🦋 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.
gh g 07:31
melezhik o/ 12:11
I've read the cro.services/community-transfer - so, I hope cro will remain, I wonder if we as a community ready to maintain the project, like I said, I have a lot of production projects which are based on it ... 12:12
just a thoughts ...
lizmat we would need more people to contribute and get acquainted with the code base 12:20
patrickb is already going ahead with this
melezhik yeah, I mean , I've been building an internal installation and configuration platform based on Spark for product in the  company I work for, and Sparky is built on cro 12:23
so if everything goes well we might have another big company using Raku and cro ), the stakes are high ) 12:25
lizmat well, the Raku IRC logs site also runs on Cro :-)
melezhik right/right , we "all" do ) 12:27
so don't make me say that Humming Bird is our last resort, this is going to be a lot of code to rewrite ha-ha ))) 12:28
xinming All my personal projects are written in raku, I believe Raku will shine :-) 12:55
xinming also when cro handles to community, I'll also use it, and still using it. 12:56
lizmat maybe a strange thought: should cro become Raku core ? 12:57
core, as in living in the lib directory, so a "use Cro::..." would always work 13:02
melezhik lizmat: imho I found cro code a bit "heavy" to be included in the core ... I might be wrong though ... 13:10
lizmat well, there's that :-) 13:10
melezhik so, you're agree or disagree? 13:11
lizmat that the code is heavy? yes 13:12
melezhik anyways, having a web framework built into a language is a good idea
ah, ok ))
lizmat whether that would matter for inclusion? not sure
one point against it would be that the spectest would run considerably longer
melezhik well, I mean code included into a core should lightweight and basic enough, so it won't get degrade with a  language inevitable changes 13:13
> one point against it would be that the spectest would run considerably longer - that's just the consequences of this )
Voldenet I do think there is worth in having as much as possible as built-ins 13:19
Voldenet built-ins don't even have to be the best in anything, but they make doing anything a lot easier 13:26
xinming lizmat: I don't think it's good to add Cro to the core. I wish the language thing should be as small as possible. probably we can also create a raku-bundle, which is the same as raku distribution but with stablized lib bundles. 13:33
lizmat is cro in Rakudo Star ?
Voldenet > github.com/rakudo/star/blob/master...odules.txt 13:34
xinming and I doubt bundle kind of thing will work
Voldenet doesn't appear so
I believe some kind of http server/client should be included with the core language, like IO::Socket 13:40
but it should be something basic, that'd fit in two files perhaps 13:45
for example, python has very basic http.server that fits in one* file 13:49
* - around 1kLOC
lizmat does it do https:// ? 13:50
Voldenet I'm not sure, I'd guess not
There's ssl.wrap_socket, which accepts the socket from http server 13:51
very similar to IO::Socket::Async::SSL, which uses openssl as dependency 13:53
golang has net/http, node has http.server, php has built-in server 13:56
Voldenet I don't think anything beats `python -m http.server` oneliner in convenience 14:00
ugexe its the yaer 2024, if a http client/server library is being included it should include ssl support out of the box 14:21
lizmat so that would mean putting OpenSSL into core 14:25
El_Che go had great success by adding a web client and server in the core library, not by providing something simple, but by providing something usefull that frameworks could easily extend. Even those with code from scratch use the same abstractions. 14:26
lizmat I for one wouldn't mind "slurp" taking a URL :-) 14:28
El_Che of course it needs to get the size/usefulness balance right 14:30
and that's hard
lizmat on that thought, I wonder whether it would make more sense to include libcurl into MoarVM
ugexe the old design docs had IO::File.new(uri => ...) 14:35
which could have taken non-file backends
guifa_ tonyo : we can pullback a fez upload if it's only been a few minutes right? Realized I left some files there because I had to change the build process -- public domain so not a huuuuuuge deal but I'm a perfectionist ha 15:30
tbrowder__ do it quickly 15:51
maybe 4 hrs, but i forget 15:52
guifa_ tbrowder: was it just a case of needing to reupload?
tbrowder__ no i was trying to remove it so i could change something and resubmit 15:53
guifa_ yeah I mean to do it did you just rerun fez upload wiwth the same version or? 15:54
tbrowder__ well, you can remove it if you haven't exceeded the time limit, but i usually have exceeded the time so i just have to bump the version 15:55
fumble fingers!! 15:56
and raku.land is quickly advertising new uploads now 15:57
guifa_ Ah I got it, just needed to use version before auth for the remove command
tbrowder__ m: class Foo { has $.a = 1;}; my $o = Foo.new; say $o.a 16:00
camelia 1
tbrowder__ m: class Foo { has $.a=1}; my $a = Foo.new; class Bar is Foo { submethod TWEAK { $!a = 2 }}; my $b = Bar.new; say $a.a; say $b.a 16:03
camelia ===SORRY!=== Error while compiling <tmp>
Attribute $!a not declared in class Bar
at <tmp>:1
------> ar is Foo { submethod TWEAK { $!a = 2 }}⏏; my $b = Bar.new; say $a.a; say $b.a
expecting any of:
horizontal whites…
tbrowder__ m: class F {has $.a = 1}; class B is F { $.a = 2}; my $b = B.new; say $b.a 16:08
camelia ===SORRY!=== Error while compiling <tmp>
Variable $.a used where no 'self' is available
at <tmp>:1
------> lass F {has $.a = 1}; class B is F { $.a⏏ = 2}; my $b = B.new; say $b.a
expecting any of:
term
tbrowder__ m: class F { has $.a =1}; class B is F { has $.a = 2}; my $a = F.new; my $b = B.new; say $a.a; say $b.a; 16:12
camelia 1
2
guifa_ vrurg & lizmat I just updated Timezones::ZoneInfo to include the fix for github.com/rakudo/rakudo/issues/5502 16:21
librasteve it's an interesting idea to bundle cro with raku --- initially I liked it, but since even ruby does not include rails, I think best to keep cro distinct as a zef install 21:07
SmokeMachine I think adding it on Star would be good 22:11
leont librasteve: That's also because ruby has tons of frameworks that are better than rails. 22:20
El_Che librasteve: part of the success of go was the included webserver, and like leont mentioned, ruby is in a different place as raku. 22:49
*due
librasteve hmmm - maybe I am coming round ... it's certainly a big positive asset to the raku capabilities and I really like the way that cro leverages core features such as react and routes ... 22:53
El_Che it's something that needs a lot of thought and vision, not a decision to take hastely