| samcv | MasterDuke, looking | 00:40 | |
| actually going to the bathroom then looking :) | |||
| MasterDuke didn't think the code was that bad... | 00:41 | ||
| notviki | :D | ||
| dalek | ast: 3cc76b8 | (Zoffix Znet)++ | S32-io/seek.t: Test IO::Handle.seek Fixes #219 It currently throws when seeking past start of file, but doesn't complain about seeking past end. I left throwage tests commented out for now, will make figuring out the failure modes as part of IO improvement stuff. |
00:44 | |
| geekosaur | um. why would seeking past end be wrong? | 00:51 | |
| is a POSIX module going to need to provide special seek calls then? | 00:53 | ||
| notviki | <notviki> hm... seek before start of file throws but seek past end of file doesn't throw. Is that normal? | 00:54 | |
| geekosaur was AFK for around an hour | |||
| en.wikipedia.org/wiki/Sparse_file | 00:55 | ||
| ah, not even limited to POSIX any more, NTFS does sparse files as well | |||
| dalek | ast: e220988 | (Zoffix Znet)++ | S32-io/seek.t: Test .seek failure modes; geekosaur++ Part of fix for #219 |
01:01 | |
| samcv | who here is an admin for rakudo? | 01:02 | |
| who do I bug to get an M-dash or a horizontal bar used for the rakudo description instead of two hyphen-minus’s on github | |||
| notviki doesn't have that access | 01:03 | ||
| [Coke] | samcv: what are you looking at? | ||
| samcv | github.com/rakudo/rakudo | ||
| so do — mdash or ― horizontal bar (usually used for quoting people ― samcv) | |||
| M-dash is probably better for this use since not a quote tho | 01:04 | ||
| [Coke] | owners are: jnthn, moritz, perlpilot, pmichaud. | 01:05 | |
| samcv | kk will bug one of them | 01:06 | |
| MasterDuke | samcv: have you taken a look at github.com/rakudo/rakudo/pull/991? | 01:07 | |
| that's another that could use a review | |||
| samcv rushes to the bathroom again /s | 01:10 | ||
| looking | |||
| dalek | kudo/nom: 119efbc | samcv++ | appveyor.yml: Use appveyor-retry for installing Strawberry Perl 5 |
01:13 | |
| MasterDuke | thanks. i may not be very available the next week and wanted to get them reviewed while i was still here and it was fresh in people's minds | ||
| samcv | MasterDuke, ok that last one, about the ${ } variables. uhm not seeing many roast tests for that | 01:18 | |
| MasterDuke | there aren't tests for any of my three PRs today | ||
| samcv | oh :( | 01:19 | |
| can you please make a test for github.com/rakudo/rakudo/pull/991/ ? | |||
| dalek | kudo/nom: ecde669 | (Zoffix Znet)++ | t/spectest.data: Add S32-io/seek.t to list of test files to run |
||
| MasterDuke | i'll add them in the next couple days | ||
| notviki notices a 3K+ tests in stresstest run :i | |||
| *:o | |||
| I mean there are 3K+ tests more than a few days ago ^_^ | 01:20 | ||
| samcv | what are they tho | ||
| notviki | no idea | ||
| Files=1213, Tests=133493, 484 wallclock secs (14.31 usr 2.61 sys + 1632.20 cusr 184.80 csys = 1833.92 CPU) | |||
| samcv | are all normal tests in the stresstest? | ||
| cause that's 1900 emoji tests | |||
| MasterDuke | samcv: didn't you add a bunch of emoji/unicode tests? | 01:21 | |
| notviki | last one from 2 days ago: Files=1211, Tests=130982, 165 wallclock secs (21.10 usr 3.39 sys + 2923.98 cusr 288.86 csys = 3237.33 CPU) | ||
| samcv | yep | ||
| notviki | \o/ | ||
| MasterDuke | notviki: same computer? cause that's a huge difference in wallclock | ||
| notviki | Nah :) The slower one is my home box, the faster one is my 24-core VM :) | 01:22 | |
| MasterDuke | good, i hoped those 3k tests didn't add 300s! | 01:23 | |
| notviki | :) | ||
| samcv | i hope not too | 01:32 | |
| MasterDuke, reviewed that more recent one | 01:34 | ||
| MasterDuke | k, i may get a test or two in today | 01:36 | |
| samcv | cool | 01:38 | |
| ooooooooooooooo | 01:40 | ||
| github.com/showcases/programming-languages rakudo is there now! | |||
| \o/ | 01:41 | ||
| timotimo | yup! | 01:57 | |
| a friend of mine hooked me up with a tiny bit of code that 1) reads the high precision timer, 2) sends events out to a thread that gathers them and 3) writes those out to a file | 01:58 | ||
| samcv | notviki, hmm seemed the instructions to override the normal perl highlighter for atom on the readme for atom-language-perl6 were wrong :) updating now | ||
| i had never actually tried them before :P | |||
| timotimo, who were you responding to? | 02:08 | ||
| timotimo | the yup was about rakudo in the showcase | 02:12 | |
| samcv | ah | ||
| but the one below that | |||
| when sometimes a normal solidus won't do, you have the BIG SOLIDUS ⧸ | 02:13 | ||
| timotimo | i just blurted that out randomly | 02:19 | |
| and now i'm going to bed i think | |||
| dalek | ast: 2667e3e | MasterDuke17++ | S (2 files): Tests for Rakudo PRs #990 and #991 and NQP PR #337 |
04:28 | |
| MasterDuke | samcv: a test for each added, todo'd for PR #991 | 04:30 | |
| samcv | nice | 05:53 | |
| MasterDuke, PR ready for merge? | 05:56 | ||
| .tell notviki seems perl6.fail has… failed. It is not showing any tickets | 06:29 | ||
| yoleaux2 | samcv: I'll pass your message to notviki. | ||
| samcv | .tell yoleaux2 thanks | ||
| yoleaux2 | samcv: Thanks for the message. | ||
| samcv | buggable, new | 06:44 | |
| buggable, help | |||
| buggable | samcv, tags | tag SOMETAG | eco | eco Some search term | speed | ||
| samcv | buggable, speed | ||
| buggable | samcv, ▇▇█▇█▇██▇▇▇█▄▆▅▃▄▄▂▃▃▄▂▁▁▂▃▂▂▅▂▂▄▂▃▂▂▁▆▄▂▂▃▂▂▂▃▃▃▃ data for 2016-12-16–2017-01-06; range: 5.137s–6.242s | ||
| ugexe | m: my $core-repo = CompUnit::RepositoryRegistry.repository-for-name(q|perl|); my $core-tap-harness-path = "$core-repo/sources/" ~ "$core-repo/dist".IO.dir.map({Distribution::Path.new($_.parent, :meta-file($_))}).head.meta<provides><TAP><lib/TAP.pm6><file>; require $core-tap-harness-path; # notviki: so it *is* possible | 07:25 | |
| camelia | ( no output ) | ||
| ugexe | ...unless someone does zef --install-to=perl install TAP::Harness | 07:34 | |
| [Tux] | This is Rakudo version 2016.12-237-gecde66939 built on MoarVM version 2016.12-71-g331a6b43 | 10:04 | |
| csv-ip5xs 3.006 | |||
| test 13.148 | |||
| test-t 5.466 | |||
| csv-parser 13.721 | |||
| lizmat | Files=1163, Tests=56515, 198 wallclock secs (11.19 usr 4.59 sys + 1176.14 cusr 130.22 csys = 1322.14 CPU) | 10:09 | |
| bartolin_ | lizmat: what do you think about this change? Are there any disadvantages in binding to natives in signatures? github.com/usev6/rakudo/commit/1b333a03a7 | 10:14 | |
| lizmat | there used to be a time that didn't work | 10:15 | |
| also, I'm not sure whether that binds or assigns | |||
| bartolin_ | yes, jnthn++ fixed that not so long ago IIRC | ||
| lizmat | and/or whether it has a performance benefit | 10:16 | |
| bartolin_ | I see. that was my main question (whether you'd have an idea about performance impacts) | 10:17 | |
| spectest passes for MoarVM an JVM | 10:18 | ||
| bartolin_ has never tried to benchmark changes before | 10:19 | ||
| I'll put it on my list :-) thanks for looking | 10:21 | ||
| lizmat | m: dd $(1, 2) Z <a b c> | 11:02 | |
| camelia | rakudo-moar ecde66: OUTPUT«(($(1, 2), "a"),).Seq» | ||
| lizmat | m: dd 1 Z <a b c> | 11:03 | |
| camelia | rakudo-moar ecde66: OUTPUT«((1, "a"),).Seq» | ||
| lizmat | hmmm... guess that makes sense :-) | ||
| m: dd <a b c> Z 1 | |||
| camelia | rakudo-moar ecde66: OUTPUT«(("a", 1),).Seq» | ||
| lizmat | jnthn: on the subject of YAGNI, I'm wondering whether push-until-lazy is a bit of a YAGNI currently | 11:42 | |
| it appears to be from a time where we also had "plan" planned | |||
| atm the only thing it does is push-all if the iterator is non-lazy | 11:43 | ||
| dalek | kudo/nom: 8655413 | lizmat++ | src/core/Rakudo/Internals.pm: Introducing R:I.OneValueIterator Create an iterator for a single given value. |
12:05 | |
| kudo/nom: c405f06 | lizmat++ | src/core/ (2 files): Use the R:I.OneValueIterator where we can |
|||
| jnthn | lizmat: push-until-lazy is the thing that gives the "mostly eager" semantics used in assignment, ain't it? | 12:28 | |
| iirc, $!todo is a list of iterables, and push-until-lazy just causes a push-all on each iterator until it encounters one that's marked lazy | 12:44 | ||
| And yes, .plan would use it, though I'm pretty sure I put it in because something *else* needed it | 12:45 | ||
| lizmat | jnthn: I think that $!todo is only the one iterator ? | 12:46 | |
| ah, yes, it's the one Reifier :-) | 12:47 | ||
| hmmm... goes into some more research :-) | 12:49 | ||
| dalek | kudo/nom: 6888f0c | lizmat++ | src/core/ (3 files): Change R:I.OneValueIterator to R:I.OneCallableIterator Which is a bit more generally useful, and the additional overhead of needing to do an extra call to get the single value, should drown in the general iterator overhead. |
13:16 | |
| kudo/nom: 012850d | lizmat++ | src/core/ (3 files): Revert "Change R:I.OneValueIterator to R:I.OneCallableIterator" This reverts commit 6888f0ce2e750af2cbd2d5e0f051c93d01c59735. Premature optimization is the root of all evil |
13:43 | ||
| masak smiles at the pair of commits above :) | 14:25 | ||
| lizmat | yeah... :-) | 14:29 | |
| dalek | kudo/nom: c0ac0e9 | lizmat++ | src/core/Rakudo/Internals.pm: Introducing R:I.ZipIterablesIterator Create an iterator of N iterables, returning a List of one pull from each iterator for each pull. |
14:31 | |
| kudo/nom: 4ab020f | lizmat++ | src/core/List.pm: Make @a Z @a 5x faster - by using the new R:I.ZipIterablesIterator - benchmarked on "my @a = ^5; for ^10000 { my @b = @a Z @a } |
|||
| lizmat | masak: some better commits :-) | ||
| timotimo | ooooooh | 14:33 | |
| this'll show up in the benchmarks | 14:34 | ||
| we do have benchmarks that have Z in them | |||
| lizmat | this is just the Z case, other cases are being worked on :-) | ||
| timotimo | fantabular | ||
| lizmat | first I'm going to check whether it makes sense to make a special case of @a Z @a (aka only 2 lists) | 14:35 | |
| timotimo | statisfiable6: core0 | ||
| statisfiable6 | timotimo, OK! Working on it… | ||
| timotimo | not sure how long those take until a newer build is available | ||
| 4 minutes may be a bit too optimistic :) | 14:36 | ||
| dalek | kudo/nom: 3d1d699 | lizmat++ | src/core/Rakudo/Internals.pm: Presize the result list for another 4% improvement |
14:51 | |
| lizmat | doesn't make sense to make a specialized 2 iterables case | 15:11 | |
| timotimo | good to know the general one is already fast enough :) | 15:13 | |
| lizmat | yeah, it may be about 1% faster | ||
| samcv | morning perl 6 * | 15:15 | |
| timotimo | statisfiable6 is taking its sweet time, isn't it? | 15:16 | |
| i thought it caches data from already-analysed versions | |||
| samcv | nope | ||
| timotimo | maybe that means it'll end up catching liz' last commit, too | 15:18 | |
| just by virtue of taking so long to reach it | |||
| though it might start from the newest commit it knows and go back in history | |||
| github.com/perl6/whateverable/blob...ble.p6#L35 - it does seem to have these files there | 15:20 | ||
| ah, looks like it only saves the last result there | 15:21 | ||
| notviki | . | 15:34 | |
| yoleaux2 | 06:29Z <samcv> notviki: seems perl6.fail has… failed. It is not showing any tickets | ||
| timotimo | i didn't expect statisfiable6 to take *this* long | 16:12 | |
| lizmat | statisfiable6: status | 16:14 | |
| statisfiable6 | lizmat, Huh? Available stats: core (CORE.setting size), install (size of the whole installation), … | ||
| lizmat | statisfiable6: core | 16:15 | |
| statisfiable6 | lizmat, OK! Working on it… | ||
| notviki | statisfiable6: install | 16:16 | |
| statisfiable6: … | |||
| lizmat | statisfiable6: status | 16:17 | |
| guess it's off into lala land | |||
| timotimo | there it goes | 16:20 | |
| notviki | statisfiable6: core | 16:21 | |
| statisfiable6 | notviki, OK! Working on it… | ||
| notviki | well, it's doing something... at the top of top | 16:23 | |
| samcv | statisfiable6, core | 17:57 | |
| statisfiable6 | samcv, OK! Working on it… | ||
| timotimo | hm | 19:56 | |
| geekosaur | now you just need stabilifiable6 :p | ||
| timotimo | totally | ||
| [Coke] | é | 20:18 | |
| AlexDaniel | statisfiable6: core | 20:26 | |
| statisfiable6 | AlexDaniel, OK! Working on it… | ||
| AlexDaniel | there's some broken tar it trips over… | 20:27 | |
| timotimo | oh damn | 20:28 | |
| AlexDaniel | statisfiable6: core | ||
| statisfiable6 | AlexDaniel, OK! Working on it… | ||
| AlexDaniel | ok now | 20:29 | |
| statisfiable6 | AlexDaniel, gist.github.com/11acd3d33f9b20c8cd...8911f29c3e | ||
| AlexDaniel | statisfiable6: install | 20:30 | |
| statisfiable6 | AlexDaniel, OK! Working on it… | ||
| AlexDaniel, gist.github.com/36ff15f9f5662a20fa...4ef93b946f | 20:31 | ||
| AlexDaniel | good | ||
| timotimo | that's quite some dip it had there | 20:32 | |
| AlexDaniel | yup | ||
| timotimo | immediately went up again for some reason | ||
| AlexDaniel | could be something with the archive itself | 20:33 | |
| timotimo | i don't see it in the listing below | ||
| AlexDaniel | you have to “view full file” | 20:34 | |
| timotimo | oh, ah | 20:38 | |
| 15a2f112ba2e78686bb6f5cc9b92b122254f2b70 11.86428 | 20:40 | ||
| c5e54efc88c5d5e481b3fcb62d412b1826bee25d 10.763264 | |||
| 1e3a325bd1db3402a0065e12e9f26e400f6b2900 11.864208 | |||
| this it? | |||
| AlexDaniel | timotimo: yea | 20:51 | |
| dalek | p: 189cbd4 | MasterDuke17++ | src/QRegex/P (3 files): Handle Nd's as regex quantifiers |
21:13 | |
| p: ab0728f | lizmat++ | src/QRegex/P (3 files): Merge pull request #338 from MasterDuke17/correctly_handle_unicode_digits_in_regex_quantifiers Handle Nd's as regex quantifiers |
|||
| kudo/nom: e40a129 | lizmat++ | tools/build/NQP_REVISION: Bump NQP to get Nd's as regex quantifiers |
21:23 | ||
| p: d626cf6 | MasterDuke17++ | src/ (4 files): Micro-optimize Grammars Make <before>'s non-capturing. |
|||
| p: 9c2f39c | lizmat++ | src/ (4 files): Merge pull request #339 from MasterDuke17/make_befores_non-capturing Micro-optimize Grammars |
|||
| kudo/nom: 06abb07 | lizmat++ | tools/build/NQP_REVISION: Bump NQP to make <before>'s non-capturing |
21:32 | ||
| lizmat | TIL that Zop is not taking the infix:<Z>(:$with!) candidate | 21:48 | |
| timotimo | oh? | 21:50 | |
| is that good or bad? probably bad | |||
| lizmat | not sure yet... I guess it codegens directly to one of the METAOP thingies | 21:52 | |
| timotimo | ah, sounds likely | 22:02 | |
| dalek | kudo/nom: eba3fe0 | MasterDuke17++ | src/Perl6/Grammar.nqp: Handle Nd's in ${} special variables |
22:44 | |
| kudo/nom: f6303d7 | samcv++ | src/Perl6/Grammar.nqp: Merge pull request #991 from MasterDuke17/correctly_error_on_unicode_digits_in_special_variable Handle Nd's in ${} special variables |
|||
| lizmat | good night, #perl6-dev! | 22:49 | |
| notviki | night |