Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6-dev, 2017-05-22

| Channels | #perl6-dev index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:27 BenGoldberg joined #perl6-dev
00:53 MasterDuke .tell jnthn if it helps, here's the MVM_dump_backtrace output at the problematic coerce_ni when running `say(100000000000000009)` https://gist.github.com/MasterDuke17/a34c63dece4f572d049a9fa67f0249db
00:53 yoleaux MasterDuke: I'll pass your message to jnthn.
00:53 MasterDuke .tell timotimo if it helps, here's the MVM_dump_backtrace output at the problematic coerce_ni when running `say(100000000000000009)` https://gist.github.com/MasterDuke17/a34c63dece4f572d049a9fa67f0249db
00:53 yoleaux MasterDuke: I'll pass your message to timotimo.
01:02 sasheto joined #perl6-dev
01:16 astj joined #perl6-dev
01:32 BenGoldberg joined #perl6-dev
01:38 sasheto joined #perl6-dev
01:55 ilbot3 joined #perl6-dev
01:55 Topic for #perl6-dev is now Perl 6 language and compiler development 2.0 | Logs at http://irclog.perlgeek.de/perl6-dev/today | For toolchain/installation stuff see #perl6-toolchain | For MoarVM see #moarvm
02:31 ugexe zef is suddently failing its appveyor tests when using an absolute path ala `zef install C:\my\repo`
02:44 ugexe must be https://github.com/rakudo/rakudo/commit/494659a1c743624493a0d009f53e7ae206340de4 somehow
02:55 samcv i'm deciding on the heirarchy of unicode property values right now
02:55 samcv we will have General_Category and Script for sure. then if we add Grapheme_Cluster_Break that cuts the number of property values with conflicts down to 30
02:56 samcv if anybody wants to see, https://github.com/perl6/specs/issues/118#issuecomment-302968322 the conflicts before resolving General_Category and Script
02:57 samcv and thinking we should resolve Numeric_Type as well. since i'm sure code that does <:Numeric> exists out there. and Numeric is a value of Numeric_Type
03:08 ugexe m: say $*CWD.parts; # regression?
03:08 camelia rakudo-moar 1c16bf: OUTPUT: «Map.new((:basename("camelia"),:dirname("/home"),:volume("")))␤»
03:09 ugexe star: say $*CWD.parts
03:09 camelia star-m 2016.10: OUTPUT: «{basename => star-2016.10, directory => /home/camelia, dirname => /home/camelia, volume => }␤»
03:11 samcv it's going to get more fun now i'm going to add in Property Names as well and how they interact with property aliases
03:18 ugexe Zoffix: ^ note the above... it switched to a Map but more importantly it does not include the "directory" key
03:20 MasterDuke ugexe: i believe the change to Map was intentional, to make it immutable. don't know about the "directory" key though
03:26 ugexe whatever the problem is causing zef to fail was introduced in the last 13 hours anyway
03:27 ugexe zef to fail on *windows*
03:37 ugexe "This type cannot unbox to a native string: P6opaque, Zef::Utils::URI::URI::File"
03:39 AlexDaniel samcv: by the way, um…
03:40 AlexDaniel samcv: why not simply error out on stuff that is ambiguous?
03:41 samcv that is fair enough. but i don't want to break everybodies code
03:41 samcv so i cannot do such a thing for 6.c
03:41 samcv there at least needs to be some heirachy when i retrofit our unicode system for me to not have conflicts and also make sure .unimatch and other random assortted things work properly :>
03:42 samcv sadly it is not that simple from 1. breaking code and 2. how things are already setup inside MoarVM rakudo/nqp etc
03:43 samcv but i'm down for only "Guaranteeing" certain properties which could be a different list
03:43 samcv but even if we don't guarantee certain things i need to at least make some heirarchy.
03:44 ugexe Zoffix: error seems to be a change in coercion. $*SPEC.join($foo, $bar, $baz) - if $foo is a match object it no longer gets stringified (hence the unbox error above)
05:45 brrt joined #perl6-dev
06:38 brrt joined #perl6-dev
06:53 Zoffix .parts is not a regression
06:54 Zoffix ugexe: are you using $*SPEC.join directly?
06:57 Zoffix Ah ok, you already committed a patch
07:01 astj_ joined #perl6-dev
07:05 Zoffix ZOFVM: Files=1242, Tests=135655, 112 wallclock secs (21.74 usr  3.25 sys + 2293.68 cusr 157.36 csys = 2476.03 CPU)
07:06 Geth ¦ rakudo/nom: 232cf190d6 | (Zoffix Znet)++ | src/core/IO/Spec/Win32.pm
07:06 Geth ¦ rakudo/nom: Add Str type constraints to IO::Spec::Win32 .join and .catpath
07:06 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/232cf190d6
07:23 Zoffix ZOFFLOP: t/spec/S17-supply/interval.t
07:25 Zoffix I need to adjust one 6.c-errata test that I'm arguing is incorrect: https://github.com/perl6/roast/blob/6.c-errata/S16-io/print.t#L84-L88
07:25 Zoffix 1) It expects different output for @array than for $array
07:26 Zoffix 2) This is due to the signature in IO::Handle.print being *@, while in IO::Handle.say (and with my patch in .put) and &say, &put, and &print it's **@
07:28 Zoffix 3) It appears the test was merely covering the current, at the time, behaviour. I don't recognize this person and they do not appear to be a member of rakudo org: https://github.com/perl6/roast/commit/d663fb5a4168e76c64f2bafc0eef486b88fabefb
07:29 Zoffix ZofBot: the milk stool is complete!
07:29 ZofBot Zoffix, A call to the scalar version "$o
07:41 nine ZofBot: FWIW zhuomingliang is more of a MoarVM developer
07:41 ZofBot nine, substr($
07:41 nine Zoffix: FWIW zhuomingliang is more of a MoarVM developer
07:41 Zoffix noted
07:42 Geth ¦ rakudo/nom: 613bdcf834 | (Zoffix Znet)++ | src/core/IO/Handle.pm
07:42 Geth ¦ rakudo/nom: [io grant] Make IO::Handle.print/.put sig consistent
07:42 Geth ¦ rakudo/nom:
07:42 Geth ¦ rakudo/nom: ... with IO::Handle.say, &say, &print, and &put
07:42 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/613bdcf834
07:42 Geth ¦ roast: 050d718a1d | (Zoffix Znet)++ | S16-io/print.t
07:42 Geth ¦ roast: Fix incorrect IO::Handle.print: @array
07:42 Geth ¦ roast:
07:43 Geth ¦ roast: 1) It expects different output for @array than for $array
07:43 Geth ¦ roast: 2) This is due to the signature in IO::Handle.print being *@,
07:43 Geth ¦ roast:     while in IO::Handle.say (and with my patch in .put) and &say,
07:43 Geth ¦ roast:     &put, and &print it's **@
07:43 Geth ¦ roast: 3) Looking at the commit that brought the test, it appears the
07:43 Geth ¦ roast:     test was merely covering the current, at the time, behaviour,
07:43 Geth ¦ roast:     rather than a product of conscious design decision to make
07:43 Geth ¦ roast:     behaviour different from subroutine forms and other methods:
07:43 Geth ¦ roast:     https://github.com/perl6/roast/commit/d663fb5a4168e76c64f2ba
07:43 Geth ¦ roast: review: https://github.com/perl6/roast/commit/050d718a1d
07:43 Geth ¦ roast/6.c-errata: e20b1008ba | (Zoffix Znet)++ | S16-io/print.t
07:43 Geth ¦ roast/6.c-errata: Fix incorrect IO::Handle.print: @array
07:43 Geth ¦ roast/6.c-errata:
07:43 Geth ¦ roast/6.c-errata: 1) It expects different output for @array than for $array
07:43 Geth ¦ roast/6.c-errata: 2) This is due to the signature in IO::Handle.print being *@,
07:43 Geth ¦ roast/6.c-errata:     while in IO::Handle.say (and with my patch in .put) and &say,
07:43 Geth ¦ roast/6.c-errata:     &put, and &print it's **@
07:43 Geth ¦ roast/6.c-errata: 3) Looking at the commit that brought the test, it appears the
07:43 Geth ¦ roast/6.c-errata:     test was merely covering the current, at the time, behaviour,
07:43 Geth ¦ roast/6.c-errata:     rather than a product of conscious design decision to make
07:43 Geth ¦ roast/6.c-errata:     behaviour different from subroutine forms and other methods:
07:43 Geth ¦ roast/6.c-errata:     https://github.com/perl6/roast/commit/d663fb5a4168e76c64f2ba
07:43 Geth ¦ roast/6.c-errata: review: https://github.com/perl6/roast/commit/e20b1008ba
07:54 pmurias joined #perl6-dev
08:18 Geth ¦ roast: 4d0f97a922 | (Zoffix Znet)++ | S16-io/put.t
08:18 Geth ¦ roast: [io grant] Add &put/IO::Handle.put tests
08:18 Geth ¦ roast: review: https://github.com/perl6/roast/commit/4d0f97a922
08:18 Geth ¦ rakudo/nom: 601cdb6d58 | (Zoffix Znet)++ | t/spectest.data
08:18 Geth ¦ rakudo/nom: Add S16-io/put.t to list of test files to run
08:18 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/601cdb6d58
08:27 robertle joined #perl6-dev
08:49 jnthn morning o/
08:49 yoleaux 00:53Z <MasterDuke> jnthn: if it helps, here's the MVM_dump_backtrace output at the problematic coerce_ni when running `say(100000000000000009)` https://gist.github.com/MasterDuke17/a34c63dece4f572d049a9fa67f0249db
09:06 Zoffix \o
09:45 |Tux| This is Rakudo version 2017.05-21-g601cdb6d5 built on MoarVM version 2017.05-9-g37460c7a
09:45 |Tux| csv-ip5xs        2.585
09:45 |Tux| test            12.491
09:45 |Tux| test-t           4.264 - 4.275
09:45 |Tux| csv-parser      12.967
09:48 lizmat Files=1192, Tests=58679, 206 wallclock secs (12.44 usr  4.81 sys + 1220.11 cusr 116.30 csys = 1353.66 CPU)
09:48 yoleaux 21 May 2017 21:36Z <Zoffix> lizmat: any idea if the Cool:D and Mu candidates can be merged together? I see the comment about optimization was deleted... do we still need to candidates with identical bodies? https://github.com/rakudo/rakudo/commit/b00d92de142f0d9b8514ca9985dc43a4b7d7625e
09:50 Zoffix buggable: speed 6
09:50 buggable Zoffix, ▁▁▇▇▂▂ data for 2017-05-20–2017-05-22; range: 4.225s–4.607s; 1% slower
09:50 Zoffix \o/
09:52 lizmat m: dd Pair.new(:a<b>)  # now it is silent ?
09:52 camelia rakudo-moar 601cdb: OUTPUT: «(Mu) => Mu␤»
09:52 Zoffix Unless the slowage is still there and got offset by .print/.put sig change
09:53 Zoffix m: sub foo (*@) { }; for ^1000_00 { $ = foo "a", "b", "c" }; say now - INIT now
09:53 camelia rakudo-moar 601cdb: OUTPUT: «0.3198904␤»
09:53 Zoffix m: sub foo (**@) { }; for ^1000_00 { $ = foo "a", "b", "c" }; say now - INIT now
09:53 camelia rakudo-moar 601cdb: OUTPUT: «0.28454634␤»
09:53 Zoffix lizmat: oops. Will fix in a sec
09:53 lizmat Zoffix: I have a fix
09:53 Zoffix ok
09:54 Zoffix Yeah, I only tested for wrong positionals :/
09:54 lizmat spectesting now
09:54 Zoffix m: dd Pair.new: <a b c>
09:54 camelia rakudo-moar 601cdb: OUTPUT: «Cannot resolve caller new(Pair: List); none of these signatures match:␤    (Pair $: Cool:D \key, Mu \value, *%_)␤    (Pair $: Mu \key, Mu \value, *%_)␤    (Pair $: Mu :$key, Mu :$value, *%_)␤  in block <unit> at <tmp> line 1␤␤»
09:54 lizmat m: dd Pair.new(:a<b>)  # will give the same error
09:54 camelia rakudo-moar 601cdb: OUTPUT: «(Mu) => Mu␤»
09:59 lizmat m: my $a = $*VM; for ^100000 { Pair.new($a,42) }; say now - INIT now  # not cool
09:59 camelia rakudo-moar 601cdb: OUTPUT: «0.13321350␤»
09:59 lizmat m: my $a = "foo:; for ^100000 { Pair.new($a,42) }; say now - INIT now  # cool
09:59 camelia rakudo-moar 601cdb: OUTPUT: «5===SORRY!5=== Error while compiling <tmp>␤Unable to parse expression in double quotes; couldn't find final '"' ␤at <tmp>:1␤------> 3new($a,42) }; say now - INIT now  # cool7⏏5<EOL>␤    expecting any of:␤        horizontal whitespace…»
09:59 lizmat m: my $a = "foo:"; for ^100000 { Pair.new($a,42) }; say now - INIT now  # cool
09:59 camelia rakudo-moar 601cdb: OUTPUT: «0.0984492␤»
09:59 lizmat feels to me the Cool candidate is still needed
10:00 lizmat m: say 0.13321350 / 0.0984492
10:00 camelia rakudo-moar 601cdb: OUTPUT: «1.3531192␤»
10:00 Zoffix OK. Perhaps could use a comment there about it being there for optimization reasons
10:01 lizmat yeah, just added
10:01 Geth ¦ rakudo/nom: 85bef661f0 | (Elizabeth Mattijsen)++ | src/core/Pair.pm
10:01 Geth ¦ rakudo/nom: Make sure that Pair.new(:foo<bar>) dies
10:01 Geth ¦ rakudo/nom:
10:01 Geth ¦ rakudo/nom: Also add comment on the need of the Cool:D candidate.
10:01 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/85bef661f0
10:02 lizmat actually, I think the above benchmark is flawed
10:03 lizmat m: my $a = Str; for ^100000 { Pair.new($a,42) }; say now - INIT now  # not Cool:D
10:03 camelia rakudo-moar 601cdb: OUTPUT: «0.102694␤»
10:03 lizmat m: my $a = "foo"; for ^100000 { Pair.new($a,42) }; say now - INIT now  # Cool:D
10:03 camelia rakudo-moar 601cdb: OUTPUT: «0.09789931␤»
10:04 lizmat m: say 0.102694 / 0.09789931
10:04 camelia rakudo-moar 601cdb: OUTPUT: «1.04897573␤»
10:04 lizmat so, more like 5% faster
10:04 lizmat still for something that is done a lot with e.g. iterating over a hash..  seems worthwhile
10:06 Geth ¦ roast: ea41dcd0d5 | (Zoffix Znet)++ | S02-types/pair.t
10:06 Geth ¦ roast: Test wrong named args to Pair.new throw
10:06 Geth ¦ roast:
10:06 Geth ¦ roast: Rakudo fix: https://github.com/rakudo/rakudo/commit/85bef661f0
10:06 Geth ¦ roast: review: https://github.com/perl6/roast/commit/ea41dcd0d5
10:07 Ven joined #perl6-dev
10:23 Zoffix ehehe, we have exactly 0 tests for &prompt
10:23 Zoffix and it's documented :)
10:24 * lizmat assumes they will get promptly added now ?
10:24 Zoffix :)
10:24 Zoffix This brings me to wanting to make $msg default to ''
10:24 Zoffix as in: my $foo = prompt; just waits for a line of text on $*IN without showing any messages
10:25 Zoffix 'cause &get uses $*ARGFILES, so you're either left with `my $foo = prompt '';` which is kinda ewww or `my $foo = $*IN.get` which is also ewwy
10:30 astj joined #perl6-dev
10:58 Geth ¦ rakudo/nom: b43db636ab | (Elizabeth Mattijsen)++ | 2 files
10:58 Geth ¦ rakudo/nom: Make (Bag|Mix).AT-KEY about 10% faster
10:58 Geth ¦ rakudo/nom:
10:58 Geth ¦ rakudo/nom: By only needing to do a lookup in the underlying hash once on a match.
10:58 Geth ¦ rakudo/nom: Keeps the original Baggy.AT-KEY implementation for subclasses: not
10:58 Geth ¦ rakudo/nom: possible to do the (Bag|Mix) implementation in the role, as we cannot
10:58 Geth ¦ rakudo/nom: initialize a class lexical in a role :-( .  At least not in the Setting.
10:58 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/b43db636ab
11:05 Zoffix w00t! Another Perl 6 book: https://www.indiegogo.com/projects/book-web-application-development-in-perl-6-website--2#/
11:05 Zoffix wow 11 backers already
11:06 Zoffix Can't say I was all that impressed with a few Gábo's articles that I've read... but it's good to have more books
11:07 Zoffix ZofBot: should I spring the $50 to get preview copies?
11:07 ZofBot Zoffix, ACCEPTS faster now Unhandled exception in code scheduled on thread 27 Unhandled exception in code scheduled on thread 30 too many open files :| ;( m: say 4
11:08 Zoffix I guess that means I should hurry the hell up with Tardigrade so that it might make the book ^_^
11:12 pmurias what's Tardigrade?
11:13 tadzik ZofBot: Are your replies a markov chain of this channel? :o
11:13 ZofBot tadzik, Array rakudo-moar 9ed89d: OUTPUT: «57␤"199:"␤» m:
11:13 tadzik I guess that answers my question
11:13 Zoffix :>
11:14 Zoffix tadzik: the original brain was a scrape of design.perl6.org
11:14 Zoffix it now evolved a bit...
11:14 tadzik hehe
11:14 tadzik ...or did it...
11:15 tadzik or are we all adhering to this design too :o
11:16 Zoffix ugh indiegogo.com doesn't take PayPal :|
11:30 lizmat .seen masak
11:30 yoleaux I saw masak 19 May 2017 08:42Z in #perl6-dev: <masak> it creates a WhateverCode of the whole expression, I guess
11:31 Geth ¦ rakudo/nom: 0646d3fa46 | (Zoffix Znet)++ | src/core/io_operators.pm
11:31 Geth ¦ rakudo/nom: [io grant] Allow no-arg &prompt
11:31 Geth ¦ rakudo/nom:
11:31 Geth ¦ rakudo/nom: It's not entirely unheard of to not want to include any text message
11:31 Geth ¦ rakudo/nom: to the user when prompting for some input. It's a bit antisocial to
11:31 Geth ¦ rakudo/nom: force the user to specify `''` arg for such cases.
11:31 Geth ¦ rakudo/nom:
11:31 Geth ¦ rakudo/nom: Worse, they might introduce a subtle bug into their code by switching
11:31 Geth ¦ rakudo/nom: to use `get` for a message-less `prompt`, but `get` defaults to
11:31 Geth ¦ rakudo/nom: `$*ARGFILES`, instead of `$*IN`
11:31 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/0646d3fa46
11:31 Geth ¦ roast: 125fe18489 | (Zoffix Znet)++ | S16-io/prompt.t
11:31 Geth ¦ roast: [io grant] Add &prompt tests
11:31 Geth ¦ roast: review: https://github.com/perl6/roast/commit/125fe18489
11:31 Geth ¦ nqp: 23cea1133c | pmurias++ | src/vm/js/Operations.nqp
11:31 Geth ¦ nqp: [js] Fix bug with neverreposses being ignore when used in void context
11:31 Geth ¦ nqp: review: https://github.com/perl6/nqp/commit/23cea1133c
11:31 Geth ¦ nqp: f2b5270337 | pmurias++ | 2 files
11:31 Geth ¦ nqp: [js] Implement nqp::loadbytecodefh
11:31 Geth ¦ nqp: review: https://github.com/perl6/nqp/commit/f2b5270337
11:32 Geth ¦ rakudo/nom: bf1d0ada35 | (Zoffix Znet)++ | t/spectest.data
11:32 Geth ¦ rakudo/nom: Run S16-io/prompt.t tests as part of spectest
11:32 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/bf1d0ada35
11:37 Zoffix Why yes, atop, close all my instances when I accidentally press CTRL+Q instead of CTRL+W -_
11:37 Zoffix *atom
11:39 ilmari chrome gets this right, by making the quit shortcut ctrl-shift-q (and there's a firefox extension to do the same)
11:41 Zoffix Yeah, I have the firefox extension.
11:49 Geth ¦ nqp: cono++ created pull request #360: Add mapping of the op code get_port_sk
11:49 Geth ¦ nqp: review: https://github.com/perl6/nqp/pull/360
11:50 Geth ¦ rakudo: cono++ created pull request #1086: Update localport if we made bind on port 0
11:50 Geth ¦ rakudo: review: https://github.com/rakudo/rakudo/pull/1086
11:53 astj_ joined #perl6-dev
11:54 cono jnthn: here is PR for what we discussed yesterday: https://github.com/MoarVM/MoarVM/pull/600 at the bottom of it you can find links to the nqp and rakudo PR's. Three is no spec-test PR yet.
12:11 lizmat jnthn: a though re native attributes in Roles: perhaps they could be handled in the Optimizer ?
12:11 lizmat *thought
12:12 lizmat *sigh*
12:15 Geth ¦ rakudo/nom: ff04a37c4b | (Elizabeth Mattijsen)++ | 3 files
12:15 Geth ¦ rakudo/nom: Streamline (Set|Bag|Mix).new(), there can only be one!
12:15 Geth ¦ rakudo/nom:
12:15 Geth ¦ rakudo/nom: Set.new() did not return set(), but it did handle subclassing.
12:15 Geth ¦ rakudo/nom: Bag.new() returned bag(), but did not handle subclassing.
12:15 Geth ¦ rakudo/nom: Mix.new() returned mix(), but did not handle subclassing.
12:15 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/ff04a37c4b
12:21 Geth ¦ roast: eb435234a2 | (Elizabeth Mattijsen)++ | 3 files
12:21 Geth ¦ roast: Set|Bag|Mix.new returns the empty set|bag|mix
12:21 Geth ¦ roast: review: https://github.com/perl6/roast/commit/eb435234a2
12:30 nine Puh...openSSL braucht da kurzfristig scho 900 MB RAM
12:30 moritz nine: wrong channel or wrong language :-)
12:36 nine moritz: argh...wrong channel indeed. As much as I love the discovery of irssi's split window mode, it does require extra care.
12:37 lizmat m: say Set (<=) Set  # awh
12:37 camelia rakudo-moar ff04a3: OUTPUT: «Invocant of method 'total' must be an object instance of type 'Setty', not a type object of type 'Set'.  Did you forget a '.new'?␤  in block <unit> at <tmp> line 1␤␤»
12:38 lizmat fixing now
12:38 AlexDaniel joined #perl6-dev
12:46 Geth ¦ rakudo/nom: accf519850 | (Elizabeth Mattijsen)++ | 3 files
12:46 Geth ¦ rakudo/nom: Get Baggy.AT-KEY optimization back into role
12:46 Geth ¦ rakudo/nom:
12:46 Geth ¦ rakudo/nom: We don't need no lexical, we can BEGIN  :-)
12:46 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/accf519850
12:46 Geth ¦ rakudo/nom: ad8fa55297 | (Elizabeth Mattijsen)++ | src/core/set_operators.pm
12:46 Geth ¦ rakudo/nom: Make sure that Setty:U is treated like any type object
12:46 Geth ¦ rakudo/nom:
12:46 Geth ¦ rakudo/nom: Fixing Set (<) Set and friends.
12:46 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/ad8fa55297
13:35 Geth ¦ rakudo/js: 0cb7e5f287 | pmurias++ | src/vm/js/perl6-runtime/runtime.js
13:35 Geth ¦ rakudo/js: [js] Hide some debugging messages
13:35 Geth ¦ rakudo/js: review: https://github.com/rakudo/rakudo/commit/0cb7e5f287
13:35 Geth ¦ rakudo/js: c176289ffe | pmurias++ | Configure.pl
13:35 Geth ¦ rakudo/js: Report a problem when the --backends is incompatible with the --with-nqp
13:35 Geth ¦ rakudo/js: review: https://github.com/rakudo/rakudo/commit/c176289ffe
13:46 Geth ¦ rakudo/nom: 5c8e3f3ab9 | (Elizabeth Mattijsen)++ | src/core/set_operators.pm
13:46 Geth ¦ rakudo/nom: Define superset ops in terms of subset ops
13:46 Geth ¦ rakudo/nom:
13:46 Geth ¦ rakudo/nom: No need to implement the same functionality twice.  Done in
13:46 Geth ¦ rakudo/nom: preparation for subset ops refactor.
13:46 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/5c8e3f3ab9
13:54 Geth ¦ nqp: f9f6667dc5 | (Jonathan Worthington)++ | src/vm/jvm/QAST/Compiler.nqp
13:54 Geth ¦ nqp: Correct decodertakeline op return type.
13:54 Geth ¦ nqp: review: https://github.com/perl6/nqp/commit/f9f6667dc5
13:56 * Zoffix ponders what IO::CatHandle should do for write methods...
13:56 Zoffix like $fh.say: "meows"
13:58 Zoffix Like in here I was thinking of mirroring it into all the given handles: https://github.com/rakudo/rakudo/commit/031aa3322570a97469144305fbbc60605b31d37a
13:58 Zoffix But now I think that's both (a) not really useful; (b) doesn't scale, since it requires all handles to be open for writing, and you hit the open file limit
13:58 lizmat isn't IO::CatHandle just a IO::Handle wrt to writing ??
13:59 lizmat ah... sorry
13:59 * lizmat shuts up again
13:59 lizmat well, no writing to IO::CatHandle
13:59 lizmat oops
13:59 lizmat :-)
14:01 Zoffix Well, writing to the currently active handle is another option, but then it's like... why use a cathandle for this then. And third option is writing X bytes/chars/lines/words and automatically switching to another file, but that feels more of a narrow use case best left for module-space
14:01 * Zoffix will go with "die X::NIY" for write methods for now
14:01 lizmat sounds like a plan to me  :-)
14:18 Zoffix .seen cygx
14:18 yoleaux I saw cygx 21 May 2017 16:40Z in #perl6: <cygx> not sure how this is supposed to work, but you could try locating you moarvm's Makefile and remove that part from CFLAGS and build manually
14:19 Zoffix .ask cygx would you recommend .seek/.tell to throw if called on a handle that isn't in binary mode? I recall you saying they are technically binary operations when I had that case where .tell was giving wrong result after some text read operation
14:19 yoleaux Zoffix: I'll pass your message to cygx.
14:24 perlpilot joined #perl6-dev
14:52 Geth ¦ rakudo/nom: 32eb285f18 | (Elizabeth Mattijsen)++ | src/core/set_operators.pm
14:52 Geth ¦ rakudo/nom: Make (<=) and (>=) about 50x faster
14:52 Geth ¦ rakudo/nom:
14:52 Geth ¦ rakudo/nom: Based on a 25/26 elem Set.  Doing a (<=) or (>=) on itself is now
14:52 Geth ¦ rakudo/nom: shortcut, as that is always True, for potentially infinite speed
14:52 Geth ¦ rakudo/nom: improvement (as the identity check is *very* cheap).
14:52 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/32eb285f18
14:53 Geth ¦ rakudo/js: df77eb6a44 | pmurias++ | 3 files
14:53 Geth ¦ rakudo/js: Add missing paren in message.
14:53 Geth ¦ rakudo/js: review: https://github.com/rakudo/rakudo/commit/df77eb6a44
15:02 hoelzro joined #perl6-dev
15:06 lizmat afk for a bit&
15:08 Ven joined #perl6-dev
15:11 cono jnthn: Huston we have a problem :D https://pastebin.com/CzhThVRa
15:12 cono somewhere in thread.finish it dies
15:12 cono any tips where to set break point?
15:13 brrt joined #perl6-dev
15:20 cono hm, in debugger it just works fine :( hate this thing. Walked from this: MVM_thread_join until the end
15:22 cono ah, looks like I was able to segfault it in debugger: https://pastebin.com/0khmkbT3
15:26 Ven joined #perl6-dev
15:30 jnthn At a guess, you'll have assumed something is thread safe that ain't
15:31 jnthn otoh
15:31 jnthn We know that sockets ain't on master :)
15:31 jnthn (the sync type, not the IO::Socket::Async type)
15:31 jnthn Which is what I'm doing in my branch
15:31 jnthn That kind of segv isn't the typical failure mode though
15:37 [Coke] ... ok, it's weird to me that -dev is filling up backscroll (I hate this irc client!) but #perl6 isn't.
15:38 mst "filling up" ?
15:40 [Coke] too many sends, this client defaults to throwing away the oldest.
15:40 [Coke] so then I have to go back to the web logs to review. if the buffer doesn't fill, I can review in-client.
15:41 [Coke] (irssi)
15:41 [Coke] (and because of how it does it, I can't even see what the last line I read was.)
15:47 Ven_ joined #perl6-dev
16:13 awwaiid joined #perl6-dev
16:20 awwaiid joined #perl6-dev
16:28 Zoffix heh
16:29 Zoffix Well, that's good. There's no such thing as too much -dev :)
16:29 Zoffix #perl6 looks quite filled up to me....
16:42 Ven joined #perl6-dev
16:49 Ven_ joined #perl6-dev
16:58 Ven_ joined #perl6-dev
17:12 Ven_ joined #perl6-dev
17:18 Ven joined #perl6-dev
17:27 Ven joined #perl6-dev
17:53 Ven_ joined #perl6-dev
18:27 Ven joined #perl6-dev
18:52 * Zoffix notices a spam comment on the IO grant: http://news.perlfoundation.org/2017/01/grant-proposal-standardization.html
18:52 Zoffix How dare they!
18:53 Zoffix Man, there's no end to this grant... I don't think I'll ever be doing another one :|
18:54 [Coke] (spam comment)... I'll delete it.
18:54 Zoffix \o/
18:55 [Coke] ... I can't, it's makoto's post.
18:55 Zoffix ah heh :)
18:55 Geth ¦ rakudo/nom: 431ed4e30e | (Elizabeth Mattijsen)++ | src/core/set_operators.pm
18:55 Geth ¦ rakudo/nom: Make (<) and (>) about 40x faster
18:55 Geth ¦ rakudo/nom:
18:55 Geth ¦ rakudo/nom: Based on a 25/26 elem Set.  Doing a (<) or (>) on itself is now
18:55 Geth ¦ rakudo/nom: shortcut, as that is always False, for potentially infinite speed
18:55 Geth ¦ rakudo/nom: improvement (as the identity check is *very* cheap).
18:55 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/431ed4e30e
19:08 Zoffix Is there any downside to using method meth (::?CLASS:D:) {} instead of method meth (SomeClass:D:) {} ?
19:09 Zoffix like perf or anything...
19:10 Ven_ joined #perl6-dev
19:25 dogbert17 joined #perl6-dev
19:25 dogbert17 commit: 2016.06 my $l = (1, 2, 3); say (0, slip $l).perl;
19:25 committable6 dogbert17, ¦2016.06: «(0, $(1, 2, 3))»
19:26 dogbert17 m: my $l = (1, 2, 3); say (0, slip $l).perl;
19:26 camelia rakudo-moar 431ed4: OUTPUT: «(0, 1, 2, 3)␤»
19:26 dogbert17 commit: 2016.08 my $l = (1, 2, 3); say (0, slip $l).perl;
19:26 committable6 dogbert17, ¦2016.08: «(0, $(1, 2, 3))»
19:26 dogbert17 commit: 2016.10 my $l = (1, 2, 3); say (0, slip $l).perl;
19:26 committable6 dogbert17, ¦2016.10: «(0, $(1, 2, 3))»
19:26 dogbert17 commit: 2016.11 my $l = (1, 2, 3); say (0, slip $l).perl;
19:26 committable6 dogbert17, ¦2016.11: «(0, 1, 2, 3)»
19:27 dogbert17 is the new behavior ^^ correct?
19:29 Zoffix yup
19:30 dogbert17 Zoffix: thx, time to update some docs then :)
19:30 Zoffix dogbert17: well, looks correct to me, though seems it got that way by accident
19:31 Zoffix c: 37d0e4661414de827de35 my $l = (1, 2, 3); say (0, slip $l).perl;
19:31 committable6 Zoffix, ¦37d0e46: «(0, 1, 2, 3)»
19:31 Zoffix c: 37d0e4661414de827de35~1 my $l = (1, 2, 3); say (0, slip $l).perl;
19:31 committable6 Zoffix, ¦37d0e4661414de827de35~1: «(0, $(1, 2, 3))»
19:31 Zoffix And that commit is a perf improvemnt: https://github.com/rakudo/rakudo/commit/37d0e4661414de827de35835de2e77a81d1d23df
19:31 Zoffix hm
19:31 * Zoffix wishes TimToady was around right now :)
19:31 Zoffix He knows this stuff well
19:32 Zoffix dogbert17: now, I'm starting to feel it's wrong lol :)
19:32 Zoffix if it were (0, |$l) then yeah, it'd be right, but I think (0, slip $l) is a different thing
19:32 dogbert17 I haven't made the update just yet :)
19:32 Zoffix or something
19:32 Zoffix I don't know this stuff
19:33 * Zoffix finally admits and shuts up :D
19:33 Zoffix m: my $l = (1, 2, 3); say (0, $(slip $l)).perl;
19:33 camelia rakudo-moar 431ed4: OUTPUT: «(0, 1, 2, 3)␤»
19:33 Zoffix well, this is definitely wrong :)
19:34 dogbert17 one bug at least then :)
19:35 dogbert17 and lizmat is writing the weekly atm I think
19:35 dogbert17 commit: 2016.10 m: my $l = (1, 2, 3); say (0, $(slip $l)).perl;
19:35 committable6 dogbert17, ¦2016.10: «(0, $(1, 2, 3))»
19:36 Zoffix She doesn't have to fix it. Just file it ¯\_(ツ)_/¯
19:37 dogbert17 will do
19:37 Zoffix m: multi foo (@a) { say "here"; dd @a }; multi foo (+a) { say "there"; dd a  }; my $l = (1, 2, 3); foo $l
19:37 camelia rakudo-moar 431ed4: OUTPUT: «here␤(1, 2, 3)␤»
19:37 Zoffix m: multi foo (+a) { say "there"; dd a  }; my $l = (1, 2, 3); foo $l
19:37 camelia rakudo-moar 431ed4: OUTPUT: «there␤($(1, 2, 3),)␤»
19:38 * dogbert17 wonders how this bug managed to slip through our tests :)
19:39 Zoffix m: my $l = (1, 2, 3); say (0, $(item $(slip $l))).perl;
19:39 camelia rakudo-moar 431ed4: OUTPUT: «(0, 1, 2, 3)␤»
19:39 * Zoffix doesn't get how that one manages to slip all the way through the item stuff :|
19:40 Zoffix /win 8
19:40 buggable Zoffix, Thank you for entering Accidental /win Lottery! The next draw will happen in 2 weeks, 3 days, 4 hours, 19 minutes, and 18 seconds
19:40 Zoffix huh
19:42 lizmat /win 6
19:42 buggable lizmat, Thank you for entering Accidental /win Lottery! The next draw will happen in 2 weeks, 3 days, 4 hours, 17 minutes, and 42 seconds
19:42 dogbert17 lizmat: backlog a screenful plz
19:42 lizmat dogbert17: yeah, please RT it
19:43 dogbert17 will do
19:43 lizmat dogbert17: fwiw, I found several bugs today in set operators that have gone undetected for years  :-(
19:43 dogbert17 oops
19:44 [Coke] lizmat: but also, unreported.
19:44 lizmat m: Date (<=) Date   # now works
19:44 camelia rakudo-moar 431ed4: OUTPUT: «WARNINGS for <tmp>:␤Useless use of "(<=)" in expression "Date (<=) Date" in sink context (line 1)␤»
19:44 lizmat m: say Date (<=) Date   # now works
19:44 camelia rakudo-moar 431ed4: OUTPUT: «True␤»
19:44 [Coke] better that you found them than a user.
19:44 lizmat m: say Set (<=) Set  # now also works
19:44 camelia rakudo-moar 431ed4: OUTPUT: «True␤»
19:44 dogbert17 commit: 2016:10 say <a b> X~ |(1, 2, 3) #
19:44 committable6 dogbert17, ¦2016:10: «Cannot find this revision (did you mean “2016.10”?)»
19:45 dogbert17 commit: 2016.10 say <a b> X~ |(1, 2, 3) #
19:45 committable6 dogbert17, ¦2016.10: «(a123 b123)»
19:45 awwaiid lizmat: is that one of those twisty things of a set that contains itself so we can break mathematics?
19:48 lizmat awwaiid: it all boils down that the Set type object is not a set, it needs to be coerced to a set first  :-)
19:49 lizmat m: dd Set.Set
19:49 camelia rakudo-moar 431ed4: OUTPUT: «set(Set)␤»
19:50 geekosaur that sounds wrong
19:50 awwaiid but oh so right
19:51 geekosaur I mean, I read that as a Set containing a Set type object. but I'm not sure what a Set type object should be, aside from a special case or not-a-set
19:52 lizmat m: dd Int.Set   # does this make sense, geekosaur?
19:52 camelia rakudo-moar 431ed4: OUTPUT: «set(Int)␤»
19:52 geekosaur (and yes, this does fall down the Russell rabbit hole)
19:52 dogbert17 lizmat: RT #131344
19:52 synopsebot6 Link:  https://rt.perl.org/rt3/Public/Bug/Display.html?id=131344
19:52 lizmat dogbert17: noted
19:52 geekosaur same thing, a Set containing a type object. which arguably is not the same as a type object _and_ not the same as whatever a Set type object would mean when used as a Set
19:53 * dogbert17 was sorely tempted to write '[REGRESSION] behavior of slip has slipped'
19:53 Zoffix $ rakudobrew switch moar-2107.04.3
19:53 lizmat geekosaur: well, anything that's not a set is coerced to a set first before being compared with e.g. (<=)
19:53 geekosaur I mean, I could see a Set type object being autocoerced to an empty set in some circumstances... but *not* all. on the other hand, that is more or less what we expect numbers and strings to do, so.
19:53 Zoffix Sorry, I have no idea what 'moar-2107.04.3' is
19:54 Zoffix k'comn, rakudobrew! 100 year language and what not!
19:54 geekosaur all roads here are wrong, some subtly and some severely
19:54 lizmat the Set type object is *not* the empty set
19:54 lizmat one could argue that the Set type object is an implementation detail
19:55 Zoffix Imagine it called FooBar. You get same behaviour FooBar.Set => set(FooBar)
19:55 geekosaur I guess that means I didn';t just say that? so I guess I'll just back out
19:58 lizmat geekosaur: dealing with sets is difficult enough with one set of eyes: we can use all the eyes we can get!
20:00 Zoffix ugexe: FWIW, I see zef here failing with "no method .as on Seq": https://ci.appveyor.com/project/ufobat/Bailador/branch/master
20:00 Zoffix ugexe: I see some method .as in zef's sauce; looks like there's some path that leads to a Seq rather than whatever has .as object
20:01 ugexe thats also moar-2016.03
20:02 Zoffix oh lol
20:12 Ven joined #perl6-dev
20:16 Geth ¦ roast: cono++ created pull request #270: Rework IO-Socket-INET test to make it more self-sufficient
20:16 Geth ¦ roast: review: https://github.com/perl6/roast/pull/270
20:19 pmurias joined #perl6-dev
20:19 cono So ran a test one at a time, looks fine, but when I run all of them it segfaults after the third one :( Somewhere in thread.finish. I'll look into it but no guarantee that I can get it to work, so rely on you about thread stuff :)
20:20 pmurias awwaiid: the build instructions for rakudo.js will be "perl Configure --backends=js --gen-nqp", I just have to set up a few things for that to work
20:35 AlexDaniel /win 0
20:35 buggable AlexDaniel, Thank you for entering Accidental /win Lottery! The next draw will happen in 2 weeks, 3 days, 3 hours, 24 minutes, and 2 seconds
20:36 AlexDaniel and 2 seconds!
20:36 AlexDaniel oh that precision!
20:36 literal joined #perl6-dev
20:39 AlexDaniel buggable: but how many milliseconds?
20:41 Zoffix It's a secret :)
20:41 Zoffix buggable: lottery
20:42 Zoffix buggable: source
20:42 buggable Zoffix, See: https://github.com/zoffixznet/perl6-buggable
20:42 Zoffix buggable: draw
20:42 buggable Zoffix, The next Accidental /win Lottery draw will happen in 2 weeks, 3 days, 3 hours, 17 minutes, and 35 seconds. Currently have 13 ballots submitted by 10 users!
20:42 Zoffix hehe
20:44 AlexDaniel /win 42
20:44 buggable AlexDaniel, Thank you for entering Accidental /win Lottery! The next draw will happen in 2 weeks, 3 days, 3 hours, 15 minutes, and 15 seconds
20:44 AlexDaniel buggable: draw
20:44 buggable AlexDaniel, The next Accidental /win Lottery draw will happen in 2 weeks, 3 days, 3 hours, 15 minutes, and 9 seconds. Currently have 14 ballots submitted by 10 users!
20:47 Geth ¦ rakudo/newer-cat-handle: c86e8c9098 | (Zoffix Znet)++ | 2 files
20:47 Geth ¦ rakudo/newer-cat-handle: Sketch of IO::CatHandle
20:47 Geth ¦ rakudo/newer-cat-handle:
20:47 Geth ¦ rakudo/newer-cat-handle: Not workable yet, but stashing it so I could work on it from
20:47 Geth ¦ rakudo/newer-cat-handle: other boxes.
20:47 Geth ¦ rakudo/newer-cat-handle: review: https://github.com/rakudo/rakudo/commit/c86e8c9098
20:48 Geth ¦ roast/newer-cat-handle: d638e18587 | (Zoffix Znet)++ | S32-io/io-cathandle.t
20:48 Geth ¦ roast/newer-cat-handle: Start sketching CatHandle tests
20:48 Geth ¦ roast/newer-cat-handle: review: https://github.com/perl6/roast/commit/d638e18587
20:48 * Zoffix is done hacking for the day
21:37 AlexDaniel c: 2014.05~4 say 42
21:37 committable6 AlexDaniel, ¦2014.05~4: «42»
21:39 AlexDaniel there's still a hole between 2014.05 and 2015.02 I think, so stuff isn't bisectable yet… but it's getting close :)
21:40 dalek joined #perl6-dev
21:40 Zoffix :)
21:42 AlexDaniel actually, I'm not sure
21:42 AlexDaniel bisect: old=2014.01 say ‘hello’
21:42 bisectable6 AlexDaniel, Bisecting by exit code (old=2014.01 new=431ed4e). Old exit code: 1
21:42 bisectable6 AlexDaniel, bisect log: https://gist.github.com/7b30ac12b4461ce7bd2b3ead0bdb5899
21:42 bisectable6 AlexDaniel, (2015-02-23) https://github.com/rakudo/rakudo/commit/a2a1afb9e91a26fe125b80e4faa4626dd4d50300
21:42 AlexDaniel bisect: old=2014.01 say ∞
21:42 bisectable6 AlexDaniel, Bisecting by exit code (old=2014.01 new=431ed4e). Old exit code: 1
21:42 bisectable6 AlexDaniel, bisect log: https://gist.github.com/ff3d1a30beb2a053311064c39a26b007
21:42 bisectable6 AlexDaniel, (2015-12-18) https://github.com/rakudo/rakudo/commit/1f09c1e4f760a138208de229e0e42648b362ab07
21:43 AlexDaniel to be honest I have no idea how much is done
21:43 AlexDaniel it's just not done yet…
21:57 Zoffix heh... "because it's Friday and I'm feeling expansive" I didn't realize we got ∞ kinda on a whim :D
21:57 * Zoffix likes ∞ a lot
22:02 Geth ¦ rakudo/nom: 379ef6f1d4 | (Zoffix Znet)++ | src/core/IO/Handle.pm
22:02 Geth ¦ rakudo/nom: Move $close handling in IO::Handle.comb/split
22:02 Geth ¦ rakudo/nom:
22:02 Geth ¦ rakudo/nom: To .slurp
22:02 Geth ¦ rakudo/nom: (Also add a TODO comment for when we get Cat)
22:02 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/379ef6f1d4
22:06 AlexDaniel bisect: say −42
22:06 bisectable6 AlexDaniel, On both starting points (old=2015.12 new=431ed4e) the exit code is 0 and the output is identical as well
22:06 bisectable6 AlexDaniel, Output on both points: «-42»
22:06 AlexDaniel bisect: old=2014.01 say −42
22:06 bisectable6 AlexDaniel, Bisecting by exit code (old=2014.01 new=431ed4e). Old exit code: 1
22:06 bisectable6 AlexDaniel, bisect log: https://gist.github.com/4df2a2749d9ab4db1b5f8a1f99891760
22:06 bisectable6 AlexDaniel, There are 2 candidates for the first “new” revision. See the log for more details
22:06 AlexDaniel https://github.com/rakudo/rakudo/commit/332ea4e3b5b7c63a7c57bdb23b04d9f29ee099a4
22:06 AlexDaniel oh, but that's very recent
22:10 Zoffix 1-day less recent than infinity :)
22:10 Zoffix And it's quite the opositite, I hate − and I want it removed :D
22:23 AlexDaniel bisect: old=2014.01 say π
22:23 bisectable6 AlexDaniel, Bisecting by exit code (old=2014.01 new=431ed4e). Old exit code: 1
22:26 bisectable6 AlexDaniel, bisect log: https://gist.github.com/34c6d5880c782814f6bdd2c86aaba5e0
22:26 bisectable6 AlexDaniel, There are 407 candidates for the first “new” revision. See the log for more details
22:26 AlexDaniel told ya, not done yet :)
22:28 nebuchadnezzar Hello
22:28 AlexDaniel o/
22:29 Zoffix \o
22:31 nebuchadnezzar I'm trying to update the Debian packages and I wonder if its normal that --libdir is not passed to gcc: https://paste.debian.net/934241/, because the moar binary ends with “libmoar.so => not found” https://paste.debian.net/934242/, any hints?
22:32 lizmat And another Perl 6 Weekly hits the Net: https://p6weekly.wordpress.com/2017/05/22/2017-21-yap6b/
22:33 lizmat that took a little longer than I expected
22:33 * lizmat decides to call it a day
22:33 lizmat good night, #perl6-dev!
22:35 Zoffix night
22:41 Zoffix nebuchadnezzar: IIRC moar recentlyish had some changes done to its build chain. Maybe worth mentioning in #moarvm to see if some eyes (timotimo) spot it
22:42 nebuchadnezzar I think I found, I'll report there
22:50 timotimo uh oh
22:50 timotimo i've done a boo boo again?
23:41 MasterDuke joined #perl6-dev
23:41 BenGoldberg joined #perl6-dev

| Channels | #perl6-dev index | Today | | Search | Google Search | Plain-Text | summary