22 Feb 2024
SmokeMachine it seems my request finds the correct route, executes it but continues trying to execute the other routes and then breaks... 22:07
This a smaller version of this (github.com/FCO/HTML-Component/blob...r.rakumod) where I got this error 22:14
lizmat: any idea? 22:15
lizmat nothing jumps out :-( 22:56
jnthn SmokeMachine: `where` blocks are not closures; they're traits attached at runtime, so there's one static where block 22:58
SmokeMachine But when I print inside of it, it prints the right values… 🤔 23:00
gist.github.com/FCO/e08b717af87627...-output-L1 23:01
jnthn: 👆
jnthn Try a `print` inside of the `where`
The block is a closure, but a signature is not copied per closure 23:02
SmokeMachine That’s a say inside the when… gist.github.com/FCO/e08b717af87627...ile-cli-L9
And it continues trying even after a True result… 23:03
On my original implementation I had made the test a block and passed the block to the where: github.com/FCO/HTML-Component/blob...akumod#L35 23:07
jnthn: 👆 23:09
jnthn: I tried doing `my &where = -> Int $i, Int $actual { say "$actual == $i -> { $actual == $i }"; $actual == $i }` and used it on when like: `post -> Int $a where &where.assuming: $i {...}` 23:18
I 'm still getting: www.irccloud.com/pastebin/bBAmBlLc/ 23:19
is there a way to do that inside a loop? 23:21
or the only solution is have a single route and do the validation inside of it? 23:23
jnthn Nol the `where` block is not a closure. You can try requesting /0 and /1 and probably observe it always compares against the same RHS
It's either that or something EVAL-y 23:24
SmokeMachine ok, I'll try that...thanks 23:25
jnthn You can completely golf Cro out of it, fwiw 23:26
raku -e 'my @x; ^3.map: -> $i { @x.push(-> $ where { say $i; $_ == $i } { say $i }) }; @x[0](0)'
SmokeMachine yes... 23:31
m: my @x; ^3 .map: -> $i { @x.push(-> $ where { say $i; $_ == $i } { say $i }) }; @x[0](2)
oh, sorry... no camelia... :(
23 Feb 2024
Have anyone seen this? it breaks when receive a request with empty body and content-type application/x-www-form-urlencoded: gist.github.com/FCO/e41f8620aa38a2...7143cadd50 23:06
24 Feb 2024
jnthn SmokeMachine: Try upgrading to the latest Cro::HTTP; I think I discovered and fixed that one last month. 18:03
SmokeMachine :) thanks, I'll try 20:01
6 Mar 2024
Geth cro-webapp: patrickbkr++ created pull request #88:
Fix using template parts via render-template
21:27
japhb I took a look at the PR, which seems pretty simple, but it's failing CI ... and as far as I can tell, that's unrelated to the PR, but rather a problem or incompatibility between Digest and the 2021.03 Rakudo release. 23:36
I want CI to actually be useful for us, so I'm going to try CI'ing with only Rakudo latest. (I've recently come across at least one module that only works in 2023.12 or later *intentionally* so it seems there's a compatibility break there of some kind.) 23:38
OK, trying by just running the ubuntu-latest/latest Rakudo job in the CI version matrix, we'll see how it goes. 23:39
12 Mar 2024
Xliff Wakey wakey? 00:11
Has anyone done anything with Cro::HTTP::Auth?
japhb Xliff: A tiny bit: github.com/Raku-MUGS/MUGS-UI-WebSi...on.rakumod 00:32
Xliff: That defines a session, and then the session-routes() sub here makes use of the definitions: github.com/Raku-MUGS/MUGS-UI-WebSi...es.rakumod 00:33
Xliff Thanks, japhb! I'll take a look, later. 00:34
japhb :+1:
patrickb Cliff: Another case: I use a JWT auth in the RakudoCIBot: github.com/Raku/RakudoCIBot/blob/m...akumod#L20 08:17
14 Mar 2024
melezhik o/ my cro application (sparky) consumes way too much CPU even being idle, is it typical for cro? 18:03
basically here the code - github.com/melezhik/sparky/blob/ma...y-web.raku
maybe web sockets do that? 18:04
github.com/melezhik/images/blob/0b...21.05.jpeg 18:06
github.com/melezhik/sparky/commit/...12e180fc4d 18:16
hopefully this will fix this 18:21
japhb melezhik: Did that work? Because no, >2.6 cores at *idle* is really not normal. 23:48
I did a lot of work on Cro's WebSockets performance, and assuming you're not sending a *continuous* stream of packets or connection requests, or really *large* packets, it should barely even notice it. 23:50
15 Mar 2024
melezhik . 12:52
hi japhb, yes this did the trick, basically I had a web socket endpoint gets called all the time from the main page, that produces constant web socket flow and executing uptime command every 10 second, and that consumed a lot of CPU, now this endpoint is not called from web UI - 12:54
github.com/melezhik/sparky/blob/ma...#L409-L428
good lesson for me )