IRCloggy #perl6-dev 2017-07-10

Logs Search ←Prev date Next date→ Channels Documentation

Provider of IRC logs since 2005.
WARNING: As Freenode became unjoinable and lost all warnings in topics, we cannot log channels on Freenode anymore.

2017-07-10

samcv crosses fingers and hopes it reduces the spectest time00:01
samcv wow00:17
MasterDuke faster?00:18
samcv 16% faster nqp::index for strings where one is 8bit one in 32 bit00:18
yes00:18
MasterDuke nice00:18
samcv i hijacked my function i use for ignorecase/ignoremark00:18
added one extra conditional00:18
and now it works for everything00:18
:P00:18
all variations of ignorecase+ignoremark or without00:19
wheeee00:19
not going to complain about 16% faster :)00:19
from 1.87seconds to 1.1600:19
may be more than 16% faster hmm00:19
the numbers *look* more than 16% faster00:20
err wiat no00:20
before 2.16 after 1.8400:20
it's called string_equal_at_ignore_case_INTERNAL_loop but now it does everything. so00:24
it could be ranamed :P00:24
Zoffix .oO( INTERNAL_loop Perl 6 )00:26
Geth ¦ nqp: f9ca4f14f0 | (Samantha McVey)++ | tools/build/MOAR_REVISION00:32
¦ nqp: Bump MoarVM for Uni fixes and nqp::index 16% speedup in many cases 00:32
¦ nqp:00:32
¦ nqp: 161ec639 Make string_index 16% faster for incompatible string types00:32
¦ nqp: e13c30b7 Use MVMint64 for MVM_string_chr (oplist/function mismatch)00:32
¦ nqp: 89a9999f Update the README to indicate Unicode 9.0 support00:32
¦ nqp: bef5802b Make some fixes to the Unicode data file downloader00:32
¦ nqp: fa5158a3 Don't break after ZWJ for Emoji=True + GCB=Other00:32
¦ nqp: <…commit message has 8 more lines…> 00:32
¦ nqp: review: https://github.com/perl6/nqp/commit/f9ca4f14f000:32
¦ nqp: version bump brought these changes: https://github.com/MoarVM/MoarVM/compare/2017.06-43-g45b008f...2017.06-56-g161ec639 00:33
¦ rakudo/nom: 4f5a1e20f9 | (Samantha McVey)++ | tools/build/NQP_REVISION00:35
¦ rakudo/nom: Bump NQP/Moar for Uni fixes and regex speed improvement 00:35
¦ rakudo/nom:00:35
¦ rakudo/nom: NQP commits brought in:00:35
¦ rakudo/nom: f9ca4f14f Bump MoarVM for Uni fixes and nqp::index 16% speedup in many cases00:35
¦ rakudo/nom: 8453fac89 Factor out part of compile for easier re-use00:35
¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/4f5a1e20f900:35
¦ rakudo/nom: version bump brought these changes: https://github.com/perl6/nqp/compare/2017.06-36-gda0dda8...2017.06-38-gf9ca4f14f 00:35
samcv not a bad weekend00:36
MasterDuke Zoffix: i was just looking into the Goodreads API, and what name do i see attached to a Perl module...00:59
tbrowder ref PR 1110: i hope someone can verify it works for them, too (linux only for the moment, requires 'install' program, could be changed to straight use of 'cp' instead i think if 'cp -rfp' is portable).02:01
geekosaur "cp -rfp"'s not going to fly on macos or *bsd02:02
tbrowder .tell llfourn please look at PR 111002:05
yoleaux tbrowder: I'll pass your message to llfourn.02:05
ugexe i'd be willing to bet it used absolute paths for windows02:06
to handle volumes02:06
llfourn .02:21
yoleaux 02:05Z <tbrowder> llfourn: please look at PR 111002:21
llfourn tbrowder: I think I02:21
I think I'm the wrong person to review your PR as I know little about how rakudo is built02:21
though I'm interested in your progress in this direction :)02:22
Geth ¦ rakudo: LLFourn++ created pull request #1111: begin_time_lexical_fixup for constants 03:31
¦ rakudo: review: https://github.com/rakudo/rakudo/pull/111103:31
llfourn #1111!!!!03:31
[Tux] This is Rakudo version 2017.06-184-g4f5a1e20f built on MoarVM version 2017.06-56-g161ec63906:38
csv-ip5xs 2.52206:38
test 13.45206:38
test-t 4.171 - 4.19606:38
csv-parser 12.66006:38
Error encoding UTF-8 string: could not encode codepoint 1548168 (0x179F88), codepoint out of bounds. Cannot encode higher than 1114111 (0x10FFFF)06:38
in block <unit> at 88.t line 2106:38
samcv, https://github.com/Tux/CSV/blob/master/88.t06:39
Geth ¦ rakudo/nom: b4b6b9409c | (Elizabeth Mattijsen)++ | src/core/set_symmetric_difference.pm09:03
¦ rakudo/nom: No need to coerce SetHashes to Sets 09:03
¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/b4b6b9409c09:03
Zoffix NeuralAnomaly: status09:07
NeuralAnomaly Zoffix, [✘] Next release will be in 4 days and 19 hours. Since last release, there are 24 new still-open tickets (3 unreviewed and 0 blockers) and 187 unreviewed commits. See http://perl6.fail/release/stats for details09:07
Geth ¦ rakudo/nom: c57590fd0c | (Elizabeth Mattijsen)++ | src/core/set_symmetric_difference.pm09:09
¦ rakudo/nom: Fix copy-pasto 09:09
¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/c57590fd0c09:09
Zoffix .tell [Coke] Looks like something busted with RT. I saw an email reply to RT#131722 yesterday and it had the 'bug follow up' address in it, but there's nothing on the ticket when vewing on Web09:11
synopsebot6 Link: https://rt.perl.org/rt3/Public/Bug/Display.html?id=13172209:11
yoleaux Zoffix: I'll pass your message to [Coke].09:11
Zoffix .tell kybr you should add yourself to CREDITS file, unless you want to be credited as all-lowercase "karl yerkes": https://github.com/perl6/doc/blob/master/CREDITS09:33
yoleaux Zoffix: I'll pass your message to kybr.09:33
Zoffix guesses seatek wants to be seatek09:34
Geth ¦ roast: ff11b6d0c0 | (Elizabeth Mattijsen)++ | 2 files09:35
¦ roast: Correct [(^)] tests 09:35
¦ roast:09:35
¦ roast: - [(^)] $a, $b, $c is not the same as $a (^) $b (^) $c09:35
¦ roast: - follows https://github.com/rakudo/rakudo/commit/0cdd6c25355f2a747a609:35
¦ roast: - and https://github.com/rakudo/rakudo/commit/c57590fd0c46c3b2da1e509:35
¦ roast: review: https://github.com/perl6/roast/commit/ff11b6d0c009:35
Zoffix Do a lot of ops have different behaviour between [op] and $a op $b? I always assumed [op] was just a way to avoid repetition between each argument09:40
This is gonna be the last release I do.09:44
.tell AlexDaniel weren't you planning to submit your CLA so you could cut releases? The August release needs a volunteer09:45
yoleaux Zoffix: I'll pass your message to AlexDaniel.09:45
Zoffix .ask [Coke] what's the usual timeframe to receive a check from TPF? Wondering if mine is lost or if it usually takes a long time.10:13
yoleaux Zoffix: I'll pass your message to [Coke].10:13
nine Zoffix: thank you so much for all the work you have done on cutting so many releases and automating the process! I guess only people who have done it before you can appreciate the worth of that work in full.10:19
Zoffix zoffix@VirtualBox~/CPANPRC/rakudo (nom)$ git log 2017.06...HEAD | grep -i temporary | wc -l10:21
510:21
I question the impact of all these "temporary" fixes. They sweep the problem under the rug, significantly reducing motivation for a proper fix. So what we end up with is a bunch of bandaids all over the codebase.10:22
zoffix@VirtualBox~/CPANPRC/rakudo (nom)$ git log 2015.12...HEAD | grep -i temporary | wc -l10:23
2410:23
Case in point. This bug was a release blocker but we kludged it with a goal of fixing it "next week". That was on May 19th and it's still kludged: https://github.com/rakudo/rakudo/commit/2f22b701d4cf44fa43500edcc3450718fa7fd46810:25
This commit excluded the entire test file from 6.c-errata tests https://github.com/rakudo/rakudo/commit/662794668b3788ccc4d2b0f364b4ed2639aa8cef10:34
https://irclog.perlgeek.de/perl6-dev/2017-05-25#i_1463544410:35
"Your branch is behind 'origin/6.c-errata' by 19 commits, and can be fast-forwarded."10:36
so much for immutable10:36
nine 6.c is immutable. 6.c-errata is a moving target. Though at times we are probably too liberal with making changes there.10:36
Zoffix Never mind. That file isn't in 6.c10:36
Geth ¦ rakudo/nom: f1fa5ac5ba | (Zoffix Znet)++ (committed using GitHub Web editor) | docs/ChangeLog10:39
¦ rakudo/nom: Log changes up to 662794668b3788ccc4d2b0f364b4ed2639aa8cef 10:39
¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/f1fa5ac5ba10:39
¦ rakudo/nom: 5b0a7f518c | (Zoffix Znet)++ (committed using GitHub Web editor) | docs/release_guide.pod10:39
¦ rakudo/nom: Remove Zoffix from 2017-07-15 release 10:39
¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/5b0a7f518c10:39
Zoffix I give up10:39
Fucking shit10:39
lizmat is not seeing the problem with https://github.com/rakudo/rakudo/commit/662794668b3788ccc4d2b0f364b4ed2639aa8cef10:56
lizmat could someone explain ?10:56
nine lizmat: 12:36 < Zoffix> Never mind. That file isn't in 6.c10:57
lizmat: I think there is no problem10:57
lizmat well, apparently it was the straw that broke Zoffix's back :-(10:58
nine Yes. I fear we have burned through another valuable core developer :(10:58
lizmat hmmm... could someone activate dalek again ?11:04
seems we've also burned a bunch of bots in the process :-(11:05
lizmat just committed https://github.com/perl6/roast/commit/1fda005b0740d4dae204111:07
moritz I've added the old dalek hooks for perl6/doc and perl6/roast11:09
lizmat moritz++11:09
lizmat just committed https://github.com/rakudo/rakudo/commit/8afbfe6f0d73184f858090b111:10
lizmat guess it missed the hook adding :-(11:10
nine is setting up Geth on camelia's VM11:10
moritz rakudo/rakudo is neither perl6/doc nor perl6/roast11:10
added rakudo/rakudo and perl6/specs hooks too11:11
lizmat moritz: duh :-)11:11
nine For some reason zef fails to find zoffix' modules in the ecosystems. But the repos are all there11:11
lizmat seems Zoffix removed all of his modules from the ecosystem11:12
tbrowder nine: see my reply to yr comments on PR 111011:12
moritz I don't have the time right now to research where the hooks need to be updated. If somebody misses commits notifications from a repo, please tell me11:12
lizmat I bet mr_chromatic is now smiling for achieving another of his goald :-(11:12
*goals11:12
moritz afk11:14
nine lizmat: nah, you can't put that on him. Zoffix burned bright and fast. I've been worried for quite a while that it may have been too fast :/11:15
Darn: Missing serialize REPR function for REPR VMException (BOOTException)11:19
lizmat ok, I guess I'll clone all of Zoffix's ecosystem repoi's into some custodian account ?11:29
and update the ecosystem to look there ?11:29
moritz maybe to http://github.com/perl6-community-modules/ ?11:42
lizmat yeah, probably11:43
hmmm. my github foo is insufficient to know how to do that :p(11:45
ah, found it I thihnk11:46
tadzik ...what on earth happened :o11:50
(a rhetorical question, just expressing my surprise)11:50
moritz people happened11:54
moritz good at senslessly answering rethorical questions11:55
nine Ok, just added a no precompilation; to lib/Geth/GitHub/Hooks/Preprocessor.pm611:55
tadzik nodnod11:55
lizmat fwiw, I'm forking atm11:55
tadzik moritz: I'll keep that in mind next time I feel like asking "what's up" :P11:55
dalek rakudo/nom: c9ee86f | (Zoffix Znet)++ | docs/release_guide.pod:12:00
rakudo/nom: Revert "Remove Zoffix from 2017-07-15 release"12:00
rakudo/nom:12:00
rakudo/nom: This reverts commit 5b0a7f518c182af38def1a72b87a9d30391aeb57.12:00
rakudo/nom: review: https://github.com/rakudo/rakudo/commit/c9ee86f45812:00
yoleaux 19 Jun 2017 19:46Z <AlexDaniel> dalek: It's still alive!!!12:00
Zoffix Geth should be placed on the hack server. So that if it goes down anyone can restart it.12:02
Same as dalek; it starts on server start and `sudo service dalek restart` works12:03
tadzik I recently put all kinds of bots from my servers behind supervisord, so that they restart when they get down for any reason12:03
I can volunteer to make that happened if anyone wants12:04
nine Zoffix: I've just got it up and running on geth.niner.name. It's managed by systemd so it will be automatically restarted.12:04
tadzik ah, I guess systemd does that too :)12:05
Zoffix nine: if it segfaults, right?12:05
nine: the hooks point to some perl6.party address. Which address is the new one? I'm gonna try to setup a redirect.12:05
nine https://gist.github.com/niner/58404dc5578f3b816a06e2b8224efba712:05
That's my service file. Yes it will restart on segfaults12:06
geth.niner.name:8888 (it's camelia's VM)12:06
lizmat Zoffix: I've forked all of your modules into perl6-community-modules12:10
I see you reverted your ecosystem withdrawal12:11
Geth_ ¦ 6.d-prep: f4bc03d072 | (Zoffix Znet)++ | TODO/README.md12:11
¦ 6.d-prep: Fix link; atom-- 12:11
¦ 6.d-prep: review: https://github.com/perl6/6.d-prep/commit/f4bc03d07212:11
nine Zoffix++ # demonstrating how easily we can become dependent on a single person and the strain it puts on them12:11
Zoffix hm, redirect didn't work.12:11
Zoffix tries somethign elese12:11
lizmat was that intended to be permanent, or just temporary until I adapted the ecosystem to use perl6-community-modules ?12:11
moritz we can slowly change all the github hook URLs12:11
Zoffix lizmat: permanent.12:12
lizmat Zoffix: ok, then I will refrain from making any ecosystem changes12:13
Zoffix GitHub doesn't like the redirect. So I'll just keep it running until we change all the URLs12:13
Is geth.niner.name owned by more than one human?12:14
nine No :(12:14
Though a couple people do have access to that VM12:14
But the VM runs on a physical machine rented by me.12:15
Zoffix I get "Internal Server Error!" from that address. I think you may have installed the latest HTTP::Server::Tiny instead of version geth likes12:15
nine: what about hack.p6c.org?12:16
It's owned by more than one human, innit? Maybe it'd make sense to put geth there. Since there are many URLs that need to be changed and it's not trivial to know which ones still point to all address12:16
nine Zoffix: what HTTP::Server::Tiny version do I need? Geth's docs are kinda non-existing ;)12:18
Zoffix nine: zef --force --/test install https://github.com/tokuhirom/p6-HTTP-Server-Tiny/archive/26df561430699ab2b769c88cb0a68ec702554102.zip12:18
nine Better?12:19
Geth ¦ 6.d-prep: f4bc03d072 | (Zoffix Znet)++ | TODO/README.md12:19
¦ 6.d-prep: Fix link; atom-- 12:19
¦ 6.d-prep: review: https://github.com/perl6/6.d-prep/commit/f4bc03d07212:19
Zoffix yup12:19
nine Thanks :)12:20
Zoffix huggable: geth12:20
huggable Zoffix, nothing found12:20
Zoffix huggable: geth :is: Set geth report URL (select application/json and send me everything): http://geth.niner.name:8888/?chan=#perl6-dev12:21
huggable Zoffix, Added geth as Set geth report URL (select application/json and send me everything): http://geth.niner.name:8888/?chan=#perl6-dev12:21
nine NeuralAnomaly: status12:24
NeuralAnomaly nine, [✘] Next release will be in 4 days and 16 hours. Since last release, there are 22 new still-open tickets (0 unreviewed and 0 blockers) and 68 unreviewed commits. See http://perl6.fail/release/stats for details12:24
nine tbrowder: frankly I would prefer if existing package descriptions would continue to work unchanged. We don't even know which Linux distributions provide rakudo packages. Therefor we'd struggle to notify them about the change.12:26
tbrowder: I'm also not sure what problem the PR actually solves?12:26
(though when I've looked at the PR this morning, I had barely started drinking my coffee)12:27
tbrowder For me the PR starts the separation between configuration, building, and final installation (the only step with the PR that requires root privileges for system-wide use). it's the way I prefer and am used to for install-from source packages. Others have asked for DESTDIR support and I believe the PR changes cause the --prefix option to provide that with the newly portable ./install directory contents.12:32
Regarding package producers, afaik, they should be reading our install and changelog docs and would know of any changes as they create new packages.12:35
We could add a new doc just for package producers.12:37
One more point: due to the tight coupling between moar, nqp, and rakudo I would not suggest independent installation for other than rakudo, moare, nqp, and jvm core developers12:41
*moar12:42
(That opinion from a linux-only user.)12:43
nine But those are already separate packages? And it tends to work out well. In fact it was quite some work to separate the different prefixes for the different parts.12:46
[Coke] if you're a rakudo-only user, you don't care about the separation, but if you're trying to use any of the other bits, you might; given they are designed to be installed separately and the combined config/install in rakudo is a convenience, I'm not sure what bundling them does to help.12:52
AlexDaniel .13:14
yoleaux 09:45Z <Zoffix> AlexDaniel: weren't you planning to submit your CLA so you could cut releases? The August release needs a volunteer13:14
Zoffix m: without Failure.new { say .Set }13:14
camelia rakudo-moar c9ee86: OUTPUT: «set((HANDLED) Failed␤)␤»13:14
Zoffix I don't get wouldn't that return a set(), based on this commit? https://github.com/rakudo/rakudo/commit/43fc751be229f526b378b4a876e27b95da25237313:14
AlexDaniel Zoffix: ok, will do13:14
Zoffix Ah, nm, it's different on HEAD https://github.com/rakudo/rakudo/blob/nom/src/core/Failure.pm13:15
AlexDaniel++13:16
AlexDaniel tell dalek It's still alive!!!13:17
.tell dalek It's still alive!!!13:17
yoleaux AlexDaniel: I'll pass your message to dalek.13:17
AlexDaniel why did dalek report something for rakudo?13:18
nine AlexDaniel: Geth was unavailable for some time13:18
AlexDaniel hmhm13:18
moritz so I switched the rakudo repo to use dalek for reporting13:18
AlexDaniel ah, I see, ok13:20
Zoffix: so you are serious about this stuff, right?13:20
Zoffix AlexDaniel: about you doing August release? Yeah.13:21
lizmat afk for another bit&13:22
Zoffix nine: I think maybe IO::Socket::SSL is missing from geth's env13:28
Geth: ver https://github.com/rakudo/rakudo/commit/1dda86222b4e786c11ef0d2d3a538890d028088713:28
^ that should work13:28
Geth_: ver https://github.com/rakudo/rakudo/commit/1dda86222b4e786c11ef0d2d3a538890d028088713:28
Geth_ Zoffix, version bump brought in these changes: https://github.com/perl6/nqp/compare/2017.06-35-g17c795c...2017.06-36-gda0dda813:28
nine Geth: ver https://github.com/rakudo/rakudo/commit/1dda86222b4e786c11ef0d2d3a538890d028088713:31
Geth nine, version bump brought in these changes: https://github.com/perl6/nqp/compare/2017.06-35-g17c795c...2017.06-36-gda0dda813:31
nine Zoffix: fixed :)13:31
Zoffix cool13:31
Geth_ ¦ geth: c4e0f48fbb | (Stefan Seifert)++ | META6.json13:39
¦ geth: Add some META data for easier installation 13:39
¦ geth: review: https://github.com/perl6/geth/commit/c4e0f48fbb13:39
¦ geth: c90ec1986b | (Stefan Seifert)++ | geth.service13:39
¦ geth: Add a systemd service file 13:39
¦ geth: review: https://github.com/perl6/geth/commit/c90ec1986b13:39
tbrowder ref PR 1110: bummer, not working--must have stayed up too late! but i think it's fixable, closing till fixed...13:39
nine tbrowder: no sweat, it's not something we'd merge so close before a release anyway :)13:40
tbrowder but in spite of the close, is the direction not worth it to anyone besides me?13:43
if i find the fix, i think a movable installation directory alone is worth it.13:44
nine tbrowder: in general I'm all in favor of untangling our installation and making it understandable for a mere mortal.13:45
tbrowder ok, just checking...13:45
nine tbrowder: I do wonder though that making rakudo relocateable is so easy. I remember seeing lots of absolute paths and mention of issues.13:47
tbrowder yes, that's true, and i thought my false fix was too easy to be true!13:48
nine Also do the precompiled files really work after moving?13:49
Well that part would not be so surprising, as I made the source file and dependency references relative to the symbolic name of the repo. And that's the mechanism CompUnit::Repository::Staging is based on.13:51
I just tend to be nervous around that area. It was so hard to get working right :)13:52
tbrowder well, i don't know because i obviously didn't do a thorough test. if/when i get it working, i need to have a good test suite for sure. and i understand yr nervousness for sure!13:53
if i can get a new make test target equiv to "make test" but for the --prefix dir would that suffice?13:56
dalek rakudo/nom: fcf8de1 | (Zoffix Znet)++ | docs/ChangeLog:13:58
rakudo/nom: Log all commits to date13:58
rakudo/nom:13:58
rakudo/nom: Logs commits:13:58
rakudo/nom: 00dc4947 04746490 05c255c1 06379113 099a84b4 0cdd6c25 10f840fc 11b02d2c13:58
rakudo/nom: 19a1caa3 19be8722 1a4d9493 1ab4fd80 1b6d048b 1c0ed61a 1dda8622 201a0bfb13:58
rakudo/nom: 25047984 2a88c20c 2a8d1e7c 2ce5b678 3058ba0d 30619e8d 327c8409 35cc7c0e13:58
yoleaux 13:17Z <AlexDaniel> dalek: It's still alive!!!13:58
dalek rakudo/nom: 36823ab1 38509227 3d2a521c 40b0169d 41ed2c03 43c1767b 43fc751b 4693ec8613:58
rakudo/nom: 47678077 4894a75d 48c18f58 495fb5f8 49b1b03b 4a37de7b 4c91b522 4da8587913:58
rakudo/nom: 4f5a1e20 5a80412c 5ab4036e 62d54c75 640404fc 66aef589 6d4691fb 6dae179a13:58
rakudo/nom: 71ffb164 72e5d614 73c3bcc6 75797af3 79ce1a99 7f109ed7 80b3e89b 8230112813:58
rakudo/nom: 86e7b2bd 8afbfe6f 92f8abe0 93d81d61 9936a3be 9a2127f2 9c04dfc4 9d14a72413:58
rakudo/nom: a2090821 a2133dbc a2a0d5c6 a95c70bd abfd7d95 ad9ed1cb ae4c04ce b2d2bf5913:58
rakudo/nom: b31b159c b3916926 b4561229 b66d8783 bacaa051 bde28209 c226b71a c4e1473113:58
rakudo/nom: c585f370 c65652d8 c727462c c76d9324 c7922f14 cb06ebac d067abf4 d0995f6813:58
rakudo/nom: d1838461 d39f7b9a d5d3bd27 d765f186 d7e10466 d8c94353 da6c6584 e1e9091f13:58
rakudo/nom: e20817fb e4f3358f f03a176c f6d4fbd2 f71cbed4 f72c97cb fe7ea124 fff4333713:58
rakudo/nom: review: https://github.com/rakudo/rakudo/commit/fcf8de15b513:58
tbrowder e.g. in the same build directory, after "make install" run "make test-site-install"13:58
dalek rakudo/nom: 373b4ea | (Zoffix Znet)++ | docs/release_guide.pod:13:59
rakudo/nom: List AlexDaniel as release manager for 2017.08 release13:59
rakudo/nom: review: https://github.com/rakudo/rakudo/commit/373b4ea4d713:59
timotimo ZofBot: i'm a fan of the word "overflop" noo13:59
ZofBot timotimo, lookup array aha finish_parameterizing has a "XXX handle possible race" comment doesn't seem to break any more with a mutex added to it have you already fixed it? it may not be the fix jnthn wanted :) :) let's see what he says too late I guess, but now I have the SEGV in gdb turns out that there are13:59
timotimo now*13:59
er Zoffix i meant13:59
nine tbrowder: it's hard to anticipate the different failure modes. If I were to test that precomp stuff, I'd run something with RAKUDO_MODULE_DEBUG=1 and check if it does indeed load the installed precomp files.14:01
tbrowder nine: roger, thnx!14:02
dalek rakudo/nom: cb43086 | (Zoffix Znet)++ | docs/ChangeLog:14:06
rakudo/nom: Fix typo; timotimo++14:06
rakudo/nom: review: https://github.com/rakudo/rakudo/commit/cb4308634814:06
Zoffix buggable: 6.d15:22
buggable Zoffix, ¯\_(ツ)_/¯15:22
stmuk w.r.t. DESTDIR above .. I think its in both rakudo & nqp but not moarvm15:23
Zoffix ????????????????????????????????????????????????????????????????????????????????????15:26
15:26
REMINDER: Next Rakudo release will be this Saturday (July 15ths).15:26
Please check the ChangeLog to ensure your work was correctly logged:15:26
https://github.com/rakudo/rakudo/blob/nom/docs/ChangeLog#L1-L8115:26
15:26
????????????????????????????????????????????????????????????????????????????????????15:26
If someone has time, there's a bunch of roast failures on Windows that I'm guessing existed for awhile. Some of them are likely bogus new tests, but a few might be real bugs. Since this R* will be based on this release, would be nice to knock those down as much as possible.15:29
When I tried to debug someone's problem a few days back, precomp was glitching when running with modules from files. it would precomp the first time but running again would give "failed to rename" error and you had to rd /q /s lib\.precomp to get it working again15:30
That was on Windows1015:30
stmuk I saw 2/3 roast failures about 3 months ago on Windows .. also there are some module test failures15:38
nine stmuk: MoarVM does support DESTDIR15:39
stmuk nine: are the exact steps doc'd anyway? there are tickets open suggesting it doesn't work and I had to hack a path to pick up the location of moarvm15:41
it's quite possible I was doing it wrong15:42
s/anyway/anywhere15:42
nine stmuk: I don't know. If nothing gets merged from tbrowder++'s changes except for a document on how to package rakudo, I'd consider that a huge success already...15:42
stmuk I was doing something like https://github.com/rakudo/star/issues/70 only with rakudo itself and not R*15:44
nine Well if the document just stated unabmiguously that R* is _not_ something that should be packaged for distros, that would already be a success...15:45
stmuk a list of current packaging attempts would also be useful15:50
. o O ( and a grade marking )15:50
lizmat m: use nqp; dd nqp::elems(nqp::list) < nqp::elems(nqp::list(42)) || 1 < 42 # expected Bool::True16:42
camelia rakudo-moar cb4308: OUTPUT: «1␤»16:42
lizmat m: use nqp; dd nqp::elems(nqp::list) < nqp::elems(nqp::list(42)) # without the || it is ok16:42
camelia rakudo-moar cb4308: OUTPUT: «Bool::True␤»16:42
Zoffix perhaps this bug? https://github.com/perl6/nqp/issues/33216:44
s: &infix:«<», \((my int $), (my int $))16:46
SourceBaby Zoffix, Sauce is at https://github.com/rakudo/rakudo/blob/cb43086/src/core/Int.pm#L35416:46
Zoffix m: dd nqp::p6bool(nqp::islt_i((my int $ = 0), (my int $ = 1)))16:46
camelia rakudo-moar cb4308: OUTPUT: «=== SORRY!=== Error while compiling <tmp>␤Could not find nqp::islt_i, did you forget 'use nqp;' ?␤at <tmp>:1␤------> ::islt_i((my int $ = 0), (my int $ = 1)) ⏏ ) ␤»16:46
Zoffix m: use nqp; dd nqp::p6bool(nqp::islt_i((my int $ = 0), (my int $ = 1)))16:46
camelia rakudo-moar cb4308: OUTPUT: «Bool::True␤»16:46
Zoffix m: use nqp; dd (my int $ = 0) < (my int $ = 1)16:46
camelia rakudo-moar cb4308: OUTPUT: «Bool::True␤»16:46
Zoffix huh16:46
Ah16:47
m: use nqp; dd (0 < 1) || (1 < 42)16:47
camelia rakudo-moar cb4308: OUTPUT: «Bool::True␤»16:47
Zoffix m: use nqp; dd (my int $ = 0) < (my int $ = 1) || (1 < 42)16:47
camelia rakudo-moar cb4308: OUTPUT: «1␤»16:47
Zoffix m: use nqp; dd ((my int $ = 1) < 2) || 316:51
camelia rakudo-moar cb4308: OUTPUT: «1␤»16:51
Zoffix Golfier. Something in optimizer, as turning it off gives True16:51
Looks like it converts QAST::Op(chain &infix:«<») to - QAST::Op(islt_i)16:54
New toaster results available17:01
buggable: toast17:01
buggable Zoffix, Between 2017.06-193-gcb43086 and 2017.06: 7 (0.84%) modules got burnt; 11 (1.32%) got unsucced; 198 (23.80%) out of 832 modules appear unusable. See https://toast.perl6.party/ for details. 17:01
dogbert17 what's the difference between burnt and unsucced?17:04
yoleaux 02:49Z <AlexDaniel> dogbert17: I fully support your work on the ticket queue, sorry I had to reopen two tickets. It's just that besides “rejected” and “resolved” there's also “open - testneeded” status :)17:04
Zoffix "Status: 7 (0.84%) of modules got burnt (switch from 'Succ' to 'Fail') and 11 (1.32%) of modules got unsucced (switch from 'Succ' to 'Kill', 'Fail', or unknown status). "17:06
dogbert17 aha17:06
Zoffix Actually a bunch of past SUCCs appear to be bogus. Like MPD::Client succing in the past, but if you look at the log zef just stopped (I think it was that Proc::Async segv bug)_17:08
dogbert17 so the toaster isn't using the latest version then?17:09
Zoffix Today I only toasted 2017.06-193-gcb43086. The other two were from last time17:10
the last 2 were from 5 days ago17:11
Yeah, don't see any burns that'd indicate something obviously wrong with rakudo17:12
lizmat Zoffix++ # toasting17:13
Zoffix Flappers in couple of modules, false SUCCs, and changes to modules made in last 5 days17:13
Gonna run one more toasting on Friday that'll retoast 2017.0617:14
dalek rakudo/nom: 4b8bc61 | lizmat++ | src/core/set_proper_subset.pm:17:18
rakudo/nom: Fix <a b>.Bag (<) <a b b>.Bag, spotted by perlawhirl++17:18
rakudo/nom:17:18
rakudo/nom: Previously, it was a bit of a mix of Setty and Baggy semantics:17:18
rakudo/nom: having the keys be identical is ok, as long as at least on of the17:18
rakudo/nom: keys on the left has a lower number of occurrences.17:18
rakudo/nom: review: https://github.com/rakudo/rakudo/commit/4b8bc617d317:18
ugexex 5 days ago I switched all the quiet procs from :out :err to :!out :!err so I would expect the segfaults to have decreased17:18
lizmat dinner&17:18
Zoffix Yeah, they probably did and FAIL modules now complete to the end and end up as FAIL instead of segfaulting which apparently confuses the toaster as to treat it as a SUCC17:21
dogbert17 Zoffix: one of your own modules is on the 'burnt' list17:27
Zoffix Yeah. Two actually, both are floppers17:28
dogbert17 IO::CatHandle::AutoLines17:28
Zoffix Oh you mean totally failing17:29
dogbert17 yeah, used the wrong term17:29
Zoffix Looks like it's the fixed Proc.exitcode used by Testo is-run is now showing up on bogus tests17:30
Zoffix will fix it some other time17:30
dogbert17 cool17:35
Zoffix sees where that optimizer bug is at17:45
Zoffix Looks like today's the day. Finally break through the optimizer barrier and fix my first optimizer bug.17:45
dogbert17 is it in C, Perl or NQP?17:47
Zoffix NQP17:52
dogbert17 impressive18:00
Zoffix buggable: geth18:00
huggable: geth18:00
huggable Zoffix, Set geth report URL (select application/json and send me everything): http://geth.niner.name:8888/?chan=#perl6-dev18:00
dogbert17 huggable: buggable18:00
huggable dogbert17, nothing found18:00
Zoffix Can someone switch rakudo's web hook to http://geth.niner.name:8888/?chan=#perl6-dev and select "application/json" format and select "send me everything"?18:00
ZOFVM: Files=1259, Tests=141156, 118 wallclock secs (20.47 usr 3.09 sys + 2290.23 cusr 181.77 csys = 2495.56 CPU)18:05
lizmat hopes nine or moritz or timotimo see Zoffix' web hook request18:10
Zoffix .in 20d Zoffix testing: https://brson.github.io/2017/07/10/how-rust-is-tested18:12
yoleaux Zoffix: I'll remind you on 30 Jul 2017 18:12Z18:12
timotimo i dont have perms for rakudo18:12
Zoffix ""All patches built in 58 configurations before landing. All patches tested in 20 configurations before landing"" I can't even think of that many configurations :/18:12
moritz Zoffix: done18:13
Zoffix moritz++ thanks18:13
Is +@ for QAST::Op for nqp::if(1, 2, 3) 2 or 3?18:14
Like does it have two children oor 3?18:14
SourceBaby: help18:14
SourceBaby Zoffix, Use s: trigger with args to give to sourcery sub. e.g. s: Int, 'base'. See http://modules.perl6.org/dist/CoreHackers::Sourcery18:14
moritz three I think18:15
Zoffix nqp: use QAST; say(+@(QAST::Op.new(:op<if>, QAST::SVal(:value<x>), QAST::SVal(:value<x>))))18:16
camelia nqp-moarvm: OUTPUT: «0␤»18:16
lizmat just committed https://github.com/perl6/roast/commit/8b700ba9527932ad6d2a18:17
lizmat hmmm.. looks like we have a Geth_ but no Geth ?18:17
or dalek ?18:17
Zoffix We have Geth, but the hook got "<h1>Service Unavailable</h1>18:17
<p>The server is temporarily unable to service your18:17
request due to maintenance downtime or capacity18:18
problems. Please try again later.18:18
Geth ¦ roast: 1fda005b07 | (Elizabeth Mattijsen)++ | S03-operators/set_proper_subset.t18:18
¦ roast: Add test for https://irclog.perlgeek.de/perl6/2017-07-10#i_1485111618:18
¦ roast: review: https://github.com/perl6/roast/commit/1fda005b0718:18
lizmat ahhh :-)18:18
Zoffix I had to click re-deliver18:18
Zoffix shrugs18:18
lizmat ah, but that was my previous roast commit18:18
Zoffix oh18:18
Don't see the new commit in the hook messages at all (on GitHub)18:20
moritz: any idea why this gives 0?18:20
nqp: use QAST; say(+@(QAST::Op.new(:op<if>, QAST::SVal(:value<x>), QAST::SVal(:value<x>))))18:20
camelia nqp-moarvm: OUTPUT: «0␤»18:20
lizmat weird18:21
Zoffix nqp: use QAST; my $op := QAST::Op.new(:op<if>, QAST::IVal.new( :value(2) ), QAST::IVal.new( :value(2) )); say(+@($op))18:21
camelia nqp-moarvm: OUTPUT: «2␤»18:21
Zoffix Oh. OK then18:21
So it's 2 not 3 :)18:21
moritz 3 if you want an else branch, no?18:23
Zoffix moritz: oh, right doh18:23
:D18:23
timotimo probably because you put NQPMu in there twice?18:30
before you put .new in there18:30
also, yeah, you only put 2 things in there. :op<if> is, of course, just a positional18:31
Geth ¦ rakudo/nom: 45aef27497 | (Elizabeth Mattijsen)++ | src/core/Rakudo/QuantHash.pm18:31
¦ rakudo/nom: Introduce R:Q.MIX-IS-PROPER-SUBSET 18:31
¦ rakudo/nom:18:31
¦ rakudo/nom: Abstracted from set_proper_subset.pm. Still needs fixing wrt to18:31
¦ rakudo/nom: <a b>.Mix (<) <a b b>.Mix returning False rather than True.18:31
¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/45aef2749718:31
¦ rakudo/nom: 008be99eb3 | (Elizabeth Mattijsen)++ | src/core/set_proper_subset.pm18:31
¦ rakudo/nom: Use R:Q.MIX-IS-PROPER-SUBSET where possible 18:31
¦ rakudo/nom:18:31
¦ rakudo/nom: This makes Mix (<) Bag and Bag (<) Mix not have to coerce before18:31
¦ rakudo/nom: checking.18:31
¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/008be99eb318:31
Zoffix timotimo: ah. Thanks18:31
timotimo er18:31
did i say positional?18:31
i meant named, of course18:31
Zoffix .ask samcv can all the TODOs in t/spec/S15-nfg/emoji-test.t now be unfudged?18:34
yoleaux Zoffix: I'll pass your message to samcv.18:34
Geth ¦ rakudo/nom: 83e157012d | (Zoffix Znet)++ | src/Perl6/Optimizer.nqp18:35
¦ rakudo/nom: Fix unwanted de-Boolification of 1 arg to nqp::if/unless 18:35
¦ rakudo/nom:18:35
¦ rakudo/nom: Only perform the optimization for if/unless that have the `else`18:35
¦ rakudo/nom: clause that would be returned. Otherwise, we can end up returning18:35
¦ rakudo/nom: the de-boolified Int instead.18:35
¦ rakudo/nom:18:35
¦ rakudo/nom: Bug find: https://irclog.perlgeek.de/perl6-dev/2017-07-10#i_1485240718:35
¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/83e157012d18:35
Zoffix lizmat: ^ that fixes the dd nqp::elems(nqp::list) < nqp::elems(nqp::list(42)) || 1 < 42 thing18:36
Achievement Unlocked: Fix an Optimizer bug18:36
Zoffix collects all the loot18:36
lizmat cool! :-) Zoffix++18:36
Zoffix Pushed the test for it too: https://github.com/perl6/roast/commit/b1f116fdd5181274cda1a520194488da82c10a7e18:39
And I think now I fixed roast hook. It was showing up as inactive, so I deleted it and put it back in and now it's showing up as active18:39
lizmat :-)18:39
refrains from ++ing Zoffix, but he deserves it nonetheless :-)18:40
jdv79 wait. we blew out another coe dev?18:42
*core18:42
ouch18:42
lizmat .oO( coed ev? )18:42
Zoffix jdv79: no, we didn't18:42
[Coke] cod veer18:43
jdv79 oh. thats a relief.18:43
lizmat indeed, it is :-)18:43
jdv79 cool18:44
Zoffix Though I don't think I'll be posting any more blog posts that involve talking to several dozens of people...18:46
Just technical stuff. Seqs part 3; then CatHandle.18:47
jdv79 i look forward to them.19:10
lizmat afk for a bit&19:24
tbrowder nine: ref rakudo, i have some success now but won't submit PR without more testing. my solution for the executable with their hardwired lib paths, etc., is to use shell wrappers scripts. that works so far for the perl6, nqp, and moar binaries showing their versions, and the perl6 executable commanded alone goes into its REPL mode and can at least do a 'say' without barfing. I'm looking at copying the test suite20:07
into the --prefix directory but am not sure if that is worth it--opinions appreciated. can you suggest a specific "something" to run to test the "precomp stuff" with RAKUDO_DEBUG_MODULE=1?20:07
lizmat and another Perl 6 Weekly hits the Net: https://p6weekly.wordpress.com/2017/07/10/2017-28-rakudo-is-hot/21:28
lizmat will fix the set_proper_subset fails in roast after some shuteye21:43
lizmat good night, #perl6-dev!21:43
Zoffix lizmat++ good weekly21:45
and good night21:45
AlexDaniel heh, well, the idea of changing the name to “P6” isn't that bad. It already says exactly that on the logo21:50
Zoffix The proposal is to tweak the name tho. "SomeWord Perl 6"21:53
If you gonna go with P6, may as well name it "Programatron 5000"21:53
AlexDaniel well, I think it counts as a tweak also21:53
dogbert17 bisect: printf("%3.3d", 16)22:03
bisectable6 dogbert17, Bisecting by output (old=2015.12 new=83e1570) because on both starting points the exit code is 022:03
dogbert17, bisect log: https://gist.github.com/a3b63d9573e8fc34b8be381ec7d3e23d22:04
dogbert17, (2017-01-21) https://github.com/rakudo/rakudo/commit/a1c7d01502249bd8bdae5a70ec85847609ca1aba22:04
AlexDaniel c: a1c7d0150^,a1c7d0150 printf("%3.3d", 16)22:04
committable6 AlexDaniel, ¦a1c7d0150^: « 16» ¦a1c7d01: «016»22:04
dogbert17 i.e. RT #12692322:05
synopsebot6 Link: https://rt.perl.org/rt3/Public/Bug/Display.html?id=12692322:05
AlexDaniel dogbert17: it has some nqp tests: https://github.com/perl6/nqp/commit/ed1b835a7a22:09
I wonder if we should have something for that in roast also22:09
Zoffix fwiw gabor's kill issue on perl6-users is something that I can consistently repro on my toast runs22:09
dogbert17 there's stuff in t/spec/S32-str/sprintf.t as well22:09
Zoffix doesn't have a perl6-users subscription22:09
dogbert17 kill issue?22:10
AlexDaniel ah right22:11
Zoffix dogbert17: http://www.nntp.perl.org/group/perl.perl6.users/2017/07/msg3973.html22:19
AlexDaniel I'm not sure if .kill(9) thingy is expected to work at all22:44
it does not kill the process tree, so may this be the reason why something is not killed?22:45
does it start working with just .kill() (which sends sighup)?22:45
like, I've had this in whateverable for ages: https://github.com/perl6/whateverable/blob/master/Whateverable.pm6#L11622:46
iirc for the same reason22:46
samcv Zoffix, yep all emoji can be untodo'd :D the whole file now passes :D23:14

Logs Search ←Prev date Next date→ Channels Documentation