Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6-dev, 2016-12-30

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

All times shown according to UTC.

Time Nick Message
00:04 dalek roast: 1238f39 | (Zoffix Znet)++ | S03-junctions/autothreading.t:
00:04 dalek roast: Remove trailing whitespace
00:04 dalek roast: review: https://github.com/perl6/roast/commit/1238f39d2d
00:05 dalek rakudo/nom: 9fc616f | (Zoffix Znet)++ | src/core/Bool.pm:
00:05 dalek rakudo/nom: Make Bool.ACCEPTS autothread
00:05 dalek rakudo/nom:
00:05 dalek rakudo/nom: Currently, Bool:U.ACCEPTS uses nqp to test for whether the passed arg
00:05 dalek rakudo/nom: istype(Bool), but that will always be false for all Junctions.
00:05 dalek rakudo/nom:
00:05 dalek rakudo/nom: Remove the Mu type constraint so we autothread and smartmatch of Junctions
00:05 dalek rakudo/nom: against Bool type object does not produce incorrect results.
00:05 dalek rakudo/nom:
00:05 dalek rakudo/nom: Fixes RT#130426: https://rt.perl.org/Ticket/Display.html?id=130426
00:05 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/9fc616f875
00:05 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=130426
00:06 dalek roast: 34ae318 | (Zoffix Znet)++ | S03-junctions/autothreading.t:
00:06 dalek roast: Test we can smartmatch Junctions against Bool:U
00:06 dalek roast:
00:06 dalek roast: RT#130426: https://rt.perl.org/Ticket/Display.html?id=130426
00:06 dalek roast: Rakudo fix: https://github.com/rakudo/rakudo/commit/9fc616f875
00:06 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=130426
00:06 dalek roast: review: https://github.com/perl6/roast/commit/34ae318971
00:08 notviki damn, this is my only fix for the day? :S
00:08 notviki And I didn't even game today....
00:08 * notviki switches IRC window away from #perl6
00:12 notviki buggable: tag REGRESSION
00:12 buggable notviki, There are 18 tickets tagged with REGRESSION; See http://perl6.fail/t/REGRESSION for details
00:18 * dugword follows notviki to #perl6-dev to ask more questions
00:41 AlexDaniel notviki just thanked himself for the bug report xD
00:42 * notviki is consistent
00:42 samcv buggable, tag UNI
00:42 buggable samcv, There are 26 tickets tagged with UNI; See http://perl6.fail/t/UNI for details
01:48 AlexDaniel m: my $i = 'a'.Int; say WHAT $i
01:48 camelia rakudo-moar 9fc616: OUTPUT«(Failure)␤»
02:13 AlexDaniel can somebody comment on this:
02:13 AlexDaniel m: my @a; @a = (2, 3,4); @a[3] = slip(2, 3); dd @a
02:13 camelia rakudo-moar 9fc616: OUTPUT«Array @a = [2, 3, 4, slip(2, 3)]␤»
02:13 AlexDaniel commit: 2015.11 my @a; @a = (2, 3,4); @a[3] = slip(2, 3); dd @a
02:13 committable6 AlexDaniel, ¦«2015.11»: Array $var = $[2, 3, 4, (2, 3)]
02:13 AlexDaniel storing a slip in an array? Does it make sense?
02:22 notviki makes sense to me. Otherwise there'd be no way to pass Slips around
02:22 notviki it's itemized
02:22 notviki committable6: 2015.11 Slip.new(<a b c>).perl.say
02:22 committable6 notviki, ¦«2015.11»: ($("a", "b", "c"),)
02:23 notviki committable6: 2015.11 slip.new(<a b c>).perl.say
02:23 committable6 notviki, ¦«2015.11»: ($("a", "b", "c"),)
02:23 notviki committable6: 2015.11 slip(<a b c>).perl.say
02:23 committable6 notviki, ¦«2015.11»: ("a", "b", "c")
02:23 notviki yup, that's just .perl output changed methinks
02:24 notviki commit: 2015.11 my @a; @a = (2, 3,4); @a[3] = slip(2, 3); dd @a[3].WHAT
02:25 committable6 notviki, ¦«2015.11»: Cannot look up attributes in a type object␤  in block <unit> at /tmp/60FImetZD9:1␤ «exit code = 1»
02:25 notviki commit: 2015.11 my @a; @a = (2, 3,4); @a[3] = slip(2, 3); dd @a[3].^name
02:25 committable6 notviki, ¦«2015.11»: Str $var = "Slip"
02:25 notviki \o/
02:46 AlexDaniel okay, great
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
03:02 notviki AlexDaniel++
03:13 samcv m: 7 ~ "\x[308]" x 100_000
03:13 camelia rakudo-moar 9fc616: OUTPUT«WARNINGS for <tmp>:␤Useless use of "~" in expression "7 ~ \"\\x[308]\" x" in sink context (line 1)␤»
03:13 samcv m: so 7 ~ "\x[308]" x 100_000
03:14 samcv m: say 'test'
03:14 samcv crap
03:14 samcv i think i filled up the bot's memory
03:14 samcv had to kill my moarvm cause computer started lagging
03:15 samcv oh nice. Konsole is using 28% of my memory
03:15 camelia rakudo-moar 9fc616: OUTPUT«WARNINGS for <tmp>:␤Useless use of "so " in expression "so 7 ~" in sink context (line 1)␤»
03:15 camelia rakudo-moar 9fc616: OUTPUT«test␤»
03:16 samcv m:  7 ~ "\x[308]" x 150_000
03:16 camelia rakudo-moar 9fc616: OUTPUT«(signal SEGV)»
03:16 samcv hm
03:17 notviki That one blows through the stack
03:17 samcv m:  7 ~ "\x[308]" x 150_000000
03:17 samcv m:  7 ~ "\x[308]" x 150000000
03:18 camelia rakudo-moar 9fc616: OUTPUT«(timeout)»
03:18 camelia rakudo-moar 9fc616: OUTPUT«(timeout)»
03:18 notviki samcv: https://irclog.perlgeek.de/moarvm/2016-09-13#i_13205266
03:19 * notviki &
03:20 samcv well it works if it's long enough. for some reason
03:20 samcv oh well
03:20 samcv m:  7 ~ "\x[308]" xx 150000000
03:20 camelia rakudo-moar 9fc616: OUTPUT«WARNINGS for <tmp>:␤Useless use of "~" in expression "7 ~ \"\\x[308]\" xx" in sink context (line 1)␤MoarVM panic: Memory allocation failed; could not allocate 1200000000 bytes␤»
03:20 samcv that is the warning one would expect
03:25 samcv i never imagined i would be working in moarvm when i started contributing to perl 6 ha
03:25 samcv first i try to add unicode properties and now i'm implementing string collation… how did this happen
03:42 bisectable6 joined #perl6-dev
03:43 AlexDaniel .tell notviki the result of this 「my @a = lazy gather { say 'we actually run' }; sink @a」 was changed due to https://github.com/rakudo/rakudo/commit/345f6a7146d6b28c37718de22c90f50d9d9b2dcc
03:43 yoleaux2 AlexDaniel: I'll pass your message to notviki.
03:44 AlexDaniel .tell notviki can you check if it's ok?
03:44 yoleaux2 AlexDaniel: I'll pass your message to notviki.
03:50 samcv notviki, what is the best way to count the number of times a string appears in some text
03:50 samcv non overlapping
03:50 samcv just need the number. i guess comb.elems would prolly work?
03:59 samcv this works fine for my use but was wondering if there was a dedicated thing for that
04:21 AlexDaniel samcv: how efficient do you want it to be? :)
04:22 samcv well i just find a single character
04:22 samcv i just want to count occurances
04:22 samcv and not be loaded like comb is
04:23 AlexDaniel m: say +‘hello world’.match(‘l’, :g)
04:23 camelia rakudo-moar 9fc616: OUTPUT«3␤»
04:23 samcv i guess that is acceptable
04:24 AlexDaniel m: dd ‘hello world’.comb.Bag
04:24 camelia rakudo-moar 9fc616: OUTPUT«(" "=>1,"r"=>1,"l"=>3,"e"=>1,"w"=>1,"h"=>1,"d"=>1,"o"=>2).Bag␤»
04:25 AlexDaniel anyway, enough for today
04:25 AlexDaniel good night o/
04:43 dalek roast: b8271d7 | samcv++ | S15-nfg/GraphemeBreakTest (2 files):
04:43 dalek roast: Correct 4 of the incorrect grapheme counts in GraphemeBreakTest.t
04:43 dalek roast:
04:43 dalek roast: We now pass these 4 tests which had incorrect expected values.
04:43 dalek roast: review: https://github.com/perl6/roast/commit/b8271d7514
04:51 samcv oh i see why we are failing some of these tests.
04:52 samcv er it's actually complicated but. i know what i need to do now :)
05:09 pyrimidi_ joined #perl6-dev
05:16 samcv m: "1F468 200D".split(' ').».parse-base(16).say
05:16 camelia rakudo-moar 9fc616: OUTPUT«(128104 8205)␤»
05:17 samcv m: "1F468 200D".split(' ').».parse-base(16).join
05:17 camelia rakudo-moar 9fc616: ( no output )
05:17 samcv m: "1F468 200D".split(' ').».parse-base(16).join.chars
05:17 camelia rakudo-moar 9fc616: ( no output )
05:17 samcv m: "1F468 200D".split(' ').».parse-base(16).join.chars.say
05:17 camelia rakudo-moar 9fc616: OUTPUT«10␤»
05:17 samcv m: "1F468 200D".split(' ').».parse-base(16).say
05:17 camelia rakudo-moar 9fc616: OUTPUT«(128104 8205)␤»
05:17 samcv m: "1F468 200D".split(' ').».parse-base(16).perl.say
05:17 camelia rakudo-moar 9fc616: OUTPUT«(128104, 8205)␤»
05:17 samcv m: "1F468 200D".split(' ').».parse-base(16).perl.join('').chars.say
05:17 camelia rakudo-moar 9fc616: OUTPUT«14␤»
05:19 samcv m: "1F468 200D".split(' ').».parse-base(16).chrs.chars
05:19 camelia rakudo-moar 9fc616: ( no output )
05:19 samcv m: "1F468 200D".split(' ').».parse-base(16).chrs.chars.say
05:19 camelia rakudo-moar 9fc616: OUTPUT«1␤»
05:29 dugword joined #perl6-dev
06:52 dalek nqp: c54dee8 | samcv++ | src/HLL/Grammar.nqp:
06:52 dalek nqp: Optimize nqp peek_delimiters when checking properties
06:52 dalek nqp:
06:52 dalek nqp: Reduces one `eq` and two property checks into only one property check
06:52 dalek nqp: for the most common case (where we are using some type of punctuation
06:52 dalek nqp: character as our delimiter. If it's not a bracket we can test and make
06:52 dalek nqp: sure it's not a colon just to be sure.
06:52 dalek nqp: review: https://github.com/perl6/nqp/commit/c54dee86e5
06:57 RabidGravy joined #perl6-dev
07:33 pyrimidine joined #perl6-dev
07:36 dalek nqp: e4cf11c | samcv++ | .gitignore:
07:36 dalek nqp: Have git ignore .html files in the root folder
07:36 dalek nqp: review: https://github.com/perl6/nqp/commit/e4cf11cf80
08:08 dalek rakudo/nom: acec1ea | samcv++ | / (3 files):
08:08 dalek rakudo/nom: Prepare uniprop for implementation of Bidi_Mirroring_Glyph in MVM
08:08 dalek rakudo/nom:
08:08 dalek rakudo/nom: Bidi_Mirroring_Glyph will be implemented as an integer property in MVM.
08:08 dalek rakudo/nom: Set uniprop to get that int and return it as a character.
08:08 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/acec1ea6cf
08:52 [Tux] This is Rakudo version 2016.12-135-gacec1ea6c built on MoarVM version 2016.12-23-g3c5253c6
08:52 [Tux] csv-ip5xs        3.149
08:52 [Tux] test            13.639
08:52 [Tux] test-t           5.332
08:52 [Tux] csv-parser      14.057
08:52 samcv sup [Tux]
08:59 samcv ugh nqp why you do this
08:59 samcv Cannot invoke this object (REPR: P6opaque; NQPMu)
08:59 samcv so unhelpful
09:06 FROGGS joined #perl6-dev
09:18 dalek nqp/bmg: 7b68aa5 | samcv++ | src/HLL/Grammar.nqp:
09:18 dalek nqp/bmg: Implement Bidi Matching brackets on MoarVM
09:18 dalek nqp/bmg:
09:18 dalek nqp/bmg: This vastly reduces the complexity of finding matching brackets. Access
09:18 dalek nqp/bmg: to this Unicode property will be landing in MoarVM fairly shortly.
09:18 dalek nqp/bmg:
09:18 dalek nqp/bmg: Instead of having to search through a long string of potential
09:18 dalek nqp/bmg: delimiters for our opening delimiter, we only need to get its matching
09:18 dalek nqp/bmg: bracket and then find the ending delimiter.
09:18 dalek nqp/bmg: review: https://github.com/perl6/nqp/commit/7b68aa5b98
09:18 samcv :D
09:19 samcv compared to the code we had before, this is like 10x better. will still have to do it the "old fashioned way" for backends that don't have access to this property yet
09:25 samcv buggable, UNI
09:25 samcv buggable, help
09:25 buggable samcv, tags | tag SOMETAG | eco | eco Some search term | speed
09:25 samcv buggable, tag UNI
09:25 buggable samcv, There are 26 tickets tagged with UNI; See http://perl6.fail/t/UNI for details
09:26 samcv would also be able to make this bug much more awesome https://rt.perl.org/Ticket/Display.html?id=130384#ticket-history
09:27 samcv '<ୈ'.uniprop('Bidi_Mirrored') would return true, but it wouldn't have any matching glyph so we could easily detect degenerates and provide a nice error for that as well
10:41 dalek roast: 41ee7e7 | samcv++ | S15-nfg/GraphemeBreakTest.t:
10:41 dalek roast: Fudge the updated GraphemeBreakTest.t’s
10:41 dalek roast:
10:41 dalek roast: We pass ~5 more of the updated tests compared to before, but I had
10:41 dalek roast: forgotten to push after refudging the generated code.
10:41 dalek roast: review: https://github.com/perl6/roast/commit/41ee7e7bc8
11:06 notviki m: my @a = lazy gather { say 'we actually run' }; sink @a
11:06 camelia rakudo-moar acec1e: OUTPUT«WARNINGS for <tmp>:␤Useless use of @a in sink context (line 1)␤»
11:06 yoleaux2 03:43Z <AlexDaniel> notviki: the result of this 「my @a = lazy gather { say 'we actually run' }; sink @a」 was changed due to https://github.com/rakudo/rakudo/commit/345f6a7146d6b28c37718de22c90f50d9d9b2dcc
11:06 yoleaux2 03:44Z <AlexDaniel> notviki: can you check if it's ok?
11:07 notviki .tell AlexDaniel looks OK to me.
11:07 yoleaux2 notviki: I'll pass your message to AlexDaniel.
11:08 notviki committable6: 2015.12 my @a = lazy gather { say 'we actually run' }; sink @a
11:08 committable6 notviki, ¦«2015.12»: WARNINGS for /tmp/6UMRyLrSBt:␤Useless use of @a in sink context (line 1)␤we actually run
11:08 notviki yeah
11:09 notviki Not sure why we warn on explicit sink, but whatever.
11:18 kalkin joined #perl6-dev
11:22 kalkin- joined #perl6-dev
11:41 dalek rakudo/nom: 696b1f4 | lizmat++ | src/core/Shaped1Array.pm:
11:41 dalek rakudo/nom: Fix for RT #130451
11:41 dalek rakudo/nom:
11:41 dalek rakudo/nom: Also fixes an unseen issue with the push-all case.  AlexDaniel++
11:41 dalek rakudo/nom: for spotting.
11:41 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=130451
11:41 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/696b1f47d5
11:45 samcv also if anybody sees any unicode implementation related bug, please CC or message me so i can add to my list
11:46 notviki \o/
11:46 lizmat m: say 0/0 cmp -Inf   # root problem of RT #130452
11:46 camelia rakudo-moar acec1e: OUTPUT«More␤»
11:46 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=130452
11:47 samcv isn't 0/0 undefined
11:48 samcv i mean in like. real life not in like perl 6 that is
11:48 lizmat m: say (0/0).WHAT
11:48 camelia rakudo-moar acec1e: OUTPUT«(Rat)␤»
11:48 samcv ah
11:48 samcv hehehehhehehehe
11:48 * samcv chuckles
11:48 lizmat m: say 0/0
11:48 camelia rakudo-moar acec1e: OUTPUT«Attempt to divide by zero using div␤  in block <unit> at <tmp> line 1␤␤Actually thrown at:␤  in block <unit> at <tmp> line 1␤␤»
11:48 lizmat I guess we need to throw on 0/0 cmp -Inf
11:49 samcv yeah
11:49 notviki (-∞^..^∞).in-range: 0/0
11:49 notviki m: (-∞^..^∞).in-range: 0/0
11:49 camelia rakudo-moar acec1e: ( no output )
11:49 samcv i finally figured a way around that unicode bug that was bugging me. about Q<? something > where ? is some combinig mark
11:49 samcv once next few moar patches get accepted, for moarvm we can use unicode properties to check brackets
11:49 notviki m: say 0/0 cmp 1/0
11:49 camelia rakudo-moar acec1e: OUTPUT«More␤»
11:49 notviki m: say 0/0 cmp -1/0
11:49 camelia rakudo-moar acec1e: OUTPUT«More␤»
11:50 notviki m: say 0/0 cmp -1/1
11:50 camelia rakudo-moar acec1e: OUTPUT«More␤»
11:50 samcv and if we find an opening bracket, and can't find an ending one, we can provide a better error message
11:50 notviki m: say 0/0 cmp 1/1
11:50 camelia rakudo-moar acec1e: OUTPUT«More␤»
11:50 notviki heh
11:50 samcv because if we can't find one (and it's a bidi_mirroring character) it's most likely either a bug or a combining character
11:50 notviki m: say NaN cmp 1/1
11:50 camelia rakudo-moar acec1e: OUTPUT«More␤»
11:50 samcv can see if they supplied multiple codeponits or something and then say what the problem is
11:50 notviki (-∞^..^∞).in-range: NaN
11:51 notviki m: (-∞^..^∞).in-range: NaN
11:51 camelia rakudo-moar acec1e: OUTPUT«Value out of range. Is: NaN, should be in -Inf^..^Inf␤  in block <unit> at <tmp> line 1␤␤»
11:51 samcv since there's no coorasponding unicode mirroring glyph
11:51 notviki m: say 0/0 cmp Inf
11:51 camelia rakudo-moar acec1e: OUTPUT«Less␤»
11:51 samcv .oO(thinks that looks like an emote)
11:51 samcv (-∞^..^∞).in-range: NaN
11:51 samcv at least in this tiny 11px high font. if it's even 11px maybe 9
11:51 notviki m: say 0/0 cmp 9999999999999999999999999999999999999999999999999999999
11:51 camelia rakudo-moar 696b1f: OUTPUT«More␤»
11:52 samcv actually. yeah it is exactly 9 px high
11:52 notviki Interesting...
11:52 samcv (-∞^..^∞ (-∞^..^∞(-∞^..^∞(-∞^..^∞
11:52 notviki m: say -0e0/0 cmp -9999999999999999999999999999999999999999999999999999999
11:52 camelia rakudo-moar 696b1f: OUTPUT«Attempt to divide by zero using /␤  in block <unit> at <tmp> line 1␤␤Actually thrown at:␤  in block <unit> at <tmp> line 1␤␤»
11:52 notviki heh, now it explodes? :)
11:53 notviki Ah, it's not a rat anymore
11:53 samcv numbers in the universe can be pretty big
11:53 samcv tho
11:53 bartolin I think about sending a CLA to the Perl Foundation and asking for a commit bit for rakudo. anyone + or - on that?
11:53 notviki bartolin: +1
11:53 lizmat bartolin: +1
11:54 bartolin (I don't plan to commit much more, but at least for some things a commit bit would be handy)
11:54 stmuk http://pl6anet.org/drop/rakudo-star-2016.12.zef-beta.tar.gz
11:55 stmuk panda is patched to warn of its removal in that beta
11:55 stmuk I may blog if I get tuits
11:55 lizmat notviki: if you want to take on RT #130452, please go ahead, will be away in the next hours
11:55 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=130452
11:55 * lizmat hands stmuk some tuits
11:55 stmuk :)
11:56 notviki lizmat: OK, I'll take it.
11:56 samcv also is it ok if ' ' ~~ /<:space>/ fails on rakudo, even if all other tests pass? i doubt anybody uses that, instead of \s or <.ws> or something. but it's the only roast test
11:56 samcv that cannot be easily fixed by me due to problems in nqp MAST
11:57 bartolin notviki: you sent an email to get your commit bit, didn't you? do you remember the recipient of that email (or could take a look)?
11:57 notviki bartolin: that recepient resigned. Ask samcv
11:57 samcv who resigned?
11:57 samcv oh yes
11:58 bartolin oh. I see
11:58 notviki Last president of TPF
11:58 samcv i sent an email
11:58 samcv to the patent email address
11:58 samcv and explained stuff and things and such reasons etc
11:58 bartolin ah, thanks!
11:58 samcv whatever put them in the email with a scanned copy of the CLA with your signature, and uh
11:58 samcv give them a contact email. of who to CC their response to you
11:59 samcv or something
11:59 samcv idk. the person replying to me just said "you can tell them we got it" but idk. seemed kinda weird
11:59 notviki stmuk: I take the final upload won't have "zef" in it and will just follow our regular naming?, I don't think the downloads page script would handle it
11:59 samcv luckily you all believed me ;P
11:59 bartolin *g*
11:59 samcv who knows maybe i never sent it!
11:59 samcv all part of my evil plan
12:00 stmuk notviki: yes its just there since is another branch currently and it won't have -beta either!
12:00 * bartolin will try something similiar soon-ish
12:00 stmuk samcv: I'm just checking Linenoise on Windows/mingw
12:00 samcv oh nice
12:01 samcv yeah. def gonna improve that message before the next release for sure. hope nobody thinks i'm ignoring it.
12:18 FROGGS joined #perl6-dev
12:24 dalek joined #perl6-dev
12:37 stmuk samcv: ah you are right .. Linenoise HEAD is broken on Windows and I had to use an earlier version
12:37 dalek nqp: eaa3eb1 | samcv++ | src/vm/moar/QAST/QASTOperationsMAST.nqp:
12:37 dalek nqp: unicmp_s add it
12:37 dalek nqp: review: https://github.com/perl6/nqp/commit/eaa3eb109c
12:37 dalek nqp: e71e517 | samcv++ | tools/build/MOAR_REVISION:
12:37 dalek nqp: Bump Moar revision to pull in the new Unicode string compare (UCA)
12:37 stmuk https://github.com/rakudo/star/commit/4b2ad3c6be746ca63ebed01e322116f2716a7faf
12:37 samcv ah
12:37 samcv dalek is bad at printing commits
12:38 samcv tho the commit message here: https://github.com/perl6/nqp/commit/003919e5bc42fecf286860596888c67dd659ca5a is the same as the PR so can be seen in both projects
12:38 dalek joined #perl6-dev
12:39 stmuk samcv: I'd suggest the REPL just suggest about Linenoise (and not Readline anyway)
12:42 FROGGS joined #perl6-dev
12:43 dalek rakudo/nom: 3383361 | samcv++ | tools/build/NQP_REVISION:
12:43 dalek rakudo/nom: Bump NQP version to bring in new nqp op for Unicode string sort (UCA)
12:43 dalek rakudo/nom:
12:43 dalek rakudo/nom: Full details on the op can be see on the nqp commit:
12:43 dalek rakudo/nom: https://github.com/perl6/nqp/commit/003919e5bc42fecf286860596888c67dd659ca5a
12:43 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/3383361189
12:46 * stmuk opens an issue against p6-linenoise
12:47 * samcv is gonna bout pass out asleep
12:47 samcv quite shortly. and this time not an exaggeration, actually going to bed within 30 mins
13:07 lucasb joined #perl6-dev
13:34 timo joined #perl6-dev
13:52 nebuchadnezzar joined #perl6-dev
14:24 FROGGS joined #perl6-dev
14:49 nebuchadnezzar joined #perl6-dev
14:59 kalkin- joined #perl6-dev
15:30 AlexDaniel joined #perl6-dev
15:37 hankache joined #perl6-dev
16:05 Ven joined #perl6-dev
16:24 lizmat Files=1159, Tests=54450, 189 wallclock secs (10.86 usr  4.41 sys + 1134.01 cusr 123.88 csys = 1273.16 CPU)  # with TEST_JOBS=10
16:32 Ven joined #perl6-dev
16:45 AlexDaniel ok, the last one is probably a duplicate
16:46 AlexDaniel of RT #128872
16:46 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=128872
16:47 AlexDaniel ok, merged
16:48 pyrimidine joined #perl6-dev
16:52 Ven joined #perl6-dev
16:53 AlexDaniel lizmat: hey! What do you think about this
16:53 AlexDaniel commit: 2015.12,HEAD say "7\x[308]".NFC ~~ /^ \d+ $/
16:53 committable6 AlexDaniel, ¦«2015.12»: 「7̈」␤¦«HEAD»: 「55」
16:53 AlexDaniel bisectable points to this: https://github.com/rakudo/rakudo/commit/e4dc8b6ed1f9b3b4846f0e371c9b8e0569b798c5
16:54 AlexDaniel hmmm that might be the wrong one though…
16:55 lizmat pretty sure that was just an nqpifaction
16:55 AlexDaniel yea it broke it but in a different way… hold on
17:00 dugword joined #perl6-dev
17:03 lizmat https://www.reddit.com/r/perl6/comments/5l42il/sharing_your_p6_program_with_others_any_plan_to/?st=ixc1sg0u&amp;sh=b406b6f9  # any takers?
17:05 notviki IIRC there was an RT for that
17:05 notviki This https://rt.perl.org/Ticket/Display.html?id=129909#ticket-history
17:06 notviki also, I recall timotimo showing off a movable nqp
17:06 notviki like a binary you could download and just run
17:08 lizmat part of the precomp design was that you should be able to get precomped code from anywhere, so also from within a jar type of file
17:16 AlexDaniel lizmat: ok, here it is
17:16 AlexDaniel committable6: 33eeb32^^^,33eeb32^^,33eeb32^,33eeb32 say "7\x[308]".NFC ~~ /^ \d+ $/
17:16 committable6 AlexDaniel, ¦«33eeb32^^^»: 「7̈」␤¦«33eeb32^^,33eeb32^»: This representation (VMArray) does not support attribute storage (for type NFC)␤  in block <unit> at /tmp/7c6yDMIEt0 line 1␤ «exit code = 1»␤¦«33eeb32»: 「55」
17:17 notviki m: use nqp; -> \a { dd nqp::getattr(a, a.WHAT, '$!denominator') }($_) given <42/1>
17:17 camelia rakudo-moar 338336: OUTPUT«1␤»
17:17 AlexDaniel committable6:  05b65d0,e4dc8b6,8c35481,33eeb32 say "7\x[308]".NFC ~~ /^ \d+ $/
17:17 committable6 AlexDaniel, ¦«05b65d0»: 「7̈」␤¦«e4dc8b6,8c35481»: This representation (VMArray) does not support attribute storage (for type NFC)␤  in block <unit> at /tmp/L9mYdkqIIM line 1␤ «exit code = 1»␤¦«33eeb32»: 「55」
17:17 notviki nm was gonna ask a question but then I realized what I did wrong
17:18 AlexDaniel lizmat: so it's a combination of 05b65d0 and 33eeb32
17:20 AlexDaniel lizmat: previously ~~ assumed that NFC is a Str-like thingy, now it attempts to match the ord value of the first character
17:26 cygx joined #perl6-dev
17:27 TimToady notviki: the sink statement prefix didn't used to warn before you did 345f6a7146d6b28c37718de22c90f50d9d9b2dcc, but I think it's probably better than an explicit sink turns into a sink method, so that the user has to put 'eager' if that's really what they mean
17:28 notviki m: my @a; sink @a
17:28 camelia rakudo-moar 338336: OUTPUT«WARNINGS for <tmp>:␤Useless use of @a in sink context (line 1)␤»
17:28 notviki committable6: 2015.12 my @a; sink @a
17:28 committable6 notviki, ¦«2015.12»: WARNINGS for /tmp/WFMUPewq9V:␤Useless use of @a in sink context (line 1)
17:28 notviki TimToady: ^ seems it warned before that commit too
17:28 TimToady on cmp, the idea was to sort NaN to the end as if it were a string, not to blow up, since it's not intended to reflect any kind of mathematical relationship, but to primarily serve as the basis for sort, which generally should produce results regardless of the input
17:29 notviki Yeah, that's what I'm making it work like. Just trying to find best performant method (I think I found it; just spectesting)
17:30 TimToady perhaps it's parsing it as a listop rather than a statement prefix there
17:31 TimToady shouldn't though
17:31 notviki --target=parse gives - statement_prefix: sink @a
17:31 notviki - sym: sink
17:31 TimToady okay, that's good
17:31 TimToady that falls out from the literal prefix rule
17:31 TimToady which basically prefers 'x' to \w
17:32 Ven joined #perl6-dev
17:32 TimToady anyway, I'm fine with explicit sink warning, though it possibly be taught to suggest 'eager' in such a case
17:33 TimToady not sure how we'd know at the time that it was explicit though
17:33 TimToady I don't think the context in the optimizer tracks that
17:33 TimToady shower &~
17:36 notviki hmmm... Stage parse      : Cannot invoke this object (REPR: Null; VMNull)
17:36 notviki at SETTING::src/core/traits.pm:132  (<ephemeral file>:trait_mod:<is>
17:36 notviki Any trick for making `is default` on  a Routine work in the setting?
17:36 notviki trying to make (Rational:D \a, Rational:D \b) tighter than (Real:D \a, Real:D \b)
17:36 lizmat on a Routine ?
17:36 notviki Yeah
17:36 lizmat ah
17:37 lizmat well, look at what trait_mod(Routine:D, :$!default!) does in terms of mop
17:37 lizmat and copy paste
17:37 lizmat ?
17:38 notviki oh, heh; all it does is mixes in does role { method default() { True } } ^_^
17:41 notviki no idea  how I'd apply that on my own tho :/
17:46 notviki ah, looks like the `does` operator isn't yet available at that point
17:47 * notviki looks for an alterantive point to stick stuff into
17:52 Ven joined #perl6-dev
17:54 lizmat notviki: does is basically a ^mixin
17:54 cygx .tell samcv I believe you misread RT #129878 - the issue is not stringifying a list generated by xx (did you miss the [~]), but the string generated by x
17:54 yoleaux2 cygx: I'll pass your message to samcv.
17:54 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=129878
17:55 notviki m: multi foo (Rational, Rational) {say "got it"}.^mixin(role { method default() { True } }); multi foo (Real, Real) {say "fail"}; foo 1.1, 1.1
17:55 camelia rakudo-moar 338336: OUTPUT«Ambiguous call to 'foo'; these signatures all match:␤:(Rational $, Rational $)␤:(Real $, Real $)␤  in block <unit> at <tmp> line 1␤␤»
17:55 lizmat notviki: obj.^mixin($role).BUILD_LEAST_DERIVED({})
17:55 notviki 0.o
17:55 lizmat perhaps in a BEGIN block  :-)
17:55 notviki eh
17:56 notviki I move it into src/core/operators and it works now (the is default)
17:56 notviki I mean I moved the thing I wanted to apply it to
17:56 lizmat ok
17:57 lizmat AlexDaniel: reverting 33eeb32 gives the VMArray error
17:58 AlexDaniel lizmat: yup, you'd have to revert 05b65d0b1 as well to get the previous behavior
17:59 AlexDaniel lizmat: I've created a ticket for this RT #130458, but to be honest I have no idea what's going on and what should this code do
17:59 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=130458
18:00 notviki ZOFVM: Files=1208, Tests=130861, 160 wallclock secs (22.40 usr  3.42 sys + 2776.65 cusr 253.22 csys = 3055.69 CPU)
18:12 Ven joined #perl6-dev
18:12 Undercover joined #perl6-dev
18:14 Undercover joined #perl6-dev
18:18 evalable6 joined #perl6-dev
18:19 evalable6 joined #perl6-dev
18:19 committable6 joined #perl6-dev
18:19 bisectable6 joined #perl6-dev
18:23 bisectable6 joined #perl6-dev
18:24 committable6 joined #perl6-dev
18:30 committable6 joined #perl6-dev
18:32 Ven joined #perl6-dev
18:34 lizmat AlexDaniel: the problem is really in Regex.ACCEPTS(@a), that's the candidate .NFC winds up in
18:35 lizmat maybe we should add a Regex.ACCEPTS(Uni:D candidate) ?
18:35 committable6 joined #perl6-dev
18:36 AlexDaniel vOv
18:36 lizmat testing that now
18:40 lizmat $ 6l 'say "7\x[308]".NFC ~~ /./'
18:40 lizmat 「7̈」
18:40 lizmat at least this restores previous behaviour and it will allow for easier changes should we decide on other behaviour
18:47 committable6 joined #perl6-dev
18:47 committable6 joined #perl6-dev
18:48 committable6 joined #perl6-dev
18:49 dalek rakudo/nom: 8d35951 | lizmat++ | src/core/Regex.pm:
18:49 dalek rakudo/nom: Fix for RT #130458
18:49 dalek rakudo/nom:
18:49 dalek rakudo/nom: Spotted by AlexDaniel++, this patch reverts to the old behaviour
18:49 dalek rakudo/nom: of interpreting a Uni as a .Str for regex matching.  Still not sure
18:49 dalek rakudo/nom: that's what we want, but since we have now isolated this behaviour
18:49 dalek rakudo/nom: into a separate Regex.ACCEPTS candidate (it was formerly taken by
18:49 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=130458
18:49 dalek rakudo/nom: the @a candidate), an adaptation of behaviour should now be simple
18:49 dalek rakudo/nom: if so needed.
18:49 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/8d359518dd
18:50 pyrimidine joined #perl6-dev
18:50 moritz does that mean I can regex-match against an Uni?
18:51 lizmat moritz: yes
18:51 moritz and how do I get from Str to Uni?
18:51 lizmat it will do a .Str on the Uni
18:51 lizmat .NFC ?
18:51 moritz :(
18:51 moritz I have an open bug report for JSON::Tiny
18:51 lizmat link?
18:51 moritz it can't parse quoted strings if the first character in the string is a combining charcter
18:52 moritz https://github.com/moritz/json/issues/25
18:52 Ven joined #perl6-dev
18:52 moritz and that's because to the regex engine, there is no ", just a " + combiner as a single grapheme
18:52 moritz and i have zero ideas how to fix it
18:53 lizmat hmmm... ok, but that's really at a completely different level
18:53 moritz that's why I hoped I could regex-match against a Uni without turning it into a Str
18:53 moritz any other ideas how to fix that?
18:53 lizmat the fix I did was basically add:  method ACCEPTS(Regex:D: Uni:D \uni) { self.ACCEPTS(uni.Str) }
18:53 notviki :/
18:54 notviki Range.in-range is 100% unspecceed
18:54 lizmat and undocumented ?
18:54 moritz the only other thing I could think of was to use :ignoremark, but then I'll lose the combiner
18:54 notviki Seems so; doesn't show up in search on docs.perl6.org
18:55 lizmat fwiw, I think I added it as an internal method really
18:55 notviki OK. I'll test it just in Rakudo's t/
19:03 Ven joined #perl6-dev
19:16 dalek rakudo/nom: afb6d65 | (Zoffix Znet)++ | t/02-rakudo/05-range-in-range.t:
19:16 dalek rakudo/nom: Add basic tests for Range.in-range
19:16 dalek rakudo/nom:
19:16 dalek rakudo/nom: It's a somewhat internal method with no presence in the roast, so
19:16 dalek rakudo/nom: I'm adding it in Rakudo's t/
19:16 dalek rakudo/nom:
19:16 dalek rakudo/nom: https://irclog.perlgeek.de/perl6-dev/2016-12-30#i_13825497
19:16 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/afb6d654ac
19:18 moritz does MoarVM support non-NFG strings?
19:21 stmuk has any analysis been done on module dependencies in the ecosystem to identify the most popular/
19:22 moritz yes
19:22 stmuk where are the results and/or the scripts?
19:23 Ven joined #perl6-dev
19:23 moritz https://github.com/finanalyst/ModuleCitation
19:23 stmuk ty
19:24 moritz http://perlpunks.de/paste/show/5866b44b.5e35.3d7 data from 2015-11
19:25 stmuk ah great
19:27 notviki hmm... t/spec/S17-procasync/stress.t hung on me :/
19:44 dalek rakudo/nom: a567eb4 | (Zoffix Znet)++ | src/core/ (2 files):
19:44 dalek rakudo/nom: Fix infix:<cmp>(Real,Real) with 0-denominator Rationals
19:44 dalek rakudo/nom:
19:44 dalek rakudo/nom: We have adopted[^1] IEEE 754-2008 semantics for 0-denominator rationals,
19:44 dalek rakudo/nom: so that <0/0> is NaN, <-1/0> is -Inf, and <1/0> is Inf when .Num'ed.
19:44 dalek rakudo/nom: However, currently infix:<cmp>(Real:D, Real:D) candidate short-cuircuits
19:44 dalek rakudo/nom: when one of the arguments is Num(±Inf), which causes incorrect comparison
19:44 dalek rakudo/nom: results when the other argument is one of the aforementioned Rats.
19:44 dalek rakudo/nom:
19:44 dalek rakudo/nom: Fix by detecting whether we have a 0-denominator Rational before trying to
19:44 dalek rakudo/nom: short-curcuit with the Num(±Inf) logic. This method is most optimal I could
19:44 dalek rakudo/nom: find, with other methods I tried giving 50%–300% slowdowns. The current
19:44 notviki rakudo/nom: method gives about a 4%–8% slowdown, depending on the types of Reals compared.
19:44 notviki rakudo/nom:
19:45 notviki rakudo/nom: The worst hit of 22% was when comparing Rational against another Rational,
19:45 notviki rakudo/nom: so I added another infix:<cmp>(Rational, Rational) candidate, which actually
19:45 notviki rakudo/nom: made Rational cmp Rational 150% *faster*. Since this candidate is ambiguious
19:45 notviki rakudo/nom: with Real,Real candidate, I had to stick `is default` on it, but because
19:45 notviki rakudo/nom: `is default` uses `does` operator that doesn't appear until later in the
19:45 notviki rakudo/nom: setting, I had to define this new infix:<cmp> candidate in src/core/operators
19:45 notviki rakudo/nom: instead of src/core/Order.
19:45 notviki rakudo/nom:
19:45 notviki rakudo/nom: This commit also 99% fixes the .in-range(0/0) ticket[^2], except for the
19:45 notviki rakudo/nom: small adjustment still needed in the exception it generates.
19:45 notviki rakudo/nom:
19:45 notviki rakudo/nom: [1] https://github.com/rakudo/rakudo/commit/1f3ca6469759c79aa809d60a7741b3a
19:45 notviki rakudo/nom: [2] https://rt.perl.org/Ticket/Display.html?id=130452#ticket-history
19:45 notviki rakudo/nom:
19:45 notviki rakudo/nom: review: https://github.com/rakudo/rakudo/commit/a567eb44cae0cdae0b83e7f37ae3ddaa2743df24
19:46 dalek roast: b5049e0 | (Zoffix Znet)++ | S03-operators/cmp.t:
19:46 dalek roast: Test infix:<cmp> with 0-denominator Rationals
19:46 dalek roast:
19:46 dalek roast: Rakudo fix: https://github.com/rakudo/rakudo/commit/a567eb44cae0cdae0
19:46 dalek roast: review: https://github.com/perl6/roast/commit/b5049e033a
19:55 dalek rakudo/nom: b2332c2 | (Zoffix Znet)++ | / (2 files):
19:55 dalek rakudo/nom: Fix Range.in-range dying when $got is a 0-denominator Rat
19:55 dalek rakudo/nom:
19:55 dalek rakudo/nom: By sticking a .perl() on it.
19:55 dalek rakudo/nom:
19:55 dalek rakudo/nom: Along with cmp fix[^1], Fixes
19:55 dalek rakudo/nom: RT#130452: https://rt.perl.org/Ticket/Display.html?id=130452
19:55 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=130452
19:55 dalek rakudo/nom:
19:55 dalek rakudo/nom: Since this is an internalish method, adding tests to t/ instead
19:55 dalek rakudo/nom: of roast.
19:55 dalek rakudo/nom:
19:55 dalek rakudo/nom: [1] https://github.com/rakudo/rakudo/commit/a567eb44cae0cdae0b83e
19:55 notviki rakudo/nom: review: https://github.com/rakudo/rakudo/commit/b2332c2172e8eaf52a0cce10266845bb61cf018a
19:56 notviki Perfect timing...
19:56 notviki I'm done working for the year!
19:56 * notviki relocates
20:02 Ven joined #perl6-dev
20:13 japhb Something has gone a bit screwy in the last few days with either Rakudo or Zef that is messing with the ecosystem.  See the issue I just filed for NativeHelpers::Blob:  https://github.com/salortiz/NativeHelpers-Blob/issues/2  -- this means DBIish won't install because NativeHelpers::Blob won't install because mysteriously one of the test files doesn't seem to run.
20:21 Ven joined #perl6-dev
20:22 notviki japhb, can you include in the issue the output of t basic? Right now the issue doen't include why it's failing
20:23 notviki "not tests run" likely means it failed to compile
20:29 pyrimidine joined #perl6-dev
20:34 japhb notviki: It compiles and runs fine with the slightly older Rakudo, and I can't see anything obvious in the test code.  (I can't immediately test again with newest Rakudo/zef because I don't have access on that machine right now.)
20:35 ugexe it installed ok for me with a week-ish old rakudo
20:38 japhb ugexe: Yeah, that's why I think something must have happened relatively recently
20:41 ugexe I just retested on blead and it also installed
20:41 japhb Huh.
20:42 japhb Maybe I got an unlucky revision.
20:42 ugexe however I'm not getting the warnings you are either
20:43 japhb ugexe: That's odd ... did you rebuild the whole stack with natural revisions (the ones specified with *_REVISION)?
20:43 japhb Also, what platform are you on?
20:46 ugexe I don't know what *_REVISION stuff is. I just rakudobrew nuke moar-blead && rakudobrew build moar-blead && perl6 -Ilib bin/zef install NativeHelpers::Blob
20:47 ugexe debian x86_64
20:48 japhb What version of gcc or clang?
20:48 ugexe gcc version 5.3.1 20160528 (Debian 5.3.1-21)
20:49 japhb My gcc is older (4.8.4 -- that machine is a slightly older Linux Mint, based on Ubuntu 14.04)
20:50 Ven joined #perl6-dev
20:51 japhb By *_REVISION, I meant NQP_REVISION and MOAR_REVISION (meaning whether you were building using the specified revisions, or forcing master).
20:51 japhb I don't use rakudobrew, I build from scratch ... I wonder if that has anything to do with it
20:52 ugexe HEAD
20:54 japhb ugexe: Is there a way to tell zef to test verbosely?
20:55 notviki you can clone the repo and run prove -e 'perl6 -Ilib' -vlr t/  in it
20:57 ugexe no, I havent looked at how to get that from TAP::Harness yet. You *could* do `--/tap-harness --/prove --perl6-test` (which just runs `perl6 t/$file` for every test file)
21:00 japhb OK, I'm going to get access on the machine that's having trouble again, switch to --per6-test, and try to rerun the test.
21:02 ugexe zef will have stored NativeHelpers::Blob in like ~/.zef/store/native-helpers.git/$some-sha1/ so you can also work from there directly
21:05 AlexDaniel m: dd Inf.Rat
21:05 camelia rakudo-moar b2332c: OUTPUT«P6opaque: get_boxed_ref could not unbox for the representation '20' of type Num␤  in block <unit> at <tmp> line 1␤␤»
21:05 AlexDaniel what's 20?
21:06 notviki rat
21:07 jnthn Rat ain't a REPR :)
21:07 jnthn Probably the internal ID for P6bigint
21:07 jnthn If that's happening, it probably means a non-Int was fed to an op that expected one
21:07 notviki well, it's sticking rat into isbig_I
21:07 jnthn Ah, or that
21:07 jnthn It won't go well
21:08 AlexDaniel m: dd <−4/2>
21:08 camelia rakudo-moar b2332c: OUTPUT«-2.0␤»
21:08 AlexDaniel notviki: what was the justification for not having a RatStr here?
21:08 AlexDaniel what: dd <−4/2>
21:08 evalable6 AlexDaniel, rakudo-moar b2332c2: OUTPUT«-2.0»
21:08 bisectable6 AlexDaniel, Bisecting by output (old=2015.12 new=b2332c2) because on both starting points the exit code is 0
21:08 notviki m: use nqp; nqp::isbig_I(1.1)
21:08 camelia rakudo-moar b2332c: OUTPUT«P6opaque: get_boxed_ref could not unbox for the representation '20' of type Rat␤  in block <unit> at <tmp> line 1␤␤»
21:08 committable6 AlexDaniel, ¦«2015.12,2016.02,2016.03,2016.04,2016.05,2016.06,2016.07.1,2016.08.1,2016.09,2016.10,2016.11»: "−4/2"␤¦«2016.12,HEAD»: -2.0
21:08 bisectable6 AlexDaniel, bisect log: https://gist.github.com/d4b3507e3f51ee42ee116226776c4219
21:08 bisectable6 AlexDaniel, (2016-11-21) https://github.com/rakudo/rakudo/commit/cb9df2bd40a10355d3bcc067943542611759f0c8
21:09 notviki AlexDaniel, I'm on the bus and yout Q scrolled outta view alreqdy :)
21:09 notviki will respond in ~30m
21:09 * notviki &
21:09 Ven joined #perl6-dev
21:09 AlexDaniel notviki: :)
21:10 jnthn Hm, the error suggest it passed a Num to a bigint op rather than a Rat though
21:10 jnthn Anyway, the answer is "fix Rakudo to not do that" :)
21:11 jnthn A Rat can't represent Inf, so a Failure is probably a reasonable answer
21:12 AlexDaniel are you sure about this?
21:12 AlexDaniel m: say (1/0).Num
21:12 camelia rakudo-moar b2332c: OUTPUT«Inf␤»
21:13 jnthn m: say (-1/0).Num
21:13 camelia rakudo-moar b2332c: OUTPUT«-Inf␤»
21:13 jnthn Hmm
21:13 AlexDaniel m: say (0/0).Num
21:13 camelia rakudo-moar b2332c: OUTPUT«NaN␤»
21:13 AlexDaniel yea…
21:13 jnthn Heh, then why ain't it producing those Rats on coercion from Num?
21:13 jnthn If it roundtrips I guess it's OK
21:14 AlexDaniel we can probably bisect by this to see a nice commit message…
21:14 AlexDaniel b: say (0/0).Num
21:14 bisectable6 AlexDaniel, Bisecting by output (old=2015.12 new=b2332c2) because on both starting points the exit code is 0
21:14 bisectable6 AlexDaniel, bisect log: https://gist.github.com/b78cc312c803dc278abdc81b768043d4
21:14 bisectable6 AlexDaniel, (2016-05-21) https://github.com/rakudo/rakudo/commit/1f3ca6469759c79aa809d60a7741b3a92fb9330f
21:14 AlexDaniel m: say Inf.Rat
21:14 camelia rakudo-moar b2332c: OUTPUT«Inf␤»
21:14 AlexDaniel m: say Inf.Rat.perl
21:14 camelia rakudo-moar b2332c: OUTPUT«P6opaque: get_boxed_ref could not unbox for the representation '20' of type Num␤  in block <unit> at <tmp> line 1␤␤»
21:15 AlexDaniel oohh so it is indeed supposed to give 1/0 and -1/0 and 0/0 !
21:15 jnthn Oh...
21:15 jnthn Wat.
21:15 jnthn So it's .perl that's bust, not the coercion?
21:15 AlexDaniel c: 1f3ca64 say Inf.Rat.perl
21:15 committable6 AlexDaniel, ¦«1f3ca64»: <1/0>
21:15 AlexDaniel b: old=1f3ca64 say Inf.Rat.perl
21:15 bisectable6 AlexDaniel, Bisecting by exit code (old=1f3ca64 new=b2332c2). Old exit code: 0
21:15 bisectable6 AlexDaniel, bisect log: https://gist.github.com/06b5dd495421121f786f316a6aad11c5
21:15 bisectable6 AlexDaniel, (2016-05-26) https://github.com/rakudo/rakudo/commit/498d0a4ae4572a84f80c0c572632e12017d779f3
21:15 jnthn These bots :D
21:16 jnthn This basically creates a special purpose Rat in which the numerator
21:16 jnthn is a Num rather than an Int.
21:16 jnthn ...what??
21:16 jnthn That sounds fragile.
21:17 geekosaur jnthn, it's a hack to roundtrip NaN and Inf
21:17 geekosaur because anything else is worse
21:18 AlexDaniel geekosaur: but it no longer roundtrips after this commit, no?
21:18 geekosaur (this was the result of a fair amount of futzing around trying to come up with solutions)
21:18 AlexDaniel commit: 498d0a4ae4^,498d0a4ae4 say Inf.Rat.Num
21:18 committable6 AlexDaniel, ¦«498d0a4ae4^,498d0a4»: Inf
21:18 AlexDaniel oh no, it does
21:18 AlexDaniel commit: 498d0a4ae4^,498d0a4ae4 say Inf.Rat
21:18 committable6 AlexDaniel, ¦«498d0a4ae4^,498d0a4»: Inf
21:18 AlexDaniel commit: 498d0a4ae4^,498d0a4ae4 say Inf.Rat.perl
21:18 committable6 AlexDaniel, ¦«498d0a4ae4^»: <1/0>␤¦«498d0a4»: P6opaque: get_boxed_ref could not unbox for the representation '20' of type Num␤  in block <unit> at /tmp/CAyYTbCMma line 1␤ «exit code = 1»
21:18 AlexDaniel you just can't .perl it
21:19 geekosaur I don't get that commit being "no longer roundtrips"; that commit *implemented* roundtrip-ability for Inf/NaN through Rat
21:19 geekosaur and required hackery elsewhere, of course
21:20 jnthn m: say Inf.Rat
21:20 camelia rakudo-moar b2332c: OUTPUT«Inf␤»
21:20 jnthn m: say Inf.Rat.WHAT
21:20 camelia rakudo-moar b2332c: OUTPUT«(Rational[Num,Int])␤»
21:21 jnthn To be honest, I'd probably have been more inclined to just say "no, this type can't represent Inf/NaN"
21:21 jnthn But I guess that ship has sailed.
21:21 AlexDaniel mc: say Inf.Rat
21:21 committable6 AlexDaniel, ¦«2015.12»: Inf
21:21 AlexDaniel mc: say Inf.Rat.WHAT
21:21 committable6 AlexDaniel, ¦«2015.12»: (Num)
21:22 jnthn Urgh. That's worse. :)
21:22 jnthn At least we return a Rat now.
21:22 AlexDaniel I don't think it has… but we have to make up our mind or something
21:22 AlexDaniel 6c: say Inf.Rat.WHAT
21:22 jnthn Even if it's a Rat of unusual shape
21:23 committable6 AlexDaniel, ¦«2015.12,2016.02,2016.03,2016.04,2016.05»: (Num)␤¦«2016.06,2016.07.1,2016.08.1,2016.09,2016.10,2016.11,2016.12,HEAD»: (Rational[Num,Int])
21:24 geekosaur something required the ability to roundtrip it (granted that "something" may have been our favorite numerical gadfly)
21:24 geekosaur (that is, Zefrem)
21:26 jnthn Anyway, I guess for now Rational.perl will need patching up some
21:30 Ven joined #perl6-dev
21:31 AlexDaniel yea, ok, RT #130459
21:31 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=130459
21:38 samcv morning everyone
21:38 yoleaux2 17:54Z <cygx> samcv: I believe you misread RT #129878 - the issue is not stringifying a list generated by xx (did you miss the [~]), but the string generated by x
21:38 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=129878
21:38 jnthn o/ samcv
21:39 samcv hmm i thought that is what i said. maybe i reversed them
21:39 samcv say ("\c[REGIONAL INDICATOR SYMBOL LETTER G]" x 2).chars #=> 2 cause this one has the problem
21:39 samcv m:  say ("\c[REGIONAL INDICATOR SYMBOL LETTER G]" x 2).chars #=> 2 cause this one has the problem
21:39 camelia rakudo-moar b2332c: OUTPUT«2␤»
21:40 samcv it is a problem with x
21:40 samcv well.
21:41 samcv .tell cygx i will look at it as soon as I wake up again. if i misinterpreted the bug
21:41 yoleaux2 samcv: I'll pass your message to cygx.
21:42 samcv m: say ("\c[REGIONAL INDICATOR SYMBOL LETTER G]" xx 2).elems #> 2
21:42 camelia rakudo-moar b2332c: OUTPUT«2␤»
21:42 samcv m: say ("\c[REGIONAL INDICATOR SYMBOL LETTER G]" x 2).elems
21:42 camelia rakudo-moar b2332c: OUTPUT«1␤»
21:42 samcv m: say ("\c[REGIONAL INDICATOR SYMBOL LETTER G]" x 2).ords
21:42 camelia rakudo-moar b2332c: OUTPUT«(127468 127468)␤»
21:42 samcv m: say ("\c[REGIONAL INDICATOR SYMBOL LETTER G]" x 2).Str.ords
21:42 camelia rakudo-moar b2332c: OUTPUT«(127468 127468)␤»
21:42 samcv m: say ("\c[REGIONAL INDICATOR SYMBOL LETTER G]" x 2).Str.chars
21:42 camelia rakudo-moar b2332c: OUTPUT«2␤»
21:43 samcv m: say ("\c[REGIONAL INDICATOR SYMBOL LETTER G]" xx 2).Str.ords
21:43 camelia rakudo-moar b2332c: OUTPUT«(127468 32 127468)␤»
21:43 samcv m: say ("\c[REGIONAL INDICATOR SYMBOL LETTER G]" x 2).Str.ords
21:43 camelia rakudo-moar b2332c: OUTPUT«(127468 127468)␤»
21:43 samcv m: say ("\c[REGIONAL INDICATOR SYMBOL LETTER G]" x 2).Str.chars
21:43 camelia rakudo-moar b2332c: OUTPUT«2␤»
21:43 samcv ok i made a mistake! yep
21:45 jnthn samcv: There's an optimized representation used by `x` that likely should not be in all cases :)
21:46 jnthn samcv: I suspect we can detect it by exposing the should_break thingy used by NFG and calling it with the string to repeat as both arguments
21:46 AlexDaniel and this concludes my second iteration! \o\ /o/ \o/
21:46 AlexDaniel only 1084 snippets left now!
21:48 jnthn AlexDaniel: Snippets of what, out of curiosity?
21:48 AlexDaniel jnthn: oh, ah!
21:49 AlexDaniel jnthn: so what I did is I've taken all m: messages from the clog of this year
21:49 AlexDaniel jnthn: and ran each of them on 2015.12,HEAD and also somewhere around the time the message was sent
21:49 AlexDaniel there are around 31.5k of them in total
21:49 AlexDaniel then I've filtered out those that had same output
21:50 AlexDaniel which left me with only 7k or something like this
21:50 AlexDaniel and… then I've been going through all of them… that's where all these latest tickets came from
21:51 notviki aitht
21:51 jnthn Wow :)
21:53 notviki "AlexDaniel │ notviki: what was the justification for not having a RatStr here?"
21:53 notviki AlexDaniel: because that's the syntax for Rat literal, not for a RatStr
21:53 AlexDaniel right
21:53 AlexDaniel okay
21:54 notviki AlexDaniel: this lets you, for example, specify a Rat in a signature as a literal
21:54 notviki m: -> <0/0> { "that looks fancy pants" }(0/0)
21:54 camelia rakudo-moar b2332c: ( no output )
21:54 notviki m: -> <0/0> { "that looks fancy pants".say }(0/0)
21:54 camelia rakudo-moar b2332c: OUTPUT«that looks fancy pants␤»
21:54 notviki m: -> <0/0> { "that looks fancy pants".say }(1/0)
21:54 camelia rakudo-moar b2332c: OUTPUT«that looks fancy pants␤»
21:54 notviki /0\
21:55 notviki well, shit
21:55 notviki m: -> <0/0> { "that looks fancy pants".say }(10/2)
21:55 camelia rakudo-moar b2332c: OUTPUT«that looks fancy pants␤»
21:55 notviki omg
21:56 notviki AlexDaniel: does that answer the question? Same goes for Complex. I guess the real question is how come <42> or <42e0> do not return an Int and Num and do the allomorphs, and that's 'cause you can get the literals for those without the angled brackets
21:56 notviki m: -> <10/3> { "that looks fancy pants".say }(10/2)
21:56 camelia rakudo-moar b2332c: OUTPUT«Constraint type check failed for parameter '<anon>'␤  in block <unit> at <tmp> line 1␤␤»
21:56 notviki Interesting
21:56 * notviki clicks "Take" on that one
21:58 AlexDaniel notviki: yea, I guess so. Doesn't change the fact that I hate everything about < > though :)
21:59 notviki heh
21:59 notviki I think it only has one bug left
21:59 notviki and I'd say "bug"
21:59 notviki m: say 1+i
21:59 camelia rakudo-moar b2332c: OUTPUT«1+1i␤»
21:59 notviki m: say WHAT 1+i
21:59 camelia rakudo-moar b2332c: OUTPUT«(Complex)␤»
21:59 notviki m: say WHAT <1+i>
21:59 camelia rakudo-moar b2332c: OUTPUT«(Str)␤»
21:59 notviki ^ that
22:09 Ven joined #perl6-dev
22:15 * japhb is back from doing a fresh reinstall test
22:16 japhb And ... it worked with the latest nom.  Not a blip.  I guess I did just get unlucky before.
22:26 committable6 joined #perl6-dev
22:27 lizmat good night, #perl6-dev!
22:28 notviki night
22:28 japhb o/
22:28 Ven joined #perl6-dev
22:29 samcv m: Uni.new(0x2640, 0xFE0E).chars.say
22:29 camelia rakudo-moar b2332c: OUTPUT«No such method 'chars' for invocant of type 'Uni'␤  in block <unit> at <tmp> line 1␤␤»
22:29 samcv m: Uni.new(0x2640, 0xFE0E).Str.chars
22:29 camelia rakudo-moar b2332c: ( no output )
22:29 samcv m: Uni.new(0x2640, 0xFE0E).Str.chars.say
22:29 camelia rakudo-moar b2332c: OUTPUT«1␤»
22:44 samcv m: '('.uniprop('MVM_COLLATION_PRIMARY').say
22:44 camelia rakudo-moar b2332c: OUTPUT«791␤»
22:44 samcv sweet!
22:45 samcv \o/
22:45 samcv m: 0x200D.uniprop('NFG_QC').say
22:45 camelia rakudo-moar b2332c: OUTPUT«N␤»
22:45 notviki m: '('.uniprop('MVM_COLLATION_PRIMARY').chr.say
22:45 camelia rakudo-moar b2332c: OUTPUT«̗␤»
22:45 notviki j: '('.uniprop('MVM_COLLATION_PRIMARY').say
22:46 camelia rakudo-jvm 8ca367: OUTPUT«uniprop NYI on jvm backend␤  in method throw at gen/jvm/CORE.setting line 27529␤  in sub die at gen/jvm/CORE.setting line 792␤  in sub uniprop at gen/jvm/CORE.setting line 15437␤  in method uniprop at gen/jvm/CORE.setting line 9661␤  in block <uni…»
22:46 notviki heh
22:46 * notviki runs a stripts on all open RTs to try to find ones that are fixed and "test needed" but not marked as such
22:47 notviki "Smartmatch is experimental at check-RT-for-tests-needed.pl line 20"
22:47 notviki ^_^
22:47 notviki "Global symbol "%RTs" requires explicit package name"
22:48 notviki no.. but I did forget parens for (@RTs) { ....
22:48 notviki That's it. I forgot how to code Perl 5 :(
22:49 Ven joined #perl6-dev
22:49 samcv m: "\c[BOY]".say
22:49 camelia rakudo-moar b2332c: OUTPUT«????␤»
22:49 samcv m: "\c[BOY]".uniprop('NFG_QC').say
22:49 camelia rakudo-moar b2332c: OUTPUT«Y␤»
22:50 samcv wait jnthn it's already nfg_qc?
22:50 samcv oh wait. nvm. it should be N
22:50 samcv still haven't been up long
23:06 notviki hoh
23:07 notviki looks like our Appvoyer config needs a tweak? It can't get perl5 to install: https://ci.appveyor.com/project/moritz/rakudo/build/1.0.1967/job/r1d4ce8i3rlbdiis
23:08 Ven joined #perl6-dev
23:09 dalek rakudo/nom: a2cc58a | skids++ | src/core/Proc/Async.pm:
23:09 dalek rakudo/nom: Make new Proc::Async constructor friendly to subclass attribs.
23:09 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/a2cc58a754
23:09 dalek rakudo/nom: 87fefa9 | (Zoffix Znet)++ | src/core/Proc/Async.pm:
23:09 dalek rakudo/nom: Merge pull request #978 from skids/procasyncsubclass
23:09 dalek rakudo/nom:
23:09 dalek rakudo/nom: Make new Proc::Async constructor friendly to subclass attribs.
23:09 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/87fefa9c70
23:26 notviki Hm, 515 ticket mentioning word "test": https://gist.github.com/zoffixznet/933a2a2c9ed3ca95790b725f81f86249
23:26 * notviki tries another run with a better heuristic
23:28 Ven joined #perl6-dev
23:44 dalek nqp: eece127 | samcv++ | docs/ops.markdown:
23:44 dalek nqp: ops.markdown: provide more accurate and detailed info on unipropcode
23:44 dalek nqp: review: https://github.com/perl6/nqp/commit/eece1271de
23:47 notviki samcv++ # fixing the atom bug with '/'
23:47 notviki Working with that same file now and glad it's not butchered anymore
23:47 samcv :)
23:47 AlexDaniel \o/
23:47 samcv yeah. it parses left to right so it will run into the quoted string
23:48 samcv since that isn't parsed yet
23:48 samcv \o/
23:48 Ven joined #perl6-dev
23:49 notviki hm
23:49 notviki m: say Rat.Range
23:49 camelia rakudo-moar 87fefa: OUTPUT«-Inf^..^Inf␤»
23:49 notviki I just now realize that's wrong and original behaviour is right
23:49 notviki for RT https://rt.perl.org/Ticket/Display.html?id=130427#ticket-history
23:50 notviki m: say [.^name, .Num] for <1/0>, <-1/0>;
23:50 camelia rakudo-moar 87fefa: OUTPUT«[Rat Inf]␤[Rat -Inf]␤»
23:51 notviki m: say Rat.Range.in-range: NaN
23:51 camelia rakudo-moar 87fefa: OUTPUT«Value out of range. Is: NaN, should be in -Inf^..^Inf␤  in block <unit> at <tmp> line 1␤␤»
23:52 samcv m: "\c[ARABIC NUMBER SIGN]".uniprop('Grapheme_Cluster_Break').say
23:52 camelia rakudo-moar 87fefa: OUTPUT«Prepend␤»
23:52 japhb .tell jnthn Found another possible oo-monitors bug for you (https://github.com/jnthn/oo-monitors/issues/9).  Hopefully this one is golfed down enough to make it easier to debug.
23:52 yoleaux2 japhb: I'll pass your message to jnthn.
23:53 japhb It seems like every bug I find ends up on jnthn's plate.  :-(
23:56 Ven joined #perl6-dev

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