🦋 Welcome to the MAIN() IRC channel of the Raku Programming Language (raku.org). Log available at irclogs.raku.org/raku/live.html . If you're a beginner, you can also check out the #raku-beginner channel!
Set by lizmat on 6 September 2022.
00:00 reportable6 left 00:01 reportable6 joined 00:02 guifa_ left 00:41 razetime joined 00:53 ab5tract left 01:12 ab5tract joined 01:20 razetime left
SmokeMachine RakuCron can start jobs now! :) 01:38
ab5tract SmokeMachine: nice :D 01:52
tellable6 2023-08-05T21:20:24Z #raku-dev <lizmat> ab5tract irclogs.raku.org/cro/live.html
guifa__ has officially commenced work on Polyglot::Regexen::Perl 01:54
ab5tract guifa__: I hope that's an additional vote in favor of removing the P5 regex "braid" from core :) 02:01
guifa__ I think we should test it first
before making the decision
ab5tract fair enough
guifa__ At the very least, there's a good argument for maintaining PCRE (if not Perl) in core
even Perl has a switch for PCRE 02:02
ab5tract I think there is a better argument for PCRE than there is for P5 02:03
happy hacking :) 02:04
guifa__ Honestly now that ECMA is done Perl (and PCRE) will be pretty easy 02:05
it's mostly update the exact definitions of escapes (easy), and handling some of the specialized (? …) blocks 02:06
the one I'm struggling figuring out how to do is the same-script one, such that backtracking occurs correctly 02:07
and then I just need to do a lot of tests for the special alternation group that adjusts how positionals are counted, and then done
well, I guess also need to figure out how to handle embedded blocks — I think the only logical choice there is to only allow Raku code. Same for any other language that allows for embedded code 02:09
02:20 evalable6 left, linkable6 left
antononcube @guifa As I threatened I complained publicly about your "laziness" - see the sub-section "Introduction :: Standard enhancements" here: rakuforprediction.wordpress.com/20...functions/ 02:21
02:22 evalable6 joined
guifa__ haha 02:23
02:23 linkable6 joined
guifa__ Polyglot::Regexen is solid now, but it does require a recent version of Rakudo 02:23
02:25 ab5tract left
antononcube Great. 02:26
02:50 guifa left 02:57 bigdata joined 03:11 cm_ joined 03:12 cm left, cm_ is now known as cm 03:31 cm left 03:32 bigdata left 03:33 gabiruh left, gabiruh_ joined 03:41 cm joined 05:10 bloatable6 left, reportable6 left, bisectable6 left, greppable6 left, committable6 left, coverable6 left, releasable6 left, benchable6 left, sourceable6 left, tellable6 left, notable6 left, statisfiable6 left, nativecallable6 left, squashable6 left, quotable6 left, linkable6 left, evalable6 left, unicodable6 left, shareable6 left, bisectable6 joined 05:11 quotable6 joined, shareable6 joined, releasable6 joined, coverable6 joined, sourceable6 joined, tellable6 joined, unicodable6 joined, statisfiable6 joined 05:12 squashable6 joined, linkable6 joined, benchable6 joined, nativecallable6 joined, reportable6 joined, notable6 joined, evalable6 joined 05:13 committable6 joined, greppable6 joined, bloatable6 joined 05:53 MoC joined 06:00 reportable6 left 06:01 reportable6 joined 07:05 abraxxa joined 07:22 japhb left, japhb joined 07:48 jpn joined 07:53 dakkar joined 08:09 squashable6 left 08:10 squashable6 joined 08:12 Sgeo left 08:16 tea3po joined 08:20 teatwo left
nemokosch Another illustration that modules rather depend on a certain version of Rakudo, not a certain language version 08:54
09:21 spacekookie left 09:41 NemokoschKiwi joined
NemokoschKiwi I wanted to test something peculiar: what if you bind-swap two variables that don't have the same scope? 09:42
m: my @outer = <ao bug>; { my @inner = <victor koji>; (@outer, @inner) := (@inner, @outer); }; dd @inner; dd @outer; 09:43
camelia ===SORRY!=== Error while compiling <tmp>
Cannot use bind operator with this left-hand side
at <tmp>:1
------> i>; (@outer, @inner) := (@inner, @outer)⏏; }; dd @inner; dd @outer;
NemokoschKiwi okay, that's still my bad 09:44
m: my @outer = <ao bug>; { my @inner = <victor koji>; :(@outer, @inner) := (@inner, @outer); }; dd @inner; dd @outer;
camelia ===SORRY!=== Error while compiling <tmp>
Variable '@inner' is not declared. Perhaps you forgot a 'sub' if this
was intended to be part of a signature?
at <tmp>:1
------> ter, @inner) := (@inner, @outer); }; dd ⏏@inner; dd @outer…
NemokoschKiwi okay, that *still* is my bad 😅
m: my @outer = <ao bug>; { my @inner = <victor koji>; :(@outer, @inner) := (@inner, @outer); dd @inner; }; dd @outer;
camelia Lexical with name '@outer' does not exist in this frame
in block <unit> at <tmp> line 1
NemokoschKiwi as expected... 09:45
m: my @outer = <ao bug>; { my @inner = <victor koji>; :(@outer, @inner) := (@inner, @outer); dd @inner; }
camelia Lexical with name '@outer' does not exist in this frame
in block <unit> at <tmp> line 1
NemokoschKiwi okay this is bullshit though
actually worse than expected
m: my @outer = <ao bug>; { my @inner = <victor koji>; dd @inner; } 09:46
camelia Array @inner = ["victor", "koji"]
NemokoschKiwi m: my @outer = <ao bug>; { my @inner = <victor koji>; dd @inner; @outer; }
camelia WARNINGS for <tmp>:
Useless use of @outer in sink context (line 1)
Array @inner = ["victor", "koji"]
NemokoschKiwi oops
m: my @outer = <ao bug>; { my @inner = <victor koji>; dd @inner, @outer; }
camelia Array @inner = ["victor", "koji"]
Array @outer = ["ao", "bug"]
NemokoschKiwi so the binding itself guards against this use, with a rather undescriptive error message 09:47
m: my @outer1 = <ao bug>; my @outer2 = <garou garashi>; { :(@outer1, @outer2) := (@outer2, @outer1); }; 09:48
camelia Lexical with name '@outer1' does not exist in this frame
in block <unit> at <tmp> line 1
NemokoschKiwi even worse when they could be swapped actually 09:49
guifa__ nemokosch: I'm depending on an experimental feature, it makes sense that right now I'm on a particular rakudo version 09:54
When 6.e is out, anything that's 6.e compatible will work fine for the module
NemokoschKiwi this is just an extremic example 09:56
in theory, those modules that mangle with the metamodel, depend on Rakudo, and quite possibly a specific version of Rakudo (i.e there are no formal guarantees that things won't break, only the "we are not changing anything" sentiment) 09:57
and in practice, **all** modules depend on Rakudo because the language is not defined in an exhaustive way
any time anything possibly user-facing changes, that can very well break a dependency; only Blin stands in the way of that 09:58
10:00 NemokoschKiwi left
nemokosch there was an issue about this github.com/Raku/problem-solving/issues/253 to be frank, I would have been dissatisfied with my response even 2 months after writing it 10:21
"user code should just depend on the language version" is easy to say but the reality is that user code does not, and mostly cannot, depend on anything but the actual runtime 10:23
actually, I might as well open a new issue for this; not because I anticipate enthusiasm towards a solution but to at least register that this issue has only been growing over time 10:32
librasteve great news - finally you can go crag 'my \λ=2.5nm; my \ν=c/λ; say ν.norm' #119.916..PHz on your command line 10:39
zef install App::Crag (calculator using raku grammars)
lizmat librasteve++ 10:47
but why stop there ? 10:48
crag 'λ=2.5nm; ν=c/λ; say ν.norm
surely should also work ?
:-) 10:52
10:57 spacekookie joined 10:58 melezhik joined
melezhik o/ 10:58
has anyone read the Raku: A language for gremlins - on HN? produced quite an attention there ...
lizmat notable6: weekly :) 11:00
notable6 lizmat, Noted! (weekly)
lizmat notable6: weekly
notable6 lizmat, 3 notes: 2023-08-08T15:34:57Z <lizmat>: news.ycombinator.com/item?id=37040681 ; 2023-08-08T15:34:59Z <lizmat>: lobste.rs/s/x5impj/raku_language_for_gremlins ; 2023-08-10T11:00:34Z <lizmat>: :)
lizmat lobste.rs is also busy
11:14 jpn left
melezhik lizmat ++ 11:38
11:59 TieUpYourCamel left 12:00 reportable6 left 12:01 reportable6 joined 12:08 Aedil left 12:12 Vyrus left 12:13 Vyrus joined 12:14 melezhik left 12:22 jpn joined 12:29 jpn left
nemokosch it might be your lucky day, somebody is wrong on the internet 😂 12:41
this is more important than some "problem solving" issue
what is lobste.rs? Can I get more Dunning-Krueger lads mad at me? 12:43
antononcube @lizmat Agreed, on crag 'λ=2.5nm; ν=c/λ; say ν.norm. Although, any opinion I might voice in that kind of functionalities is going to be very biased by Mathematica and Wolfram|Alpha. 12:45
nemokosch okay this lobste.rs is much more chill 12:48
12:58 Aedil joined
antononcube At some point you realize that time is a precious resource, not worth spending on games, TV, and programming. You can spent all that time arguing with people in The Internet. 13:02
nemokosch makes sense right?? 13:05
13:09 spacekookie left
antononcube I read the lobste.rs thread — mildly interesting… 13:11
… it makes sign/log in and tell them that Julia sucks. 13:12
nemokosch 🤣 13:15
I know, Julia is like Obama...
13:16 jpn joined 13:21 spacekookie joined
antononcube Yes, thank you! 13:21
nemokosch it's always the worst if you don't choose a side though 🥲 13:22
librasteve well it's only a wrapper on raku -e ' ...' not setting out to make yet another language ... and it should be easy to remember too! 13:24
nemokosch anyway, the idea is quite cool 13:25
librasteve ... if ονλυ there was a use nostrict΄ 13:29
checks docs - aha 13:30
how about
crag '$λ=2.5nm; $ν=c/$λ; say $ν.norm'
v0.0.4 on the way
antononcube @nemokosch Reading your comments in ycombinator: I am not sure how generic or scientific the name “produce” (for the routine produce) is. But I have been aware of that operation “very early on” reading/learning Functional Programming (FP). 13:36
My main FP language is Mathematica / WL. So, WL has Fold corresponding to reduce and FoldList corresponding to produce. WL also has Accumulate which is a shorthand of FoldList[Plus, …]. 13:38
@nemokosch LOL : > My point is that Raku may be legitimately confusing and you may be illegitimately confused. 13:43
librasteve huh? 13:46
13:47 Sgeo joined
nemokosch this is what Mr. Antonov always says 13:53
like here ^^ (for IRC folks: quote reply to a 20th March 2023 message) 13:54
oh yes, it's accumulate in Python as well, and that's how the author called it, too 13:55
I think the point is much less what it's called in Raku because that didn't even appear in the article
it just made sense for me to use the Raku name when referencing it 🤷‍♂️ 13:56
and this all leads to the same thing: it's bothering, if not outright crippling, that the article could apparently contain whatever because the audience (at least the loud part of it) will just echo their own prejudices, even if they don't fit the concrete topic 13:59
You may say it's ironic but in this I see the reassurance of my own prejudice: namely that there isn't much point in trying to show people "Raku, the good parts" 14:01
jgaz I may have found a bug 14:09
if I set a `$x = Buf[int8].new(-1)` the value is 0xff (correct) but if I coerce $x [0xff] to an Int with `say $x.Int` I get "1", it should be "-1".
lizmat nemokosch github.com/Raku/problem-solving/issues/377 I wonder if the word Perl hadn't occurred in the article, how the reactions would have been
jgaz That would be correct, given `Buf[uint8]` 14:10
lizmat jgaz: could you make a gist with the *exact* code and what you expect and what isn't ? 14:11
jgaz Can I use pastebin?
lizmat jgaz sure 14:12
nemokosch lizmat: I guess "we will never know that" but it's rather worrying that this wasn't an article very loud about the "Perl" part 14:13
it very slightly mentioned it
lizmat the slightest mention of Perl will set a lot of people off :-( 14:14
nemokosch That is quite possible but I don't think we can really help with that. We can't ban people to every now and then reference "Perl" with regards to something that has spent 19 years "in its infancy" as Perl 6 14:15
lizmat true, but we can limit the exposure as much as we can 14:16
14:17 squashable6 left
nemokosch Also, I think "Perl" is to a large extent just a symbol/strawman to the "write-only" sentiment so my impression is that people would keep making the Perl comparison even if it had no "genetic" basis 14:18
14:18 squashable6 joined
I actually wonder if a comparison to Ruby is much better 14:21
jgaz lizmat: pastebin.com/j7tvNd9k
please met me know if I am having a sudden attack of dumb, but I don't think I am. :D 14:22
lizmat jgaz: when you coerce a Buf to an Int like that, you get the number of elements in the Buf 14:23
jgaz Because Buf is a container not a scalar?
lizmat m: my $x = Buf[int8].new(0xff, 0xff); say $x.Int
camelia 2
lizmat m: my $x = Buf[int8].new(0xff, 0xff); say $x[0].Int
camelia -1
lizmat jgaz ^^
so not a bug :-)
nemokosch see, this may not be a very fortunate conversion, for example... 14:24
and this one isn't even about "write-only" and TIMTOWTDI
jgaz yeah, had I picked something other than 1/-1 I probably would have spotted it.
lizmat how is this different from doing .Int on a native int array ?
jgaz lizmat: thanks
Sorry to bother.
lizmat jgaz: no worries, thanks for asking :-) 14:25
nemokosch the whole "listy things can be coerced to Int as the number of their elements" is shady imo
and I feel some sort of socially reinforced internal pressure to use "clever tricks" like this that will bite everyone at least once 14:26
this particular "clever trick" is too much for my taste - but I kind of keep using it, "since it's already there anyway"
lizmat and that sorta gives the trick legitimacy, as you *do* tend to use it :-) 14:27
but I see your point: some decisions made for Perl, and later Raku, were prematurely huffmanized because of: less typing in oneliners 14:28
14:30 anthk_ joined 14:31 voidzero joined
nemokosch This is a sociocultural thing in a sense. If several mainstream languages did it (like say, implicit coercion to string with the + operator if the other argument is a string), then it wouldn't be worth mentioning probably 14:32
14:37 bigdata joined 14:39 heartburn left 14:49 heartburn joined 15:02 abraxxa left
antononcube Julia got an some numerical software award given every 3-5 years, well, just because Julia's team released version 1.0. It is like Obama getting the Nobel Peace Prize just for "showing up." 15:19
15:30 bigdata left
jdv most of the huff stuff is useful once one learns that's how stuff works 15:31
or so i think
15:39 razetime joined 15:41 razetime left
librasteve ooo 15:44
15:44 voidzero left 15:51 ab5tract joined
ab5tract The thing with lobste.rs is that it is invitation only -- lobste.rs/about#invitations 16:19
It would be great if we could get a few invites for some community members to chime in on Raku discussions 16:20
I'm also fairly sure the website name is a reference to Charles Stross' Accelerando, but I can't be certain
antononcube @ab5tract I have not heard of Charles Stross' Accelerando before. (And I should just because the name of my company sounds similar.) 16:23
16:33 dakkar left
anthk_ hello, I can't use zef under limited constraints 16:48
1GB of RAM, 1.2GB of SWAP, i386, OpenBSD 7.3 16:49
I tried to install Linenoise but it always fails with being out of memory
and, yes, I raised the datasize limits to 4GB
lizmat that feels.... strange 16:53
reading big JSON files doesn't help, but I haven't seen it grow beyond 600MB or so during an install 16:55
tonyo that's probably a big part of it ^, i'm currently working on an indexer that would reduce memory usage 16:58
ugexe: ^^ 16:59
su.shi anyone know if it's possible to host a git server using raku? im trying to use the Cro library
lizmat su.shi I don't know of anybody who's done that, but I think all of the basic building blocks are there in Cro 17:02
su.shi i'll be checking it out, then. thanks! 17:03
anthk_ tonyo: thanks, hope it helps 17:04
ugexe hard to say: obviously zef installed fine, they are using it. But with Linenoise being a native library and OpenBSD being one of the least used OS for rakudo (I would guess anyway) 17:05
who knows
anthk_ I can use rlwrap instead of Linenoise, but not having access to p6's "CPAN" it's lmiting
tonyo anthk_: does `zef update` work OK? 17:06
su.shi stuck on git requesting from Cro, when I try to fetch from the remote, Cro says "Malformed request line", but I've no idea what that means, and it doesn't show anything about the request so I'm not sure how to debug this 17:15
git tells me: 'OpenSSL SSL_connect: SSL_ERROR_SYSCALL in connection to localhost:8080', I'm not sure what this means 17:16
anthk_ tonyo: yes 17:17
su.shi im dumb, it's because i was hosting an http server and tried accessing it as an https server 17:20
lizmat su.shi I think you've misunderstood me
I have no idea what the git protocol does, but I guess it's not really http:// is it ?
su.shi ok , shutting up now :) 17:21
anthk_ ok trying Zef install Linenoise with fetch-degree and test-degree to 1 17:23
nemokosch so that's why lobste.rs is a more decent place, ha 17:26
then I better stay outside :DD 17:27
anthk_ MoarVM panic: Memory allocation failed; could not allocate 524288 bytes
:| 17:29
japhb anthk_: It is possible to tune down MoarVM's various memory sizes -- this has been used extensively in the past to validate memory management and garbage collection correctness. But it's *default* configuration is for a machine with more available RAM than that. 17:31
anthk_ japhb: at runtime or while compiling it? 17:32
japhb anthk_: #moarvm would be the place to ask about how to configure MoarVM memory smaller.
17:32 jpn left
japhb anthk_: I think there are some dials that are compile time, and other that are runtime. Part of the GC config, IIRC. 17:32
Been quite a while since I went spelunking in that code. 17:33
Please note: being really aggressive in tuning those parameters down will result in MoarVM having to GC a LOT more often, which will have "non-trivial performance effects". :-P 17:34
17:35 jpn joined
su.shi it looks like Cro does not support PROPFIND, which it seems to want to use for pushing, so I'm thinking it won't be able to host a git server 17:36
any other good http server hosting libraries?
anthk_ japhb: as long I can still stuff, it's fine. The best would be to finetune it at runtime, enough to not OOM on installing packages.I tried even with --serial so it tries to install things one at a time, but no luck
s,still, install 17:37
nemokosch what is special about a "git server"? 17:39
17:40 jpn left
su.shi I'm writing a server that hosts a git repository through http 17:40
im not sure there is anything special about it, though idk what PROPFIND is yet, so idk if that is something special for git 17:41
looks like it's something to do with a thing called WebDAV 17:42
ugexe If you built rakudo from source in the last week or so it might be worth setting RAKUDO_NO_PRECOMPILATION=1 when installing Linenoise to potentially workaround precompilation itself causing the memory issue 17:43
anthk_ the problem it's when trying to compile LibraryMake 17:54
18:00 reportable6 left, reportable6 joined
nemokosch haven't heard of PROPFIND either 18:05
lizmat yeah, that's pretty ancient WebDAV stuff 18:09
jdv webdav. a blast fron the svn flavored past. 18:18
su.shi i just want to write a server that takes git commits then test compilation on the new code that was committed, but it seems to be so difficult to do 18:22
im not sure how to get it to not use webdavb
antononcube @sushi You might look-up what Melezhik is doing. E.g. "Sparrow". 18:24
www.youtube.com/watch?v=FZBqRRgkDxE 18:36
18:44 PipStuart joined
ugexe anthk_: it would probably help to post the full output using the --debug flag. you also might want to try and explictly install LibraryMake to see if that works in isolation 18:47
su.shi is there anyway to set environment variables in raku? 18:50
nemokosch so that they are visible to the invoking shell? 18:54
anthk_ ugexe: I tried that, installing it to set up the rest later 18:55
but same error
ok, now launcher with --debug 2> err
su.shi the git http backend needs me to set env variables for the script, so they need to be visib le in a script that i will be invoking later
nemokosch I mean that's how it sounds useful but also that's how it sounds so hard to imagine...
anthk_ s,launcher,launched
su.shi oh, i think i can just set env variables when i call the script, right?
nemokosch that you can always do 😉 18:56
su.shi great
anthk_ Enabled extracting backends [git path unzip] don't understand /tmp/.zef/1691693717.54319/1691693874.54319.9076.856737109147/LibraryMake%3Aver%3C1.0.0%3E%3Aauth%3Cgithub%3Aretupmoca%3E.tar.gz 19:02
You may need to configure one of the following backends, or install its underlying software - [tar psunzip]
it's just a gzip file, file from unix got it right, it doesn't look corrupted 19:04
tonyo is tar in your path?
anthk_ yes 19:05
19:09 SuspiciousSwan joined
tonyo iirc, does tar in openbsd not support gzip ? 19:11
ugexe since its showing up in that underlying software listing, that means the `probe` check is returning false - github.com/ugexe/zef/blob/710034fe...od#L81-L90 19:13
i think i remember something about openbsd not liking % in filenames... but maybe that was just for make files 19:14
does `tar --help` give exitcode 0? 19:15
it might be worth updating zef as well... you posted output of zef mentioning `psunzip` which was removed in february 19:22
tonyo it was either freebsd or openbsd doesn't like the .gz stuff, can't remember which 19:25
ugexe if you are thinking of the same thing as me, it was .zip 19:32
maybe not though, bsdtar is the one that actually handles .zip unlike gnutar 19:33
thing is, its not even getting that far 19:35
the probe check on `tar --help` from my link above is returning false
so zef won't even try to use tar to extract anything
nemokosch there was definitely an issue with the % stuff, there is an issue in LibraryMake now that you mention it 19:38
github.com/retupmoca/P6-LibraryMake/issues/42 19:39
tonyo wonder if it just needs escaped with the linker 19:40
anthk_ sorry, I was outside
SuspiciousSwan Please let me know if this would be better in the beginners channel. 19:41
Supposed that I receive objects or types of objects, and a function with a type signature, is there a good way to ask if the function can take in an object?
Clearly this is wrong, but I'm not sure if it's just the use of a role instead of specifying it in the type signature, or something deeper.
And if it matters, the classes will be defined during runtime with the MOP
class A {
    has Int $.a
anthk_ openbsd tar doesn't have long switches
SuspiciousSwan };
class B {
    has Int $.a;
    has Int $.b;
class C {
    has Date $.a
role hasIntA {
    has Int $.a
anthk_ sorry
SuspiciousSwan: cat file "
cat file | nc termbin.com 9999
that way the channel won't be flooded
SuspiciousSwan Sorry about that 19:42
tonyo m: sub a ($p1, $p2) { }; say a.^signature; 19:43
camelia ===SORRY!=== Error while compiling <tmp>
Calling a() will never work with declared signature ($p1, $p2)
at <tmp>:1
------> sub a ($p1, $p2) { }; say ⏏a.^signature;
tonyo m: sub a ($p1, $p2) { }; say &a.^signature;
camelia No such method 'signature' for invocant of type
in block <unit> at <tmp> line 1
tonyo ahh
nemokosch signature is a normal method, no?
m: sub a ($p1, $p2) { }; say &a.signature;
Raku eval ($p1, $p2)
ugexe anthk_: how does one print out what would otherwise be `tar --help`? Like `tar -h` or some such? 19:44
anthk_ tar -h tries to open a non existant tape device, /dev/rst0 19:45
tonyo doh 19:46
tar --help?
anthk_ unknown option -
echo $? returns 1 ofc
nemokosch man tar? 😄 19:47
anthk_ ^exactly what the bot says
ugexe ideally the probe works the same across all os, and does not require additional external software to figure it out (otherwise we're now probing for man`) 19:49
perhaps the exitcode / text of just running `tar` by itself could provide a commonality 19:50
if you were to clone zef and checkout v0.16.2 i suspect things might work for you, because it would still have access to the perl script tar extraction fallback. but i'd also like to put out a fix for the `tar` detection in zef to work correctly 19:53
gfldex m: sub f(Int $, Str $) {}; &f.cando(\(42, 'answer'));
camelia ( no output )
gfldex m: sub f(Int $, Str $) {}; &f.cando(\(42, 'answer')).say;
camelia (&f)
gfldex SuspiciousSwan: ^^
SuspiciousSwan Is the difference between cando and ACCEPTS if it takes a type vs an instance? 19:55
anthk_ ugexe: which git checkout it's v0.16.2? 19:56
ugexe yeah
anthk_ ah ok git checkout v0.1.6.2
ugexe you'll probably need to `raku -I. bin/zef uninstall zef` first 19:57
no its v0.16.2
anthk_ that was a typo, sorry
nemokosch cando returns the candidates that can utilize the given arguments 19:59
and not sure how ACCEPTS is related. That would be the method to back up any sort of smartmatching 20:00
anthk_ ugexe: this way? .raku -I. bin/zef uninstall zef ; raku -I. bin/zef install . 20:01
ugexe yeah that should do it
what does `raku -e 'say $*VM.osname'` say on your system? i might just have to special case the detection in this case 20:02
gfldex SuspiciousSwan: Your question isn't really valid, because in Raku there is no clear destinction between values and types. Types are objects and can be used as values. 20:10
anthk_ wait. raku ... uninstall lasted a lot, so I just rmdir'ed ~/.raku and I'm trying to reinstall it
SuspiciousSwan gfldex - ok so I can detect if it has an a with a destructuring subsignature, but how to I allow it to also accept things from class B?  - sub doit((Int:$a)) { say $a } works for A, but rejects B due to the extra argument 20:13
20:15 teatwo joined
ugexe anthk_: it might not be installed to ~/.raku. `raku -I. bin/zef nuke site home` would delete everything though 20:15
(everything = all installed modules)
anthk_ osname returns openbsd
well, still installing 20:16
20:16 teatwo left, teatwo joined
su.shi not a huge fan of Cro's documentation 20:17
how am I to get the data sent with a post request?
20:18 tea3po left
gfldex m: class C { has $.a; has $.b; }; sub f((:a($a), |)) {}; f(C.new); 20:19
camelia ( no output )
gfldex SuspiciousSwan: you can use a capture to consume any extra attributes
nemokosch cro.services/docs/reference/cro-ht...est_bodies 20:20
I guess this can be useful
20:20 MoC left
anthk_ MoarVM panic: Memory allocation failed; could not allocate 327680 bytes 20:21
SuspiciousSwan Is that better than using a slurpee like sub doit((Int:$a, *%_))? 20:22
su.shi im confused on what this syntax is doing: request-body -> %json-object { # Save it, and then... created 'product/42', 'application/json', %json-object; } is this calling something called request-body, then passing a callable to it ?
gfldex SuspiciousSwan: might be a wee bit faster 20:23
SuspiciousSwan Ok, thank you - I think I was able to navigate through the documentation with people's help and I have something that I thinks works now 20:26
nemokosch As a convenience, the router also exports the subs request-body, request-body-text, and request-body-blob, which take a block. 20:30
so it seems you are right
su.shi ah nice
ugexe I created github.com/ugexe/zef/pull/521 to fix tar detection on openbsd 20:33
20:42 SuspiciousSwan left
anthk_ MoarVM panic: Memory allocation failed; could not allocate 40 bytes 20:46
nemokosch lmfao
japhb anthk_: You know, given the usual OpenBSD paranoia, it occurs to me that it might be necessary to specify that MoarVM is a JIT, and thus should be allowed to write code and then later execute it? Somewhere deep in my memory banks I remember OpenBSD being even stricter about that than basic W^X. 20:50
ab5tract antononcube: It's a great novella that was slow-rolled onto the net for free back in the early 2000s during the slashdot era. Charlie Stross is a helluva writer, from what I remember
yes, USE_WXALLOWD = yes or something llike that in the port 20:51
antononcube @ab5tract Thanks! I will try to find / read it.
anthk_ but the configure args don't compile it with jit support
I can always recompile raku with jit and wxneeded support, but woudn't a JIT need more memory? 20:53
ab5tract antononcube: www.antipope.org/charlie/blog-stati...intro.html (would have sent as a PM but you are on discord :) ) 21:03
antononcube Great thanks! 21:08
ab5tract anthk_: are you compiling from git HEAD? or from a specific release tag? 21:14
anthk_ I used the openbsd package 21:15
ab5tract ouch 21:19
oh wait, sorry, that's not so bad
. o O ( what is time, anyway? )
anthk_: it might be worth trying recompiling with jit, just to see what happens. I'm not aware of any breakage in non-jit builds but it is also not a very common way to build rakudo 21:20
anthk_ ab5tract: ok, brb, reboot (wxallowed on port dir) 21:21
21:21 rypervenche left, Tirifto left, kaskal left, eseyman left, vrurg left, camelia left, bdju left, tadzik left, eseyman joined, Tirifto joined, rypervenche joined, rypervenche left, rypervenche joined 21:22 vrurg joined, tadzik joined, bdju joined 21:27 camelia joined 21:28 kaskal joined
anthk_ ok, recompiling, it will last a lot 21:33
ab5tract anthk_: ack. thanks for trying this again 21:35
22:28 linkable6 left, evalable6 left 22:29 evalable6 joined, linkable6 joined 23:29 squashable6 left, evalable6 left, bisectable6 left, statisfiable6 left, coverable6 left, nativecallable6 left, releasable6 left, linkable6 left, unicodable6 left, notable6 left, quotable6 left, reportable6 left, tellable6 left, benchable6 left, shareable6 left, sourceable6 left 23:30 coverable6 joined, releasable6 joined, evalable6 joined, unicodable6 joined, bisectable6 joined 23:31 quotable6 joined, linkable6 joined 23:32 notable6 joined, tellable6 joined, statisfiable6 joined, squashable6 joined, sourceable6 joined, shareable6 joined, benchable6 joined 23:33 nativecallable6 joined, reportable6 joined