Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6-dev, 2016-12-13

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

All times shown according to UTC.

Time Nick Message
00:06 samcv nice, more work on uniprops imminent. getting Titlecase_Mapping, Uppercase_Mapping, Lowercase_Mapping and 'Name' working
00:16 dalek roast: 17a5f0e | samcv++ | S15-unicode-information/uniprop.t:
00:16 dalek roast: Add uniprop tests for 'Name' property and some rewording desc.
00:16 dalek roast: review: https://github.com/perl6/roast/commit/17a5f0e278
00:53 dalek roast: 82f679d | samcv++ | S15-unicode-information/uniprop.t:
00:53 dalek roast: Add tests for uniprop 'Numeric_Value' and additional others
00:53 dalek roast:
00:53 dalek roast: Add tests 'Dash' and 'Script' as well, and keep a list
00:53 dalek roast: of the ones we test the actual value of at the top.
00:53 dalek roast: Eventually we should test all Unicode properties
00:53 dalek roast: on at least one codepoint per property
00:53 dalek roast: review: https://github.com/perl6/roast/commit/82f679dee3
01:04 phatbot in $uniprop-type = 'B' if ~$type eq 'Binary'; and one line below it, no need for ~$type (it's already done that a few lines above and eq would work fine even if it weren't)
01:11 dalek roast: 1dde4a5 | samcv++ | S15-unicode-information/uniprop.t:
01:11 dalek roast: uniprop: test "Bidi_Paired_Bracket"
01:11 dalek roast: review: https://github.com/perl6/roast/commit/1dde4a528f
01:26 b2gills Perhaps *.uniprop('Numeric_Value') could just redirect to *.unival . at least for the time being.
01:32 unicodable6 joined #perl6-dev
01:35 unicodable6 joined #perl6-dev
01:40 pyrimidine joined #perl6-dev
01:45 phatbot t/spec/S15-unicode-information/uniprop.rakudo.moar              (Wstat: 256 Tests: 0 Failed: 0)
01:45 phatbot Non-zero exit status: 1
01:45 phatbot Parse errors: No plan found in TAP output
01:47 samcv b2gills, that's what my PR does
01:47 samcv hmm did i mess up
01:48 b2gills samcv: Oh I didn't look
01:48 dalek roast: f0f06aa | samcv++ | S15-unicode-information/uniprop.t:
01:48 dalek roast: Test 'Bidi_Mirroring_Glyph' for uniprop
01:48 dalek roast: review: https://github.com/perl6/roast/commit/f0f06aad5b
01:49 samcv i was talking to phatbot
01:49 samcv but yeah the PR i made does unival for 'nv'/'Numeric_Value'
01:51 phatbot Ah.
01:51 phatbot samcv: todo fudges need a message
01:51 samcv oh
01:51 samcv k
01:51 phatbot well, all fudge I guess.
01:51 phatbot ------> <BOL>⏏todo(); isa-ok "½".uniprop('Numeric_Valu
01:51 phatbot Calling todo() will never work with any of these multi signatures:
01:52 phatbot ($reason, $count = 1)
01:52 samcv ah
01:56 dalek roast: b171e24 | samcv++ | S15-unicode-information/uniprop.t:
01:56 dalek roast: Add messages for the `todo`'s
01:56 dalek roast: review: https://github.com/perl6/roast/commit/b171e24ca6
02:00 samcv u: LRM
02:00 unicodable6 samcv, Found nothing!
02:00 samcv u: directional
02:00 unicodable6 samcv, U+202C POP DIRECTIONAL FORMATTING [Cf] (‬)
02:00 unicodable6 samcv, U+2069 POP DIRECTIONAL ISOLATE [Cf] (⁩)
02:00 samcv u: direction
02:00 unicodable6 samcv, U+202C POP DIRECTIONAL FORMATTING [Cf] (‬)
02:00 unicodable6 samcv, U+2069 POP DIRECTIONAL ISOLATE [Cf] (⁩)
02:00 unicodable6 samcv, U+1DA07 SIGNWRITING FACE DIRECTION POSITION NOSE FORWARD TILTING [Mn] (◌????)
02:00 unicodable6 samcv, https://gist.github.com/52e1ea8f4e7a29627049c6b5a7081404
02:01 samcv u: rtl
02:01 unicodable6 samcv, U+2EF1 CJK RADICAL TURTLE [So] (⻱)
02:01 unicodable6 samcv, U+2EF2 CJK RADICAL J-SIMPLIFIED TURTLE [So] (⻲)
02:01 unicodable6 samcv, U+2EF3 CJK RADICAL C-SIMPLIFIED TURTLE [So] (⻳)
02:01 unicodable6 samcv, https://gist.github.com/5235caa780b878f9103d406b44816fa4
02:04 jdv79 how many bots do we have now?
02:04 phatbot not enough :)
02:04 kalkin-_ joined #perl6-dev
02:05 * phatbot counts 13 in #zofbot
02:15 dalek roast: cb977cc | (Zoffix Znet)++ | S15-unicode-information/uniprop.t:
02:15 dalek roast: Quote fudge messages
02:15 dalek roast: review: https://github.com/perl6/roast/commit/cb977ccf3e
02:19 gfldex joined #perl6-dev
02:21 dalek rakudo/nom: 6fc7698 | (Zoffix Znet)++ | src/Perl6/Actions.nqp:
02:21 dalek rakudo/nom: Throw useful message when <sym> is not used in a proto regex
02:21 dalek rakudo/nom:
02:21 dalek rakudo/nom: Bug finding: https://irclog.perlgeek.de/perl6/2016-12-12#i_13729253
02:21 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/6fc76989c0
02:22 dalek roast: 81cd6a6 | (Zoffix Znet)++ | S05-grammar/action-stubs.t:
02:22 dalek roast: Test incorrect use of <sym> throws useful message
02:22 dalek roast:
02:22 dalek roast: Bug finding: https://irclog.perlgeek.de/perl6/2016-12-12#i_13729253
02:22 dalek roast: Rakudo fix: https://github.com/rakudo/rakudo/commit/6fc76989c0
02:22 dalek roast: review: https://github.com/perl6/roast/commit/81cd6a6c04
02:23 phatbot Enough for the day.
02:23 * phatbot drops to bed
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
02:49 pyrimidine joined #perl6-dev
02:55 pyrimidine joined #perl6-dev
03:31 pyrimidine joined #perl6-dev
05:10 dalek roast: bc9e619 | samcv++ | S15-unicode-information/uniprop.t:
05:10 dalek roast: uniprop: add tests for 6 more Unicode properties.
05:10 dalek roast:
05:10 dalek roast: ASCII_Hex_Digit
05:10 dalek roast: Hex_Digit
05:10 dalek roast: Case_Ignorable
05:10 dalek roast: Bidi_Class
05:10 dalek roast: Word_Break
05:10 dalek roast: Line_Break
05:10 dalek roast: review: https://github.com/perl6/roast/commit/bc9e619a8b
05:15 samcv tests for 17 so far.. out of ~110 unicode properties total...
05:15 samcv :(
05:25 pyrimidine joined #perl6-dev
06:09 dalek roast: 2fd1806 | samcv++ | S15-unicode-information/uniprop.t:
06:09 dalek roast: uniprop add tests for Age, na1, Soft_Dotted and Quotation_Mark
06:09 dalek roast: review: https://github.com/perl6/roast/commit/2fd1806653
06:31 pyrimidine joined #perl6-dev
06:50 RabidGravy joined #perl6-dev
07:36 pyrimidine joined #perl6-dev
07:55 [Tux] This is Rakudo version 2016.11-221-g6fc76989c built on MoarVM version 2016.11-44-g8e24145d
07:55 [Tux] csv-ip5xs        3.235
07:55 [Tux] test            14.638
07:55 [Tux] test-t           6.459
07:55 [Tux] csv-parser      15.976
08:30 pyrimidine joined #perl6-dev
09:31 pyrimidine joined #perl6-dev
09:55 pmurias joined #perl6-dev
09:56 dalek nqp: 8db31c6 | (Pawel Murias)++ | src/vm/js/nqp-runtime/container-specs.js:
09:56 dalek nqp: [js] Refactor to use STable.addInternalMethods(class {...})
09:56 dalek nqp:
09:56 dalek nqp: Also fix some method signatures.
09:56 dalek nqp: review: https://github.com/perl6/nqp/commit/8db31c6b79
09:56 dalek nqp: 2987189 | (Pawel Murias)++ | t/moar/02-qast-references.t:
09:56 dalek nqp: Fix test descriptions.
09:56 dalek nqp: review: https://github.com/perl6/nqp/commit/298718986b
09:56 dalek nqp: c4cf849 | (Pawel Murias)++ | src/vm/js/nqp-runtime/container-specs.js:
09:56 dalek nqp: [js] Make nqp::iscont work on NativeRefs.
09:56 dalek nqp: review: https://github.com/perl6/nqp/commit/c4cf8499a5
09:56 dalek nqp: b9612c5 | (Pawel Murias)++ | src/vm/js/Operations.nqp:
09:56 dalek nqp: [js] Implement nqp::xor.
09:56 dalek nqp: review: https://github.com/perl6/nqp/commit/b9612c5747
10:14 dalek nqp: f5ceea7 | (Pawel Murias)++ | t/nqp/060-bigint.t:
10:14 dalek nqp: [js] Stop skipping a test.
10:14 dalek nqp:
10:14 dalek nqp: The bug in bignum that necessiated that has been fixed by hoelzro++.
10:14 dalek nqp: review: https://github.com/perl6/nqp/commit/f5ceea7a7d
10:44 lucasb joined #perl6-dev
10:56 pyrimidine joined #perl6-dev
11:01 phatbot samcv: I don't think you have to worry about adding fudges for neizca
11:01 phatbot samcv: can't that be somehow automatically pre-generated? RE tests for 17 so far.. out of ~110 unicode properties total...
11:11 cognominal joined #perl6-dev
11:12 dogbert17 joined #perl6-dev
13:37 pyrimidine joined #perl6-dev
13:38 dalek rakudo/nom: 7e1ede4 | Altai-man++ | src/Perl6/Metamodel/RoleToClassApplier.nqp:
13:38 dalek rakudo/nom: Fix of RT #130211 with creating a new hash that connects method names to their respective roles and using that hash to get role name in case of error.
13:38 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/7e1ede4e68
13:38 dalek rakudo/nom: 2e487e2 | Altai-man++ | src/Perl6/Metamodel/RoleToClassApplier.nqp:
13:38 dalek rakudo/nom: Move method-role gathering code into error block
13:38 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=130211
13:38 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/2e487e2ff6
13:38 dalek rakudo/nom: c254931 | Altai-man++ | src/Perl6/Metamodel/RoleToClassApplier.nqp:
13:38 dalek rakudo/nom: We don't need to build a table for all methods, we only need particular method role.
13:38 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/c2549310d5
13:38 dalek rakudo/nom: bff0510 | (Zoffix Znet)++ | src/Perl6/Metamodel/RoleToClassApplier.nqp:
13:38 dalek rakudo/nom: Merge pull request #954 from Altai-man/rt130211-fix
13:38 dalek rakudo/nom:
13:38 dalek rakudo/nom: Display role requiring stubbed method
13:38 dalek rakudo/nom:
13:42 |Tux| joined #perl6-dev
14:02 Ven joined #perl6-dev
14:11 [Coke] There is no point in fudging anything other than rakudo at this point
14:19 dalek rakudo/nom: 1d46004 | (Zoffix Znet)++ | src/core/Range.pm:
14:19 dalek rakudo/nom: Fix crash on Rat Ranges constructed with Range ops
14:19 dalek rakudo/nom:
14:19 dalek rakudo/nom: When using range ops, we clone the Range. However, the code
14:19 dalek rakudo/nom: makes a mistaken assumption that Mu.clone can tweedle anything,
14:19 dalek rakudo/nom: but only things with accessors can be passed as tweedles. As a
14:19 dalek rakudo/nom: result, the private $!is-int attribute wasn't being set, leading
14:19 dalek rakudo/nom: to an explosion when a Rat range was being treated as an Int Range.
14:19 dalek rakudo/nom:
14:19 dalek rakudo/nom: Fix by setting $!is-int attr in the clone ourselves.
14:19 dalek rakudo/nom:
14:19 dalek rakudo/nom: Fixes RT#128887: https://rt.perl.org/Ticket/Display.html?id=128887
14:19 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=128887
14:19 phatbot review: https://github.com/rakudo/rakudo/commit/1d46004dd3575a02f3656b4567ff5189caf180b9
14:20 dalek roast: d23534b | (Zoffix Znet)++ | S02-types/range.t:
14:20 dalek roast: Test Rat Ranges made with Range ops do not crash
14:20 dalek roast:
14:20 dalek roast: RT#128887: https://rt.perl.org/Ticket/Display.html?id=128887
14:20 dalek roast: Rakudo fix: https://github.com/rakudo/rakudo/commit/1d46004dd3575a0
14:20 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=128887
14:20 dalek roast: review: https://github.com/perl6/roast/commit/d23534b2d7
14:22 phatbot nom nom nyam...
14:22 phatbot Tasty bug
14:27 phatbot Not 100% sure $clone.WHAT is OK. Should it be ::?CLASS or Range? Not sure how that will play out in subclasses
15:20 MasterDuke_ joined #perl6-dev
15:22 dalek roast: 815e0cd | lizmat++ | S15-unicode-information/uniprop.t:
15:22 dalek roast: Unfudge passing tests
15:22 dalek roast: review: https://github.com/perl6/roast/commit/815e0cd356
15:36 MasterDuke_ jnthn: i tried removing first the sink call (just having $past in the QAST::Want), then removing the Want completely (in my change -n to 'while get()' attempt). no errors when run, but running '-ne ".say"' doesn't print anything
15:39 MasterDuke_ i tried just blindly calling tweak_loop on my $past, but it dies at https://github.com/rakudo/rakudo/blob/nom/src/Perl6/Actions.nqp#L1741 complaining about trying to get an attribute from a NQPMu (or something like that)
15:40 MasterDuke_ i haven't yet tried just pulling out bits of tweak_loop manually
15:43 MasterDuke_ but i'm pretty sure an END phaser worked without doing anything tweak_loop related, are they handled differently? (i.e., if i tried LAST it wouldn't work?)
15:43 pyrimidine joined #perl6-dev
15:54 dogbert17 .seen jnthn
15:54 yoleaux2 I saw jnthn 12:50Z in #moarvm: <jnthn> o/ brrt
15:56 dogbert17 jnthn: does the included gist contain anything useful from a bug hunting perspective, if so I'll report it. https://gist.github.com/dogbert17/8e0194e19bcb48477b7500f86a3c8f33
16:00 [Coke] m: say "½".uniprop('Numeric_Value').WHAT
16:00 camelia rakudo-moar 1d4600: OUTPUT«(Str)␤»
16:03 phatbot m: dd "½".uniprop('Numeric_Value')
16:03 camelia rakudo-moar 1d4600: OUTPUT«"0.5"␤»
16:03 phatbot neat
16:03 phatbot u: half zero
16:03 unicodable6 phatbot, U+0F33 TIBETAN DIGIT HALF ZERO [No] (༳)
16:03 phatbot m: dd "༳".uniprop('Numeric_Value')
16:03 camelia rakudo-moar 1d4600: OUTPUT«"-0.5"␤»
16:06 jnthn dogbert17: Yes, somewhat, glancing at it quickly. Was MVM_GC_DEBUG turned on at all?
16:06 dogbert17 jnthn: I can check when I get home. Did you expect more output?
16:08 jnthn Well, if there isn't more then it's a less obvious bug...
16:08 dogbert17 *if* it was on it was set to 1
16:08 jnthn OK; try it on 2 also if that doesn't slow stuff down enough to hide the bug :)
16:08 dogbert17 will do, running with 64k nursery
16:11 [Coke] looks like we have a bug report from mt1957 on perl6-users
16:24 * phatbot read title only so far...
16:24 phatbot Doesn't sound good :}
16:24 phatbot For those not subbed: http://www.nntp.perl.org/group/perl.perl6.users/2016/12/msg3094.html
16:25 lizmat I think there's an RT for that now as well
16:26 phatbot ah yeah https://rt.perl.org/Ticket/Display.html?id=130338#ticket-history
16:30 lizmat RabidGravy++ # nice post
16:30 lizmat RabidGravy: some paragraphs seem to have formatting issues at the end
16:31 lizmat e.g.: It couldn’t really be much more
16:31 lizmat simple.
16:31 lizmat simple being on the next line without it needing to be
16:34 pyrimidine joined #perl6-dev
16:41 phatbot the version in that ticket is 2016.11-238-g2f502b4 but when I try https://github.com/rakudo/rakudo/commit/2f502b4 I get a 404, presumably because there's more than one commit with that short part?
16:41 * phatbot wonders what's the easy way around that issue
16:42 phatbot https://github.com/rakudo/rakudo/commit/2016.11-238-g2f502b4 is a blank page entirely :o
16:43 pyrimidine joined #perl6-dev
16:43 phatbot error: pathspec '2016.11-238-g2f502b4' did not match any file(s) known to git.
16:44 * phatbot is puzzled
16:44 phatbot I'm on head and I'm only on 2016.11-226-g1d46004 wtf
16:45 phatbot I guess it's proof that time travel exists...
16:46 geekosaur usually get it to emit a longer hash
16:46 phatbot geekosaur: how?
16:46 phatbot m: $*PERL.compiler.version.say
16:46 camelia rakudo-moar 1d4600: OUTPUT«v2016.11.226.g.1.d.46004␤»
16:46 geekosaur I doubt the full hash is in the perl config, sadly
16:46 phatbot Ohhh... that's a puzzle piece actually.
16:47 phatbot Oh wait, no. It would've read 2016.*10* for what I was thinking.
16:47 geekosaur if it was that, git fetch --tags would fix it
16:48 * phatbot replies asking for more info
16:53 geekosaur might be worth figuring out how to embed the full hash (from e.g. git log)
16:54 geekosaur won't help now of course
16:55 phatbot I guess it's not needed, since there's a tag too and that has the distance from the tag in it too, doesn't it?
16:55 phatbot like -226- is 226th commit from the tag
16:55 phatbot in 2016.11.226.g.1.d.46004
16:57 geekosaur but not everything takes git describe output
16:57 phatbot ah
16:58 geekosaur a full hash should work everywhere
16:59 geekosaur ok, not a branch
16:59 * geekosaur had wondered in particular if it was the lexical modules stuff...
17:27 jnthn MasterDuke_: Sorry, was tied up with a $dayjob task before. END is handled quite differently from the loop phasers, so wouldn't be affected by tweak_loop not being called. It's LAST/NEXT/FIRST that would be at issue. Try added a QAST::Op.new( :op('say'), QAST::SVal.new( :value('here') ) ) or so just above the loop to see you reach that code path though
17:29 dogbert2 joined #perl6-dev
17:30 dogbert2 jnthn: gist updated, 64k nursery, MVM_GC_DEBUG=2. https://gist.github.com/dogbert17/8e0194e19bcb48477b7500f86a3c8f33
17:31 * phatbot Takes RT130205
17:31 * phatbot Takes RT130213
17:31 phatbot Sometime to hack on for the day.
17:31 phatbot *thing
17:33 jnthn dogbert2: ooh! :)
17:33 jnthn That's more interesting
17:34 dogbert2 cool, it's possible to break quite a few tests with these settings
17:38 MasterDuke_ jnthn: thanks, i'll play around with it some more this evening
17:40 dogbert2 should I report a MoarVM issue with the data from the gist?
17:40 jnthn dogbert2: Please. I'll take a look tomorrow. :)
17:41 jnthn dinner &
17:42 pyrimidine joined #perl6-dev
17:52 b2gills m: (supply loop { emit Bool.pick }).tap: *.say # Should the supply "keyword" be able used like this
17:52 camelia rakudo-moar 1d4600: OUTPUT«No such method 'phasers' for invocant of type 'Code'␤  in block <unit> at <tmp> line 1␤␤»
17:53 b2gills m: say supply loop {...}
17:53 camelia rakudo-moar 1d4600: OUTPUT«Supply.new␤»
17:57 phatbot m: react { whenever (supply {loop { emit Bool.pick }}).tap { .say } }
17:57 camelia rakudo-moar 1d4600: OUTPUT«(timeout)»
17:58 phatbot m: react { whenever (supply {emit Bool.pick}).tap { .say } }
17:58 camelia rakudo-moar 1d4600: OUTPUT«Tap.new␤»
17:58 phatbot hm, I thought we could whenever a tap
17:59 phatbot m: react { whenever supply {loop {emit Bool.pick}} { .say } }
17:59 camelia rakudo-moar 1d4600: OUTPUT«(timeout)»
18:00 phatbot m: react { whenever supply {for ^10 {emit Bool.pick}} { .say } }
18:00 camelia rakudo-moar 1d4600: OUTPUT«True␤True␤True␤True␤True␤False␤True␤True␤True␤False␤»
18:02 phatbot rakudobugged in RT#130342
18:02 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=130342
18:05 pyrimidine joined #perl6-dev
18:13 pyrimidine joined #perl6-dev
18:14 jnthn Hmm, I thought it always wanted a block.
18:14 jnthn Will have a look at it when I get chance
18:14 jnthn Oh...
18:14 jnthn I can guess though
18:14 jnthn If it parses as a blorst then the thunked statement will be a Code
18:15 jnthn But a Code cannot have phasers, those are in the Block class (a subclass of Code)
18:15 jnthn Thus the error.
18:15 jnthn So probably just turning a .phasers into a .?phasers or so
18:17 dogbert2 jnthn, reported as MoarVM issue 452
18:20 phatbot m: say "foo" for 42.?dasdasdsa
18:20 camelia rakudo-moar 1d4600: OUTPUT«foo␤»
18:20 phatbot hmm
18:22 jnthn Ah, it turns Nil and Nil will iterate :)
18:22 jnthn May be safter to nqp::istype(&thing, Block)
18:22 phatbot ok
18:27 phatbot Is (supply loop { emit Bool.pick }).tap: *.say; supposed to produce some output tho? It just exits with my patch
18:28 phatbot m: (supply for ^10 { emit Bool.pick }).tap: *.say;
18:28 camelia rakudo-moar 1d4600: OUTPUT«True␤True␤False␤False␤False␤False␤False␤False␤True␤True␤No such method 'phasers' for invocant of type 'Code'␤  in block <unit> at <tmp> line 1␤␤»
18:28 phatbot This works.
18:28 b2gills m: supply { loop { emit Bool.pick }}).tap: *.say
18:28 camelia rakudo-moar 1d4600: OUTPUT«===SORRY!=== Error while compiling <tmp>␤Unexpected closing bracket␤at <tmp>:1␤------> supply { loop { emit Bool.pick }}⏏).tap: *.say␤»
18:28 b2gills m: (supply { loop { emit Bool.pick }}).tap: *.say
18:28 camelia rakudo-moar 1d4600: OUTPUT«(timeout)False␤False␤True␤False␤False␤True␤False␤False␤True␤False␤False␤False␤True␤False␤True␤False␤False␤False␤True␤False␤False␤False␤True␤True␤True␤False␤True␤True␤False␤False␤False␤False␤Tr…»
18:29 phatbot eh... why did I run that last one on my box
18:30 [Coke] jnthn: did you see the weird behavior with doc's htmlify --parallel=3? (uses some async stuff and then dies horribly or segvs)
18:30 phatbot b2gills: so I take it the blockless version is meant to work the same?
18:31 b2gills That is why I started this line of thought, I didn't know. I would suspect yes though.
18:32 FROGGS joined #perl6-dev
18:32 phatbot weird that (supply { for ^10 { say "picking"; emit Bool.pick } }).tap: *.say; works but not with loop {}
18:32 b2gills I once ran some Perl6 code on my computer that used so many resources I couldn't even open another terminal to run `pkill moar`, so I used an old laptop to ssh in to do that.
18:33 phatbot :)
18:44 lizmat phatbot: you looking at RT #130342 ?  otherwise I will
18:44 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=130342
18:44 ugexe m: (supply lazy loop { emit Bool.pick }).tap: *.say;
18:44 camelia rakudo-moar 1d4600: OUTPUT«(timeout)False␤True␤True␤True␤True␤False␤True␤True␤False␤False␤True␤True␤True␤False␤True␤True␤False␤True␤False␤True␤False␤True␤True␤False␤True␤False␤False␤True␤True␤True␤False␤False␤False␤Tr…»
18:45 phatbot lizmat: go ahead. I'm not skilled enough to crack this one yet. I've added this if() and the error went away, but (supply loop { emit Bool.pick }).tap: *.say; doesn't produce any output and just exits https://gist.github.com/zoffixznet/bf7c5a84c4d82854f95b8347901982fd
18:46 phatbot and I've added that debug stuff now before the if() but I'm just fumbling in the dark :)
18:46 ugexe it produces output if you put `lazy` in front of loop
18:47 phatbot it produces output if you put a for ^5 { instead of loop
18:47 phatbot ugexe: so why is it needed if it's not needed for the block version of supply?
18:48 ugexe dunno. I just know there have been other weird things like that where explicitly making it lazy makes it work
18:48 phatbot hm, ok, the debug statement shown in the diff just prints the messages and exits immediatelly, so I &!block ends up not being the loop but something else
18:49 phatbot *I suspect
18:49 * phatbot gives up
18:59 dalek rakudo/nom: a980eb1 | lizmat++ | src/core/Supply.pm:
18:59 dalek rakudo/nom: Fix RT #130342
18:59 dalek rakudo/nom:
18:59 dalek rakudo/nom: Only blocks can haz phasers.
18:59 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/a980eb125d
18:59 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=130342
19:00 phatbot :/
19:01 lizmat fwiw, I'm double testing this one
19:01 lizmat I think the problem is actually in the parsing of supply loop { }
19:01 lizmat supply { loop { } } works as expected
19:02 lizmat m: (supply { loop { emit Bool.pick } }).tap: *.say;
19:02 camelia rakudo-moar 1d4600: OUTPUT«(timeout)False␤False␤False␤False␤True␤True␤True␤False␤False␤False␤False␤False␤False␤False␤False␤True␤False␤True␤True␤False␤False␤False␤False␤False␤False␤True␤True␤False␤False␤True␤True␤False␤F…»
19:07 jnthn [Coke]: It's in RT, but plenty of better-golfed things also are, so I'll pick those off first
19:07 jnthn [Coke]: With a bit of luck, some of them are also the issue afflicting that
19:08 lizmat jnthn: what can create a Code object ?
19:09 lizmat a blorst?
19:09 jnthn lizmat: Something that directly has type Code? A load of things.
19:09 jnthn Yeah, a blorst will do it
19:09 lizmat ok, so I guess we need to do something special for the
19:09 lizmat supply loop { }
19:09 lizmat case
19:10 jnthn I'd expect for any `supply EXPR` case, no?
19:10 lizmat cause that apparently stores a Code in the &!block
19:10 jnthn Ah
19:10 lizmat guess so, yes
19:10 jnthn It's only a problem in so far as we check to see if it has phasers, though?
19:10 lizmat well, *that* problem I fixed
19:11 jnthn OK, what other problems are there? I didn't think there were any further significant ones :)
19:11 lizmat but then the example: (supply loop { emit Bool.pick }).tap: *.say
19:11 lizmat doesn't work
19:11 phatbot it just exits. even tho, say, supply for ^6 { .... } works
19:11 lizmat as it does when it is: (supply { loop { emit Bool.pick  }}).tap: *.say
19:12 jnthn What if it's supply eager loop ... ?
19:12 jnthn May be a case of missing sinking
19:12 lizmat ah, perhaps, recomping atm
19:12 jnthn Gotta go for a bit, but check how we handle this with gather
19:13 jnthn Since I guess it would also have to worry about the same thing
19:13 jnthn m: .say for gather loop { take 10 }
19:13 camelia rakudo-moar a980eb: OUTPUT«(timeout)10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10␤10…»
19:13 jnthn Yeah, that handles it
19:14 jnthn bbiab :)
19:14 lizmat yeah, eager makes it work
19:15 lizmat phatbot: not seeing anything special in the parsing of gather or actions
19:15 lizmat so I guess it lives in GATHER
19:18 [Coke] jnthn: I'll see if I can golf it better. thanks.
19:18 lizmat ok, my flu ridden head doesn't see it right away
19:19 lizmat perhaps later, otherwise tomorrow
19:19 phatbot :)
19:33 RabidGravy lizmat, there's something funny about the way wordpress handles the html formatting :)
19:45 pyrimidine joined #perl6-dev
20:05 phatbot Hm. Apparently if you install readline and then uninstall it (with panda) your REPL starts showing this giant message: https://gist.github.com/zoffixznet/499106957d4af7e374e2aca6b20cb222
20:06 AlexDaniel joined #perl6-dev
20:07 phatbot wow... that's just laziness
20:08 phatbot I thought warning in REPL printing a completely unrelated message was an accident, but there's actualyl a test for it :S
20:15 phatbot well or some other reason :(
20:21 phatbot m: take
20:21 camelia rakudo-moar a980eb: OUTPUT«take without parameters doesn't make sense␤  in block <unit> at <tmp> line 1␤␤»
20:21 phatbot m: emit
20:21 camelia rakudo-moar a980eb: OUTPUT«===SORRY!=== Error while compiling <tmp>␤Calling emit() will never work with declared signature (\value --> Nil)␤at <tmp>:1␤------> <BOL>⏏emit␤»
20:27 unicodable6 joined #perl6-dev
20:32 dalek rakudo/nom: 42b7147 | (Zoffix Znet)++ | / (2 files):
20:32 dalek rakudo/nom: Improve warnings in REPL
20:32 dalek rakudo/nom:
20:32 dalek rakudo/nom: Currently, the CONTROL {} block reports "Control flow commands not
20:32 dalek rakudo/nom: allowed in toplevel" message for any control exception, which includes
20:32 dalek rakudo/nom: warnings, resulting in very confusing messages any time a warning
20:32 dalek rakudo/nom: is emited.
20:32 dalek rakudo/nom:
20:32 dalek rakudo/nom: Fix by adding checks for warnings and print those instead. Also improve
20:33 dalek rakudo/nom: emit/take exceptions by reporting that emit/take was outside an
20:33 dalek rakudo/nom: expected enclosure.
20:33 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/42b714749d
20:40 phatbot samcv: where does .history-file attribute come from? I see it's used in a bunch of places
20:40 samcv it's in the REPL line
20:40 samcv err it's in the file
20:40 samcv it's a method
20:40 samcv of the class they inherit from
20:41 samcv (both linenoise and readline)
20:42 samcv i tested and it works and doesn't hang with that PR
20:43 phatbot oh right, I'm blind
20:43 dalek rakudo/nom: b716b00 | samcv++ | src/core/REPL.pm:
20:43 dalek rakudo/nom: Fix Readline REPL not initializing properly.
20:43 dalek rakudo/nom:
20:43 dalek rakudo/nom: Would break since the last commits to add a history file in.
20:43 dalek rakudo/nom: Fix this.
20:43 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/b716b007cc
20:43 dalek rakudo/nom: 772e2f2 | (Zoffix Znet)++ | src/core/REPL.pm:
20:43 dalek rakudo/nom: Merge pull request #955 from samcv/REPL-readline
20:43 dalek rakudo/nom:
20:43 dalek rakudo/nom: Fix Readline REPL not initializing properly.
20:43 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/772e2f28b0
20:45 samcv phatbot, i had the same thought until i did ctrl + f
20:45 phatbot I was ctrl+f '.history-file' thinking it was an attribute :P
20:47 phatbot m: class A { method foo {say "hi"} }; role B { method hi {…}}; my $na = A but B; say $na
20:47 camelia rakudo-moar a980eb: OUTPUT«Method 'hi' must be implemented by A+{B} because it is required by a role B␤  in any compose_method_table at gen/moar/Metamodel.nqp line 2827␤  in any apply at gen/moar/Metamodel.nqp line 2841␤  in any compose at gen/moar/Metamodel.nqp line 3013␤  …»
20:47 phatbot .oO( and it is? )
20:49 phatbot oh :}
20:49 phatbot m: class A { method foo {say "hi"} }; role B { method foo {…}}; my $na = A but B; say $na
20:49 camelia rakudo-moar a980eb: OUTPUT«(A+{B})␤»
20:52 dalek rakudo/nom: 7b84adc | jnthn++ | src/Perl6/Metamodel/RoleToClassApplier.nqp:
20:52 dalek rakudo/nom: Grammar fix to error message.
20:52 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/7b84adcec5
21:01 bartolin joined #perl6-dev
21:17 nine The good news is that I've already found 2 bugs in dyncall. Bad news is that it's only in comments and that I'm no closer to solving the x86 NC callback issue dogbert17++ reported
21:24 RabidGravy joined #perl6-dev
21:25 dogbert2 nine: let me know if there's something you want me to test on 32-bit
21:28 nine dogbert2: the really odd part are the ??? stackframes where the invalid read seems to happen. I don't know what exactly they are and cannot even get a proper disassembly. The next frame should be dcCall_x86_cdecl.
21:36 dogbert2 nine: I can set a breakpoint in dc_callvm_call_x86_cdecl if that is of any use
21:37 nine It can't hurt in any case :)
21:37 dogbert2 anything I should look at now that the program is stopped?
21:37 dogbert2 Breakpoint 1, 0xb7dcbd1a in dc_callvm_call_x86_cdecl () from //home/dogbert/repos/rakudo/install/lib/libmoar.so
21:37 dogbert2 (gdb) bt
21:37 dogbert2 #0  0xb7dcbd1a in dc_callvm_call_x86_cdecl () from //home/dogbert/repos/rakudo/install/lib/libmoar.so
21:37 dogbert2 #1  0xb7dcb925 in dcCallVoid () from //home/dogbert/repos/rakudo/install/lib/libmoar.so
21:37 dogbert2 #2  0xb7d01bbe in MVM_nativecall_invoke (tc=0x804c450, res_type=0x8389a40, site=0xa1294a0, args=0xb7802694) at src/core/nativecall_dyncall.c:587
21:39 nine Is this the first or the second time you hit the break point? The first would be the "simple callback" test where everything seems to be alright
21:41 dogbert2 the first
21:42 dogbert2 #0  0xb7dcbd1a in dc_callvm_call_x86_cdecl () from //home/dogbert/repos/rakudo/install/lib/libmoar.so
21:42 dogbert2 #1  0xb7dcb925 in dcCallVoid () from //home/dogbert/repos/rakudo/install/lib/libmoar.so
21:42 dogbert2 #2  0xb7d01bbe in MVM_nativecall_invoke (tc=0x804c450, res_type=0x8389a40, site=0xa12a460, args=0xb780b030) at src/core/nativecall_dyncall.c:587
21:42 dogbert2 #3  0xb7cd3bce in MVM_interp_run (tc=0x804c450, initial_invoke=0xb7dc9137 <toplevel_initial_invoke>, invoke_data=0x80b3cc0) at src/core/interp.c:3984
21:42 dogbert2 #4  0xb7dc9266 in MVM_vm_run_file (instance=0x804c008, filename=0xbffff2ea "/home/dogbert/repos/rakudo/perl6.moarvm") at src/moar.c:309
21:42 dogbert2 y
21:43 dogbert2 that should be the second time
21:45 ugexe m: my @a = [".LTA" => 1]; say "xxx" ~~ any(flat @a);
21:45 camelia rakudo-moar 7b84ad: OUTPUT«No such method '.LTA' for invocant of type 'Str'␤  in block <unit> at <tmp> line 1␤␤»
21:45 nine Next step should be a call into dcCall_x86_cdecl followed by a call to TakeIntCallback
21:46 pyrimidine joined #perl6-dev
21:47 dogbert2 nine: here's a more complete gist, i.e. bt full. https://gist.github.com/dogbert17/1b347d5473177856d5545aaa7aa67dfe
21:50 phatbot <Zoffix> m: my @allowed-types = 'uc', 'lc', 'tc', 'na', 'B', 'S', 'nv'; for ^100_000 { my $x = 'x' eq any @allowed-types }; say now - INIT now
21:50 phatbot <camelia> rakudo-moar 6fc769: OUTPUT«10.7136853␤»
21:50 phatbot <Zoffix> m: my @allowed-types = 'uc', 'lc', 'tc', 'na', 'B', 'S', 'nv'; for ^100_000 { my $x = @allowed-types.first: 'x' }; say now - INIT now
21:50 phatbot <camelia> rakudo-moar 6fc769: OUTPUT«1.14418706␤»
21:50 phatbot <Zoffix> samcv, ^ an example of how slow our Junctions are ATM
21:50 phatbot samcv: not sure if you saw in the other channel.
21:50 dogbert2 nine: I have updated the gist after typing advance dcCall_x86_cdecl, new info at the bottom
21:50 phatbot No idea about unicode stuff, so I'll wait for someone else to review and merge https://github.com/rakudo/rakudo/pull/953
21:51 samcv oh
21:51 samcv yeah i am
21:52 samcv did not see it, trying to get some way to benchmark the atom syntax highlighter...
21:52 samcv and trying horribly to write in coffeescript for node
21:52 samcv if anybody knows javascript please help
21:52 phatbot hehe
21:52 phatbot like what kindahelp?
21:53 phatbot and well, I don't know coffeescript
21:53 samcv like should be easy
21:53 samcv i think it's mostly the same thing
21:53 samcv all i'm trying to do is gett all the files in a folder
21:53 phatbot ohh.. I only done web dev with JS so no idea :)
21:54 samcv https://cry.nu/p/q9jk if anybody can help me.. it's working except for
21:54 samcv trying to get the files
21:56 AlexDaniel jdv79: we have less bots than humans. That's unhealthy.
21:57 AlexDaniel 50/50 would be a good ratio
21:57 dogbert2 I'm now in TakeIntCallback
21:57 samcv oh i think i got it
21:57 samcv finallyyyy
21:58 samcv sweet i did it!
21:59 samcv nice.
21:59 Ven joined #perl6-dev
22:00 jdv79 AlexDaniel: maybe in 100 years.  more humans is more normal today though.
22:00 AlexDaniel jdv79: I think we can solve it in one month or so
22:00 AlexDaniel there are at least 2 proposals for new whateverables
22:01 jdv79 great!
22:01 phatbot Why do we spawn a new bot for every feature tho?
22:01 phatbot And not make it the normal way: one bot does all the feautres
22:02 phatbot I wonder how long it'll take Freenode staff to notice you have a gazillion connections all comming from one host
22:02 jdv79 what does a phatbot do?
22:02 phatbot I eat food and complain a lot.
22:02 AlexDaniel it pretends to be a human
22:03 jdv79 food sounds good but i have to cook first.  not so enticing.
22:05 nine Oh boy it's late! Sorry, have to leave now. Need to get up early tomorrow
22:05 dogbert2 night
22:46 pyrimidi_ joined #perl6-dev
23:22 AlexDaniel joined #perl6-dev
23:41 phatbot The REPL error I showed today that occurs after uninstalling Readline... That appears to be just panda failing to uninstall correctly. Works fine with zef
23:55 unicodable6 joined #perl6-dev

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