Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6-dev, 2016-12-22

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

All times shown according to UTC.

Time Nick Message
00:32 japhb joined #perl6-dev
00:42 nebuchadnezzar joined #perl6-dev
00:47 travis-ci joined #perl6-dev
00:47 travis-ci Rakudo build passed. Zoffix Znet 'Fix IO::ArgFiles.lines treating non-Int numerics as Inf'
00:47 travis-ci https://travis-ci.org/rakudo/rakudo/builds/185934560 https://github.com/rakudo/rakudo/compare/541d12776135...a9c0480c61a7
00:47 travis-ci left #perl6-dev
00:47 camelia joined #perl6-dev
01:07 timotimo joined #perl6-dev
01:22 japhb_ joined #perl6-dev
01:23 travis-ci joined #perl6-dev
01:23 travis-ci Rakudo build passed. Zoffix Znet 'Revert "Fix IO::ArgFiles.lines treating non-Int numerics as Inf"
01:23 travis-ci https://travis-ci.org/rakudo/rakudo/builds/185934870 https://github.com/rakudo/rakudo/compare/a9c0480c61a7...6873d7db635a
01:23 travis-ci left #perl6-dev
01:37 pyrimidine joined #perl6-dev
01:58 travis-ci joined #perl6-dev
01:58 travis-ci Rakudo build passed. Zoffix Znet 'Fix IO::ArgFiles.lines treating non-Int numerics as Inf'
01:58 travis-ci https://travis-ci.org/rakudo/rakudo/builds/185935056 https://github.com/rakudo/rakudo/compare/6873d7db635a...c42b6bf2f564
01:58 travis-ci left #perl6-dev
02:00 pyrimidine joined #perl6-dev
02:50 kalkin- joined #perl6-dev
03:30 pyrimidi_ joined #perl6-dev
04:30 samcv ack i keep running into this bug trying to add tests for unicode properties https://rt.perl.org/Ticket/Display.html?id=130384
04:30 samcv well i guess I can just use the codepoint hex instead of putting the letter in the text of the file
04:31 samcv otherwise it won't even compile
04:36 samcv guys where do I ask questions related to implementation?
04:36 samcv I am trying to write roast tests for all the Unicode properties, and most only have on value. But the Name_Alias one, some of them have multiple values
04:37 samcv 0085;NEXT LINE;control
04:37 samcv 0085;NEL;abbreviation
04:37 samcv this one has a 'control' name and an 'abbreviation' alias names
04:37 samcv these are the formal name aliases, not the ones that are more like also known as names
04:41 samcv and there is a _much_ more important one, Script_Extensions which specifies which scripts a language is used with, in addition to the one that is the value of the 'Script' property
04:41 samcv that can hold multiple values
04:44 samcv and is also not supported in MoarVM or Rakudo at this time
04:50 MasterDuke man, just watched jonathan blow's latest video about his language Jai. from-scratch compile of a 33k line program in less than 1s
04:50 yoleaux2 18 Dec 2016 21:46Z <Zoffix> MasterDuke: Hi. I see this commit says it fixes that ticket but the ticket is still open. Is it OK to close? Do you recall whether tests were added as well? https://github.com/rakudo/rakudo/commit/376950acf2 https://rt.perl.org/Ticket/Display.html?id=128097#ticket-history
05:07 dalek roast: 2759cf2 | MasterDuke17++ | S32-io/io-path.t:
05:07 dalek roast: Add tests for RT #128097
05:07 dalek roast:
05:07 dalek roast: "-".IO gists correctly
05:07 dalek roast: https://rt.perl.org/Ticket/Display.html?id=128097
05:07 dalek roast: review: https://github.com/perl6/roast/commit/2759cf2fce
05:07 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=128097
05:09 MasterDuke .tell Zoffix good catch. tests added for https://rt.perl.org/Ticket/Display.html?id=128097 in https://github.com/perl6/roast/commit/2759cf2fce and ticket resolved
05:09 yoleaux2 MasterDuke: I'll pass your message to Zoffix.
05:18 dalek roast: 4e9476d | samcv++ | S15-unicode-information/uniprop.t:
05:18 dalek roast: uniprop: add a test for Indic_Positional_Category
05:18 dalek roast: review: https://github.com/perl6/roast/commit/4e9476dde5
06:49 RabidGravy joined #perl6-dev
08:51 dalek rakudo/nom: 73797b7 | MasterDuke17++ | src/core/IO/ArgFiles.pm:
08:51 dalek rakudo/nom: Make IO::ArgFiles.lines about 10% faster
08:51 dalek rakudo/nom:
08:51 dalek rakudo/nom: Turn it into a multi so the usual use case can be optimized.
08:51 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/73797b7591
08:51 dalek rakudo/nom: 96f5fba | lizmat++ | src/core/IO/ArgFiles.pm:
08:51 dalek rakudo/nom: Merge pull request #965 from MasterDuke17/more_IO-ArgFiles.lines_optimization
08:51 dalek rakudo/nom:
08:51 dalek rakudo/nom: Make IO::ArgFiles.lines about 10% faster
08:51 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/96f5fba743
09:01 dalek roast: 5c148ed | lizmat++ | S03-feeds/basic.t:
09:01 dalek roast: Fix faulty feed check
09:01 dalek roast:
09:01 dalek roast: Trying to feed an infinite list into an array *should* fail, and
09:01 dalek roast: not just create an empty array silently.
09:01 dalek roast: review: https://github.com/perl6/roast/commit/5c148ed2dd
09:06 dalek roast/6.c-errata: cd83e00 | lizmat++ | S03-feeds/basic.t:
09:06 dalek roast/6.c-errata: Fix faulty feed check
09:06 dalek roast/6.c-errata:
09:06 dalek roast/6.c-errata: Trying to feed an infinite list into an array *should* fail, and
09:06 dalek roast/6.c-errata: not just create an empty array silently.
09:06 dalek roast/6.c-errata: review: https://github.com/perl6/roast/commit/cd83e00844
09:19 lizmat grrrr  two more Terminal crashes
09:19 lizmat getting moderately tired of it
09:41 nine Are there no alternative terminals for Mac OS?
09:42 nine I guess reporting the bug and hoping for constructive work and a speedy fix would be in vain with a huge company like Apple :)
11:04 notviki samcv: does Unicode declare which chars those chars can be combined with?
11:05 notviki samcv: I don't get; what's actually the question about the multiple properties?
11:06 notviki samcv: this is the place the ask questions about impl. Or open an RT ticket.
11:07 notviki m: dd '١'.uniprop('Numeric').^name
11:07 camelia rakudo-moar 96f5fb: OUTPUT«"Str"␤»
11:07 notviki m: dd '١'.uniprop('Number').^name
11:07 camelia rakudo-moar 96f5fb: OUTPUT«"Int"␤»
11:09 notviki m: dd '༳'.uniprop('Number')
11:09 camelia rakudo-moar 96f5fb: OUTPUT«1␤»
11:09 notviki m: dd '٢'.uniprop('Number')
11:09 camelia rakudo-moar 96f5fb: OUTPUT«1␤»
11:10 * notviki doesn't get it
11:21 notviki .tell samcv if the question was how to return the multi props... don't know, but one thing I'd note is it's preferable the methods/sub return predictable values. Right now, I notice if I have $char and $prop, I can no longer know what type .uniprop will return to me. For some reason it returns Int 1 or Int 0 for property 'Number', yet it returns Strs for other properties. If we add Lists to the mix it'd be
11:21 yoleaux2 notviki: I'll pass your message to samcv.
11:21 notviki even more unpredictable
11:21 notviki .tell samcv even more unpredictable.
11:21 yoleaux2 notviki: I'll pass your message to samcv.
11:32 * notviki sighs
11:34 notviki Just yesterday I first noted in #perl6 and then on the PR that we can't coerce an Inf to an Int and so can't have an Int() coercer in the sig and less than 24 hours later a PR doing exactly that gets merged -_-
11:34 notviki m: $*ARGFILES.lines(Inf)
11:34 camelia rakudo-moar 96f5fb: OUTPUT«P6opaque: get_boxed_ref could not unbox for the representation '20' of type Failure␤  in block <unit> at <tmp> line 1␤␤»
11:34 notviki star: $*ARGFILES.lines(Inf)
11:34 camelia star-m 2016.10: ( no output )
11:34 notviki m: "a\nb".lines(Inf)
11:34 camelia rakudo-moar 96f5fb: ( no output )
11:34 timotimo yeah, that also confused me
11:38 notviki .tell MasterDuke FWIW, $*ARGFILES.lines(Inf) no longer works after #965
11:38 yoleaux2 notviki: I'll pass your message to MasterDuke.
12:20 lizmat nine: I think it's a conspiracy to get me to use emacs, because it only happens if you're using vi
12:20 nine Do not give in!
12:29 lizmat :-)
12:30 dalek rakudo/nom: 24f0eeb | MasterDuke17++ | src/core/IO/ArgFiles.pm:
12:30 dalek rakudo/nom: Handle IO::ArgFiles.lines(Inf)
12:30 dalek rakudo/nom:
12:30 dalek rakudo/nom: Base everything off the common case (much more similar to IO::Handle.lines now).
12:30 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/24f0eeb84b
12:30 dalek rakudo/nom: 660b0b9 | (Zoffix Znet)++ | src/core/IO/ArgFiles.pm:
12:30 dalek rakudo/nom: Merge pull request #966 from MasterDuke17/patch-1
12:30 dalek rakudo/nom:
12:30 dalek rakudo/nom: Handle IO::ArgFiles.lines(Inf)
12:30 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/660b0b9bd2
13:42 stmuk_ joined #perl6-dev
13:47 go|dfish joined #perl6-dev
14:24 pyrimidine joined #perl6-dev
14:32 notviki m: my @a = (1..Inf)[^1000]; say now - INIT now
14:32 camelia rakudo-moar 660b0b: OUTPUT«0.383827␤»
14:32 notviki m: my @a = (1..<Inf>)[^1000]; say now - INIT now
14:32 camelia rakudo-moar 660b0b: OUTPUT«3.0618037␤»
14:32 notviki ^^^ is there a reason iterators in Range compare to Infs with === and not just == ?
15:05 notviki hm, I'm still getting "/home/zoffix/.rakudobrew/bin/../moar-nom/install/share/perl6/site/bin/zef" in usage output: https://gist.github.com/zoffixznet/a33ea166a13f4ecb3281384f83ead9e0
15:06 notviki I thought we made it do something better :/
15:06 notviki oh, I guess the old one had SHA hashes in it?
15:07 * notviki would rather have the "command" thing we had for a short bit than a giant ass path that makes everything unreadable
15:09 notviki bleh
15:09 notviki I hate everything today :[
15:09 * notviki crawls into a hole and waits for the holidays
15:16 notviki m: constant HNY = BEGIN DateTime.new(year => .year + (1 if .month <= 6)).utc given DateTime.now.utc; dd HNY
15:16 camelia rakudo-moar 660b0b: OUTPUT«===SORRY!===␤QAST::Block with cuid 4 has not appeared␤»
15:17 notviki :(
15:22 pyrimidi_ joined #perl6-dev
15:28 nine notviki: nine@sunshine:~> zef uinstall blah
15:28 nine Usage: zef [--force] fetch [<identities> ...] -- Download specific distributions
15:28 nine notviki: I see you are using rakudobrew. A tool for people who know what they are doing and who can live with a LTA user experience.
15:30 notviki nine: ah, OK, cool
15:30 notviki Yeah, if it's just in rakudobrew, then it doesn't matter.
15:36 nine I strongly suspect rakudobrew is installing shims for commands like zef or panda to redirect the call to the currently selected rakudo. So it's a shim for a shim for a script that just loads a single module.
15:36 nine We just love our indirections ;)
15:36 notviki :)
16:21 AlexDaniel joined #perl6-dev
16:49 notviki m: my $i = 2; say ++$i**2
16:49 camelia rakudo-moar 660b0b: OUTPUT«9␤»
16:49 notviki m: my $i = 2; say ++$i²
16:49 camelia rakudo-moar 660b0b: OUTPUT«===SORRY!=== Error while compiling <tmp>␤Operators '++' and '²' are non-associative and require parentheses␤at <tmp>:1␤------> my $i = 2; say ++$i²⏏<EOL>␤    expecting any of:␤        postfix␤»
16:50 notviki This feels like a bug to me.
16:51 notviki .ask TimToady is it by design that ++$i² throws, while ++$i**2 works? I kinda expected the two to be equivalent. See https://github.com/perl6/roast/issues/200
16:51 yoleaux2 notviki: I'll pass your message to TimToady.
16:59 dalek joined #perl6-dev
17:24 notviki :o
17:25 pyrimidine joined #perl6-dev
17:57 TimToady well, simply changing the precedence to exponentiation causes the operator precedence parser to get confused, currently
17:57 yoleaux2 16:51Z <notviki> TimToady: is it by design that ++$i² throws, while ++$i**2 works? I kinda expected the two to be equivalent. See https://github.com/perl6/roast/issues/200
17:58 TimToady and if we changed it to methodcall precedence like most other postfixes, it wouldn't work anyway
17:58 TimToady but I don't think much "design" was put into it, more like copy/pasta
18:00 notviki hm, OK
18:01 notviki I interpret that as: I'm free to try to make it parse the same :)
18:01 TimToady STD had some code to deal with dueling precedences in affixes, but I don't know if nqp's EXPR has the same code
18:01 TimToady it might also be an interaction with the fact that the EXPR action method is turning it into a binary op internally
18:02 TimToady anyway, just changing the prec right now results in:
18:02 TimToady MVMArray: Can't pop from an empty array at gen/moar/stage2/NQPHLL.nqp:1099  (/home/larry/nom/install/share/nqp/lib/NQPHLL.moarvm:EXPR_reduce)
18:02 TimToady yes, feel free :)
18:03 * TimToady has to go out shopping...
18:12 TimToady yes, EXPR has the interleaving code, so probably it's the postfix-to-infix transform that is the issue
18:13 TimToady afk &
19:31 dalek rakudo/nom: 0db0810 | lizmat++ | src/core/Date (2 files):
19:31 dalek rakudo/nom: Make clearer that Date is timezone agnostic
19:31 dalek rakudo/nom:
19:31 dalek rakudo/nom: The code for initializing $*TZ was living in the Date module, whereas
19:31 dalek rakudo/nom: this is only used by DateTime.  So moving it to the source file of
19:31 dalek rakudo/nom: DateTime seemed appropriate.
19:31 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/0db0810c02
19:34 lucasb joined #perl6-dev
19:44 lucasb I grepped 'nqp::add_i' and 'nqp::sub_i' and noticed most uses of it are to add or subtract 1
19:45 lucasb how come nqp/moarvm doesn't have a specific OP to increment/decrement 1?
19:45 lucasb wouldn't this make things more performant? :)
19:45 notviki Would it really?
19:46 lucasb idk, would it? It was a question :)
19:46 notviki I recall nqp::ord(...) stuff compiles to more performant stuff; wouldn't same be the case for add/sub?
19:46 notviki Like, I mean this stuff can probably done automagically instead of adding more ops and doing it manually
19:47 notviki m: use nqp; my int $n; nqp::add_i($n, 1) for ^1000_000; say now - INIT now
19:47 camelia rakudo-moar 660b0b: OUTPUT«0.2257413␤»
19:48 notviki I wouldn't call it the slowest piece of code :)
19:48 notviki m: use nqp; my int $n; nqp::add_i($n, 10) for ^1000_000; say now - INIT now
19:48 camelia rakudo-moar 660b0b: OUTPUT«0.2267643␤»
19:48 lucasb ok, then. we can say it's already fast enough :)
19:55 * lizmat goes off to see some Passengers
19:59 AlexDaniel joined #perl6-dev
21:31 FROGGS joined #perl6-dev
21:34 [Coke] lizmat: I like iTerm2 for the mac, if you're looking for a replacement.
23:13 vendethiel joined #perl6-dev
23:23 pyrimidine joined #perl6-dev
23:56 buggable joined #perl6-dev

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