00:50 sena_kun joined 00:52 Altai-man_ left 01:12 Kaeipi left, Kaeipi joined 01:19 Altai-man_ joined 01:21 sena_kun left 02:49 Kaiepi joined, Kaeipi left 03:04 Kaiepi left 03:11 Kaiepi joined 03:20 sena_kun joined 03:21 Altai-man_ left 05:19 Altai-man_ joined 05:22 sena_kun left 06:09 j3nnn1 left 06:25 Kaiepi left 06:26 Kaiepi joined
Geth nqp: 6f190e52d4 | (Christian Bartolomäus)++ | 2 files
[JVM] Fix some error messages
07:20 sena_kun joined 07:21 Altai-man_ left 09:19 Altai-man_ joined 09:21 sena_kun left 09:35 lichtkind joined
lizmat Files=1306, Tests=111242, 214 wallclock secs (28.76 usr 8.36 sys + 3005.82 cusr 273.84 csys = 3316.78 CPU) 09:40
10:11 [Tux] left 10:18 lichtkind_ joined 10:20 lichtkind left 11:10 [Tux] joined 11:16 MasterDuke left 11:20 sena_kun joined 11:21 Altai-man_ left 12:04 Voldenet left 12:10 Voldenet joined, Voldenet left, Voldenet joined
[Coke] greppable6: qlist 12:10
greppable6 [Coke], Found nothing!
Geth ¦ nqp: coke self-assigned Remove qlist opcode github.com/Raku/nqp/issues/620 12:12
12:15 patrickb joined 12:35 MasterDuke joined
tbrowder just saw an annnouncement about the upcoming gcc 10 will have a built-in static analyzer which caught an exploitable error in openssl 12:44
Geth nqp/rm-qlist: 425e92f7c2 | Coke++ | 2 files
Remove qlist opcode
nqp: coke++ created pull request #621:
Remove qlist opcode
[Coke] jnthn, timotimo: made a PR for nqp to remove qlist
Is there a data structure I can poke at from inside a running nqp to find out what opcodes are available? 12:48
[Tux] Rakudo version 2020.02.1-345-gfa3c0dba9 - MoarVM version 2020.02.1-107-gab089e947
csv-ip5xs0.813 - 0.862
csv-ip5xs-208.112 - 8.997
csv-parser25.047 - 25.836
csv-test-xs-200.383 - 0.383
test7.542 - 7.880
test-t1.902 - 2.010
test-t --race0.842 - 0.954
test-t-2031.881 - 32.170
test-t-20 --race9.688 - 9.809
[Coke] timotimo, jnthn: do you see any other deprecated opcodes in gist.github.com/coke/325a98074f317...7266df7da4 ? 12:59
(sorry, the list is 270 long.) 13:00
anyone have an opinion on github.com/Raku/nqp/issues/613? Guessing it's mostly "not worth it" 13:02
It's an easy enough set of PRs for nqp/rakudo if we decide to change them.
nqp: class EEK {method Numeric() {return 3}}; my $a:= 13:05
camelia Malformed binding at line 2, near ""
at gen/moar/stage2/NQPHLL.nqp:1029 (/home/camelia/rakudo-m-inst-1/share/nqp/lib/NQPHLL.moarvm:panic)
from <unknown>:1 (/home/camelia/rakudo-m-inst-1/share/nqp/lib/nqp.moarvm:initializer)
from gen/moar/stage2/NQP.…
[Coke] nqp: class EEK {method Numeric() {return 3}}; my $a:= EEK.new(); say(+$a);
camelia Cannot intify this object of type P6opaque (EEK)
at <tmp>:1 (<ephemeral file>:<mainline>)
from gen/moar/stage2/NQPHLL.nqp:1916 (/home/camelia/rakudo-m-inst-1/share/nqp/lib/NQPHLL.moarvm:eval)
from gen/moar/stage2/NQPHLL.nqp:2121 (/home/camelia/raku…
MasterDuke nqp: class EEK {method Numeric() {return 3}}; my $a:= EEK.new(); say(nqp::numify($a)); 13:10
camelia Cannot numify this object of type P6opaque (EEK)
at <tmp>:1 (<ephemeral file>:<mainline>)
from gen/moar/stage2/NQPHLL.nqp:1916 (/home/camelia/rakudo-m-inst-1/share/nqp/lib/NQPHLL.moarvm:eval)
from gen/moar/stage2/NQPHLL.nqp:2121 (/home/camelia/raku…
13:19 Altai-man_ joined 13:22 sena_kun left 13:23 Kaiepi left
Geth nqp: 425e92f7c2 | Coke++ | 2 files
Remove qlist opcode
nqp: 933d7bd673 | (Jonathan Worthington)++ (committed using GitHub Web editor) | 2 files
Merge pull request #621 from Raku/rm-qlist

Remove qlist opcode
[Coke] jnthn++
down to 269 undocumented opcodes. :)
nine [Coke]: great work!
[Coke] nine: just nibbling around the edges at the easy stuff, but thank you. :) 13:36
nine [Coke]: doesn't matter if stuff is easy. If no one's doing it, it's still remains an open task 13:37
timotimo i believe we can start off just making desugars in the qast compiler for underscore-less forms
MasterDuke [Coke]: as someone who permanently has github.com/Raku/nqp/blob/master/do...s.markdown open in a browser tab, it is appreciated
timotimo and whenever we make the next rebootstrap, switch them over properly
nine Frankly, I'd rather not rename existing nqp ops. What user would be helped by this? I know NQP comes with no guarantee and user's have to expect stuff to break. Still I think we shouldn't go out of our way to break it unless we have a good reason. 13:39
13:39 Kaiepi joined
nine nqp::force_gc is the only way to even have a chance for writing tests for destructors. I'd be surprised if I was the only one to use it 13:40
MasterDuke greppable6: nqp::force_gc 13:41
greppable6 MasterDuke, 4 lines, 2 modules: gist.github.com/16710d1c80fca345e1...cbbb46ec35
MasterDuke well, you're not *quite* the only one....
timotimo Telemetry could offer force-gc functionality so you don't have to nqp it 13:49
jnthn Or as a method on $*VM even 13:56
lizmat jnthn: or VM ? 14:02
jnthn Well 14:03
m: say $*VM.WHAT
camelia (VM)
jnthn So implicitly, yes :)
timotimo should it throw a method not found on other platforms or just no-op? 14:05
jnthn No-op or warning is maybe kinder 14:08
Though I know we can impl it for JVM 14:09
14:10 cognomin_ joined 14:11 cognominal left
nine $*VM.request-garbage-collection? 14:12
jnthn request is probably a good level of expectation management, yes
On the JVM it seems to mean "well, it'd be nice if you did a GC now, if you feel like it" 14:13
And I guess the answer under a concurrent scheme might be "well, I am all the time anyway, so..." :)
lizmat commit coming up 14:20
Geth rakudo: e223be5669 | (Elizabeth Mattijsen)++ | src/core.c/VM.pm6
Implement VM.request-garbage-collection

For the MoarVM backend only for now
patrickb I just tried building rakudo with the jvm backend on Windows. After some buildsystem fixups the build dies at the very end in the "Preparing installation" phase with an StringIndexOutOfBoundsException: fpaste.scsys.co.uk/588696 14:40
The line pointed at is: "nqp::bindhllsym('Raku', '$!ARGITER', 0);", which looks rather innocent. 14:41
Any ideas what might be wrong?
ping bartolin_ 14:42
Geth nqp: patrickbkr++ created pull request #622:
Fix building JVM on Windows
14:49 NicoleMejia joined 14:51 NicoleMejia left
[Coke] based on nine's comment, I'll reject the opcode rename ticket and instead add a note to the ops docs. 15:12
Geth nqp: 180a93652b | Coke++ | docs/ops.markdown
Note concern raised in #613
linkable6 NQP#613 [closed]: github.com/Raku/nqp/issues/613 force_gc opcode name inconsistent
15:20 sena_kun joined 15:21 Altai-man_ left
lizmat patrickb: if you replace "Raku" with "perl6" , does that help ? 15:28
patrickb lizmat: small letter "p"? 15:29
lizmat yeah, I think so...
b787095358bcbac4c6326 15:30
linkable6 (2020-01-24) github.com/rakudo/rakudo/commit/b787095358 Change the internal HLL language to "Raku"
Geth nqp: 7cdeb0550c | (Patrick Böker)++ | 3 files
Fix building JVM on Windows
nqp: eddead51ae | (Patrick Böker)++ (committed using GitHub Web editor) | 3 files
Merge pull request #622 from patrickbkr/jvm-win-fix

Fix building JVM on Windows
16:17 Kaiepi left 16:18 Kaiepi joined
patrickb lizmat: Changing "Raku" to "perl6" made no difference. 16:36
16:36 patrickb left 16:53 cognominal joined 16:57 cognomin_ left 17:19 Altai-man_ joined 17:21 sena_kun left
gfldex raku: my $s = Set(1, 2, 3, 4); for $s { .say }; 17:26
evalable6 Set(1 2 3 4)
gfldex is that intentional?
17:34 MasterDuke left 17:35 patrickb joined
AlexDaniel gfldex: I think so, yeah 17:40
c: all my $s = Set(1, 2, 3, 4); for $s { .say };
committable6 AlexDaniel, gist.github.com/87a1bfb6562d1ddb49...8d0d76235d 17:41
nine it is
It's because you put the set into a Scalar container. 18:19
m: my $s := Set(1, 2, 3, 4); for $s { .say };
camelia 4 => True
2 => True
3 => True
1 => True
lizmat patrickb: then I have no idea 18:21
18:34 cognomin_ joined 18:38 cognominal left 18:52 lichtkind_ left
Geth nqp: patrickbkr++ created pull request #623:
Enable running precompiled scripts
patrickb I'd like to have an opinion on the above PR. It's a really small change (measured in LOC). Any reason this shouldn't be merged? Should pamplemousse be contacted first or is it OK to just go forward? 19:05
19:07 MasterDuke joined
patrickb nine, MasterDuke: colabti.org/irclogger/irclogger_lo...04-24#l172 19:07
nine Is that really all that's missing? 19:13
patrickb nine: I've skimmed though the respective blog posts by pamplemousse and did some small tests that look promising. 19:17
19:20 sena_kun joined 19:22 Altai-man_ left 19:24 lucasb joined
Geth rakudo: feb036719e | (Patrick Böker)++ | tools/create-release-announcement.raku
Fix link to non-existent rakudo.org/how-to-help page

That page has been missing for a very long time. So just pick a better page to link to.
patrickb .tell rba Can you add a redirect to rakudo.org forwarding 'rakudo.org/how-to-help' to rakudo.org/community ? 19:45
tellable6 patrickb, I'll pass your message to rba
rba patrickb: You mean in the nginx config? 19:46
patrickb rba: Yes.
rba patrickb: Sure. Give me a sec.
patrickb rba: The nginx.config file in the rakudo.org repo is not actually used, correct? 19:49
Should we just remove that file?
If not, is github.com/perl6/rakudo.org/pull/39/files a faithful representation of the change you are doing on the server? 19:54
19:58 finsternis joined 19:59 maggotbrain left 20:01 maggotbrain joined 20:05 cognominal joined 20:09 cognomin_ left
rba patrickb: I’ve added your change to the nginx config manually, as there is no automatic update from the repo. Should work now. 20:19
patrickb rba: Thanks! 20:26
rba: What's your opinion about the nginx.config file in the rakudo.org repo? Try to keep synced to server or just remove? 20:27
rba please let it there. will merge your pr later. 20:28
Idea is to dockerize rakudo.org too and then nginx might be in play again...
patrickb ok 20:32
nine: You were right. The --bytecode PR does not work with modules. 20:36
21:18 patrickb left 21:19 Altai-man_ joined 21:22 sena_kun left 21:37 Xliff joined
Xliff \o 21:37
If I need to write a native call sub to take a STRUCT (not a struct pointer) how would I do that? 21:38
Or can I just use the struct name and nativecall will figure it out on its own?
nine I don't think you can
Xliff :(
Why is that?
That's basic C and I'm faced with such routine right now.
It's funny. The more I use NativeCall, the more I find that it really isn't as ready for production as the docs would lead you to believe. 21:39
nine: Might this be an FFI limitation? 21:41
nine Well it's producing just fine for us
Xliff jnthn: ^^
nine: That is not a valid answer. 21:42
NativeCall is missing a basic portion of C functionality. It may be working for you, but it is not working for me.
When I'm trying to bring new users to the language, such a response is not acceptable.
I will see if this is noted in MoarVM and make a bug if it is not. 21:43
nine Well, you're asking on a channel that's about _developing_ Raku questions about using Raku and go on to call stuff that's used in production quite successfully "not as ready for production as the docs would lead you to believe". 21:44
Xliff nine: Is not the passing of structs a basic function of C? Is not NativeCall billed as a way to allow Raku code to interoperate with C? If so then how am I mistaken in my assertion? 21:46
nine Since you feel so strongly about this functionality, I kindly ask you to participate in the solution. A great first step would be to make a suggestion on how we could differentiate the pass by pointer and pass by value situations 21:47
Xliff The fact that is working for certain situations in production is to its credit, but it does not support a basic function of C.
That, in my opionion is a serious oversite that is not mentioned in the docs.,
nine As to this being a basic function of C, I've somehow managed to get by for 20 years of coding in C on and off without noticing that this was actually possible. I've yet to come across a code base where this is used and I really only know of it because of NativeCall questions 21:48
Xliff nine: I am not in the business of designing solutions for you. If I have one, I will pass it on. But for you to suggest otherwise is pure laziness. For the record, I have already bugged this for MoarVM. And as a user that is the extent of my responsibility. The rest is up for you as a develoment community to either support or not.
nine MoarVM was the first place where I've seen structs returned from a function and where I've learned that the calling convention is very platform dependent. 21:49
Xliff If the answer is not, then the users of your work deserve an explanation.
nine: Again, citing your experience is not an answer!
It's in the C spec!
nine Solutions for me? But I don't need that functionality. So how is this a solution for me?
Xliff Your experience on this particular issue is irrelevant! 21:50
nine: You. Are. A. Core. Developer.
I shouldn't need to explain that any further.
21:50 scovit joined
Xliff So again. I have left a bug on MoarVM. Based on that response, I will also add a bug on nqp and rakudo. 21:51
nine Apparently I am. I am so exactly because I want to use Raku and there are issues in the way and fields where I want it to do better.
scovit Xliff, I once hacked up a solution for the same issue github.com/rakudo/rakudo/pull/2648 , you might want to check that out
Xliff scovit: Oh! Thank you. I will look at that now.
scovit hold on
it implied modify MoarVM quite a bit 21:52
nine scovit: oh, `is copy`, that's clever
scovit nine : in MoarVM only the FFI backend could get this functionality
nine Xliff: me being a core developer means just that I'm a Raku user who acknowledges that he doesn't deserve anything and has no rights other than to work on the things he needs to be worked on 21:53
scovit also I was not an expert of MoarVM I might have introduced bugs 21:54
just here to say that it is possible to get this done
Xliff nine: As a "Core Developer" you have a responsibility to do better by your user than throwing their question away based on "what works for you".
If you don't know then ask them to submit a bug. Don't start an argument.
scovit: And that's all I need, for now. 21:55
Especially as the code I am working on will have test code, shortly.
And I can probably assist in fixing those bugs, if necessary.
scovit: For the record, this is for Raku bindings for GStreamer.
gstreamer.freedesktop.org/ 21:56
And I would assume that the Raku community would not mind the work I am putting in of my own volition.
40,000 lines of code and I am looking at trashing at least much of the video functionality because of this missing feature.
scovit what is the function? there might be a fast hack to work around this issue :) 21:58
Xliff Sure thing. One sec.
nine I have not thrown away any question.
Xliff gstreamer.freedesktop.org/document...enter_rect
nine: Yes. You. Did.
And if you still think strongly that you did not, we will have to agree to disagree. 21:59
nine You repeating that assertion does not make it true.
I solely commented on your comment about readiness for production or the lack thereof 22:00
Xliff If you aren't supporting something that's in the C spec and this lack of support is not documented, then you are DAMN RIGHT that you have a problem. 22:01
To say so otherwise is outright lying.
nine No, you have that problem, not me. You seem to think that just because I give 100s of hours of my time foe rakudo that somehow puts some responsibilities on me. 22:02
[Coke] I think folks might want to step away from the keyboard for a bit, take a breath.
nine It does not. I'm scratching my own itch, nothing moee. 22:03
I certainly would appreciate a more civil tone. 22:04
[Coke] Xliff: If the issue is bugged, that's great. tests, prs, even better. I understand this issue is important to you. If you're really interested in getting it fixed for production work you're interested in, as the Grants Committe Secretary at TPF, I always suggest contributing to the development fund as a concrete way to contribute. 22:05
As a member of the Raku community, please go easy on the volunteers that are contributing a great deal of time and effort to corners of the products that not in your particular area of interest. 22:06
Xliff [Coke]: I am trying very hard to be civil. 22:07
nine Xliff: I very much appreciate the work you do on all these libraries and am glad to help wherever I can. I am however a volunteer, same as you.
Xliff nine: Then please don't use the "Works for me!" issue as justification for why a certain feature does not exist.
And why it's not likely to.
nine I did not.
I said its used in production successfully. 22:08
Xliff Again. We will have to agree to disagree. Because that is certainly the take I got from your comments.
nine so it kinda seems to be ready for production
[Coke] Having been using some form of chat since the 80s, I'd like to remind folks that you're always likely to read in emotion that isn't there in someone's sends. 22:09
Xliff nine: If you can show me where this issue is documented as not implemented, then I will happily retract. Until then, I will not rest to point out where offhanded comments go a long way to turn off volunteers like yourself who do NOT have the ability to contribute to core.
[Coke] Xliff: If theres a doc issue, that's not on nine.
Xliff [Coke]: No. It's on the documentation. 22:10
Again, I am more turned off by the offhanded comments, than anything else.
[Coke] Yes. And that also is a volunteer effort. Please, if you have time or inclination, since this is an important subject for you, provide a PR for Raku/docs
Xliff Again ** 2, for a core developer to say "Works for me" when a feature that is in the C spec is not implemented is a huge turn off.
[Coke]: When I have the time and inclination to search the entire document repository for a mention and correct the issue, I will do so. 22:11
[Coke] Xliff: Great. Thank you for your time and effort, in advance.
docs.raku.org/language/nativecall#...nativecall is probably a good place to start. 22:12
I reiterate my suggestion for folks to step away and/or consider that tone on IRC is hard to read. I myself and going to step away for some family time. Cheers. 22:13
Xliff [Coke]: Thank you for your suggestion. As I've said, when I have the time and inclination, I will submit a PR for the docs. 22:17
scovit: Were you able top look at the function I linked? 22:22
scovit Xliff yes, I'm checking that out 22:25
Xliff scovit++: Thanks so much! 22:28
jnthn Xliff: There's a difference between "in the C spec" and "in the ABI so you can reliably do it" and "actually provided by the FFI library we use". 22:42
nine Xliff: have you considered that you claiming that the docs are overselling NativeCall's production worthiness may be a huge turn off, too?
jnthn Xliff: I don't know about the second of those (I'm pretty sure returning structs is a little dubious in that sense); last I checked, the dyncall library doesn't support it.
Xliff jnthn: I'm not looking to return structs. Just pass them by value. 22:43
nine: In this case, I happen to have a point.
nine: I was trying to back off on the subject. Why are you reopening it? 22:44
nine Yeah, apparently I do everythjng wrong. 22:46
jnthn Xliff: I didn't say you were; I was pointing it out as an example of something that you can do in C but is potentially compiler-defined behavior, and so not so easy as one would hope to support. 22:47
Xliff: Also, given from what I can see, dyncall doesn't look to support this (I'd really love to find otherwise, since it'd be easier to implement then...)
Xliff: Which means that there's a whole lot of NativeCall-alikes for other languages that have precisely the same limitations. 22:48
www.dyncall.org/docs/manual/manualse8.html fwiw 22:50
Xliff jnthn: Ack! 22:52
jnthn: Alright. I will have to scour the docs to make sure this limitation is documented.
Thank you for looking in to it.
In the meantime, what would you suggest as a workaround?
Currently, I have that method NYI'd
scovit Xliff 22:53
Xliff Yes?
scovit It might be very system/compiler dependent (not sure), can you check this script give sensible output on your workstation? gist.github.com/scovit/5b87d9a17ed...2702528097 22:54
Xliff Sure! One sec.
scovit something like the following
GstVideoRectangle $src = GstVideoRectangle.new(x => 50, y => 51, w => 53, h => 54)GstVideoRectangle $dst = GstVideoRectangle.new(x => 70, y => 71, w => 73, h => 74)Pointer $result = NativeCall::Types::Pointer.new(140550574225616)Int $scaling = 1src: x: 51 y: 50 w: 54 h: 53dst: x: 71 y: 70 w: 74 h: 73result: 140550574225616scaling: 1
Xliff Yup! That works. 22:55
Oh... HAH! 22:56
scovit: Wow! I should have thought of that. 22:57
I'll have to abstract that out, but that works.
jnthn omg... :)
scovit glad to hear about that :) 22:58
jnthn libffi apparently can do this... github.com/libffi/libffi/blob/mast....texi#L406
scovit: Does it break down once the struct gets beyond a certain size, ooc, or does it end up working 'cus the ABI has structs split over registers and stack?
Xliff scovit++: Well, it's a strategy to use. I hope I only need to deploy it rarely. 22:59
scovit jnthn: I am not sure it will break down, I do not think so 23:03
normally ABI architect are very smart
I would be more worried about endianness
another thing that is missing from the ffi libraries are support for variadic functions. But normally everything work as you would expect if you were just pushing values on a stack 23:05
the use of registers is almost transparent 23:06
jnthn Smart maybe, but there's multiple ABIs even for one hardware architecture. 23:08
(ABI for Windows and POSIX on x64 differs.)
scovit Yes, and not for really good reasons 23:10
Xliff varadic functions do have a workaround in NativeCall. 23:11
nine: So to ammend my comments. NativeCall can be production ready if the right conditions are met. 23:12
nine: Fortunately, both MoarVM AND Rakudo can be bent to the task.
My task is now to look at where to appropriately place this workaround in the documentation. Hopefully that will be this weekend.
nine Xliff: so we are on the same page after all
Xliff nine: with reservations, yes.
Still, the whole struct-by-value limitation needs to be documented. I hope to get to that this weekend. 23:13
jnthn Xliff: Also, I think there's a degree to which libraries that are likely to be bound by different languages via. FFI mechanisms are written so as to avoid things that FFIs often miss handling.
Xliff And I still need to revisit the whole .precomp thing.
I have too much on my plate, right now.
scovit jnthn but that is such a real pity 23:14
Xliff jnthn: You have a point. It is a lot to take in.
jnthn Xliff: And ultimately, there's always the escape-hatch of "write a little bit of C that maps the awkward things"
Xliff I am trying to focus on my work, ATM.
scovit it hurts eyes to see a Point function passed by a pointer
Xliff I am 16k out from 400k of Raku in 2 years.
scovit it hurts the eyes
Xliff I have Just overt 6 weeks in which to do it.
Actually.,.. that's seven weeks. 23:16
jnthn scovit: It's a fair point. I wasn't so much saying the state of things is ideal, just nothing why running into the limitation isn't as common as it might be otherwise. 23:17
Xliff jnthn: BTW - How are things with LessVM going? 23:18
I think after June 15th, I will take an extended break from all things p6-GLib and work on something else. Macros would be a nice project to assist on...assuming I can get my nqp in order.
Yes. I have a vested interest in using it. It would fit all things p6-GLib, nicely. 23:19
jnthn "less is moar"... :-) Well, RakuAST is the current focus. There's a rakuast branch in the Raku repository. I push the odd thing there. And probably a lot more soon, 'cus this was the last busy week on a $dayjob project.
23:20 sena_kun joined
jnthn There's an "if you want to help" section in github.com/rakudo/rakudo/blob/raku.../README.md 23:20
Xliff jnthn: Cool!
How would you think that would fit github.com/Raku/problem-solving/issues/98
Considering the focus is on AST. Could you build a Parameter object that would serve this purpose? 23:21
23:21 Altai-man_ left
Xliff Or would the whole NativeCall def need to be built? And is that currently possible with the planned features of RakuAST? 23:22
jnthn I think you'd probably be able to write a macro that takes a constant initialized with a signature literal and a point block or sub, and have it replace the block/sub's signature 23:24
Or if you prefer do it in a trait 23:25
is sig(the-sig-constant) or some such
timotimo that's sig, bro 23:28
jnthn I had to look up how "sig" can be used as slang :P 23:30
timotimo can it? i was punning off of "sick" 23:31
jnthn Oh!
Well, "sick" is a fun one too, 'cus its meaning has inverted in my lifetime. :) 23:32
When I was at school it meant "disgusting", and apparently now it means "awesome" :P
Though maybe either could be applied to what I described, so... :) 23:33
23:38 scovit left
Xliff jnthn: Hah! I got that the first time around. :) 23:40
timotimo++ -- I lurfed.
jnthn: Although I'd be more interested in a macro that would just be a reusable signature to be used in a NC definition. 23:41
jnthn Also possible, I'd think 23:42
Since it's a less abstract version of what I described
Xliff jnthn: OK. When RakuAST is ready, I'd like to discuss this further! :) 23:45
jnthn Yes, I'll be interested in problems folks actually want to solve, beyond my own pile of them. :) 23:48
Xliff :)
jnthn In the meantime, I should get some rest. 'night o/
Xliff Well, I think RakuAST is a really neat idea, so I'd like to contribute where I can.
23:48 Xliff left