Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6-dev, 2017-11-04

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

All times shown according to UTC.

Time Nick Message
00:15 committable6 joined #perl6-dev
00:39 BenGoldberg joined #perl6-dev
01:28 * Zoffix chalks another -1 for "Perl 6" name
01:28 Zoffix Doesn't justify sanely: https://i.imgur.com/hkPI2XP.png
01:29 timotimo Zoffix: what happens if you use a non-breaking sace or something fancy like that?
01:29 yoleaux 3 Nov 2017 23:57Z <MasterDuke> timotimo: followed jnthn's advice and got a general protection fault in valgrind with azawawi's test code: https://gist.github.com/MasterDuke17/ee516fde9916dc4dfc6ab74ca4e72e82
01:29 Zoffix timotimo: yeah, I did use nbsp
01:39 Geth ¦ rakudo: MasterDuke17++ created pull request #1231: Convert the rest of the `{ * }` to `{*}`
01:39 Geth ¦ rakudo: review: https://github.com/rakudo/rakudo/pull/1231
01:39 synopsebot RAKUDO#1231 [open]: https://github.com/rakudo/rakudo/pull/1231 Convert the rest of the `{ * }` to `{*}`
01:42 Geth ¦ rakudo: 102fbd5179 | (Daniel Green)++ | 47 files
01:42 Geth ¦ rakudo: Convert the rest of the `{ * }` to `{*}`
01:42 Geth ¦ rakudo: review: https://github.com/rakudo/rakudo/commit/102fbd5179
01:42 Geth ¦ rakudo: 33d476d1be | (Zoffix Znet)++ (committed using GitHub Web editor) | 47 files
01:42 Geth ¦ rakudo: Merge pull request #1231 from MasterDuke17/more_normalizing_proto_bodies
01:42 Geth ¦ rakudo:
01:42 Geth ¦ rakudo: Convert the rest of the `{ * }` to `{*}`
01:42 Geth ¦ rakudo: review: https://github.com/rakudo/rakudo/commit/33d476d1be
01:51 Zoffix m: use NativeCall; sub strnlen (Str is encoded('ascii') --> size_t) is native {}; say strnlen 'I NativeCall Perl 6';
01:51 camelia rakudo-moar 33d476d1b: OUTPUT: «19␤»
01:51 Zoffix m: use NativeCall; sub strnlen (Str is encoded('ascii') --> size_t) is native {}; say strnlen 'I NativeCall Perl 6';  say chars 'I NativeCall Perl 6'
01:51 camelia rakudo-moar 33d476d1b: OUTPUT: «2␤19␤»
01:51 Zoffix Why'd it change to "2" after added more unrelated code afterwards?
01:54 Zoffix Oh, I'm calling the wrong function; meant `strlen`
02:02 Zoffix m: use NativeCall; sub strlen (Str --> size_t) is native {}; say strlen 'I love Perl 6!';
02:03 camelia rakudo-moar 33d476d1b: OUTPUT: «14␤»
02:03 Zoffix Would love a more interesting NativeCall example that takes about the same amount of code.
02:57 ilbot3 joined #perl6-dev
02:57 Topic for #perl6-dev is now Perl 6 language and compiler development 2.0 | Logs at http://irclog.perlgeek.de/perl6-dev/today | For toolchain/installation stuff see #perl6-toolchain | For MoarVM see #moarvm
04:32 Geth ¦ rakudo: MasterDuke17++ created pull request #1232: Add some INTERPOLATE variations
04:32 Geth ¦ rakudo: review: https://github.com/rakudo/rakudo/pull/1232
04:32 synopsebot RAKUDO#1232 [open]: https://github.com/rakudo/rakudo/pull/1232 Add some INTERPOLATE variations
04:32 samcv MasterDuke++
04:33 samcv this makes me happy :}
04:35 MasterDuke heh, glad to hear it. now i'm off to get some sleep
04:45 ugexe joined #perl6-dev
09:13 AlexDaniel joined #perl6-dev
09:18 pmurias joined #perl6-dev
10:05 evalable6 joined #perl6-dev
10:27 AlexDaniel m: say Order.^mro
10:27 camelia rakudo-moar 33d476d1b: OUTPUT: «((Order) (Int) (Cool) (Any) (Mu))␤»
10:27 AlexDaniel isn't it weird that there's no Enum?
11:01 evalable6 joined #perl6-dev
11:26 patrickz joined #perl6-dev
12:22 patrickz joined #perl6-dev
12:52 [Tux] Rakudo version 2017.10-116-g33d476d1b - MoarVM version 2017.10-47-gf88ab169d
12:52 [Tux] csv-test-xs-20      0.453 -  0.459
12:52 [Tux] csv-ip5xs           1.120 -  1.130
12:52 [Tux] test-t              3.265 -  3.325
12:52 [Tux] test               12.011 - 12.092
12:52 [Tux] csv-ip5xs-20       13.024 - 13.077
12:52 [Tux] csv-parser         12.696 - 13.205
12:52 [Tux] test-t-20 --race   20.325 - 20.711
12:52 [Tux] test-t-20          60.137 - 60.155
12:54 MasterDuke hm, looks like all the test-t's got a little slower
12:54 jnthn AlexDaniel: No, enum types are done as subtypes of the enum's base type, which is Int. They compose a few roles into that, such as the Enumeration role
13:00 ilmari joined #perl6-dev
13:11 AlexDaniel ok
13:11 MasterDuke m: my $s = 1_000_000; my $b = ^$s; my $p = (^$s).map({$b[$_]}); say $p[*-1]; say now - INIT now
13:11 camelia rakudo-moar 33d476d1b: OUTPUT: «999999␤3.4725444␤»
13:11 MasterDuke m: my $s = 1_000_000; my $b = ^$s; my $p = $b[^$s]; say $p[*-1]; say now - INIT now
13:11 camelia rakudo-moar 33d476d1b: OUTPUT: «999999␤4.0715540␤»
13:12 MasterDuke on my machine i consistently see that second form ~0.7s slower
13:13 MasterDuke oh, List vs Seq
13:14 MasterDuke i still would have expected the slice to be faster
13:28 Geth ¦ roast: 284ff78be6 | usev6++ | S05-mass/properties-general.t
13:28 Geth ¦ roast: [jvm] Re-fudge unicode tests, fixed on Moar only
13:28 Geth ¦ roast: review: https://github.com/perl6/roast/commit/284ff78be6
13:40 Geth ¦ roast: 8e5f7252d2 | usev6++ | S11-modules/export.t
13:40 Geth ¦ roast: [jvm] Fudge tests for RT #129215
13:40 Geth ¦ roast: review: https://github.com/perl6/roast/commit/8e5f7252d2
13:40 synopsebot RT#129215 [open]: https://rt.perl.org/Ticket/Display.html?id=129215 Passing a Hash[...] as a parameter to an imported sub
13:52 patrickz joined #perl6-dev
14:28 Zoffix m: say Order.^roles
14:28 camelia rakudo-moar 33d476d1b: OUTPUT: «No such method 'roles' for invocant of type 'Perl6::Metamodel::EnumHOW'. Did you mean any of these?␤    does␤    roll␤␤  in block <unit> at <tmp> line 1␤␤»
14:39 llfourn joined #perl6-dev
14:41 Zoffix $ echo "" | perl6 -e '$*IN.encoding: Nil; say $*IN.read: 10000000; say $*IN.eof;'
14:41 Zoffix Buf[uint8]:0x<0a>
14:41 Zoffix False
14:46 Zoffix Doesn't seem to happen with normal handles. Perl gives True
14:47 * Zoffix investigamates
14:57 Zoffix The answers on this one demonstrate just how awful the strategy to program using SO answers is. 3 experts answered. All wrong in some way or another: https://stackoverflow.com/questions/47046761/does-perl-6s-eof-give-up-too-quickly/47106348
15:03 Zoffix ehehe see it... I see the bug!
15:03 * Zoffix readies the sandwich bread
15:14 pmurias Zoffix: SO is sometimes usefull a source of ideas if you assume the answers are likely wrong ;)
15:14 Zoffix Yeah :D
15:26 Geth ¦ rakudo/js: 6 commits pushed by pmurias++
15:26 Geth ¦ rakudo/js: f434e34065 | Update tools/build/makeNATIVE_SHAPED_ARRAY.pl6
15:26 Geth ¦ rakudo/js: 44ef8fc0c9 | Parameterize using type objects instead of native ints
15:26 Geth ¦ rakudo/js: 994940a824 | [js] Also type check when resetting a Scalar to a default with Nil
15:26 Geth ¦ rakudo/js: 40348ae491 | Remove some leftover multidim disablement
15:26 Geth ¦ rakudo/js: 8909cff195 | [js] Avoid pushing labels as a second child of a p6sink
15:26 Geth ¦ rakudo/js: 34f1b310b2 | [js] Avoid passing a :cps named argument that's not accepted anymore
15:26 Geth ¦ rakudo/js: review: https://github.com/rakudo/rakudo/compare/580a232e2f...34f1b310b2
15:29 Zoffix ZOFFLOP: t/spec/S17-procasync/basic.t
15:37 Geth ¦ nqp: a25acd1e82 | (Zoffix Znet)++ | tools/build/MOAR_REVISION
15:37 Geth ¦ nqp: Bump MoarVM
15:37 Geth ¦ nqp: review: https://github.com/perl6/nqp/commit/a25acd1e82
15:37 Geth ¦ nqp: version bump brought these changes: https://github.com/MoarVM/MoarVM/compare/2017.10-47-gf88ab16...2017.10-56-g2f0672f
15:37 Geth ¦ rakudo: a8c789ecac | (Zoffix Znet)++ | tools/build/NQP_REVISION
15:37 Geth ¦ rakudo: Bump NQP
15:37 Geth ¦ rakudo: review: https://github.com/rakudo/rakudo/commit/a8c789ecac
15:38 Zoffix ZOFVM: Files=1283, Tests=152838, 157 wallclock secs (21.96 usr  3.53 sys + 3394.35 cusr 178.67 csys = 3598.51 CPU)
15:39 Geth ¦ roast: c08ef07a3b | (Zoffix Znet)++ | S16-io/eof.t
15:39 Geth ¦ roast: Test EOF on TTY isn't too lately-set
15:39 Geth ¦ roast:
15:39 Geth ¦ roast: Fix: https://github.com/MoarVM/MoarVM/commit/2f0672f74523c6ddfe8a216
15:39 Geth ¦ roast: review: https://github.com/perl6/roast/commit/c08ef07a3b
15:44 evalable6 joined #perl6-dev
15:51 travis-ci joined #perl6-dev
15:51 travis-ci NQP build failed. Zoffix Znet 'Bump MoarVM'
15:51 travis-ci https://travis-ci.org/perl6/nqp/builds/297242592 https://github.com/perl6/nqp/compare/20536d98a51e...a25acd1e82e0
15:51 travis-ci left #perl6-dev
16:21 b2gills That eof "fix" makes `.eof` return True at the end of a line of input if you use it interactively. Which breaks briandfoy's code on Stack Overflow in new and interesting ways.
16:22 b2gills So is `.eof` supposed to be the end of the input, or the end of currently available data?
16:26 Zoffix input
16:30 Geth ¦ roast: 9c900e56f7 | (Zoffix Znet)++ | S16-io/eof.t
16:30 Geth ¦ roast: Revert "Test EOF on TTY isn't too lately-set"
16:30 Geth ¦ roast:
16:30 Geth ¦ roast: This reverts commit c08ef07a3b05c7d59637c5f53c0611d835d6fc8b.
16:30 Geth ¦ roast:
16:30 Geth ¦ roast: https://irclog.perlgeek.de/perl6-dev/2017-11-04#i_15401308
16:30 Geth ¦ roast: review: https://github.com/perl6/roast/commit/9c900e56f7
16:31 Zoffix gotta $work; just reverting everything now. Will take a look later if there's some trick can be done
16:33 Zoffix But the "If I do the same thing in Perl 6 (Rakudo 2017.07) the program stops right away:" part of bdfoy's question is very likely the older TTY EOF bug on MacOS. And for "I'm really after a Supply that can give me one line of input as it arrives" part. That'd just be .lines.Supply innit?
16:38 Geth ¦ nqp: f263b9c8a6 | (Zoffix Znet)++ | tools/build/MOAR_REVISION
16:38 Geth ¦ nqp: Bump MoarVM
16:38 Geth ¦ nqp: review: https://github.com/perl6/nqp/commit/f263b9c8a6
16:38 Geth ¦ nqp: version bump brought these changes: https://github.com/MoarVM/MoarVM/compare/2017.10-56-g2f0672f...2017.10-57-gd257df2
16:38 b2gills Here is a quick test that I created that shows the issue with that code https://gist.github.com/b2gills/354aed9849d8a0f5fc537882f28f896c
16:38 Geth ¦ rakudo: 629e8684e1 | (Zoffix Znet)++ | tools/build/NQP_REVISION
16:38 Geth ¦ rakudo: Bump NQP
16:38 Geth ¦ rakudo: review: https://github.com/rakudo/rakudo/commit/629e8684e1
16:38 Zoffix ZOFFLOP: t/spec/S11-modules/require.t
16:42 Zoffix Yeah. It'd just mark as EOF any time we read off TTY anything smaller than 0x100000 bytes :o
16:43 b2gills Unless we want to make `.eof` blocking on interactive input, I think it is working correctly.
16:43 Zoffix Our TTY handle tests need some beefing up.
16:44 b2gills Well I'd run the test in the gist I just posted, except git isn't pulling down any data.
16:55 travis-ci joined #perl6-dev
16:55 travis-ci NQP build failed. Zoffix Znet 'Bump MoarVM'
16:55 travis-ci https://travis-ci.org/perl6/nqp/builds/297261642 https://github.com/perl6/nqp/compare/a25acd1e82e0...f263b9c8a60f
16:55 travis-ci left #perl6-dev
17:00 Geth ¦ nqp: b21e0e25f6 | pmurias++ | t/nqp/023-named-args.t
17:00 Geth ¦ nqp: [jvm] Skip not yet fixed test
17:00 Geth ¦ nqp: review: https://github.com/perl6/nqp/commit/b21e0e25f6
17:06 Zoffix "<Zoffix> [...] Will take a look later if there's some trick can be done"   Won't. Can't think of anything that wouldn't make .eof block at some point and current behaviour isn't that terrible.
17:06 * Zoffix &
17:17 Zoffix /o\
17:17 travis-ci joined #perl6-dev
17:17 travis-ci NQP build passed. pmurias '[jvm] Skip not yet fixed test'
17:17 travis-ci https://travis-ci.org/perl6/nqp/builds/297268451 https://github.com/perl6/nqp/compare/f263b9c8a60f...b21e0e25f6de
17:17 travis-ci left #perl6-dev
17:19 Zoffix IO::CatHandle.read's design is crap. It switches to next handle if it didn't read enough stuff, but that doesn't quite work right, since .read on, say, TTY STDIN could return only some bytes and still not be at EOF
17:50 lizmat Zoffix: good point
17:50 lizmat hmmm... though, is it?
17:51 lizmat isn't read supposed to block until it has enough bytes ?
17:51 dogbert17 wrt the latest ZOFFLOP: not ok 31 - got correct STDOUT
17:51 dogbert17 # Failed test 'got correct STDOUT'
17:52 dogbert17 # at t/spec/S17-procasync/basic.t line 85
17:52 dogbert17 # expected: 'Perl 6'
17:52 dogbert17 #      got: 'Perl'
17:52 dogbert17 seems to happen under load
18:01 lizmat .oO( is that a sign ?  )
18:06 Zoffix lizmat: no idea. it does block until at least one byte was received. Were it to block until 0x100000 bytes (the .read size our .lines stuff uses) were read from STDIN then it'd expect the user to bang out quite a few lines before .lines would give up the first line, wouldn't it?
18:08 Zoffix dogbert17: good spot. Looks like second write promise is never awaited: https://github.com/perl6/roast/blob/master/S17-procasync/basic.t#L75
18:08 Zoffix dogbert17: need to await it before this line: https://github.com/perl6/roast/blob/master/S17-procasync/basic.t#L80
18:29 Geth ¦ rakudo: 6bca84fa1a | MasterDuke17++ | 2 files
18:29 Geth ¦ rakudo: Add some INTERPOLATE variations
18:29 Geth ¦ rakudo:
18:29 Geth ¦ rakudo: Create INTERPOLATE_ASSERTION and some multis to pull functionality out
18:29 Geth ¦ rakudo: of the regular INTERPOLATE multis.
18:29 Geth ¦ rakudo:
18:29 Geth ¦ rakudo: Makes `/ $rx /` and `/ <$rx> /` a bit faster.
18:29 Geth ¦ rakudo: review: https://github.com/rakudo/rakudo/commit/6bca84fa1a
18:29 Geth ¦ rakudo: ea3ffa2ee3 | (Aleks-Daniel Jakimenko-Aleksejev)++ (committed using GitHub Web editor) | 2 files
18:29 Geth ¦ rakudo: Merge pull request #1232 from MasterDuke17/add_some_INTERPOLATE_variations
18:29 synopsebot RAKUDO#1232 [closed]: https://github.com/rakudo/rakudo/pull/1232 Add some INTERPOLATE variations
18:29 Geth ¦ rakudo:
18:29 Geth ¦ rakudo: Add some INTERPOLATE variations
18:29 Geth ¦ rakudo: review: https://github.com/rakudo/rakudo/commit/ea3ffa2ee3
18:30 AlexDaniel huggable: cla
18:30 huggable AlexDaniel, nothing found
18:30 AlexDaniel huggable: CLA
18:30 huggable AlexDaniel, http://www.perlfoundation.org/contributor_license_agreement  print it out, sign it, scan it in (resize your scan to 250KB or less). Email to trademark -at- perlfoundation.org
18:36 MasterDuke i've just never gotten around to it. guess i will eventually
18:40 Geth ¦ rakudo: 2f12bea114 | (Elizabeth Mattijsen)++ | lib/Telemetry.pm6
18:40 Geth ¦ rakudo: Telemetry tweaks
18:40 Geth ¦ rakudo:
18:40 Geth ¦ rakudo: - make sure order of @format *is* actually alphabetical
18:40 Geth ¦ rakudo: - add blurb about supervisor thread in report header
18:40 Geth ¦ rakudo: review: https://github.com/rakudo/rakudo/commit/2f12bea114
18:50 AlexDaniel MasterDuke: please do
18:52 MasterDuke lizmat: did i see you're registered for the london perl workshop?
18:52 lizmat yes, that's correct :-)
18:53 MasterDuke cool, i just registered yesterday. passing through london for work, decided to fly a day early and catch the conference
18:54 Geth ¦ roast: 6e68b2c884 | (Zoffix Znet)++ (committed using GitHub Web editor) | S17-procasync/basic.t
18:54 Geth ¦ roast: Add missing await
18:54 Geth ¦ roast: review: https://github.com/perl6/roast/commit/6e68b2c884
18:54 Zoffix dogbert17: ^ that should fix it
18:54 lizmat MasterDuke: looking forward to meeting you!
18:54 MasterDuke likewise
19:08 Zoffix Well, that was a waste of a day. Made a throttler/scheduler in Perl with Minion for a thing at $work... and it doesn't actually run on a target server because it's Windows and windows don't support fork. In Perl 6 it'd just be Supply.throttle -_-
19:17 Geth ¦ rakudo: 4ed91ed644 | (Elizabeth Mattijsen)++ | lib/Telemetry.pm6
19:17 Geth ¦ rakudo: More Telemetry tweaks
19:17 Geth ¦ rakudo:
19:17 Geth ¦ rakudo: - default for :header-repeat is now %*ENV<RAKUDO_REPORT_HEADER_REPEAT> // 32
19:17 Geth ¦ rakudo: - default for :legend is now %*ENV<RAKUDO_REPORT_LEGEND> // 1
19:17 Geth ¦ rakudo: - corrected report header for cpu/wallclock: we want seconds, not microseconds
19:17 Geth ¦ rakudo: review: https://github.com/rakudo/rakudo/commit/4ed91ed644
19:24 dogbert17 Zoffix++, so it was the test itself which was broken
19:24 dogbert17 one ZOFFLOP hopefully gone :)
19:26 Zoffix :)_
19:42 evalable6 joined #perl6-dev
20:43 Geth ¦ rakudo: c4d373c514 | (Elizabeth Mattijsen)++ | src/core/Kernel.pm
20:43 Geth ¦ rakudo: Expose basic CPU usage information: Kernel.cpu-usage
20:43 Geth ¦ rakudo:
20:43 Geth ¦ rakudo: For situations like the Benchmark module, we don't want to have to load
20:43 Geth ¦ rakudo: a lot of stuff.  This gives the bare-bones of the stuff we need there.
20:43 Geth ¦ rakudo: review: https://github.com/rakudo/rakudo/commit/c4d373c514
21:11 Geth ¦ rakudo: fea5612cb8 | (Elizabeth Mattijsen)++ | lib/Telemetry.pm6
21:11 Geth ¦ rakudo: Allow for "use Telemetry <cpu wallclock>"
21:11 Geth ¦ rakudo:
21:11 Geth ¦ rakudo: Allow selective import of all subs that are available with :COLUMNS.
21:11 Geth ¦ rakudo: review: https://github.com/rakudo/rakudo/commit/fea5612cb8
21:20 Geth ¦ rakudo: 360eb2289f | (Elizabeth Mattijsen)++ | lib/Telemetry.pm6
21:20 Geth ¦ rakudo: Oops, we want the default exports available as well
21:20 Geth ¦ rakudo:
21:20 Geth ¦ rakudo: - so not only slice through EXPORT::COLUMNS, but also EXPORT::DEFAULT
21:20 Geth ¦ rakudo: review: https://github.com/rakudo/rakudo/commit/360eb2289f

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