Perl 6 language and compiler development 2.0 | Logs at irclog.perlgeek.de/perl6-dev/today | For toolchain/installation stuff see #perl6-toolchain | For MoarVM see #moarvm
Set by moderator on 23 May 2018.
00:57 stmuk_ joined 01:27 MasterDuke joined 01:58 ilbot3 joined
moderator Perl 6 language and compiler development 2.0 | Logs at irclog.perlgeek.de/perl6-dev/today | For toolchain/installation stuff see #perl6-toolchain | For MoarVM see #moarvm
samcv ok still getting some failing tests 03:03
t/spec/S26-documentation/multiline-leading.t t/spec/S26-documentation/block-leading.t t/spec/S26-documentation/block-trailing.t t/spec/S26-documentation/why-leading.t
Geth roast: c2a0fb2caa | (Samantha McVey)++ | S26-documentation/multiline-leading.t
Add test descriptions for multiline-leading.t
 03:19
skids samcv: the pod tests have been failing most probably since 2f4b2f428 05:05
06:20 shareable6 joined 06:51 robertle_ joined 07:19 brrt joined
AlexDaniel samcv: rakudo release/2018.05 and roast 1bbfa7eb314ac should be clean 07:32
couldn't care less about master atm :) I want a release 07:33
By the way, does anybody look over JJ's doc commits? 07:34
most of his commits are OK but *sometimes* things can be improved a bit more 07:35
I used to look at them but there's so much work going on that I simply can't keep up anymore
07:48 shareable6 joined
lizmat hmmm.. how do I do a runtime lookup of a sub *without* using PseudoStashes 08:10
Geth rakudo: 472f6e4844 | (Elizabeth Mattijsen)++ | src/core/traits.pm6
Look up DEPRECATED at the *end* of compilation

Since traits.pm6 is about the first file in settings being compiled, it did not know about &DEPRECATED yet. And somehow it is being forgotten so the precomp got a VMnull there. Mitigated by having the coderef for
  &DEPRECATED looked up at end of setting compilation by using CHECK.
Finally fixes R #1219 . skids++ for the nudge to get me looking at this again.		 08:38
synopsebot RAKUDO#1219 [open]: github.com/rakudo/rakudo/issues/1219 is DEPRECATED doesn't work on Routines in precomped modules
09:02 shareable6 joined 09:16 bartolin joined 09:29 Zoffix joined
Zoffix jdv79: don't recall anyone working on them other than jnthn++ mentioning that it might be an area he might soon might make faster. 09:30
jnthn I did reduce code size a bit for some cases, and got a quite big speedup for protoregexes/alternations with a large number of different literal prefixes (which was a useful speedup for parsing categories like infix/postfix/prefix in the Perl 6 grammar) 09:32
Got more plans, though a bit distracted with something else at these weeks. :-) 09:33
s/at//
Zoffix .tell lucasb "Just to conclude the subject of :nth"... Were there a ticket for it? All your comments on the subject are split across two IRC channel across several days and are impossible to follow. 09:34
09:35 literal joined
lizmat grrr... I keep forgetting how to do this and now can't find it: 09:58
wasn't there a simple way to mix in a role with a single attribute and set that attribute on the fly ? 09:59
Zoffix s: &infix:<o> 10:05
SourceBaby Zoffix, Sauce is at github.com/rakudo/rakudo/blob/472f...s.pm6#L781
Zoffix lizmat: like that maybe? github.com/rakudo/rakudo/blob/472f...#L789-L794 10:06
Though I think this is a bad way if the parametarization value changes a lot, 'cause it makes a new role each time or something
lizmat that's parameterizing, so each time we'll create a different class
well, for any combination of arity/count/of I guess 10:07
yup
lizmat gives up on making Attribute is DEPRECATED work for now again 10:09
jnthn lizmat: $foo but TheRoleName($value) 10:13
lizmat ah, duh... you cannot do this with a anonymous role (which I was trying to do) 10:14
Zoffix jnthn: is that different from parametarization? 10:18
oh yeah, cool
m: role Foo { has $.bar; method z { dd $!bar } }; say (42 does Foo(10)).z 10:19
evalable6 Int $!bar = 10
Nil
Zoffix m: role Foo { has $!bar; method z { dd $!bar } }; say (42 does Foo(10)).z
evalable6 (exit code 1) Can only supply an initialization value for a role if it has a single public attribute, but this is not the case for 'Foo'
in block <unit> at /tmp/gns9aTlBqE line 1
10:21 nine joined 10:23 camelia joined
jnthn Zoffix: Yes, it just sets the mixed in attribute 10:23
10:28 p6lert joined 10:37 Zoffix left
tbrowder_ i have a pr that fixes current pod test failures. i’m happy to merge it but want to know protocol. does it need at least one reviewer, or does it depend on complexity or something else? 11:01
11:04 yoleaux joined
tbrowder_ samcv: the pr passes yr new pod tests, so i’m merging it. 11:08
Geth rakudo: e1e7e4c61b | (Tom Browder)++ | src/Perl6/Pod.nqp
fix nbsp-fix problem by adding vertical and other missing breaking and non-breaking space chars for a complete unicode solution
 11:09
rakudo: 8a62e5e08c | (Tom Browder)++ (committed using GitHub Web editor) | src/Perl6/Pod.nqp
Merge pull request #1859 from tbrowder/fix-nbsp-probs

fix nbsp-fix problem by adding vertical and other missing breaking an…
12:18 stmuk joined 12:53 robertle joined 13:25 lizmat joined 13:31 skids joined 13:35 lucasb joined
lucasb . 13:35
hehe :) 13:36
Thanks for caring, Zoffix!
I know you suggested to open a rakudo issue, but I refrained from open it, because it isn't a bug, it's working perfectly as specced: github.com/perl6/roast/blob/master...#L212-L213
13:36 shareable6 joined
lucasb And even if this is decided to change, I think someone would be reluctant to do so in 6.c, since Xmas release shipped with that specced behavior already. 13:36
so maybe a better place would be in perl6/6.d-prep 13:38
but then... What will happen to those issues when 6.d is out?
13:39 Zoffix joined
lucasb Do you all plan to resolve all those issues before 6.d? I don't think there's any need to be so strict 13:40
Zoffix lucasb: TBH, when I started perl6/6.d-prep it was to keep planned and scheduled 6.d work. I see people have taken it as a debate forum for language changes, which IMO is not an appropriate place. A change can be done in later language versions and the Issues would then need to be moved out of that repo.
lucasb and so, I suggest to rename that repo or create a new one in perl6 GH org, with the name like perl6/{proposals,rfcs,ideas,design,etc.} 13:41
Zoffix github.com/rakudo/rakudo/issues/new is the place for everything. No need to have a billion repos to track and get confused about.
lucasb ^^ this new repo would then be used for all things from 6.d up to 6.z
Zoffix: ah, ok. I didn't know this was the current way. I'll suggest things in rakudo/rakudo then. Thanks! 13:42
If someone knows how to make GitHub migrate/export a issue from one repo to another... (I submitted 3 in perl6/6.d-prep, and wish they would land in rakudo/rakudo instead) 13:47
Geth 6.d-prep: 9d1898c259 | (Zoffix Znet)++ | 2 files
Document this repo isn't meant for language change discussions
 13:48
timotimo github.com/IQAndreas/github-issues-import ?
lucasb thanks timotimo 13:49
Just 3, I can do by hand, but I didn't want to lose the comments from others 13:50
Geth 6.d-prep: b4cc4624b8 | (Zoffix Znet)++ | 2 files
Clarify discussion of FEATURES.md stuff is OK
lucasb seems this tool posts the comments as another user: github.com/IQAndreas-testprojects/.../issues/10
*re-posts
Zoffix :o I hope it requires an API key.
lucasb: I'd just copy-paste the text and prefix it "moved from [URL of original Issue]", and close the original 13:51
timotimo it can hardly post a comment or bug under someone else's name
lucasb yeah, I mean the tool re-posts the comments as a predefined user, not just any random one :D 13:52
Like that "IssueImportBot commented on ..."
Geth 6.d-prep: 5b5b30161d | (Zoffix Znet)++ | 2 files
Fix markdown syntax error
 13:53
samcv tbrowder_++ 14:11
tbrowder_ samcv: whew, glad it worked for you! 14:13
14:22 eater joined 14:28 stmuk_ joined
Geth rakudo: 154d985518 | (Elizabeth Mattijsen)++ | src/core/Attribute.pm6
We don't need HLL Bools here
 14:37
15:10 skids joined 15:29 lucasb joined
lucasb so, "language design proposals" together with this specific implementation bugs queue? 15:30
I think this is practical, and we know that rakudo is the only implementation. but wasn't the idea to keep things separated?
the same way I was said to fill proposals in perl6/6.d-prep in the past. I worry sometime in the future I'll be said to stop putting design issues together with this specific implementation 15:31
but I'm ok with whatever is decided to be the best place :) 15:32
Zoffix lucasb: ah, right, I forgot that rakudo ain't the only implementation. Then I guess perl6/roast is a better place 15:41
lucasb would be better to keep perl6/roast/issues for... testy issues :) 15:42
Geth 6.d-prep: 3a197d52b9 | (Zoffix Znet)++ | 2 files
Don't list AN implementation as location for language design Issues
Zoffix "testy"?
lucasb therefore the suggestion of a separated perl6/{proposals,rfcs,ideas,design,etc.} repo
Zoffix lucasb: but roast *is* THE Perl 6 Specification. 15:43
lucasb: and my personal problem with roast as location for issues, is I never really bother to look at the list of its issues. If you're adding Yet Another Repo, that just multiplies the problem.
lucasb Zoffix: ok, agreed
but people normally submit issues there about different things than language design proposals: github.com/perl6/roast/issues 15:44
Zoffix we could have a label
IIRC rakudo/rakudo uses "RFC" as label for proposals. 15:45
lucasb I noticed that, and I was going to use that...
Zoffix lucasb: actually my explanation was likely inaccurate 16:12
Looking at p6for desugar, it don't care about Iterable at all: github.com/rakudo/rakudo/blob/mast...#L561-L578 16:13
But, Iterable provides the `.item` method that shoves things into container: github.com/rakudo/rakudo/blob/154d...le.pm6#L16
While \\Mu's version is a no-op: github.com/rakudo/rakudo/blob/154d...u.pm6#L574 16:14
m: .say for $ = \\(1, 2, 3)
camelia \\(1, 2, 3)
lucasb Maybe it's Mu.item is not a no-op. did you noticed the "is raw"? 16:15
Zoffix lucasb: what does that provide?
lucasb actually I dunno why I said it 16:18
but it seems to have made a difference in the Capture case: github.com/rakudo/rakudo/commit/71...64399c883c
Zoffix I think that's the defaults (| 16:19
for like (|c) and stuff 16:21
m: sub z(\\c) { use nqp; dd nqp::iscont(c) }($ = \\()) 16:22
camelia 1
Zoffix m: sub z(\\c is raw) { use nqp; dd nqp::iscont(c) }($ = \\())
camelia 1
Zoffix c: 7116059675fa7423,7116059675fa7423^1 sub z(|c) { use nqp; dd nqp::iscont(c) }(\\())
committable6 Zoffix, ¦7116059: «0␤» ¦7116059675fa7423^1: «1␤»
Zoffix yeah
m: dd $_ for $(set <a b c>) 16:27
camelia :b
:a
:c
Zoffix .ask lizmat I recall like half a year ago you asking why QuantHash objects aren't Iterable. Did you ever figure it out? I notice lack of Iterable makes them get .item from Mu, which is a no-op, and thus you can't iterate over, say, a set as a single item: `dd $_ for $(set <a b c>); # OUTPUT: «:b␤:a␤:c␤»` 16:28
yoleaux Zoffix: I'll pass your message to lizmat.
Zoffix ZOFVM: Files=1281, Tests=153301, 149 wallclock secs (20.86 usr 3.51 sys + 3140.33 cusr 181.83 csys = 3346.53 CPU) 16:42
Geth rakudo/Mu-item-cont-R1862: 4dcea18610 | (Zoffix Znet)++ | 2 files
Always containerize stuff when asked to with .item

Fixes R#1862 github.com/rakudo/rakudo/issues/1862
Always cont stuff when asked to by calling .item or $() contextualizer. Otherwise, we get no-op effects from those constructs even on objects where conting would make sense and does affect how they get iterated over.		 16:44
synopsebot R#1862 [open]: github.com/rakudo/rakudo/issues/1862 [@LARRY] Some objects don't .item'ize even though they probably should 16:45
rakudo: zoffixznet++ created pull request #1863:
Always containerize stuff when asked to with .item
17:05 Zoffix left
jdv79 cool 17:38
18:46 shareable6 joined 19:29 lucasb joined, brrt joined 19:40 dct joined
samcv moarvm is released now 20:14
[Coke] waves from Chicago 20:18
AlexDaniel alright! samcv++ 20:19
[Tux] Rakudo version 2018.04.1-132-g154d98551 - MoarVM version 2018.05
csv-ip5xs0.871 - 0.885
csv-ip5xs-208.918 - 8.956
csv-parser25.518 - 26.017
csv-test-xs-200.458 - 0.465
test8.845 - 8.860
test-t2.393 - 2.443
test-t --race0.994 - 0.995
test-t-2042.744 - 43.165
test-t-20 --race14.797 - 14.914
 20:29
20:34 releasable6 joined, greppable6 joined, coverable6 joined, benchable6 joined, unicodable6 joined, statisfiable6 joined 20:41 squashable6 joined 20:43 undersightable6 joined, greppable6 joined, releasable6 joined 20:44 undersightable6 joined 20:52 bisectable6 joined 21:05 AlexDaniel joined 21:20 Kaiepi joined
Geth rakudo/release/2018.05: 2dd85e8c82 | (Aleks-Daniel Jakimenko-Aleksejev)++ | 2 files
Log all changes (+ announcement)

Deliberately not logged: 4120b226 50f76831 df0b95df 8c617cfa c40fc5d7 93baa290 6c1983a9 94b4783d a51112eb 29b092c9 3d6e25c6 dbc527d4 305c1d6c 086980d4 40d887c8 ed740406 eac5ae1d a1aa1b62 a957b712 1447f41e 6ae0526f 852464a6 df4de34a f8c48450 e5b10f2b 4264c9e5 755627cb 16bc0cce a16eeac7 c66a6f0e 284f16da 79af50f6 f72a6a49 fb86392c 20e2f1d2 9aa6747f 35afcc9c a36adb2e b8318b8e 5c7f7295 1a425640 b15af2c9 60ee289a ec0422e7 450663bc 3642a879 fb0654cf f93ee2d1 4537a8e5 f98ea64d faaa2f24 791c3e35		 21:33
21:38 shareable6 joined
Geth rakudo/release/2018.05: 9d6eda0660 | (Aleks-Daniel Jakimenko-Aleksejev)++ | docs/release_guide.pod
Actual date, claim next release
 21:38
rakudo: 89a4cf0d98 | (Elizabeth Mattijsen)++ | 9 files
Move &DEPRECATED to Rakudo::Deprecations.DEPRECATED

The main issue is that in setting compilation, we are referring to
  &DEPRECATED *long* before it actually gets defined. This is normally
not an issue (afaik), but it apparently *is* in the core setting, especially in combination with precompiled modules doing deprecations.
This provides an alternate fix for 472f6e48444b5ed157658 , and it paves the way to making "has $.a is DEPRECATED<b>" work.		 21:55
tbrowder_ .ask samcv should unicode chars 0x2000 and 0x2001 rountrip. they don’t appear to (both are space chars) and AlexDaniel thinks that may be normalization. 22:39
yoleaux tbrowder_: I'll pass your message to samcv.
jnthn m: say Uni.new(0x2000, 0x2001).NFC.list 22:40
camelia (8194 8195)
jnthn m: say Uni.new(0x2000, 0x2001).NFC.list.map(*.base(16))
camelia (2002 2003)
jnthn Yeah, those ones change under NFC
There's a number of such singleton decompositions in NFC 22:41
tbrowder_ jnthn: thnx, as the guy in tv ads here says, “who knew.” 22:42
uh, “who knew?” 22:43
22:44 Kaiepi joined 22:49 travis-ci joined
travis-ci Rakudo build failed. Elizabeth Mattijsen 'Move &DEPRECATED to Rakudo::Deprecations.DEPRECATED 22:49
travis-ci.org/rakudo/rakudo/builds/383425858 github.com/rakudo/rakudo/compare/1...a4cf0d980c
22:49 travis-ci left
buggable [travis build above] ☠ All failures are due to: failed make test (6 failures). Across all jobs, only t/02-rakudo/11-deprecated.t test file failed. 22:50
Geth nqp/release/2018.05: bf27e6f985 | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/build/MOAR_REVISION
[release] Bump MoarVM revision to 2018.05
 22:55
nqp/release/2018.05: 577b265e45 | (Aleks-Daniel Jakimenko-Aleksejev)++ | VERSION
[release] Bump VERSION to 2018.05
rakudo/release/2018.05: 1977659733 | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/build/NQP_REVISION
[release] Bump NQP revision to 2018.05
rakudo/release/2018.05: efa05942ad | (Aleks-Daniel Jakimenko-Aleksejev)++ | VERSION
[release] Bump VERSION to 2018.05
Kaiepi shit i forgot about making tickets about the tests that fail 23:49
BeastieBot, stresstest 23:50
BeastieBot [freebsd] Running Roast's stress test suite (this will take a while)...
23:59 stmuk joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!