seamusb yeah, that is what I thought. I was poking around in Cro and trying to see if the localhost value is set somewhere but nothing obvious. This is a Cro problem though, I have other simple services running in the same setup. I suspect there is some magic with IP addresses I am missing. I don't think I am doing anything wrong as I took out my own app and dropped in an example CRO app. That did not work either :) 00:03
It is binding to localhost in the container and that is it... it does not seem to bubble up through to the docker host. I tried host vs bridge networking too and no luck. 00:04
00:19 abraxxa-home left 00:23 seamusb left 00:24 seamusb joined
timo do you have the service.raku that it autogenerated for you if you used "cro stub"? 00:24
it takes NAME_OF_YOUR_APP_HOST and NAME_OF_YOUR_APP_PORT from the environment
but you should be able to search for a call to Cro::HTTP::Server.new 00:25
seamusb cool, I will take a look, cheers! 00:26
timo cro stub also generates a Dockerfile for you that has the env vars set in it and also the correct EXPOSE directive 00:28
at least i assume it did that for me for this project because i don't remember doing any of this by hand 00:29
00:34 eseyman left 01:27 Manifest0 left 01:52 MasterDuke joined 02:24 kylese left, kylese joined 03:15 kylese left, kylese joined 03:33 nine left 03:34 camelia left 03:48 silug5 joined 03:49 MasterDuke left, silug left, silug5 is now known as silug 04:05 camelia joined 04:10 nine joined 05:03 gabiruh left, skaji__ left, skaji__ joined 05:04 gabiruh joined, Ekho left 05:05 Aedil joined 05:08 Ekho- joined 05:57 elcaro left 06:59 eseyman joined 08:01 Sgeo left 08:41 seamusb left, seamusb joined 09:07 zenmov left 09:12 seamusb left 09:13 eseyman left 09:18 zenmov joined 09:24 dakkar joined 09:38 Aedil left 09:52 Manifest0 joined 09:56 sena_kun joined
Geth advent/main: 4ecef5234a | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | raku-advent-2024/authors.md
6 -> 24, 19 -> 6
 10:27
lizmat 6 -> 23 actually 10:37
11:53 mscha joined
mscha m: for <a b c>.pairs.combinations(2) -> ($a, $b) { say $a; say $b; } 11:53
camelia Too few positionals passed to '<anon>'; expected 2 arguments but got 0 in sub-signature
in block <unit> at <tmp> line 1
mscha What's wrong with this?
antononcube Interesting... 🤔 13:05
@mscha More containerization is needed. 13:09
Or flattening. 13:10
@mscha Here is code that kind of explains what is going on: for <a b c>.pairs.combinations(2).flat -> $a, $b { say $a; say $b; } 13:11
Basically, <a b c> has odd number elements, that is why the message "Too few positionals ..." 13:12
Another illustration code: for <a b c>.pairs.combinations(2) -> $a, $b { say (:$a, :$b); } 13:13
librasteve seamus’s: are you going docker run -p 8888:8888 -it name. .?
13:27 seamusb joined
seamusb just wanted to thank timo for the headup on the cro stub command... the docker images it pulls are out of date, but the dockerfile struture was a help. I also removed Docker Desktop and just used docker-ce. 13:28
That actually worked :) so two issues - Cro does not seemm to like some level of indirection that happens in Docker Desktop (it has a more complex internal IP routing thingy, compared to standard docker-ce) and also Cro seems to need to be the primary/first thread started in the app. I switched my app logic to a background thread and started Cro first. Bingo! Popped up at localhost:10000 using docker-cd. Nice one timo, thanks! 13:31
(docker-ce is the simpler community edition cli available for Linux)
lizmat seamusb++ 13:32
perhaps turn that into a Cro PR ?
seamusb Well, I might actually do that :) But I am struggling through building a prototype for a deadline at the moment. It will be January before I get to it. It needs a proper write up really, so on the todo list. 13:34
lizmat ++seamusb 13:48
[Coke] github.com/tony-o/perl6-data-dump/issues/40 14:17
14:26 seamusb left
[Coke] timo: github.com/timo/json_fast/pull/93 14:27
nahita3882 > Basically, <a b c> has odd number of elements, that is why the message "Too few positionals ..." but if you try with, e.g., <a b c d>, same error appears; the oddness would be a problem in the code you shared afterwards (for <a b c>.pairs.combinations(2) -> $a, $b { say (:$a, :$b); }) because the signature now tries to take 2 elements at a time 14:33
14:33 Sgeo joined
but with ($a, $b) it's supposed to take 1 item at a time, then destructure it to 2 things 14:33
when one (or both) of those things is a Pair, the mysterious error mscha showed appears (and I think it's a bug)
14:46 Guest89 joined 14:47 Guest89 left
[Coke] weird. my PR to json-fast seems to fix my windows issue ('zef test .') in a checkout of my PR passes... 9 times out of 10? I got one test failure with a --debug run... which did not recur. 15:03
so maybe there's a flapper in there?
bah, the next 10 didn't fail either. 15:04
timo: if you can get a release out for that would be appreciated. 15:05
... heh. I just tried to do 'zef install .' from my PR - failed a test. rerun. passed. wondering if there's still some comp issue that happens first time? 15:07
ugexe, japhb : github.com/japhb/CBOR-Simple/issues/4 - there is a 'use lib' in those tests, but it's pulling in libs that are JUST for testing, not the installable libs. The error message, however, is identical to the one I was seeing on JSON::Fast which was resolved by removing the 'use lib 'lib'' from the test files. 15:14
(this on my win11 box)
(and that's the only failure left from my windows cert revocation yak shaving yesterday)\ 15:17
thanks to ugexe++ and ab5tract++ for their help on that.
ugexe seems like it might be hitting the path length issue trying to precopile t/lib/CodecMatches.rakumod 15:25
not sure how it could pass its tests with the `use lib ...` removed though since presumably it is needed to use that test helper module in the tests 15:29
15:32 thundergnat joined
thundergnat Actually, mschas example fails due to trying to pass in a list of pairs. When the list is destructured, the pairs are treated as named parameters so there are no positionals to slot into the $a and $b variables. 15:36
tellable6 2024-02-16T00:50:40Z #raku <[Coke]> thundergnat Terminal::ReadKey does not work here, no
thundergnat Unfortunately, there doesn't seem to be an easy way to override the behavior, other than destructuring inside the block. 15:37
m: for <a b c>.pairs.combinations(2) { my ($a, $b) = @_; say $a,' and ',$b; } 15:38
camelia 0 => a and 1 => b
0 => a and 2 => c
1 => b and 2 => c
thundergnat So, not really a bug, but maybe LTA. 15:39
15:41 thundergnat left 15:46 donaldh_ is now known as donaldh 15:55 abraxxa-home joined 16:01 Aedil joined 16:05 Aedil left 16:08 yewscion joined 16:13 Aedil joined
[Coke] patrickb: The path on that is actually slightly longer because I switched out my work ID for "coke". :| 16:21
Thankfully it didn't reduce it under the critical limit.
16:32 seamusb joined 17:33 dakkar left 17:50 eseyman joined
ugexe each time you run `zef install foo` it will precompile everything into a new temporary staging comp unit repository, so a "first time" precompilation failure that gets reused for a second attempt that can succeed wouldnt be a thing 17:50
[Coke] ok. then it 17:51
's probably "just" a flapper.
ugexe yeah
[Coke] I can try to setup a stress test and see if I can get it to trigger with more details.
ugexe im still not sure how you were able to remove that `use lib "lib"` and have tests pass though 17:53
[Coke] .. that's the direction you've been giving for ages? 17:54
perhaps you're refering to the CBOR issue as opposed to the JSON::Fast issue? 17:55
ugexe i should have said `use lib "t/lib"`
which is fine
use lib $*PROGRAM.sibling('lib');
evalable6 Saw 1 occurrence of deprecated code.
==========…
ugexe, Full output: gist.github.com/663eb4c1eb44900458...4a59145527
ugexe which gives the tests access to CodecMatches (github.com/japhb/CBOR-Simple/blob/...3C5-L3C17) 17:56
17:57 mscha left
ugexe ah i misunderstood your message earlier. i thought you had removed the `use lib`s similar to what you did in JSON::Fast, not that you were just seeing a similar situation but hadnt attempted to correct it 17:58
[Coke] right, that's it. :) 18:00
so if it's the path length, will hopefully be able to test patrickb's branch after he makes a build for me.
18:19 mscha joined
mscha A bit late, but thanks all for the explanations. The last one (thundergnat) seems correct. ("for <a b c>.combinations(2) -> ($a, $b)" works fine.) LTA indeed. 18:19
lizmat if you feel this should be fixed, please make an issue for it 18:27
patrickb [Coke]: Since the new CI pipeline, we have now artifacts for every commit, including relocatable builds for Windows. :-) 18:42
dev.azure.com/Rakudo/rakudo/_build...dArtifacts
19:12 mscha left 19:50 hellwolf left, hellwolf joined
ab5tract patrickb++ and timo++ and. MasterDuke++ and everyone else++ who helped get the CI back into pristine shape! 20:01
That was some serious work that you all put in and it’s deeply appreciated 20:04
[Coke] looks like #20241205.6 • Hopefully address $*USER / $*GROUP on Windows 20:15
failed on win?
20:18 Aedil left
[Coke] (I tried to rerun it but I don't have privs - looks like that test might be a flapper?) 20:20
20:36 seamusb left 21:03 tjr left 21:04 tjr joined, whatsnew joined 21:05 whatsnew left, whatsnew joined 21:06 whatsnew left 21:20 yewscion left 21:26 yewscion joined 21:33 seamusb joined 21:48 yewscion left 21:56 sena_kun left 22:20 seamusb left
patrickb [Coke]: Did you manage to download and use one one of the CI builds? 22:35
22:48 perlbot_ joined 22:49 perlbot left, perlbot_ is now known as perlbot
[Coke] I didn't realize that was an option. 23:03
I'll give it a shot when back on the windows box
patrickb 👍🏼 23:04
23:31 rakufan1 joined
Geth advent/main: 865595c29d | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | raku-advent-2024/authors.md
Scheduled day 7
 23:41

Powered by Cro and the Raku® Programming Language.

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

Thank you!