why can't classes where .say will use self as the "where" to implement their own .say? speed wise or something like that?" 11:40
m: class A { method say(|c) { dd c; say "my own say" } }; A.say("foo")
camelia \("foo")
my own say
lizmat m: class A { method say(|c) { dd c; say "my own say" } }; say A 11:41
camelia (A)
lizmat m: class A { method say(|c) { dd c; say "my own say" } }; A.say
camelia \()
my own say
lizmat hmmmm
SmokeMachine but it also "prepares" say for cases like when calling it on file handles... couldn't them have theres own say? 11:43
I mean this: github.com/rakudo/rakudo/blob/dc03...#L859-L868 11:44
lizmat I'm still not following what you mean *exactly*
SmokeMachine I mean, why not, instead of this: github.com/rakudo/rakudo/blob/dc03...#L843-L852 don't, we have a role with that that (for example) IO::Handle would "does"()? 12:00
(but of course I may be misunderstanding everything) 12:01
lizmat and then remove the code that you just highlighted ? 12:02
pretty sure that would have ecosystem fallout
SmokeMachine my question was more about the reason that was made that way, not a request for changing... 12:25
lizmat the reason was that if a class would like to provide its own output methods 12:31
it would only need to provide a .print method
and .put and .say would automatically be supported as well, with the expected semantics 12:32
I'm not sure adding a Cool.print method would make sense to prevent the error that you saw 12:33
the BootStr in that error is caused by nqp::concat, which is an optimization for infix ~
timo nqp::concat and nqp::join and such "just" return a native str, what it's boxed to when being passed would be a choice of whatever dispatcher it goes through on the call path i think? maybe the dispatcher is hllizing / autoboxing the native string without the right type object from the raku hll, or maybe only the error reporting path doesn't know or something else 12:38
SmokeMachine thanks 12:48
lizmat and yet another Rakudo Weekly News hits the Net: rakudoweekly.blog/2024/11/11/2024-...ternaries/ 13:32
SmokeMachine Odd... I released ASTQuery yesterday and it is not listed at raku.land... 13:42
lizmat SmokeMachine: the REA harvester didn't see it either, so I guess the upload didn't work ? 13:45
SmokeMachine I didn't see any error (I may have missed...). I'll try to pay more attention on the next release... 13:46
I think I'm also going to release this module: usercontent.irccloud-cdn.com/file/.../image.png 14:21
El_Che SmokeMachine: add mofo before the exclamation marks :) 14:34
SmokeMachine mofo? 14:35
El_Che SmokeMachine: short for mother f* 14:36
something that people say when they overreact :)
SmokeMachine :) I like it!
El_Che you need a switch for profanity :) 14:37
SmokeMachine There is a alias on that that only brazilians will understand... :) 14:38
because that function can accept an argument... usually $*CU, and that has a very distinctive meaning in portuguese... 14:39
librasteve SmokeMachine: I usually go zef install My::Module --force-install immediately after a fez upload ... then sense check the the new version made it. Why? Well the fez upload works instantly to the zef repo - then the raku.land indexer will take a while to pick it up - but you know it's on the road. 14:41
SmokeMachine and in brazilian portuguese we say that when someone is overreacting they are with "fogo no <bad word>"...
librasteve: thanks for the suggestion! I'll try that next time 14:42
librasteve yw 14:50
15:18 japhb left
tbrowder howdy. this morning i was thinking about a modules i’m working on and i will need to add test files to check wrappers for system commands—ugh. 15:55
imho, we need a version of Raku’s built-in lives/dies test that auto-uses “run” and just tests the error code 15:58
without any special setup by the use 15:59
*user
unless i get a definite “not feasible” or “no” i will file an issue 16:00
[Coke] Would a module that did that be acceptable?\ 16:05
That seems like an easy module to implement.
16:24 kjp left 16:25 guifa left, kjp joined 16:32 tejr left 16:34 tejr joined
lizmat weekly: dev.to/lizmat/a-path-to-paths-21b5 16:47
notable6 lizmat, Noted! (weekly)
16:52 orangebot joined
lizmat tbrowder: t/packages/Test-Helpers/lib/Test/Helpers.rakumod might be a good source of inspirationm 16:56
16:59 manu_ joined 17:01 eseyman left, manu_ is now known as eseyman 17:07 guifa joined 17:10 japhb joined
ab5tract I don't understand our reticence to expose clearly usable and working implementations that already exist in core with the world 17:16
lizmat ab5tract: fwiw, I don't either 17:22
ab5tract glad I'm not the only one! 17:25
antononcube @lizmat Is there an easy way to get the individual directory names of paths? I.e. using "IO:Paths" metho .parts that gives a list of the (sub-)directories and basename. 17:31
Of course, I can split the directory path string with "/" (or similar) but maybe there is a more direct way.
guifa ab5tract : the ENTER wouldn't work, because it'd be before the object itself existed. You could probably move it to BUILD to beat the TWEAK, but then you're back to the problem of what if they override BUILD 17:35
17:39 japhb left
guifa I guess if COMPOSE were to get implemented it could catch that kind of thing 17:39
ab5tract antononcube: what's IO::Path.parts not doing for you? 17:42
paths(".").map({ $_ => IO::Path.new($_).parts }) 17:44
maybe I didn't understand correctly
guifa: ah, yeah makes sense. It would have been cool if it worked symmetrically like that, but at least your approach does the trick! 17:45
I wonder if/what is a good location for this in the docs...
17:46 japhb joined
guifa I'm actually taking advance of Raku's async and it's really easy 17:46
basically my project ingests an article, and then has several modules that all need to run, some are not processor intensive and can be done simultaneously, others are (because launching multiple LLMs would fry my machine) and need to be loaded one at atime. Meanwhile, want to give status updates for which ones are awaiting start, in progress, and finished. And it's really surprisingly few lines of code to get all that logic 17:47
ab5tract that sounds really awesome 17:48
antononcube @ab5tract I want to get from the directory "/p/h/alf/" the list <p h alf>. I can do that with string splitting, but I was hoping for some unknown to me option to speedup that. 17:51
Or for the package "paths" (by @lizmat) to have some sort of related functionality.
lizmat .say for paths>>.split("/") 17:54
antononcube Right, of course. 17:55
18:06 leah2 left 18:10 leah2 joined
ab5tract Isn't that roughly equivalent to: paths(".").map({ $_ => IO::Path.new($_).parts }) 18:10
without using the original path as the key 18:11
lizmat pretty much, but the paths>>.split("/") bypasses the creation of the IO::Path object completely 18:12
ab5tract right, but in a non-portable manner
anyway, just wanted to make sure I wasn't as confused as I feared I was 18:13
lizmat fwiw, I think on Windows paths are also reported with "/" instead of "\" nowadays 18:18
sadly I haz no way to test, other than starting some stupid CI job :-) 18:19
18:27 japhb left
ugexe i dont think one can assume they can just split on / 18:30
to show that in a CI you'd have to already know in which scenarios that is relevant
otherwise it is trivial to show examples where it is fine 18:32
ab5tract missed the part where antononcube mentioned wanting a speedup, so splitting is probably the best option in that case 18:34
ugexe i'd suggest saving yourself the trouble and using $*SPEC.splitdir("/p/h/alf") 18:35
antononcube đź‘Ť
ugexe as for an example of when it is relevant: windows long paths don't work with / as they bypass the internal windows api that otherwise does that conversion 18:37
dont forget that windows paths often contain a volume e.g. C:\ when splutting 18:39
18:40 sena_kun joined
antononcube @ugexe Right, that is why I was asking for a more bureaucratic way of get the file parts than using split. 19:14
lizmat didn't realize that antononcube was of the windows persuasion :-) 19:17
19:27 guifa left 19:33 japhb joined 19:35 Aedil left 19:36 abraxxa-home joined
antononcube 🤣 19:44
tbrowder antononcube: i hear you’re a heavy Raku and Windows user. the chocolatey packager folks could sure use your talents! 20:00
antononcube @tbrowder My talents a graphed for now... 20:04
Or sparsed... 20:06
20:30 guifa joined 20:33 Guest64 joined 21:11 zetaaaa joined 21:16 Guest64 left 22:01 Geth left 22:06 orangebot left
tbrowder lizmat: ref “inspiration” did you mean for a PR for core Raku or my own tests? 22:53
22:54 abraxxa-home left 22:57 Geth joined
lizmat your own tests 22:59
tbrowder ok 23:10
no interest in possible raku core use? 23:11
23:13 sena_kun left 23:44 yewscion joined 23:54 yewscion left 23:57 Sgeo joined

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!