Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6-dev, 2016-08-14

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

All times shown according to UTC.

Time Nick Message
00:15 dalek roast: 06f9c5d | TimToady++ | S04-statements/for.t:
00:15 dalek roast: nowadays next does not produce a loop value
00:15 dalek roast: review: https://github.com/perl6/roast/commit/06f9c5d010
01:48 ilbot3 joined #perl6-dev
01:48 Topic for #perl6-dev is now Perl 6 language and compiler development 2.0 | Logs at http://irclog.perlgeek.de/perl6-dev/today
03:13 skids joined #perl6-dev
06:03 JimmyZ m: my $l = (for ^5 { 41; next if $_ == 2; $_; }); say $l
06:03 camelia rakudo-moar fe90be: OUTPUT«WARNINGS for <tmp>:␤Useless use of constant integer 41 in sink context (line 1)␤(0 1 3 4)␤»
08:48 RabidGravy joined #perl6-dev
08:54 [Tux] CSV tests fail. /me investigates
08:54 [Tux] something with utf8
08:57 [Tux] $!text ~~ m{^ <[ \x00       .. \x7F ]>+ $} or
08:57 [Tux] $!is_utf8   = True;
08:57 [Tux] did the meaning of that regex change?
09:00 [Tux] it is about this field: "a\r\n"
09:00 [Tux] so my guess wouyld be that \r\n is now returned as a single grapheme (even if it is NOT a newline)
09:01 geekosaur behavior of csets changed, not related to \r\n
09:01 geekosaur [13 19:20] <dalek> nqp: 2df0a06 | TimToady++ | src/vm/moar/QAST/QASTRegexCompilerMAST.nqp:
09:01 geekosaur [13 19:20] <dalek> nqp: charrange without :m should avoid ordat/ordfirst
09:02 [Tux] geekosaur, can you summarize the change in one and - and hand me the "fix" to what now fails?
09:02 [Tux] s/one/one line/
09:04 geekosaur Synthetics wrongly look like they're in a character range if we strip the combining information, so instead we'll just use the raw char code, and the synthetic will naturally fall out of the range by virtue of being negative.
09:05 geekosaur I do not have a fix, I'm reporting that something happened, not that I understood the details and can tell you exactly how it relates to things
09:06 geekosaur that said, that description doesn't seem like it should affect this, unless somehow it was integrating the \r and \n in the charrange into a synthetic newline...
09:07 [Tux] This is Rakudo version 2016.07.1-181-gfe90be0 built on MoarVM version 2016.07-18-g2f269d8
09:07 [Tux] csv-ip5xs       10.143
09:07 [Tux] test            16.156
09:07 [Tux] test-t           7.433
09:07 [Tux] csv-parser      16.624
09:09 geekosaur alternately the change doesn't work with non-synthetic nonprintables for some reason...
10:10 pmurias joined #perl6-dev
10:10 dalek nqp/no-eval: b25a452 | (Pawel Murias)++ | src/vm/js/Compiler.nqp:
10:10 dalek nqp/no-eval: [js] Don't emit the setting setup code.
10:10 dalek nqp/no-eval: review: https://github.com/perl6/nqp/commit/b25a45271d
10:10 dalek nqp/no-eval: b307758 | (Pawel Murias)++ | src/vm/js/ (2 files):
10:10 dalek nqp/no-eval: [js] Load modules using a hacked in module loader rather then special casing QAST nodes.
10:10 dalek nqp/no-eval: review: https://github.com/perl6/nqp/commit/b307758c41
10:10 dalek nqp/no-eval: aa27971 | (Pawel Murias)++ | src/vm/js/Compiler.nqp:
10:11 dalek nqp/no-eval: [js] Remove the old way of loading in the setting.
10:11 dalek nqp/no-eval: review: https://github.com/perl6/nqp/commit/aa27971103
10:11 dalek nqp/no-eval: 6a01f6f | (Pawel Murias)++ | src/vm/js/Compiler.nqp:
10:11 dalek nqp/no-eval: [js] Remove useless check.
10:11 dalek nqp/no-eval: review: https://github.com/perl6/nqp/commit/6a01f6f685
10:11 dalek nqp/no-eval: f093174 | (Pawel Murias)++ | src/vm/js/ (2 files):
10:11 dalek nqp/no-eval: [js] Stop specially marking symbols from setting.
10:16 dalek nqp/no-eval: 02f9804 | (Pawel Murias)++ | src/vm/js/Compiler.nqp:
10:16 dalek nqp/no-eval: [js] Stop special casing forceouterctx in post_deserialize.
10:16 dalek nqp/no-eval: review: https://github.com/perl6/nqp/commit/02f9804417
10:59 AlexDaniel joined #perl6-dev
13:29 Zoffix joined #perl6-dev
13:31 dalek rakudo/nom: 191dc0a | lizmat++ | src/core/Array.pm:
13:31 dalek rakudo/nom: Fix my @a; @a.splice(0,0,42)
13:31 dalek rakudo/nom:
13:31 dalek rakudo/nom: As noted by masak++ on
13:31 dalek rakudo/nom:
13:31 dalek rakudo/nom:   http://irclog.perlgeek.de/perl6-dev/2016-08-13#i_13017831
13:31 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/191dc0a715
13:38 Zoffix left #perl6-dev
13:41 Zoffix joined #perl6-dev
13:52 AlexDaniel joined #perl6-dev
14:09 MasterDuke while there are some @LARRY here, i've been looking at RT #126264 on and off for a while, but i'm not finding it quite as straightforward as i first thought it would be
14:09 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=126264
14:09 MasterDuke suggesting 'EVAL' for 'eval' happens because EVAL is a function, but BEGIN is a phaser
14:12 MasterDuke would it be terrible to add all the phasers as things to compare the undeclared routine against?
14:15 dalek rakudo/nom: 6d97bad | lizmat++ | src/core/Rakudo/Internals.pm:
14:15 dalek rakudo/nom: Introduce MappyIterator-values class
14:15 dalek rakudo/nom:
14:15 dalek rakudo/nom: General purpose iterator class that returns the values of a hash.
14:15 dalek rakudo/nom: To be used in Map, Baggy and typed hashes.
14:15 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/6d97bad13f
14:15 dalek rakudo/nom: 9a495d4 | lizmat++ | src/core/ (3 files):
14:15 dalek rakudo/nom: Use the MappyIterator-values class where we can
14:15 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/9a495d4986
14:15 lizmat MasterDuke: if the unfound string is allcaps, that would make sense, I think
14:16 lizmat MasterDuke: not so sure otherwise
14:17 MasterDuke lizmat: well, the use case in the ticket is the user writing something lowercase that should be upper
14:17 MasterDuke e.g., begin instead of BEGIN
14:17 lizmat hmmmm
14:18 lizmat well, I guess by the time we find we have an error, we might as well add the extra overhead of checking all the phasers
14:18 lizmat MasterDuke: consider me convinced
14:20 MasterDuke lizmat: cool, i'll try to figure that out. is there any sort of built-in datastructure that has all the phasers, or will i have to hardcode a list?
14:22 lizmat not that I know
14:22 lizmat perhaps a Block.possible-phasers method would be nice to have
14:23 lizmat that would return <BEGIN END ENTER LEAVE FIRST LAST NEXT  etc...
14:24 lizmat the actual list can be found it src/Perl6/Grammar, line 1622 and following
14:24 lizmat I *think* we can actually remove COMPOSE there
14:25 lizmat m: COMPOSE my $a = 42; say $a   # doesn't seem to do a lot
14:25 camelia rakudo-moar 191dc0: OUTPUT«(Any)␤»
14:36 lizmat afk&  # let's see if I can still ride a bicycle
14:39 Zoffix * three hours later * https://batorego.files.wordpress.com/2007/09/bandages.jpg
14:40 MasterDuke lizmat: ha, i was scouring Block, Calleable, etc looking for something exactly like that (.possible-phasers)
14:48 lizmat MasterDuke: just realized, it should probably be Mu.possible-phasers
14:48 lizmat as BEGIN / END are not tied to a block, e.g.
14:49 lizmat really afk&
14:50 AlexDaniel joined #perl6-dev
15:55 Zoffix 0.o  strangest error ever:   Variable '$c' is not declared  at blah blah ------> sub prep-commit (⏏$c) {
15:58 Zoffix And using $^c in the body instead gives Placeholder variable $^c may not be used here because the surrounding block takes no signature
16:00 Zoffix oh gee
16:02 Zoffix m: my $out = qq:to/END/;␤blah␤END;␤for ^10 { my $x = 'just some random code' }␤for ^10 { my $x = 'moar random code' }␤for ^10 { my $x = 'even moar random code' }␤␤sub prep-commit ($c) {}
16:02 camelia rakudo-moar 9a495d: OUTPUT«===SORRY!=== Error while compiling <tmp>␤Variable '$c' is not declared␤at <tmp>:8␤------> sub prep-commit (⏏$c) {}␤»
16:02 Zoffix Supremely LTA error, though I can't think of how to fix that stuff
16:03 Zoffix (hint: it's the spurious semicolon on the closing END, so it never gets found)
16:33 gfldex AlexDaniel: i got a response, unsurprisingly they refuse to help me
16:34 gfldex wc,sorry
16:35 dalek rakudo/nom: 8e5f5e7 | (Zoffix Znet)++ | docs/ChangeLog:
16:35 dalek rakudo/nom: Prep 2016.08 section in ChangeLog
16:35 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/8e5f5e7554
16:52 gfldex i was under the impression that assinging a lazy list to Array (and thuse @-sigiled countainers) would not explode a lazy list. Was that changed with GLR?
16:52 gfldex m: my @a = gather { say 1; take 1; say 2; take 2; say 3; take 3 };
16:52 camelia rakudo-moar 8e5f5e: OUTPUT«1␤2␤3␤»
16:52 gfldex or is that the sink context of the my declarator?
16:59 ugexe m: my @a := cache gather { say 1; take 1; say 2; take 2; say 3; take 3 }; @a[1] # eh
16:59 camelia rakudo-moar 8e5f5e: OUTPUT«1␤2␤»
17:03 TimToady m: my @a = lazy gather { say 1; take 1; say 2; take 2; say 3; take 3 }
17:03 camelia rakudo-moar 8e5f5e: ( no output )
17:04 TimToady gather is not assumed to be lazy by default
17:05 TimToady (and my ignores sink context, basically)
17:05 ugexe docs lied: `C<gather/take> generates values lazily`
17:05 gfldex it does, that's why I asked :)
17:06 TimToady it does, but that is not the same thing as "must be lazy or you'll die"
17:06 TimToady array assignment is eager unless "must be lazy"
17:09 Rele6sr joined #perl6-dev
17:09 Zoffix Rele6sr, reminder
17:09 Rele6sr ???????????? Friends! I bear good news! Rakudo's release will happen in just 6 days! Please update the ChangeLog with anything you worked on that should be known to our users. ????????????
17:09 Rele6sr ???????????? Here are still-open new RT tickets since last release: http://bug.perl6.party/1471194589.html And here is the git log output for commits since last release: http://bug.perl6.party/1471194591.html ????????????
17:09 Zoffix :D :D :D \o/
17:18 timotimo fantastic
17:19 Zoffix (FWIW, I'm going over the commits and will add all of 'em to changelog; just check I got all right after I'm done)
17:24 mst right, this should be hilarious
17:24 Zoffix Wat?
17:24 mst time to convince Rakudo's Configure.pl to find the moar and nqp installs
17:25 Zoffix ah
17:26 mst because of COURSE it's going to look in the prefix instead of where I told it to
17:26 * mst sulks at Configure.pl
17:32 travis-ci joined #perl6-dev
17:32 travis-ci Rakudo build failed. Zoffix Znet 'Prep 2016.08 section in ChangeLog'
17:32 travis-ci https://travis-ci.org/rakudo/rakudo/builds/152224223 https://github.com/rakudo/rakudo/compare/9a495d4986cc...8e5f5e75546a
17:32 travis-ci left #perl6-dev
17:32 Zoffix your momma
17:32 Zoffix (github timeout error)
17:39 mst right, open sesame hack time
17:44 masak whoa, Rele6sr++
17:47 mst augh
17:59 * mst upgrades Configure.pl from 'dumpster fire' to 'rolling dumpster fire'
17:59 Zoffix :)
18:07 dalek nqp/no-eval: 512345a | (Pawel Murias)++ | / (6 files):
18:07 dalek nqp/no-eval: [js] Load modules and setting using the ModuleLoader.
18:07 dalek nqp/no-eval: review: https://github.com/perl6/nqp/commit/512345a2de
18:07 dalek nqp/no-eval: a64a92e | (Pawel Murias)++ | src/vm/js/nqp-runtime/runtime.js:
18:07 dalek nqp/no-eval: [js] Remove module loading that doesn't use the ModuleLoader.
18:07 dalek nqp/no-eval: review: https://github.com/perl6/nqp/commit/a64a92ea0b
18:12 dalek rakudo/nom: b548e43 | (Zoffix Znet)++ | docs/ChangeLog:
18:12 dalek rakudo/nom: Populate ChangeLog
18:12 dalek rakudo/nom:
18:12 dalek rakudo/nom: - Includes all Rakudo commits since last release
18:12 dalek rakudo/nom: - NQP/MoarVM fixes that were added to Rakudo as merele a version bump
18:12 dalek rakudo/nom:   might have been missed
18:12 dalek rakudo/nom: - All of jnthn++'s async work is marked "Numerous stability
18:12 dalek rakudo/nom:   improvements in async," because I don't know any better
18:12 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/b548e437e5
18:13 Zoffix .tell jnthn I wasn't sure what to add to ChangeLog for all of your async work this release. Please improve if needed: https://github.com/rakudo/rakudo/blob/nom/docs/ChangeLog#L19
18:13 yoleaux2 Zoffix: I'll pass your message to jnthn.
18:15 MasterDuke lizmat: it's not pretty, but it does work. https://gist.github.com/MasterDuke17/eb8e42371424528d6ca8d8312ad0692b
18:16 Rele6sr joined #perl6-dev
18:18 MasterDuke Zoffix: "such us on sinks"
18:20 dalek rakudo/nom: ee8a25e | (Zoffix Znet)++ | docs/ChangeLog:
18:20 dalek rakudo/nom: Fix typo MasterDuke++
18:20 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/ee8a25e886
18:58 mst fuck. I'm not even sure this is going to be sane
18:58 mst half the makefile's going to need redoing
18:58 * mst sulks
19:00 mst (don't mind me, I know how to deal with this, it's just annoying)
19:00 mst (also I could totally do this an easier way but then I wouldn't end up with patches for you)
19:01 arnsholt mst++ # Fixing the build system
19:01 mst nqp and rakudo both *really* want to be in the same prefix as moar
19:01 mst I disagree :)
19:05 arnsholt Yeah, that seems to be the opinion of most people who do packaging =)
19:05 arnsholt And I can see why, TBH
19:06 mst actually, I figured out a plan B that would fit all of my original requirements and not actually require a separate prefix at all
19:06 mst but I've already put enough effort into getting this to work that that's NOT THE POINT ANYMORE
19:12 travis-ci joined #perl6-dev
19:13 travis-ci Rakudo build passed. Zoffix Znet 'Populate ChangeLog
19:13 travis-ci https://travis-ci.org/rakudo/rakudo/builds/152236925 https://github.com/rakudo/rakudo/compare/8e5f5e75546a...b548e437e58b
19:13 travis-ci left #perl6-dev
19:13 mst also - having moar, nqp and rakudo as three packages all going into $PREFIX, with a $DESTDIR for staging into, works fine, which is what OS packagers basically need
19:13 mst I'm just being difficult, and having three different $PREFIX
19:33 brrt joined #perl6-dev
19:40 Zoffix Another example of new Rust's error awesomeness: https://twitter.com/ozkriff/status/764730852509151232
19:41 Zoffix Sounds like an article should be made for Perl 6's errors' awesomeness.
19:47 mst that was why I was asking about examples :P
19:55 Zoffix Yeah, I can't think of anything groundbreaking off the top of my head. Gonna peek into Exception.pm to see what's what.
19:57 Zoffix m: class { has $.foo; method bar { 'foo' ~~ m/$!foo/ } }.new.bar
19:57 camelia rakudo-moar ee8a25: OUTPUT«===SORRY!=== Error while compiling <tmp>␤Attribute $!foo not available inside of a regex, since regexes are methods on Cursor.␤Consider storing the attribute in a lexical, and using that in the regex.␤at <tmp>:1␤------> has $.foo; method ba…»
20:02 travis-ci joined #perl6-dev
20:02 travis-ci Rakudo build passed. Zoffix Znet 'Fix typo MasterDuke++'
20:02 travis-ci https://travis-ci.org/rakudo/rakudo/builds/152237755 https://github.com/rakudo/rakudo/compare/b548e437e58b...ee8a25e88658
20:02 travis-ci left #perl6-dev
20:12 timotimo "rustacean" is a pretty cool term as far as members of programming language communities go
20:20 dalek nqp/no-eval: 0720516 | (Pawel Murias)++ | src/vm/js/nqp-runtime/deserialization.js:
20:20 dalek nqp/no-eval: [js] Remove dead code.
20:20 dalek nqp/no-eval: review: https://github.com/perl6/nqp/commit/07205168d2
20:20 dalek nqp/no-eval: 21d0ddd | (Pawel Murias)++ | src/vm/js/nqp-runtime/ (4 files):
20:20 dalek nqp/no-eval: [js] Fix style errors found by make js-lint
20:20 dalek nqp/no-eval: review: https://github.com/perl6/nqp/commit/21d0ddd75a
20:40 travis-ci joined #perl6-dev
20:40 travis-ci Rakudo build errored. Elizabeth Mattijsen 'Use the MappyIterator-values class where we can'
20:40 travis-ci https://travis-ci.org/rakudo/rakudo/builds/152207115 https://github.com/rakudo/rakudo/compare/191dc0a715ec...9a495d4986cc
20:40 travis-ci left #perl6-dev
20:40 buggable ☢ [travis build above] Build log missing from at least one job. Check results manually.
23:35 skids joined #perl6-dev

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