Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6-dev, 2016-11-26

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

All times shown according to UTC.

Time Nick Message
00:59 seatek joined #perl6-dev
01:06 dalek nqp: f2c0405 | (Zoffix Znet)++ | .gitignore:
01:06 dalek nqp: Add tools/build/install-jvm-runner.pl to .gitignore
01:06 dalek nqp:
01:06 dalek nqp: So it does not accidentally get committed
01:06 dalek nqp: review: https://github.com/perl6/nqp/commit/f2c0405aeb
01:20 dalek nqp: 1db2689 | (Zoffix Znet)++ | / (2 files):
01:20 dalek nqp: Remove trailing whitespace
01:20 dalek nqp: review: https://github.com/perl6/nqp/commit/1db2689a51
01:27 dalek nqp: ddebbfa | (Zoffix Znet)++ | / (2 files):
01:27 dalek nqp: Make sprintf %f, %e, and %g handle negative zero
01:27 dalek nqp:
01:27 dalek nqp: Negative zeros in those formats produce a string starting with -, as corroborated by
01:27 dalek nqp: Perl 5's sprintf.
01:27 dalek nqp:
01:27 dalek nqp: Fixes RT#128821: https://rt.perl.org/Ticket/Display.html?id=128821
01:27 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=128821
01:27 dalek nqp: review: https://github.com/perl6/nqp/commit/ddebbfa959
01:34 dalek rakudo/nom: f1263ab | (Zoffix Znet)++ | tools/build/NQP_REVISION:
01:34 dalek rakudo/nom: (bump nqp) Make sprintf %f, %e, and %g handle negative zero
01:34 dalek rakudo/nom:
01:34 dalek rakudo/nom: Negative zeros in those formats produce a string starting with -, as corroborated by
01:34 dalek rakudo/nom: Perl 5's sprintf.
01:34 dalek rakudo/nom:
01:34 dalek rakudo/nom: Fixes RT#128821: https://rt.perl.org/Ticket/Display.html?id=128821
01:34 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=128821
01:34 dalek rakudo/nom: NQP commit: https://github.com/perl6/nqp/commit/ddebbfa959a67e250099fc685
01:34 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/f1263abc27
02:00 dalek roast: 795fc1f | (Zoffix Znet)++ | S32-num/negative-zero.t:
02:00 dalek roast: Test sprintf can handle negative zero
02:00 dalek roast:
02:00 dalek roast: RT#128821: https://rt.perl.org/Public/Bug/Display.html?id=128821
02:00 dalek roast:
02:00 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=128821
02:00 dalek roast: N.B.: I have more negative zero tests to write, which is why I
02:00 dalek roast: started this separate file for negative zero.
02:00 dalek roast: review: https://github.com/perl6/roast/commit/795fc1f3b3
02:01 dalek rakudo/nom: 2efeda0 | (Zoffix Znet)++ | t/spectest.data:
02:01 dalek rakudo/nom: Add S32-num/negative-zero.t to list of test files to run
02:01 dalek rakudo/nom:
02:01 dalek rakudo/nom: Tests negative zeros are correctly handled and propagated by
02:01 dalek rakudo/nom: various routines.
02:01 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/2efeda0597
02:19 travis-ci joined #perl6-dev
02:19 travis-ci Rakudo build passed. Zoffix Znet '(bump nqp) Make sprintf %f, %e, and %g handle negative zero
02:19 travis-ci https://travis-ci.org/rakudo/rakudo/builds/178976598 https://github.com/rakudo/rakudo/compare/15f51a52bd41...f1263abc2741
02:19 travis-ci left #perl6-dev
02:24 kalkin-_ joined #perl6-dev
02:49 ilbot3 joined #perl6-dev
02:49 Topic for #perl6-dev is now Perl 6 language and compiler development 2.0 | Logs at http://irclog.perlgeek.de/perl6-dev/today
02:55 travis-ci joined #perl6-dev
02:55 travis-ci Rakudo build passed. Zoffix Znet 'Add S32-num/negative-zero.t to list of test files to run
02:55 travis-ci https://travis-ci.org/rakudo/rakudo/builds/178979087 https://github.com/rakudo/rakudo/compare/f1263abc2741...2efeda059771
02:55 travis-ci left #perl6-dev
03:05 nchambers joined #perl6-dev
03:20 ugexe m: "xxx".subst(/no second argument/); # infinite loop from Cool subst candidate https://github.com/rakudo/rakudo/blob/nom/src/core/Cool.pm#L225
03:20 camelia rakudo-moar 2efeda: OUTPUT«(timeout)Potential difficulties:␤    Space is not significant here; please use quotes or :s (:sigspace) modifier (or, to suppress this warning, omit the space, or otherwise change the spacing)␤    at <tmp>:1␤    ------> "xxx".subst(/no⏏ secon…»
03:22 dalek rakudo/nom: 4663d43 | (Zoffix Znet)++ | src/Perl6/Grammar.nqp:
03:22 dalek rakudo/nom: Reject synthetics in numbers in colonpairs in :42foo format
03:22 dalek rakudo/nom:
03:22 dalek rakudo/nom: Currently, a colon pair like ":7\x[308]a".EVAL would appear to parse correctly,
03:22 dalek rakudo/nom: but in reality, the number would get parsed only until the first synthetic.
03:22 dalek rakudo/nom:
03:22 dalek rakudo/nom: Fix by going over each char of the numeral and checking it `eq` to its .ord.chr
03:23 dalek rakudo/nom:
03:23 dalek rakudo/nom: Fixes RT#128551: https://rt.perl.org/Ticket/Display.html?id=128551
03:23 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=128551
03:23 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/4663d43130
03:23 dalek roast: b362890 | (Zoffix Znet)++ | S02-literals/pairs.t:
03:23 dalek roast: Test synthetic numerals in colon pairs in :42foo format throw
03:23 dalek roast:
03:23 dalek roast: RT#128551: https://rt.perl.org/Ticket/Display.html?id=128551
03:23 dalek roast:
03:23 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=128551
03:23 dalek roast: Commit with fix: https://github.com/rakudo/rakudo/commit/4663d43130
03:23 dalek roast: review: https://github.com/perl6/roast/commit/b36289063e
03:35 viki ugexe: spectesting a fix
03:44 dalek rakudo/nom: 0a874fb | (Zoffix Znet)++ | src/core/ (2 files):
03:44 dalek rakudo/nom: Fix infinit loop in .subst with no arguments
03:44 dalek rakudo/nom:
03:44 dalek rakudo/nom: If .subst method is called without arguments, the matching multi is one
03:44 dalek rakudo/nom: provided by Cool. It coerces the arg to Stringy and calls .subst method
03:44 dalek rakudo/nom: on it, passing args via a Capture. However, if that Capture doesn't satisfy
03:44 dalek rakudo/nom: Stringy candidates, we get the Cool candidate again, and the loop repeats.
03:44 dalek rakudo/nom:
03:44 dalek rakudo/nom: Fix by making Cool.subst an only method and moving the proto to Str.subst,
03:44 dalek rakudo/nom: making only Str's candidates available.
03:44 dalek rakudo/nom:
03:44 dalek rakudo/nom: ugexe++ for spotting:
03:44 dalek rakudo/nom: https://irclog.perlgeek.de/perl6-dev/2016-11-26#i_13630780
03:44 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/0a874fb6a4
03:45 dalek roast: 7c3a384 | (Zoffix Znet)++ | S05-substitution/subst.t:
03:45 dalek roast: Remove trailing whitespace
03:45 dalek roast: review: https://github.com/perl6/roast/commit/7c3a3841eb
03:46 dalek roast: 4eede1c | (Zoffix Znet)++ | S05-substitution/subst.t:
03:46 dalek roast: Test .subst() with no args does not infiniloop
03:46 dalek roast:
03:46 dalek roast: Test for fix https://github.com/perl6/roast/commit/7c3a3841eb
03:46 dalek roast: IRC: https://irclog.perlgeek.de/perl6-dev/2016-11-26#i_13630780
03:46 dalek roast: review: https://github.com/perl6/roast/commit/4eede1cba0
03:46 viki ugexe: done
03:46 ugexe viki++
04:19 dalek rakudo/nom: e2587cd | (Zoffix Znet)++ | src/core/allomorphs.pm:
04:19 dalek rakudo/nom: Teach val() to handle negative zero nums correctly
04:19 dalek rakudo/nom:
04:19 dalek rakudo/nom: Part of fixing RT#128897:
04:19 dalek rakudo/nom: https://rt.perl.org/Ticket/Display.html?id=128897
04:19 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=128897
04:19 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/e2587cd012
04:20 dalek roast: 7fed1c0 | (Zoffix Znet)++ | S32-num/negative-zero.t:
04:20 dalek roast: Test val() can parse a negative zero num
04:20 dalek roast:
04:20 dalek roast: Implemented in commit:
04:20 dalek roast: https://github.com/rakudo/rakudo/commit/e2587cd012
04:20 dalek roast: review: https://github.com/perl6/roast/commit/7fed1c00a0
04:38 ugjka joined #perl6-dev
04:48 dalek roast: f0269f4 | (Zoffix Znet)++ | S32-num/negative-zero.t:
04:48 dalek roast: Test val() can handle U+2212 minus with negative zero
04:48 dalek roast: review: https://github.com/perl6/roast/commit/f0269f46fb
05:02 dalek rakudo/nom: a9654c0 | (Zoffix Znet)++ | src/core/Str.pm:
05:02 dalek rakudo/nom: Teach Str.Num to handle negative zeros
05:02 dalek rakudo/nom:
05:02 dalek rakudo/nom: While val will return[^1] negative zeros when the string contains a num,
05:02 dalek rakudo/nom: it won't do so for other types. Since we do have the sign information
05:02 dalek rakudo/nom: available in the string itself and the method requests a Num, attach the
05:02 dalek rakudo/nom: correct sign to zeros, even if the string does not contain a num.
05:02 dalek rakudo/nom:
05:02 dalek rakudo/nom: Fixes RT#128897: https://rt.perl.org/Ticket/Display.html?id=128897
05:02 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=128897
05:02 dalek rakudo/nom:
05:02 dalek rakudo/nom: [1] https://github.com/rakudo/rakudo/commit/e2587cd012
05:05 dalek roast: cbc959f | (Zoffix Znet)++ | S32-num/negative-zero.t:
05:05 dalek roast: Test Str.Num correctly handles negative zeros
05:05 dalek roast:
05:05 dalek roast: RT#128897: https://rt.perl.org/Ticket/Display.html?id=128897
05:05 dalek roast: Fix commit: https://github.com/rakudo/rakudo/commit/e2587cd012
05:05 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=128897
05:05 dalek roast: review: https://github.com/perl6/roast/commit/cbc959fc22
05:06 viki https://github.com/rakudo/rakudo/commit/a9654c0d8ecb64db3bd7c465f81aea84cef00577
05:11 TimToady joined #perl6-dev
05:29 MasterDuke hmm, how do i open a subprocess for writing in nqp...?
05:29 yoleaux2 25 Nov 2016 11:52Z <viki> MasterDuke: can you look at this ticket? IIRC we decided this was a DITHWIDT and wasn't worth a fix, right? https://rt.perl.org/Ticket/Display.html?id=128551
05:30 viki disregard that
05:51 dalek rakudo/nom: c797d3f | (Zoffix Znet)++ | src/core/Date.pm:
05:51 dalek rakudo/nom: Fix sleep() with huge values failing to sleep
05:51 dalek rakudo/nom:
05:51 dalek rakudo/nom: Imagine you get into a cryogenic chamber, hoping to be whisked into the distant
05:51 dalek rakudo/nom: future exactly 317,097,919 years, 50 weeks, 2 days, 17 hours, 46 minutes, and
05:52 dalek rakudo/nom: 40 seconds in the future. The control software is written in Perl 6 and the lazy
05:52 dalek rakudo/nom: programmer just stuck a sleep() function in there. Unfortunately, you will be
05:52 dalek rakudo/nom: defrosted instantly, because there's a bug with such huge numbers and sleep()
05:52 dalek rakudo/nom: doesn't sleep at all! Of course, another lazy programmer may assume such huge
05:52 dalek rakudo/nom: amounts of time are practically infinity and fix the bug with such assumption,
05:52 dalek rakudo/nom: but then you will be frozen... forever! Therefore, we must fix the sleep()
05:52 dalek rakudo/nom: function to handle huge chunks of time by splitting them into several sleeps in
05:52 dalek rakudo/nom: intervals of 1e16 seconds (approximatelly max that JVM can handle), thereby
05:52 dalek rakudo/nom: guaranteeing all cryogenic chambers will function correctly.
05:52 viki rakudo/nom:
05:52 viki rakudo/nom: Fixes RT130170: #https://rt.perl.org/Ticket/Display.html?id=130170
05:52 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=130170
05:52 viki rakudo/nom:
05:52 viki rakudo/nom: Review: https://github.com/rakudo/rakudo/commit/c797d3ff5082c6738
05:53 dalek roast: 0fc0919 | (Zoffix Znet)++ | S29-context/sleep.t:
05:53 dalek roast: Test sleep() with huge values actually sleeps
05:53 dalek roast:
05:53 dalek roast: RT130170: #https://rt.perl.org/Ticket/Display.html?id=130170
05:53 dalek roast:
05:53 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=130170
05:53 dalek roast: Fix commit: https://github.com/rakudo/rakudo/commit/c797d3ff5082c6738
05:53 dalek roast: review: https://github.com/perl6/roast/commit/0fc0919c7f
05:58 MasterDuke nqp: my $f := nqp::open("|sqlite3 tmp.sqlite", :w); nqp::sayfh($f, "CREATE TABLE allocators(id INTEGER PRIMARY KEY ASC, name TEXT);"); nqp::closefh($f);
05:58 camelia nqp-moarvm: OUTPUT«Invalid open mode for file |sqlite3 tmp.sqlite: 1␤   at <tmp>:1  (<ephemeral file>:<mainline>)␤ from gen/moar/stage2/NQPHLL.nqp:1437  (/home/camelia/rakudo-m-inst-2/share/nqp/lib/NQPHLL.moarvm:eval)␤ from gen/moar/stage2/NQPHLL.nqp:1640  (/home/camelia/rakud…»
05:58 timotimo yeah, you spell that very differently in nqp-land
05:59 MasterDuke ah, do you know the incantation?
05:59 timotimo nope, but i can tell you where to look
06:00 timotimo actually ...
06:00 MasterDuke that would help
06:01 timotimo it might be a tad difficult to tear out the code from Proc::Async and throw away the rakudo-y bits
06:01 timotimo hm. actually ... spawnprocasync might be easy-ish
06:01 timotimo yeah
06:02 timotimo set a few options in a hash, grab the $!process_handle, then you can nqp::asyncwritebytes as well as nqp::closefh later
06:03 MasterDuke ahhh. i saw spawnprocasync in the nqp docs, but it had no explanation
06:03 MasterDuke thanks, i'll give that a go
06:03 timotimo it's rather a little bit involved
06:04 timotimo seems like you need to also create a queue for the events to go into
06:05 timotimo events like "oi, there's some bytes for ye!"
06:05 timotimo but! you're not required to capture stdout and stderr at all, so you can just put in a random ConcurrentBlockingQueue and never look at it again
06:06 timotimo we can't have a process that we write stuff into without using Proc::Async, can we?
06:06 timotimo i think we're even going to re-implement the non-async proc ops based on the async one, it was rumored
06:07 MasterDuke hm, nqp::spawnprocasync isn't ever used in nqp
06:07 timotimo yeah
06:07 timotimo why would it be :)
06:07 timotimo i don't think anything in nqp uses asynchronous stuff at all. though it could very well be that we have tests in nqp's test suite for that stuff
06:08 timotimo hah, it's on the TODO list that i didn't even know existed
06:08 MasterDuke well, there are no tests that use nqp::spawnprocasync
06:08 timotimo yeah
06:08 MasterDuke ha, yeah, i see that TODO
06:09 timotimo the neat thing is, that the interface to spawnprocasync doesn't require you to have an implementation of Promise or of Supply
06:09 timotimo because all it does is push callbacks into the scheduler's queue when things happen
06:10 MasterDuke this looks annoying to try and use in nqp-land
06:11 timotimo *shrug*, it could be much worse
06:12 viki Hm, looks like in the middle of making that joke I also found a bug in polymod
06:12 viki m: my $seconds = 1e31; dd $seconds.Num.polymod: 1e16 xx *;
06:12 camelia rakudo-moar c797d3: OUTPUT«(0e0, 1e+15).Seq␤»
06:12 viki m: my $seconds = 1e32; dd $seconds.Num.polymod: 1e16 xx *;
06:12 camelia rakudo-moar c797d3: OUTPUT«(0e0, 0e0, 1e0).Seq␤»
06:12 timotimo https://github.com/rakudo/rakudo/blob/nom/src/core/ThreadPoolScheduler.pm#L42 - here's how you're expected to use the queue, i.e. what kind of data lands in these $task pieces
06:12 viki Or maybe I need to go bed
06:13 timotimo well, the second result does look wrong, right?
06:13 viki oh "If the divisors are given as a lazy list, runs until the remainder is 0 or the list of divisors is exhausted"
06:13 timotimo because 1e0 * 1e16 isn't 1e32
06:13 timotimo oh, you mean that there's 0e0 in front
06:14 viki Well, I thought this would give me the total back:
06:14 viki m: my $seconds = 1e33; say sum gather { .say, take $_*1e16 for grep *.so,  $seconds.Num.polymod: 1e16 xx *; }
06:14 camelia rakudo-moar c797d3: OUTPUT«10␤1e+17␤»
06:14 viki It kinda bugs me my fix for sleep doesn't actually sleep for correct number of seconds, regardless of how unrealistic that scenario is...
06:15 timotimo you're forgetting to increase the multiplicand
06:15 timotimo like, you're trying to calculate 512 as 5*10+1*10+2*10
06:21 ZzZombo joined #perl6-dev
06:51 viki right, that won't help me then, 'cause I need to sleep in intervals of at most 1e16 seconds.
06:51 * viki goes for a gather/take thing
06:57 dalek rakudo/nom: 7c5ea31 | (Zoffix Znet)++ | src/core/Date.pm:
06:57 dalek rakudo/nom: sleep() for correct number of seconds for values above 1e16 seconds
06:57 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/7c5ea31800
06:58 viki and speaking of sleep....
06:58 * viki drops to bed
08:00 geekosaur joined #perl6-dev
08:02 RabidGravy joined #perl6-dev
09:49 vendethiel joined #perl6-dev
09:51 FROGGS joined #perl6-dev
09:52 FROGGS o/
11:26 vendethiel \o
12:17 nine What do you people think about merging lexical_module_load right now? Next release is a couple weeks off giving us quite some time to determine whether it should be reverted or not.
12:17 nine I fear we won't make any progress any other way.
12:19 lizmat nine: +1
12:26 jnthn Think I've asked this before but forgot the answer: did we try an ecosystem test with it on?
12:27 jnthn One other prerequisite is that we have a migration guide that we can point people at.
12:27 jnthn That explains the change and what they need to do.
12:29 jnthn Preferably on doc.perl6.org
12:30 dalek roast: 08ffd2f | usev6++ | S29-context/sleep.t:
12:30 dalek roast: Fudge test for RT #130170 for JVM
12:30 dalek roast: review: https://github.com/perl6/roast/commit/08ffd2f682
12:30 dalek roast: 4cb14ff | usev6++ | S32-io/open.t:
12:30 dalek roast: Unfudge test for $fh.get after $fh.seek on JVM
12:30 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=130170
12:30 dalek roast: review: https://github.com/perl6/roast/commit/4cb14ff76e
12:50 MasterDuke before sleeping i spent a little time looking at nqp::[spawnprocasync, syncpipe, shell, spawn, etc.] and never managed to find a simple way to open a subprocess for writing
12:51 MasterDuke spawnprocasync is complicated, shell/spawn would block, and open("|sqlite3 tmp.sqlite", :w) just isn't implemented
12:52 MasterDuke any suggestions before i take the time to figure out spawnprocasync?
12:54 nine jnthn: Task::Star installs successfully and my Bailador example works. It's been a while since I've seen issues running this branch. But I'm not much of a Perl 6 _user_ :)
12:54 nine jnthn: I'll write something up for docs
12:57 jnthn nine: At some point we had something that could smoke the whole module ecosystem with a particular Rakudo change.
12:58 timotimo well, what happens when we just run "panda smoke" or whatever it was called
12:58 timotimo or do you need to actually grab emmentaler for that?
12:58 jnthn Ah, emmentaler...
12:58 jnthn I've forgotten what the state of the art was there
12:58 timotimo the fate of the start
12:58 nine That sounds quite handy!
12:58 jnthn Indeed
12:58 jnthn lunch &
13:09 nine First draft: https://gist.github.com/niner/70f7b46eefb7e22af78d896bea11efeb
13:09 nine Comments/suggestions/patches very welcome :)
13:14 timotimo nine: it says "will complain about 'Foo' already being defined", but isn't this about 'Bar' already being defined?
13:14 timotimo or am i Just Wrong?
13:14 AlexDaniel joined #perl6-dev
13:19 nine timotimo: I should rename Foo in the second example so it is not confused with the Foo in the first
13:20 nine fixed it
13:21 timotimo oh, yes
13:37 RabidGravy nine, looks good
13:39 RabidGravy nine, is that actually in nom now?  I was going to test against  the branch but time got away from me
13:58 dalek rakudo/nom: bdc7a91 | (Zoffix Znet)++ | src/core/Date.pm:
13:58 dalek rakudo/nom: Document the 1e16 magic number
13:58 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/bdc7a91100
14:05 viki +1 on lexical_module_load merge
14:23 nine RabidGravy: not yet merged
14:23 RabidGravy right
14:40 nine emmentaler seems to be severely bitrotted
14:40 timotimo i can imagine
14:42 nine Ah, got it to do something now :)
14:43 nine But....it installs everything into my rakudo installation instead of installlib
14:44 nine ah, better
14:57 nine After building Linenoise, every module fails with X::AdHoc.new(payload => "setcodeobj needs a code ref")
15:06 * TimToady sneaks back into the sausage factory via the back door
15:06 yoleaux2 17 Nov 2016 14:36Z <viki> TimToady: was − (fancy minus) meant to work everywhere a regular minus works or just with literals? "−42".Numeric and sub (−4) {} fail for example. Should those be fixed?
15:06 yoleaux2 21 Nov 2016 12:25Z <viki> TimToady: never mind. japhb said the reason is mostly due to val() being done prior to addition of fancy pants minus ( https://irclog.perlgeek.de/perl6-dev/2016-11-18#i_13587383 )
15:06 yoleaux2 21 Nov 2016 12:51Z <lizmat> TimToady: any thoughts on val() also processing enums in scope ?
15:06 yoleaux2 21 Nov 2016 12:54Z <lizmat> TimToady: aka: having val("True") returning Bool::True
15:06 TimToady darn, tripped the alarm...
15:06 viki :D
15:14 viki TimToady: for the enum thing... We were trying to solve the issue of sub MAIN being able to take enums(). We did with https://github.com/rakudo/rakudo/blob/d1ef82ac189ead6dab42fe246277f22ecd3e53a6/src/core/Main.pm#L18-L22  but now we can't have Str $foo that can take, say, value "False", since it'd be converted to the Bool and not a Str. So as a solution to that, we thought of EnumStr allomorph, but that
15:14 viki also has a problem that my $foo = set <foo bar ber>; say so $foo{'foo'} would break the minute you bring an enum `foo` into scope, since it'll now become EnumStr instead of Str. So... that's where we're at ATM.
15:16 TimToady yeah, enums as text literals are...fraught...
15:16 viki Maybe just limit them to handling sub MAIN stuff?
15:19 * TimToady is unfortunately still recovering from the turn-brain-to-snot virus he came home from Portugal with, so is not entirely compos mentis
15:46 diakopter yeck
16:05 [TuxCM] This is Rakudo version 2016.11-65-gbdc7a91 built on MoarVM version 2016.11-20-g0f7277a
16:05 [TuxCM] csv-ip5xs        3.285
16:05 [TuxCM] test            14.610
16:05 [TuxCM] test-t           6.588
16:05 [TuxCM] csv-parser      14.504
16:11 diakopter I still don't know what "test" or "test-t" or "csv-parser" refer to. I can guess what csv-ip5xs is, and why it's the baseline
16:16 viki .ask bartolin what's the failure in this test (sleep huge values)? Seems to work fine in camelia. https://github.com/perl6/roast/commit/08ffd2f682bc376f1219f1802a96fb67fb20b397
16:16 yoleaux2 viki: I'll pass your message to bartolin.
16:16 viki bbrrr
16:16 viki It'd help if I run the test with j: and not m: >>:(
16:16 viki .tell bartolin never mind. Reproduced :)
16:16 yoleaux2 viki: I'll pass your message to bartolin.
16:18 [TuxCM] diakopter, test is the first pure-perl6 draft, test-t is the full-featured Text::CSV in perl6, csv-parser is CSV::Parser from Tony (perl6)
16:20 diakopter [TuxCM]: ok thanks. how big are the csv inputs and how complex/conformat are they
16:28 nine diakopter: for i in $(seq 1 10000); do echo 'hello,","," ",world,"!"'; done > /tmp/hello.csv
16:30 diakopter giggle
16:30 diakopter so I guess the timings are in nanoseconds
16:31 nine if only
16:40 diakopter a meaningful comparison would compare to the 45 fastest json readers in npm
16:40 diakopter er. csv.
16:44 [TuxCM] diakopter, http://tux.nl/Talks/CSV6/speed5.html
16:44 [TuxCM] additions welcome
16:44 [TuxCM] all tests and sources therefor available in https://github.com/Tux/CSV
16:47 diakopter 76 distributions mention CSV in CPAN; 898 packages mention CSV in NPM (11.8 ratio); 34,500 total distributions in CPAN; 360,000 total packages in NPM (10.4 ratio).  Similar ratio seems similar
16:48 [TuxCM] patches welcome :)
16:49 diakopter fsdo patches, which may or may not be feasible or useful or worthwhile
16:55 nine Seems like emmentaler stumbles over modules that need LibraryMake. On rakudo nom btw.
16:58 dalek rakudo/nom: 085145f | (Zoffix Znet)++ | src/core/Num.pm:
16:58 dalek rakudo/nom: Teach infix:<===> to handle signed zeros correctly
16:58 dalek rakudo/nom:
16:58 dalek rakudo/nom: And since cmp uses ===, we also need make it watch for zeros, since it
16:58 dalek rakudo/nom: needs to handle both zeros as same.
16:58 dalek rakudo/nom:
16:58 dalek rakudo/nom: Fixes RT#128395: https://rt.perl.org/Ticket/Display.html?id=128395
16:58 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=128395
16:58 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/085145f3a5
16:58 dalek roast: f67de6c | (Zoffix Znet)++ | S32-num/negative-zero.t:
16:58 dalek roast: Test cmp and === on signed num zeros
16:58 dalek roast:
16:58 dalek roast: RT#128395: https://rt.perl.org/Ticket/Display.html?id=128395
16:58 dalek roast: Fix commit: https://github.com/rakudo/rakudo/commit/085145f3a5
16:58 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=128395
16:58 dalek roast: review: https://github.com/perl6/roast/commit/f67de6cf58
17:10 japhb Zoffix++  # Edge cases matter too
17:11 viki m: dd [ atan2(-0e0, -1), atan2(0e0, -1) ]; # since they can give wildly different results
17:11 camelia rakudo-moar 085145: OUTPUT«[-3.14159265358979e0, 3.14159265358979e0]␤»
17:17 dalek rakudo/nom: 938d4d8 | (Zoffix Znet)++ | src/core/Date.pm:
17:17 dalek rakudo/nom: Use smaller value as upper edge of super-large sleeps
17:17 dalek rakudo/nom:
17:17 dalek rakudo/nom: nqp::sleep(1e16) actually exits instantly on JVM and the only reason
17:17 dalek rakudo/nom: sleep(Inf) worked was because it was continuously looping.
17:17 dalek rakudo/nom:
17:17 dalek rakudo/nom: Fix by using smaller 1e15 value.
17:17 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/938d4d8cd0
17:17 dalek roast: db4efab | (Zoffix Znet)++ | S29-context/sleep.t:
17:17 dalek roast: Revert "Fudge test for RT #130170 for JVM"
17:17 dalek roast:
17:17 dalek roast: This reverts commit 08ffd2f682bc376f1219f1802a96fb67fb20b397.
17:17 dalek roast:
17:17 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=130170
17:17 dalek roast: Issue fixed in commit: https://github.com/rakudo/rakudo/commit/938d4d8cd0
17:17 dalek roast: review: https://github.com/perl6/roast/commit/db4efabe74
17:37 viki ZOFVM: Files=1204, Tests=130206, 137 wallclock secs (20.37 usr  2.93 sys + 2414.20 cusr 202.46 csys = 2639.96 CPU)
17:54 dogbert17 hmm: # Failed test 'huge values to sleep() work'
17:54 dogbert17 # at /home/dogbert/repos/rakudo/t/spec/packages/Test/Util.pm (Test::Util) line 76
17:54 dogbert17 #      got out: "Fail\n"
17:54 dogbert17 # expected out: ""
17:56 dogbert17 This is Rakudo version 2016.11-67-g938d4d8 built on MoarVM version 2016.11-20-g0f7277a
17:57 viki dogbert17: what happens when you run perl6 -e 'use nqp; nqp::sleep(1e15)'
17:57 viki and try with 1e14 and 1e13, and etc, until it actually doesn't exit immediately
17:57 dogbert17 it exists immediately
17:58 dogbert17 as do 14 and 13
17:58 viki Which exponent doesn't exit immediatelly? This is on a 32-bit box, right?
17:58 dogbert17 1e9 stops the program from exiting immediately
17:58 dogbert17 now it just hangs
17:58 viki good
17:59 dogbert17 classic note :) I'm on 32 bit
17:59 geekosaur sounds about right...
18:01 dogbert17 precision wise?
18:02 geekosaur well, overflow-wise
18:02 dogbert17 ah, of course
18:02 lizmat_ joined #perl6-dev
18:04 * dogbert17 wonders if he's the only one still running a vm in 32 bit
18:04 viki my dev box at $work is 32-bit
18:05 dogbert17 viki: perfect, please try t/spec/S32-num/power.rakudo.moar and t/spec/S04-declarations/constant.rakudo.moar when you get a chance
18:06 dogbert17 they fail on 32 bit
18:07 dalek rakudo/nom: 2f72fa0 | (Zoffix Znet)++ | src/core/Date.pm:
18:07 dalek rakudo/nom: Use smaller base interval for ultra-large sleeps
18:07 dalek rakudo/nom:
18:07 dalek rakudo/nom: 1e15 causes overflow on 32-bit boxes[^1]. Fix by using 1e9 intervals
18:07 dalek rakudo/nom: that don't.
18:07 dalek rakudo/nom:
18:07 dalek rakudo/nom: dogbert17++ for spotting
18:07 dalek rakudo/nom:
18:07 dalek rakudo/nom: [1] https://irclog.perlgeek.de/perl6-dev/2016-11-26#i_13633257
18:07 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/2f72fa09cf
19:57 ___crlz___ joined #perl6-dev
20:22 ___crlz___ joined #perl6-dev
21:33 gfldex lizmat: for the next weekly it may be noteworthy to point as trait_mod declarator for those who want to define their own traits https://docs.perl6.org/type/Sub#index-entry-trait__mod_%28declarator%29
21:34 lizmat gfldex: is it noteworthy because it is new?
21:37 AlexDaniel lizmat: if changes to our exponentially increasing number of bots are still worth mentioning, unicodable6 is a new bot
21:38 lizmat unicodable6: help
21:38 unicodable6 lizmat, Just type any unicode character or part of a character name. Alternatively, you can also provide a code snippet or a regex
21:38 lizmat unicodable6: set
21:38 unicodable6 lizmat, U+0088 CHARACTER TABULATION SET [Cc] (ˆ)
21:38 unicodable6 lizmat, U+008A LINE TABULATION SET [Cc] (Š)
21:38 unicodable6 lizmat, U+0093 SET TRANSMIT STATE [Cc] (“)
21:38 unicodable6 lizmat, https://gist.github.com/5587a8d9afd8860dfee2995e7a84273b
21:38 AlexDaniel lizmat: https://irclog.perlgeek.de/perl6/2016-11-21#i_13608245
21:39 lizmat unicodable6:  VIDEOCASSETTE    # :-)
21:39 unicodable6 lizmat, U+0056 LATIN CAPITAL LETTER V [Lu] (V)
21:39 unicodable6 lizmat, U+0049 LATIN CAPITAL LETTER I [Lu] (I)
21:39 unicodable6 lizmat, U+0044 LATIN CAPITAL LETTER D [Lu] (D)
21:39 unicodable6 lizmat, https://gist.github.com/6cc4c05b1e2b5544956bd56a6bae499f
21:40 jnthn Hm, how does it decide whether to search by char name, or to show the chars in the string?
21:40 lizmat unicodable6: U+1F4FC
21:40 unicodable6 lizmat, U+0055 LATIN CAPITAL LETTER U [Lu] (U)
21:40 unicodable6 lizmat, U+002B PLUS SIGN [Sm] (+)
21:40 unicodable6 lizmat, U+0031 DIGIT ONE [Nd] (1)
21:40 unicodable6 lizmat, https://gist.github.com/cf0944f3a416904e1ad3d1ac5b457ac3
21:40 AlexDaniel lizmat: ok, that's a TODO
21:40 lizmat ah, ok  :-)
21:40 AlexDaniel jnthn: a-zA-Z\s
21:40 jnthn Ah
21:42 jnthn unicodable6: Ḍ̇
21:42 unicodable6 jnthn, U+0307 COMBINING DOT ABOVE [Mn] (◌̇)
21:42 unicodable6 jnthn, U+1E0C LATIN CAPITAL LETTER D WITH DOT BELOW [Lu] (Ḍ)
21:42 jnthn Hm, curious, out of order? :)
21:43 AlexDaniel lizmat: the main idea here was that yoleaux2 does not know about the new unicode stuff (e.g. BUTTERFLY)
21:43 lizmat but shouldn't yoleaux2 be updated then ?
21:43 AlexDaniel jnthn: yes, possibly a bug in IRC::Client https://github.com/zoffixznet/perl6-IRC-Client/issues/32
21:43 jnthn Ah, OK :)
21:43 AlexDaniel lizmat: sure, but dogfood :)
21:44 lizmat AlexDaniel: true, which reminds me to run a few spectests with HARNESS_TYPE=6
21:44 AlexDaniel lizmat: and it will never allow stuff like this anyway
21:44 AlexDaniel u: {.uniname.words == 13}
21:45 unicodable6 AlexDaniel, U+0753 ARABIC LETTER BEH WITH THREE DOTS POINTING UPWARDS BELOW AND TWO DOTS ABOVE [Lo] (ݓ)
21:45 unicodable6 AlexDaniel, U+FBF9 ARABIC LIGATURE UIGHUR KIRGHIZ YEH WITH HAMZA ABOVE WITH ALEF MAKSURA ISOLATED FORM [Lo] (ﯹ)
21:45 unicodable6 AlexDaniel, U+FBFA ARABIC LIGATURE UIGHUR KIRGHIZ YEH WITH HAMZA ABOVE WITH ALEF MAKSURA FINAL FORM [Lo] (ﯺ)
21:45 unicodable6 AlexDaniel, https://gist.github.com/f9b0e78b684968fdc40bcf9165ee6c75
21:46 AlexDaniel .u Me
21:46 yoleaux2 U+0019 END OF MEDIUM [Cc] (␙)
21:46 yoleaux2 U+0095 MESSAGE WAITING [Cc] (<control>)
21:46 yoleaux2 U+009E PRIVACY MESSAGE [Cc] (<control>)
21:46 AlexDaniel .u [Me]
21:46 yoleaux2 U+004D LATIN CAPITAL LETTER M [Lu] (M)
21:46 yoleaux2 U+005B LEFT SQUARE BRACKET [Ps] ([)
21:46 yoleaux2 U+005D RIGHT SQUARE BRACKET [Pe] (])
21:46 AlexDaniel u: Me
21:46 unicodable6 AlexDaniel, U+0489 COMBINING CYRILLIC MILLIONS SIGN [Me] (◌҉)
21:46 unicodable6 AlexDaniel, U+0488 COMBINING CYRILLIC HUNDRED THOUSANDS SIGN [Me] (◌҈)
21:46 unicodable6 AlexDaniel, U+1ABE COMBINING PARENTHESES OVERLAY [Me] (◌᪾)
21:46 unicodable6 AlexDaniel, https://gist.github.com/42af7c4d01723094630cf2f45add7906
22:30 gfldex lizmat: it is new because we forgot to show how users can define traits
22:35 RabidGravy it is in the documentation, but a bit thin
22:36 RabidGravy it probably wants a whole page as the basic thing is easy, it's what you do with them gets more interesting
22:36 ___crlz___ joined #perl6-dev
22:53 seatek joined #perl6-dev
23:23 Ven joined #perl6-dev
23:29 dalek rakudo/nom: eaa8e2d | lizmat++ | src/core/Rakudo/Internals.pm:
23:29 dalek rakudo/nom: Make R:I:ShapeLeafIterator support native arrays
23:29 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/eaa8e2d647
23:34 dalek rakudo/nom: e0c0ae5 | lizmat++ | / (2 files):
23:34 dalek rakudo/nom: Make copying 1+dimmed native arrays 6x faster
23:34 dalek rakudo/nom:
23:34 dalek rakudo/nom: - create a separate STORE candidate for this
23:34 dalek rakudo/nom: - benchmarked for 2x2 array, higher sizes / dimensions probably faster still
23:34 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/e0c0ae5d7e
23:46 lizmat good night, #perl6-dev!
23:48 viki night
23:49 timotimo gniet lizmat

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