timo is just three digits for the FD enough? 00:46
ugexe 3 levels of ternary 😬 01:08
4 01:10
patrickb quick googling gives I'd need 5 digits 04:53
I have tests failing on x86 macos with MoarVM panic: Internal error: zeroed target thread ID in work pass t/02-rakudo/12-proto-arity-count.t 05:14
MoarVM panic: Internal error: zeroed target thread ID in work pass t/02-rakudo/12-proto-arity-count.t 05:15
that blocks the precomp release.
timo please try with a moar built with --asan 05:37
check the --help of Configure.pl for the additional env var you need to set
patrickb I don't have a Mac. That's on Azure 06:51
09:06 finanalyst joined 10:05 finanalyst left
Geth nqp/main: 5bf8ae18ea | (Elizabeth Mattijsen)++ | tools/templates/MOAR_REVISION
Bump MoarVM for libuv bump, dogbert++
 10:22
rakudo/main: a0626a20c7 | (Elizabeth Mattijsen)++ | tools/templates/NQP_REVISION
Bump NQP for libuv bump on MoarVM, dogbert++
 10:35
rakudo/main: 83fc678835 | (Elizabeth Mattijsen)++ | t/12-rakuast/xx-fixed-in-rakuast.rakutest
Add test for #5649

To make sure we don't revert to the legacy behaviour		 10:39
11:45 sena_kun joined
jdv somehow the last blin container ended up 11GB. i think its normally more like 3GB. wonder what happened... 13:40
timo do you still have the files? 14:15
jdv nope. was cleaning up my disk. 15:22
maybe next month ill take a gander
ugexe tbrowder: have you tried just copying what the dependency using that library does for its github action? github.com/pdf-raku/Font-FreeType-.../macos.yml 15:24
notably it looks like you are not setting LD_LIBRARY_PATH to where ever homebrew installs libraries 15:25
so probably need export LD_LIBRARY_PATH=/opt/homebrew/opt/freetype/lib/ 15:26
Geth rakudo/main: 6edab9f9e0 | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | src/core.c/Order.rakumod
Add :by to .min/.max/.minmax

The sub versions of min/max/minmax take a :by named argument to indicate the comperator. However, the method versions take an optional positional argument and ignore any :by specification (because of the implicit *%_ in the method signature.
... (6 more lines)		 16:00
lizmat ugexe: it was my understanding that setting LD_LIBRARY_PATH is *not* enough MacOS 16:07
?
ugexe all i did was look at the actions on that repo and they are passing on macOS, so it certainly seems to be what needs to be done 16:11
the latest commit even calls out fixing macos specifically - github.com/pdf-raku/Font-FreeType-...d62204676c 16:12
lizmat that *copies* the native lib to the current directory ? 16:15
ugexe among other things sure, but that is getting off track. it would need to be set regardless and that is not done in their existing github action. my originally suggestion was to copy what was done in the demonstratably working repo of the dependency itself 16:21
lizmat ok, so I recently worked on GD::Raw 16:22
brew install GD
ugexe to be clear i am not trying to debug this issue. i just asked myself if anyone else had this working, checked, and saw they indeed do. so i suggest others to look at that for installation
lizmat then run "ensure-symlink-for libgd.3" from MacOS::NativeLib 16:23
and the tests of GD::Raw pass on MacOS
no need to set LD_LIBRARY_PATH or amything else 16:24
ugexe well to be fair i dont know what "ensure-lylink-for" is or what it does. i can see what the code I linked to does
lizmat it puts a symlink in the install/lib dir of the executable to the library in /opt/homebrew/... 16:25
ugexe first as someone either reading the CI stuff or writing it for my modules that may not be raku, I don't want to learn some special tool. native library stuff like this is going to be mostly the same across different languages so i would want to do things in the more generalized way. next, i dont think the description you gave is even correct (which i could verify when just writing the actual 16:29
commands instead) since i thought homebrew uses different paths for x86 and arm
lizmat for the CI of GD::Raw, I added: raku -e 'symlink "/opt/homebrew/lib/libgd.3.dylib".IO, $*EXECUTABLE.parent.sibling("lib/libgd.3.dylib")' 16:30
I didn't want to add another dependency 16:31
ugexe yeah i just dont really know enough about or have enough experience with the underlying issue to give any actual insights 16:36
timo i'm not having the greatest success with comma-2.0 haha 17:16
[Coke] jdv - there was a bug in a module that would have corrupted the local zef listing and made a bunch of stuff break. Possible that after the zef fix now you're just testing more modules. 17:45
jdv wut now 17:49
if thats true maybe the tprf should shoulder the blin cost 17:51
11 to 12h is a bit much
a huge aws box can probably do it in < 3h 17:52
for ~$30 a run or so 17:53
i used to do it in ~2h for ~$20 17:58
but im poor now...
and if the tprf could get back to me anytime this year thatd be great *grumble* 18:01
*sigh*
timo it would be interesting to know how much of the time is spent in the "automatically bisect rakudo commits for changes in models" step, maybe there's a lot of time to save there until someone actually needs to know where exactly the change happened, like sometimes it's obvious
18:05 rba_ joined 18:06 summerisle joined
ab5tract timo: what trouble are you having with comma? 18:07
Also, I take comma problems seriously, they are no laughing matter ;) 18:08
18:10 bloatable6__ left, rba left, eof left, jdv left, rba_ is now known as rba 18:11 bloatable6__ joined, jdv joined
[Coke] timo;almost no time is spent there. 18:11
it's drowned out by the sheet number of tests we run, esp. on those that are failing both ends 18:12
jdv youz guyz... 18:13
timo ab5tract: not sure how much you can reasonably do about the things i see here haha 18:22
ab5tract I don’t get what’s so funny, so maybe you could share? 18:24
I’ve fixed a lot of bugs, I am pretty confident I can fix more
jdv youz guyz... 18:26
timo well, right now i'm looking at a frozen IDEA :) 18:28
the jvisualvm next to it is for some reason also frozen now 18:29
when i open something that automatically adds a pair of parenthesis and hit enter, it seems like the closing paren gets 18:31
put on column 0, and when i put my cursor in front and hit tab, it does the "jump out of parens" action, which is expected and would be correct if i didn't explicitly want to (/ have to) indent the closing paren 18:32
in the grammar previewer panel i get "failed to parse grammar" or so and i don't think i see a way to get the actual error. which is extra un-fun because rakudo itself compiles that grammar just fine
gist.github.com/timo/578ba3a54764b...d64ee565a9 maybe this gives a hint, maybe not 18:35
ab5tract Regarding the grammar panel, I have to admit I haven’t had a chance to look at that yet 18:38
I’ve never used it even I. The original comma, so I need to go back and see what’s actually supposed to work
The parents thing I’m familiar with, I don’t think it’s a comma specific thing but i will take a look at it 18:41
timo the grammar preview thing might also be unhappy about "unit grammar" 18:42
ab5tract Did it freeze while engaging with the grammar preview? 18:43
timo no, there i was following a use statement to the source of a module 18:51
ab5tract Got it 18:58
Would you mind filing some issues? An example broken grammar would be helpful
timo gonna have to circle back to you on that 19:27
19:35 doomlord joined 19:36 finanalyst joined 20:13 doomlord left, doomlord joined 20:18 doomlord left
timo oh, Pair isn't Cool 20:24
timo sent two pull requests to Data::Dump::Tree 20:51
ab5tract timo: your intellij version is 2024.2+, right? I think it should fail to install otherwise but I wanted to double check 20:52
timo 2024.2.3 20:59
ab5tract Excellent, thanks for checking 21:07
21:46 finanalyst left 22:29 sena_kun left

Powered by Cro and the Raku® Programming Language.

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

Thank you!