🦋 Welcome to Raku! raku.org/ | evalbot usage: 'p6: say 3;' or /msg camelia p6: ... | irclog: colabti.org/irclogger/irclogger_log/raku
Set by ChanServ on 14 October 2019.
00:00 peteretep joined 00:01 spycrab0 joined, kawaii joined 00:08 samebchase- left 00:09 kawaii left, brass left, samebchase- joined, skaji_ left, brass_ joined, skaji_ joined 00:12 kawaii joined 00:13 peteretep left 00:14 spycrab0 left 00:16 spycrab0 joined 00:18 cpan-raku left 00:19 cpan-raku joined, cpan-raku left, cpan-raku joined, peteretep joined 00:26 aborazmeh left 00:51 Altai-man joined 00:52 dogbert17 joined 00:54 sena_kun left 00:55 dogbert11 left 00:57 dogbert11 joined 01:00 dogbert17 left, dogbert12 joined 01:01 dogbert11 left 01:18 dogbert17 joined 01:19 dogbert12 left 01:22 dominix joined
dominix I have a Windows10 m&chine to setup for some dev. 01:25
I am surprised to see rakudostar to be only at 2019.01 version
how do you folks install a current raku on a windows 10 machine ? 01:26
other question do you know about a team that work on raku for windows10 ? 01:27
01:31 wintre37 joined
[Coke] there's a more recent installer for windows than 2019.01. one sec 01:38
I build from source on win10, but I am probably weird. 01:39
rakudo.org/downloads - has 2020.07
01:40 wintre37 left, molaf left
dominix I will ask for a package update on chocolatey 01:52
01:53 molaf joined 02:05 Manifest0 left, Manifest0 joined 02:18 stoned75 left 02:19 stoned75 joined 02:28 brass_ left, brass joined 02:31 Cabanossi left 02:32 Cabanossi joined 02:55 dogbert17 left, dogbert17 joined 02:57 dominix left 03:20 hungryd40 joined 03:21 hungrydonkey left 03:28 squashable6 left, squashable6 joined 03:31 xinming_ left 03:34 xinming_ joined 03:45 Mawile left 03:47 gnufr33dom joined 03:48 hungryd40 left 03:49 hungrydonkey joined 04:03 dogbert11 joined 04:05 dogbert17 left 04:08 Kaiepi left, hungrydonkey left, Kaiepi joined 04:10 hungrydonkey joined 04:13 dogbert17 joined 04:17 dogbert11 left 04:21 xinming_ left 04:22 xinming_ joined 04:33 dogbert11 joined 04:34 dogbert11 left, dogbert11 joined 04:37 dogbert17 left 04:41 xinming_ left 04:42 xinming_ joined 04:52 sena_kun joined 04:53 Altai-man left 05:00 dogbert17 joined, ponbiki joined 05:02 dogbert11 left 05:04 ponbiki left 05:08 dogbert17 left, dogbert17 joined 05:15 dogbert11 joined 05:17 dogbert11 left 05:19 dogbert11 joined, dogbert17 left 05:20 dogbert11 left, dogbert11 joined 05:21 xinming_ left 05:22 xinming_ joined 05:24 dogbert17 joined 05:26 dogbert11 left 05:40 bocaneri joined 05:41 camelCaser joined 05:43 rindolf joined, dogbert11 joined 05:47 dogbert11 left, dogbert17 left 05:48 dogbert11 joined 06:02 dogbert11 left, dogbert11 joined 06:08 dogbert17 joined 06:10 dogbert11 left 06:12 dogbert11 joined 06:13 dogbert11 left, dogbert11 joined 06:15 dogbert12 joined, dogbert17 left 06:20 dogbert11 left 06:25 dogbert12 left 06:26 dogbert12 joined 06:38 dogbert17 joined 06:40 guifa left 06:41 dogbert12 left 06:43 AlexDaniel joined, HarmtH joined, Celelibi joined, gfldex joined, agentzh joined, telex joined, veesh joined, m_athias joined, Util joined, erdic joined, stux|RC-only joined 06:44 dogbert11 joined 06:45 dogbert11 left, Sgeo left 06:46 dogbert11 joined 06:49 dogbert17 left 06:51 dogbert11 left, dogbert11 joined 06:55 leont joined 06:58 Archenoth joined 07:05 stoned75 left
El_Che [Coke]: do you follow specific instructions to build on Windows? 07:12
07:13 dogbert17 joined 07:15 dogbert11 left, dogbert12 joined 07:18 dogbert11 joined, dogbert17 left 07:19 leont left 07:20 hungryd67 joined 07:21 dogbert12 left 07:22 hungrydonkey left, dogbert17 joined 07:25 dogbert11 left 07:30 wamba joined 07:32 dogbert11 joined 07:35 dogbert17 left
Geth advent: 9dae2750a8 | (JJ Merelo)++ | raku-advent-2020/authors.md
Revising advent RFC
07:35
timotimo that's the difference between rakudo star and rakudo isn't it? 07:46
oh, that conversation was a few hours ago
07:49 molaf left 07:50 bocaneri left, Kaiepi left, samebchase- left 07:53 bocaneri joined 08:00 dogbert17 joined 08:03 dogbert11 left, dogbert11 joined 08:04 redhands joined 08:06 dogbert17 left 08:22 Kaiepi joined, samebchase- joined, raydiak joined, [ptc] joined, Bucciarati joined, KotH joined 08:27 hungryd67 left 08:33 BHANU joined 08:36 BHANU left 08:45 ChoppedBacon joined 08:47 stoned75 joined 08:50 gnufr33dom left 08:51 Altai-man joined 08:52 finanalyst joined 08:53 sena_kun left 09:17 Kaiepi left 09:19 dogbert17 joined 09:20 Kaiepi joined 09:23 dogbert11 left 09:26 Kaiepi left, Kaeipi joined 09:34 rindolf left
Geth doc/typo: 438575b385 | (Stoned Elipot)++ | doc/Type/List.pod6
typo
09:45
doc: stoned++ created pull request #3573:
typo
doc: 5826022f24 | stoned++ (committed using GitHub Web editor) | doc/Type/List.pod6
typo (#3573)
09:49
linkable6 Link: docs.raku.org/type/List
DOC#3573 [closed]: github.com/Raku/doc/pull/3573 typo
09:51 dogbert11 joined 09:55 dogbert17 left 09:58 Black_Ribbon left 10:10 Cabanossi left 10:24 Cabanossi joined 10:36 dogbert17 joined 10:41 dogbert11 left 10:43 Kaeipi left 10:44 Kaiepi joined 10:49 Kaiepi left 10:51 molaf joined 10:56 redhands left 11:08 hungrydonkey joined 11:09 dogbert11 joined 11:12 dogbert17 left
Geth doc/ref-ops: f904d64a48 | (Stoned Elipot)++ | doc/Type/List.pod6
xref operators
11:21
doc: stoned++ created pull request #3574:
xref operators
11:25 dogbert11 left 11:26 dogbert11 joined 11:27 rindolf joined
codesections m: class A { method f() {} }; say A.^lookup('f').signature 11:32
camelia (A: *%_)
codesections When you create a method, it seems to implicitly get a signature of (Self: *%_). I understand why it gets `Self:` – that's what makes it a method. But why does it get *%_ (allowing it to be called with arbitrary named arguments)?
lizmat codesections: that's been a source of debate and a source of frustration for yours truly 11:33
there's a nice name for this behaviour, which I forget right now, (traumatic?) 11:34
but it's to allow nextsame and friends to have access to *any* named parameters that have been passed, to be passed along without any additional boilerplate 11:35
stoned75 "automatic signature" ?
hum no 11:37
lizmat no, that'd be more for something like: 11:41
m: dd ({ $^a + $^b }).signature
camelia :($a, $b)
Geth doc/list-eg-output: 6bf9432527 | (Stoned Elipot)++ | doc/Type/List.pod6
Fix examples' output

if there is a ␤, don't break the line :)
11:49
doc: stoned++ created pull request #3575:
Fix examples' output
11:50
codesections lizmat: Thanks. So, related set of questions: Is that feature documented anywhere? If not, should it be? And why do *none* of the method signatures in the docs show that parameter?
Are all they docs wrong? :D (Or, rather, would it be worth adding that to all method docs?) 11:51
11:52 dogbert17 joined 11:55 dogbert11 left
codesections «Methods will ignore extra named arguments where other types of Routine will throw at runtime. Extra arguments will be forwarded by nextsame and friends.» 192.168.0.235:3000/type/Method 11:55
er, docs.raku.org/type/Method 11:56
So, I guess it's _kind of_ documented
lizmat I guess the term I was looking for was "interface consistency"
yeah, but I would definitely *not* go documenting that for every method :-) 11:57
stoned75 like roast/S12-class/interface-consistency.t ?
and similar
lizmat yup :-) 11:58
a lemma about "interface consistency" would be nice
codesections lizmat: Makes sense. I'll open a docs issue. I guess that would live in the /type/Method docs? 12:02
12:05 zacts joined 12:14 dogbert11 joined
Geth doc: codesections self-assigned expand documentation of implicit *%_ parameter/interface consistency for Methods github.com/Raku/doc/issues/3576
6bf9432527 | (Stoned Elipot)++ | doc/Type/List.pod6

if there is a ␤, don't break the line :)
12:16
12:17 dogbert17 left 12:19 dogbert17 joined 12:23 dogbert11 left
doc: 1a607959db | (Juan Julián Merelo Guervós)++ (committed using GitHub Web editor) | doc/Type/List.pod6
Merge pull request #3575 from Raku/list-eg-output

Fix examples' output
linkable6 Link: docs.raku.org/type/List
12:34 zacts left
kawaii Can the values stored in a hash be type restrained? i.e. `my Rat %results = ...`, so that it can only store whatever you want to cast? 12:40
tellable6 hey kawaii, you have a message: gist.github.com/b395d1ea6ec3cd6580...680f50aa95
12:41 andrzejku joined
tobs kawaii: I don't understand what you mean by "cast". What exactly is wrong with `my Rat %results`? 12:44
stoned75 kawaii: yes they can. see docs.raku.org/language/hashmap#Con...alue_types 12:45
kawaii tobs: I didn't test that what I wrote is actually valid, but am I right in thinking it would give me a hash capable of storing only Rat's? :)
stoned75: thanks! :D
Geth doc: f904d64a48 | (Stoned Elipot)++ | doc/Type/List.pod6
xref operators
12:48
doc: 53588f5199 | stoned++ (committed using GitHub Web editor) | doc/Type/List.pod6
Merge pull request #3574 from Raku/ref-ops

xref operators
linkable6 Link: docs.raku.org/type/List
12:52 sena_kun joined 12:53 Altai-man left 12:57 sena_kun left, sena_kun joined 13:04 skids joined 13:15 aborazmeh joined, aborazmeh left, aborazmeh joined, hungrydonkey left 13:17 hungrydonkey joined, wamba left 13:18 hungryd86 joined
[Coke] El_Che: to build on windows? I use the x64 native tools command prompt, have perl 5 installed (can dig up which version if you need), git, and I think it just works. 13:20
looks like strawberry perl 5 13:21
13:21 hungrydonkey left
[Coke] and I have visual studio 2017, but didn't do any particular config that I recall to make it work. 13:21
13:22 cpan-raku left
Geth ¦ doc: codesections self-assigned List.pick with Callable and .pick/pickpairs/grab/grabpairs for Baggy and Setty with Callable are not documented github.com/Raku/doc/issues/3564 13:22
13:23 cpan-raku joined, cpan-raku left, cpan-raku joined 13:24 wamba joined 13:26 wamba left
xinming_ m: multi a (:$a where "a") { "a".say; }; multi a (:$b where "b") { "b".say; }; a(:a<a>); 13:33
camelia a
xinming_ Is there a shorter version of :$a where "a" in multi signatures?
Something like, multi a (:$a<a>) { }; 13:34
Something like, multi a (:$a where <a>) { };
tadzik m: multi foo(:a("a")) { say "OK" }; foo("a")
camelia 5===SORRY!5=== Error while compiling <tmp>
Missing block
at <tmp>:1
------> 3multi foo(:a(7⏏5"a")) { say "OK" }; foo("a")
tadzik that'
...not the errorr I thought I'd see, hmm
oh pfft 13:35
m: multi foo(:$a("a")) { say "OK" }; foo(a => "a")
camelia 5===SORRY!5=== Error while compiling <tmp>
Shape declaration with () is reserved;
please use whitespace if you meant a subsignature for unpacking,
or use the :() form if you meant to add signature info to the function's type
at <tmp>:1…
13:36 gnufr33dom joined 13:38 wamba joined 13:42 wamba left 13:43 wamba joined 13:45 dogbert11 joined 13:48 dogbert17 left 13:56 dogbert17 joined, dogbert17 left 13:57 dogbert17 joined 14:00 dogbert11 left 14:02 aborazmeh left 14:15 wamba left
jdv79 but why shoudln't the actual signature of all methods be documented that way? 14:21
14:21 andrzejku left
jdv79 (including the *%_ i mean 14:21
codesections jdv79: it *would* just be a simple s/// to do so :) 14:22
14:23 andrzejku joined
jdv79 yeah. i'm just curious why some think omitting it from docs is a better idea. 14:23
lizmat jdv79: should we document that each method automatically has a `self` defined that contains a deconted version of the invocant ? 14:25
with each method documentation?
or should we do that somewhere more general? 14:26
docs.raku.org/language/objects#index-entry-self # current state
14:27 ponbiki joined, ponbiki left
codesections Well, the docs already have signatures like «multi method lc(Str:D: --> Str:D)». That `Str:D:` shows what the invocant is (but I maybe I'm misunderstanding what you mean, lizmat?) 14:28
lizmat sub a { self }
m: sub a { self }
camelia 5===SORRY!5=== Error while compiling <tmp>
'self' used where no object is available
at <tmp>:1
------> 3sub a {7⏏5 self }
expecting any of:
term
lizmat m: method a { self }
camelia Potential difficulties:
Useless declaration of a has-scoped method in mainline (did you mean 'my method a'?)
at <tmp>:1
------> 3method7⏏5 a { self }
lizmat so, "self" is defined inside *each* method 14:29
do we document that with each method? I think not, nor should they be
in my opinion, the same goes for %_: it is added to the signature of *each* method unless it has an explicit slurpy hash 14:30
codesections Oh, yeah, I agree *that* doesn't need to be called out on each method. It's something that's generally true of all methods. And, analogously, I think we don't need to say anything about the method taking *%_ in the _discription_ of the method
I guess I'm less sure of that when it comes to the _signature_ 14:31
jdv79 what does this have to do with self? that's not in the sig. the sig is what is featured on every method doc.
[Coke] things that are common like that (self) should be in overall OO documentation, we could have 'method' definition link to >something<. shouldn't be part of each individual method, though
codesections imo, there's an argument that the documented signature should be the same as the .signature Signature
lizmat jdv79: ok, I got the impression you wanted it documented with each method, but this is just about being mentioned in the sig 14:32
I'm fine with just showing it in the sig
14:32 wamba joined
jdv79 well, the use case i'm thinking of is a new user looking at hte docs for a method and seeing the sig and thinking that's *exactly* what it is 14:32
lizmat sorry for the misunderstanding :-)
jdv79: gotcha 14:33
[Coke] we could add an xt/ to docs saying that if it's a method declaration, it has to have a parameter that ends with a :
xkcd.com/2347/ reminds of a few projects. :| 14:35
jdv79 oof 14:36
lizmat Ahhh... ImageMagick, my old friend :-)
14:37 wamba left
[Coke] sort of project that you'd hope some company would support the person/team and have a plan to provide backup. 14:38
14:38 wamba joined 14:43 dogbert11 joined 14:46 dogbert17 left 14:52 wamba left 15:00 raku-bridge joined
SmokeMachine m: my method a { self } 15:10
camelia ( no output )
15:10 andrzejku left 15:12 dogbert17 joined 15:15 dogbert12 joined 15:16 dogbert11 left 15:17 dogbert17 left 15:19 dogbert17 joined 15:20 dogbert12 left
codesections I created a docs issue that I'd appreciate people's input on: github.com/Raku/doc/issues/3577 15:29
15:30 patrickb joined
codesections It's about deciding which methods should be documented in the /type/* docs 15:30
(apologies for the length)
15:39 patrickb left 15:41 gnufr33dom left 15:49 dogbert11 joined 15:51 dogbert17 left, patrickb joined 15:54 stoned75 left 15:56 stoned75 joined 16:00 Geth joined 16:08 patrickb left 16:09 patrickb joined 16:22 dogbert17 joined 16:25 dogbert11 left
xinming_ SmokeMachine: termbin.com/zapo <-- Should this example work in Red? :-) 16:28
SmokeMachine: I added my $*RED-FALLBACK = True for that example, doesn't work either 16:30
It's kind of database table inheritance
with Red
termbin.com/f42so <--- This is the updated version which expose the real problem. :-) 16:33
16:39 patrickz joined 16:40 patrickb left
rypervenche Does anyone know what I'm doing wrong with my grammar here? It's able to match a single line, but it doesn't match the <eol> (tried it with Grammar::Debugger, fails to match <eol>): repl.it/@rypervenche/Asterisk-file#main.raku 16:42
I should say, when I try to run it against the whole file, it doesn't match the \n at the end of each line. 16:43
16:43 zacts joined 16:48 dogbert11 joined 16:51 Altai-man joined 16:52 dogbert17 left 16:53 sena_kun left 17:01 dogbert17 joined 17:04 dogbert12 joined 17:05 zxcvz left 17:06 dogbert11 left 17:08 dogbert17 left 17:23 patrickb joined 17:26 patrickz left 17:28 zacts left 17:37 dogbert17 joined 17:41 dogbert12 left 17:58 stoned75 left 18:05 greppable6 joined 18:06 reportable6 joined 18:07 tellable6 joined, shareable6 joined, vrurg_ left 18:08 dogbert11 joined 18:09 vrurg joined 18:10 dogbert17 left 18:11 vrurg left 18:12 vrurg joined 18:16 patrickz joined 18:18 patrickb left
Geth ¦ doc: codesections self-assigned Add *%_ to method signature documentation github.com/Raku/doc/issues/3578 18:19
18:21 patrickz left
codesections Hmm, does Rakudo not put any limits on its own memory consumption? I just triggered a bug that causes an infinite hang, an the REPL is now at > 10 GiB of mem, with no sign of stopping yet. 18:31
(I'll kill the process soon, but I was just interested in seeing it go for a bit) 18:32
moritz no 18:33
you can use something on the outside like ulimit to that effect
codesections Interesting/surprising
and good to know 18:34
(it made it to 15 GiB before I got bored. Pretty big, for one-line) 18:35
18:39 patrickb joined 18:41 zacts joined, dogbert17 joined
codesections m: class A is Setty {}; A.grab 18:42
camelia Cannot resolve caller grab(A:U); Routine does not have any candidates. Is only the proto defined?
in block <unit> at <tmp> line 1
codesections What's the point of having Setty define a `grab` method, but just declare the proto? 18:43
Is it saying «if your class `does` Setty, it *might* want a `grab` method, but I won't insist on it»
18:43 dogbert12 joined
codesections or is there some other point I'm missing? 18:43
18:44 dogbert11 left 18:47 dogbert17 left 18:49 dogbert17 joined 18:51 dogbert12 left 18:56 dogbert11 joined 18:58 andinus joined
rypervenche Gah! I figured out my issue. The space at the end of a rule matters. 19:00
19:00 dogbert17 left, dogbert17 joined 19:01 finanalyst left 19:03 dogbert11 left 19:06 tigerpaws joined 19:08 natrys joined 19:11 dogbert17 left, dogbert17 joined 19:13 finanalyst joined 19:17 dogbert11 joined 19:20 stoned75 joined 19:21 dogbert17 left 19:24 stoned75 left 19:26 leont joined 19:29 Mawile joined 19:32 ChoppedBacon left, Archenoth left, pecastro joined, bocaneri left 19:34 tellable6_ joined, ChoppedBacon joined, tellable6 left, vrurg left 19:35 vrurg joined, skaji__ joined
[Coke] you can put some strictures about what it has to look like with the proto, I think. 19:37
19:38 brass_ joined, reportable6_ joined, brass_ left, brass_ joined, brass_ left
[Coke] m: dd Setty.can("grab")[0] 19:38
camelia Method grab = proto method grab (Setty: |) {*}
[Coke] ... not that that does.
codesections [Coke]: but the source doesn't. I'ts just `proto method grab(|) {*}` 19:39
well, *that* was faster than checking the source code :D
[Coke] there used to be a multi method grab in that file. 19:40
19:40 pecastro left
[Coke] see if you can find out which commit removed the multi itself, maybe there's a clue there. Might have accidentally been left in. 19:40
(ah. the old version did this: 19:41
X::Immutable.new( method => 'grab', typename => self.^name ).throw
codesections the multi got moved to things that `does` (things that do?) Setty
?
[Coke] why bother throwing that error when you now get a very similar one "for free"
19:42 brass_ joined, skaji_ left, reportable6 left, hungryd86 left, brass left, camelCaser left, skaji__ is now known as skaji_, brass_ left, ccamel joined
codesections m: Set.new.grab 19:42
camelia Cannot call 'grab' on an immutable 'Set'
in block <unit> at <tmp> line 1
[Coke] (I think lizmat is the one that knows all about the various container types, I'd defer to her.) 19:43
codesections Oh, it's not really "for free": that exact X::Imut... line is now in Set 19:44
19:46 edk_ joined, bocaneri joined
codesections This is exactly the sort of thing that makes me conflicted about the "document all public methods" approach for that GitHub issue I opened. `grab` is now a public method on `Set` (that does nothing but error) when it used to be a public method on `Setty` 19:48
Is that a change that should be/should have been reflected in the docs?
It *really* feels like a Rakudo implementation detail
but documenting it falls out of the «document all public methods» rule
19:50 dogbert17 joined
moritz rules are meant to be broken sometimes 19:52
19:53 dogbert11 left 19:54 Sgeo joined, dogbert2 joined
codesections moritz: And I'm fine with that :) But I want *these* rules to be tool-enforced, so if they're going to be broken, I want them to be broken in ways I can explain to the tools! 19:54
19:57 dogbert17 left
[Coke] codesections: rakudo isn't the source of truth there. it's a convenient place to pull the method list, but it's not the source of truth (as you noted in your list of options) 20:05
I would treat the work like an exercise ending up with everything either being in roast or tagged as implementation detail. 20:06
codesections But what _is_ the source of truth about _where_ methods are implemented?
Roast is intentionally agnostic about that
[Coke] but I wouldn't expect it to be right as is.
codesections: If roast doesn't care, why do you?
Roast says where you can call them from - that's the part that's spec.
unless there are specific tests, doesn't matter if it's done in a parent class or a role. 20:07
codesections Ok, so how do we decide where to put the documentation?
[Coke] so part of the doc effort can be "this method on this role, 8 classes inherit it, but we only test that you can call it on 3 of them"
s/inherit/use/ 20:09
codesections So, where should `SetHash.kv` be documented? It's currently doc'ed in `Setty` but implemented (in Rakudo) in `SetHash` (and also in `Set`)
Since Roast doesn't care, we shouldn't either? 20:10
[Coke] It's a good question. I'm saying there's no right answer today. 20:11
aren't we setup so that methods available from outside our class are displayed on the class? 20:12
but it sounds like right now it's not matching roast nor rakudo
codesections Yeah, on the SetHash docs, `kv` shows up under the heading «Routines supplied by class Setty» 20:14
so the docs *do* display
it's just a question of if we care that the info doesn't match the Rakudo implementation (and, relatedly, how we detect missing docs if we don't follow Rakudo – e.g., SetHash would *also* inherit the `kv` method from Any, but that's not the correct implementation, so just saying "does this method show up on the page" isn't enough 20:16
)
20:18 dogbert2 left 20:20 dogbert2 joined
[Coke] we've said since day one that rakudo is *an* implementation, not *the* implementation. So that should inform this question on the docs site. 20:27
(we can have rakudo specific stuff in there, and do, but we try to call it out, like for 'dd', which is not spec.)
(day one) of the project, not the doc site.
codesections Yeah, I get that and support it 100%
I'm just kind of lost on how to apply it here 20:28
Like, we organize the Type docs by listing which methods are supplied by each Class/Role 20:29
and that's fundementally something Roast doesn't have an opinion on
20:29 Black_Ribbon joined
[Coke] I'd treat it as incremental progress. let's get the docs in the site as a first approximation, and we can true up the location as we go. 20:29
i don't think we need to finalize everything about step 2 before completing step 1. (but I get wanting to settle it as much as possible to reduce moves later) 20:30
codesections sighs
You're probably right...
[Coke] Sorry. also: I'm just me. I'm sure JJ has a strong opinion, you can do what he says. :)
codesections Well, you have me mostly convinced. But it pretty much means I need to give up on the goal of having all our docs checked against a tool that introspects Raku code 20:32
and that the work I did on the list-missing-methods script to get it to be that tool was ... well, I'll say a Good Learning Experience rather than "a waste" 20:33
20:33 natrys left
[Coke] I think it's a reasonable start, and everything in rakudo should be documented. you can still use that part of the tool. 20:44
and it may be used to highlight those differences between rakudo & roast and allow us to true things up, in either direction.
20:47 rindolf left, zacts left
codesections Coke: Yeah, all fair points :). Just to clarify, in your view should `Mixy` even be documented? It's not in Roast and is arguably an implementation detail, right? 20:48
[Coke] we can have additionally a mapping config that says "rakudo defines this here, and in docs, we've intentionally put it here.".
if it's not in Roast, I wouldn't bother, no.
assuming all documentation time is fungible, don't document the things that aren't spec.
(it's not, and people will work on what they want)
codesections Well, taken literally, that would mean *large* changes 20:49
Mixy isn't in Roast, Setty and Baggy are in ~2 tests, but that looks like an more of an accident than an intentional decision to include them 20:51
20:52 sena_kun joined
codesections Dateish isn't either 20:52
rypervenche Question: Is it possible to get the output of a Match object with the tree that it creates, but without the text that it matched against at the very top?
20:54 Altai-man left
rypervenche Ahh, looks like it's .caps, although it doesn't do the indenting. Ok, I can make a method to do that, similar to the Match gist. 20:56
[Coke] codesections: I am sure there are large gaps in the docs, yes.
only way to keep it up to date (and get it there in the first place) is testing.
so codesections++ for pursuing this and all the work on the rakudo introspector 20:57
codesections :D "large gaps" is a whole different thing from "many pages that shouldn't exist because we're following the implemention too closely"
It's a false-positive instead of a false negative 20:58
[Coke] add a failing xt test and just keep making it pass more. :) 20:59
"just"
codesections :D 21:00
[Coke] that's what we did with examples compilation, spellcheck... 21:01
there was a huge pile, we wrote a test so we'd know when we were done, and just hacked away
codesections Sounds like a plan. Or at least a plan-to-have-a-plan 21:02
But I guess we still need to reach consensus on this
21:15 hungrydonkey joined, MasterDuke joined 21:25 dogbert11 joined 21:28 dogbert2 left 21:32 chloekek joined
chloekek . 21:35
21:40 patrickb left 21:41 pecastro joined 21:48 dogbert17 joined, wamba joined 21:49 dogbert2 joined 21:51 dogbert11 left 21:52 dogbert17 left 22:01 dogbert2 left 22:03 dogbert2 joined 22:08 Manifest0 left 22:09 edk_ is now known as e 22:10 Cabanossi left 22:17 Cabanossi joined 22:21 dogbert11 joined 22:22 dogbert2 left
SmokeMachine xinming_: yes it should (if add the column c) 22:26
22:31 dogbert17 joined 22:34 dogbert11 left
lizmat codesections: perhaps Set.grab should be marked as "implementation detail " ? 22:35
and similar candidates that are just there to provide better error messages ?
22:43 dogbert17 left 22:45 dogbert17 joined 22:53 wamba left 22:54 lucasb joined 23:03 oneeggeach joined 23:07 molaf left 23:10 japhb left 23:11 japhb joined 23:13 dogbert2 joined 23:15 dogbert17 left 23:21 leont left 23:22 dogbert11 joined 23:23 dogbert17 joined 23:24 dogbert2 left 23:25 dogbert2 joined 23:26 dogbert12 joined 23:27 dogbert11 left 23:28 dogbert17 left 23:29 dogbert2 left 23:30 aborazmeh joined, aborazmeh left, aborazmeh joined 23:37 oneeggeach left 23:42 dogbert17 joined 23:45 dogbert12 left 23:46 dogbert2 joined 23:47 chloekek left 23:50 dogbert17 left 23:52 rir joined 23:53 dogbert11 joined 23:54 maggotbrain joined 23:55 dogbert2 left