🦋 Welcome to the IRC channel of the core developers of the Raku Programming Language (raku.org #rakulang). This channel is logged for the purpose of history keeping about its development | evalbot usage: 'm: say 3;' or /msg camelia m: ... | Logs available at irclogs.raku.org/raku-dev/live.html | For MoarVM see #moarvm
Set by lizmat on 8 June 2022.
00:08 Guest72 left, Guest72 joined
ab5tract nine: any further thoughts on R#5875? 00:15
linkable6 R#5875 [closed]: github.com/rakudo/rakudo/issues/5875 [good first issue][enumerations][objects] Success in enum resolution during &MAIN invocation depends on declaration scope of the enum.
ab5tract oops, wrong ticket
R#5879 00:16
linkable6 R#5879 [open]: github.com/rakudo/rakudo/pull/5879 Enhance feedback around named subset parameters
ab5tract I tried chasing down those undefined value exceptions but I've lost patience. It doesn't happen with add-sorry instead 00:19
00:44 [Coke]_ joined 00:46 [Coke] left
Geth ¦ rakudo: coke assigned to MasterDuke17 Issue Unable to build 2025.04 on Alpine Linux github.com/rakudo/rakudo/issues/5864 00:57
02:19 Guest72 left 03:46 Guest72 joined 07:20 librasteve_ joined 09:25 finanalyst left
Geth nqp/main: 136a90a3d4 | (Elizabeth Mattijsen)++ | tools/templates/MOAR_REVISION
Bump NQP to get latest MoarVM fixes
09:34
lizmat argh... Bump MoarVM ... *sigh* 09:39
09:50 librasteve_ left
Geth rakudo/main: fdd3bdc560 | (Elizabeth Mattijsen)++ | tools/templates/NQP_REVISION
Bump NQP to get latest MoarVM fixes
09:56
nine ab5tract: when I had to track down where an exception was coming from, I added an nqp::die to the Resolver's build-exception method. Or if it's not generated by RakuAST code, the exception's constructor 10:17
ab5tract Sounds reasonable.. 10:22
[Coke]_ Now that we've done the bump, rakudo might be broken on alpine again 12:12
12:13 [Coke]_ is now known as [Coke]
ab5tract R#5882 — is this a bug or just natural fallout from disabling the special casing of single args for meta op left? 12:24
linkable6 R#5882 [open]: github.com/rakudo/rakudo/issues/5882 "Triangular comma" metaoperator [\,] unexpectedly puts first element inside a list
12:31 librasteve_ joined
ab5tract unrelated, but bummed to find out that CLion does absolutely nothing interesting for highlighting MoarVM source code 12:46
I have to imagine that normally it knows how to highlight user-defined types and constants 12:47
ah, nice, these aren't exceptions per se but uses of warn in Mu.rakumod 13:05
13:10 Guest72 left
ab5tract nope, don't have the tuits for chasing this, sorry 13:24
right now *nothing* is output unless I do `use fatal`, at which point it does barf but not in any useful way 13:30
it does look like print_exception is doing something silly with stringification 13:34
aside: with any luck R#5882 will turn out to be simply removing a compensation that was added to work around the previous special-casing :) 13:37
linkable6 R#5882 [open]: github.com/rakudo/rakudo/issues/5882 "Triangular comma" metaoperator [\,] unexpectedly puts first element inside a list
ab5tract ahhhhhh 13:46
it's the line number generator
and the Exception details dilly
13:48 finanalyst joined
Geth rakudo/enhance-named-param-subset-feedback: ddc8a40f71 | ab5tract++ | 3 files
Enhance feedback around named subset parameters

Adding a worry here provides users with a bit of helpful feedback when they define an optional named parameter with a subset type but fail to provide a valid default value.
14:32
ab5tract nine: okay, I *think* that covers it
only question left is whether we would want to use a different wording than `at line <anon>`
nine Well the actual line number would be nice :) 14:49
ab5tract there is no line number 14:51
unless we want to default to 1 when it is passed as undefined 14:52
honestly, I'm out of patience with it
not a good justification for merging, but hey, it's honest 14:53
nine But the problem is in source code and thus there is a line number? Do you mean that the proper line number doesn't get set on the Exception object?
ab5tract this is happening while testing via a -e invocation 14:54
so there is no proper line number
nine Even there, the line number is usually just 1
> rakudo -e 'die "here"'
here in block <unit> at -e line 1
ab5tract well, it's not set here 14:55
14:55 finanalyst left
ab5tract the code is essentially the same between add-worry and add-sorry, though, so I'm not grokking why there would be issues with one and not the other 14:56
and this is perhaps the single most boring issue I've chased down in half a decade or more :)
nine Well I've done a few hundred of these so I can't offer very much sympathy :) 14:58
ab5tract :D 14:59
Geth rakudo/enhance-named-param-subset-feedback: b1dda4ae31 | ab5tract++ | 2 files
Add the defaults to add-worry/add-sorry
15:18
ab5tract nine: ^^ reasonable?
nine The problem is that RakuAST::Parameter objects don't get their origin set. The parameter Actions method doesn't call the usual self.attach (which calls SET-NODE-ORIGIN) and instead just does make $parameter; The signature method just collects those objects and passes them to the RakuAST::Signature constructor 15:27
Probably method parameter doesn't call attach to avoid a premature call to .to-begin-time. Ostensibly that would be the correct time for such a call, but setting up parameters and signatures and routines is a very, very, *very* intricate dance, so I wouldn't experiment with the timing in any way. 15:28
Thus the proper course of action seems to be to just call SET-NODE-ORIGIN in method parameter like we do in statement-control:sym<use>. 15:29
ab5tract ok 15:32
Geth rakudo/enhance-named-param-subset-feedback: bc85c88ec2 | ab5tract++ | 2 files
Call SET-NODE-ORIGIN from Actions' parameter method

This was probably a simple oversight, as `SET-NODE-ORIGIN` is usually called "automatically" when an action method calls `attach`.
Due to intricacies in setting up parameters, signatures, and routines, we won't add an `attach` call but will instead call `SET-NODE-ORIGIN` directly just prior to calling `make`.
15:46
ab5tract done and done :) 15:58
nine Does the bootstrap-rakuast branch still build with these changes? 16:00
ab5tract checks 16:21
nine: it does indeed 16:58
16:58 librasteve_ left
nine \o/ 17:15
Geth rakudo/enhance-named-param-subset-feedback: 17 commits pushed by (Tim Nelson)++, rir++, (Elizabeth Mattijsen)++, (Jimmy Zhuo)++, (Justin DeVuyst)++, ab5tract++
review: github.com/rakudo/rakudo/compare/b...e5057d23e5
17:28
rakudo/enhance-named-param-subset-feedback: 8add471ef1 | ab5tract++ | t/12-rakuast/xx-fixed-in-rakuast.rakutest
Fix expected type name
17:44
17:45 notna joined
Geth rakudo/enhance-named-param-subset-feedback: 360a598b60 | ab5tract++ | 3 files
Enhance feedback around named subset parameters

Adding a worry here provides users with a bit of helpful feedback when they define an optional named parameter with a subset type but fail to provide a valid default value.
17:58
rakudo/enhance-named-param-subset-feedback: 4453801552 | ab5tract++ | 3 files
Call SET-NODE-ORIGIN from Actions' parameter method

This was probably a simple oversight, as `SET-NODE-ORIGIN` is usually called "automatically" when an action method calls `attach`.
Due to intricacies in setting up parameters, signatures, ... (5 more lines)
ab5tract sometimes these bots feel a bit too expressive
ab5tract if there's any way to recognize a forced push, it would be super helpful
I guess it's somewhere in the webhooks settings. Which of course don't fall under Actions and therefore invisible to the likes of me 18:07
lizmat github.com/raku/Geth :-) 18:12
nine I don't think you can limit this in the settings. There's a ton of events to select from, but just one for Push in general 18:13
ab5tract boo 18:18
lizmat: I was looking at the source. got there via `/msg Geth help` , which mentions 'Send me everything' 18:19
lizmat yeah, it would need to be handled in Geth, perhaps there's a settting in the JSON that can be interrogated
ab5tract detecting it on Geth's side sounds near impossible
lizmat why?
Geth doesn't push all of the events it gets
ab5tract because there is no differentiation on GH's side
lizmat if there isn't, then tough luck 18:20
I was thinking maybe there is?
18:20 notna left
ab5tract partly it seems tricky to subscribe locally to the hose 18:20
but I haven't looked at Geth's internal docs yet 18:21
lizmat fwiw, it's been a while since I looked at that as well 18:22
ab5tract lizmat: did you see the latest note on R#5882?
linkable6 R#5882 [open]: github.com/rakudo/rakudo/issues/5882 "Triangular comma" metaoperator [\,] unexpectedly puts first element inside a list
ab5tract I'd love to wow this newcomer with a quick fix 18:23
(the original reporter of the issue, I mean)
Geth ¦ rakudo: lizmat self-assigned "Triangular comma" metaoperator [\,] unexpectedly puts first element inside a list github.com/rakudo/rakudo/issues/5882
lizmat will look at it tomorrow 18:24
18:59 Guest72 joined
Geth rakudo/main: 5f6d6e05d2 | ab5tract++ | 3 files
Enhance feedback around named subset parameters

Adding a worry here provides users with a bit of helpful feedback when they define an optional named parameter with a subset type but fail to provide a valid default value.
19:31
rakudo/main: 46ed7b4037 | ab5tract++ | 3 files
Call SET-NODE-ORIGIN from Actions' parameter method

This was probably a simple oversight, as `SET-NODE-ORIGIN` is usually called "automatically" when an action method calls `attach`.
Due to intricacies in setting up parameters, signatures, ... (5 more lines)
22:03 Guest72 left, Guest72 joined 22:55 Guest72 left 22:56 Guest72 joined