Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6-dev, 2016-12-28

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

All times shown according to UTC.

Time Nick Message
00:20 facetious joined #perl6-dev
00:43 bartolin joined #perl6-dev
00:59 notviki m: dd [0, False].sort
00:59 camelia rakudo-moar ca919d: OUTPUT«(0, Bool::False)␤»
00:59 notviki \o/
01:34 notviki I'm getting a ton of spectest failures :/
01:34 notviki https://gist.github.com/zoffixznet/29301168d1fbbfeb2f003d3c7f515802
01:35 notviki Unrecognized character name LINE FEED (LF)
01:35 notviki at /home/zoffix/CPANPRC/Standard-Rakudo/t/spec/S02-literals/char-by-name.rakudo.moar:16
01:36 notviki ------> is("\c[LINE FEED (LF)⏏]", "\c10", '\c[LINE FEED (LF)
01:38 notviki ------> ok("\c[NULL⏏]" ~~ m/^<:InBasicLatin>$/, q{Match <:In
01:38 notviki Dubious, test returned 1 (wstat 256, 0x100)
01:39 notviki ------> ok("\c[FORM FEED (FF)⏏]" ~~ m/^<:bc<WS>>$/, q{Match (Whitespac
01:39 notviki Dubious, test returned 1 (wstat 256, 0x100)
01:40 notviki This can be avoided by running spectest before pushing commits to nom
01:40 samcv notviki, can you try this branch https://github.com/samcv/MoarVM/tree/working-merged
01:40 samcv err
01:40 samcv wait
01:40 samcv not that branch i think
01:40 samcv here https://github.com/samcv/MoarVM/commits/fix_control_names_temp
01:41 samcv should hopefully fail less spectests, though it will still fail the LINE FEED (LF) ones
01:46 samcv i'm going to fudge the tests now, but if you can checkout that moar branch
01:46 samcv and let me know if anything seems amiss
01:47 notviki Why are we fudging tests instead of reverting the commit that broke all of them?
01:47 samcv err
01:48 samcv yeah yeah we can revert it. that is fine with me
01:50 samcv though it remains unclear if we want to be able to specify characters with unicode 1 names and not the official aliases, which we have tests for already like \c[LF], and \c[LINE FEED (LF] is awkward anyway. but i think reverting would be good until that is sorted out
01:50 samcv i will table the names problem until after we work out the property issues
01:51 samcv # TODO: name aliases (see http://www.unicode.org/reports/tr18/#Name_Properties)
01:51 samcv # for U+0009 the implementation could accept the official name CHARACTER TABULATION, and also the aliases HORIZONTAL TABULATION, HT, and TAB
01:51 samcv hmm says this at the bottom, so that seemed to be intended. don't know if we intended to test for unicode 1 names on purpose though.
01:52 samcv just has been that way in moarvm forever
01:53 samcv idk maybe it would be easier to just manually edit the file afterward for breaking things (the unicode database) :\
01:55 samcv wow the size of the data structure went down hugely once i got rid of dupes
01:55 samcv which I _finally_ suceeding in doing
01:59 notviki Looks like many of those fail in 6.c-errata too: https://gist.github.com/zoffixznet/29f443ead147b43382756278a9892b79
02:00 notviki And t/spec/S11-modules/nested.t fails consistently with "No appropriate parametric role variant available for 'RoleB'"
02:00 * notviki wonders what's that about
02:01 notviki :S and only on 6.c-errata :/
02:01 samcv hmm
02:01 samcv yeah all the \c[ ] ones that have unicode 1 names specified fail
02:02 samcv and also all the ones that are unicode alias names
02:03 notviki hmmm
02:03 samcv but my patch i linked to, it only fails the unicode 1 names
02:03 samcv and passes all the other ones
02:04 samcv (there is only 1 test in 6.c-errata testing unicode 1 name)
02:04 samcv we have like a billion in master roast for some reason
02:04 notviki .tell nine 6.c-errata version of t/spec/S11-modules/nested.t now fails on HEAD. with "No appropriate parametric role variant available for RoleB". What do we do?
02:04 yoleaux2 notviki: I'll pass your message to nine.
02:05 samcv there's something wrong with QAST Regex tho
02:05 samcv argh
02:05 NeuralAnomaly joined #perl6-dev
02:05 notviki NeuralAnomaly: status
02:05 NeuralAnomaly notviki, [✘] Next release will be in 3 weeks and 3 days. Since last release, there are 24 new still-open tickets (23 unreviewed and 0 blockers) and 116 unreviewed commits. See http://perl6.fail/release/stats for details
02:26 samcv nqp: say('test')
02:26 camelia nqp-moarvm: OUTPUT«test␤»
02:26 samcv nqp: say('so') if 'a' ~~ /<:Latin>/
02:26 camelia nqp-moarvm: OUTPUT«so␤»
02:32 NeuralAnomaly joined #perl6-dev
02:48 ilbot3 joined #perl6-dev
02:48 Topic for #perl6-dev is now Perl 6 language and compiler development 2.0 | Logs at http://irclog.perlgeek.de/perl6-dev/today
03:31 samcv my constant π := pi; #> there must be somethnig special about how we look this up
03:31 samcv the pi symbol
03:32 samcv or maybe i just really borked something :\
03:32 geekosaur s: pi
03:32 SourceBaby geekosaur, Something's wrong: ␤ERR: Cannot resolve caller sourcery(Num); none of these signatures match:␤    ($thing, Str:D $method, Capture $c)␤    ($thing, Str:D $method)␤    (&code)␤    (&code, Capture $c)␤  in block <unit> at -e line 6␤␤
03:32 samcv with my moarvm that is
03:32 samcv not sure why that would be broken, nqp compiles fine.
03:32 * geekosaur may figure out how to use that thing eventually...
03:33 samcv rror while compiling, type X::Syntax::Missing
03:33 notviki samcv: special how?
03:33 samcv idk
03:33 samcv different than non special unicode things
03:34 samcv because it's failing at parsing at the unicode constants like e and tau and pi
03:34 notviki hm
03:34 samcv commented those three lines out and it is fine
03:34 notviki I know e was failing on JVM before
03:34 samcv geekosaur, are you any good at QAST Regex?
03:34 geekosaur nope
03:34 samcv ahh
03:35 samcv jnthn, explained some things, but i really am not sure what happens in there, as far as how it resolves property names to actual values
03:35 samcv the ops are confusing
03:36 samcv but there is a bug in there.
03:36 samcv i fixed MoarVM i *think* i will run a spectest and see. if i only fail regex unicode matches
03:37 samcv and all the properties now have the SAME property code
03:37 samcv \o/
03:37 samcv which now makes bugs show up that relied on them being incorrect before ;P
03:38 samcv wow this is really nice. i've never gotten all of them having the correct numbers… in the database
03:44 samcv oh it must rely on unimatch
03:44 notviki .ask lizmat I see comment "canary until 2014". Does it mean it can be removed now? https://github.com/rakudo/rakudo/commit/66cd275b018722057b58da796a9e314b3492155e
03:44 yoleaux2 notviki: I'll pass your message to lizmat.
04:00 notviki m: <fix-combinations-and-regret-staying-late-in-the-morning  go-to-sleep>.roll.say
04:00 camelia rakudo-moar ca919d: OUTPUT«go-to-sleep␤»
04:00 notviki You're the boss!
04:17 notviki .ask lizmat for RT#130418 (chdir), the commit says "so we don't lose failures", so maybe if we `return` them instead of throwing that'd work out? https://github.com/rakudo/rakudo/commit/7dec9a9540a8a4b265e70a67d741f70c83492b12
04:17 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=130418
04:17 yoleaux2 notviki: I'll pass your message to lizmat.
04:55 dalek roast: 08aa3ea | samcv++ | S05-mass/properties-script.t:
04:55 dalek roast: Fix properties-script.t to use FF instead of FORM FEED (FF)
04:55 dalek roast:
04:55 dalek roast: Makes it work on the latest Rakudo, and this file is testing mostly
04:55 dalek roast: regex script matches, so it is fine to use one of the other notations.
04:55 dalek roast: We test \c[FORM FEED (FF)] and others in S05-mass/named-chars.t
04:55 dalek roast: review: https://github.com/perl6/roast/commit/08aa3ea371
04:56 samcv FF works on jvm and moar current and before the most recent moar bump so this should be fine, we already test \c[stuff] in another script, but the 3 occurrences in this file stopped it from compiling
04:57 samcv m: "💩".unimatch("Miscellaneous Symbols and Pictographs")
04:57 camelia rakudo-moar ca919d: ( no output )
04:57 samcv m: "💩".unimatch("Miscellaneous Symbols and Pictographs").say
04:57 camelia rakudo-moar ca919d: OUTPUT«False␤»
04:57 samcv aha. hmm maybe i can get this working while i'm at it
05:00 samcv bisectable6, use Test, is ༳, -0.5, "Tibetan number literal produces a negative";
05:00 bisectable6 samcv, On both starting points (old=2015.12 new=ca919d9) the exit code is 1 and the output is identical as well
05:00 bisectable6 samcv, Output on both points: ===SORRY!=== Error while compiling /tmp/C1V7HoLu_6␤Confused␤at /tmp/C1V7HoLu_6:1␤------> use Test⏏, is ༳, -0.5, "Tibetan number literal pr
05:00 samcv committable6, is ༳, -0.5, "Tibetan number literal produces a negative";
05:00 committable6 samcv, ¦«is»: Cannot find this revision
05:00 samcv committable6, all is ༳, -0.5, "Tibetan number literal produces a negative";
05:00 committable6 samcv, https://gist.github.com/a7ec249f6a8afa7485457797632060d2
05:01 samcv committable6, all use Test; is ༳, -0.5, "Tibetan number literal produces a negative";
05:01 committable6 samcv, https://gist.github.com/4f623327c4d7755bd9747646cf9a5ccf
05:03 samcv ok unival is still broke for me. not sure why
05:12 samcv missing the denominator and numerator for some reason. that would do it
05:18 pyrimidi_ joined #perl6-dev
05:26 geekosaur joined #perl6-dev
06:00 dalek rakudo/nom: 3baffe7 | samcv++ | / (2 files):
06:00 dalek rakudo/nom: uniprop: return Bool's for Emoji properties to prepare for MoarVM adding
06:00 dalek rakudo/nom:
06:00 dalek rakudo/nom: With this pull: https://github.com/MoarVM/MoarVM/pull/468
06:00 dalek rakudo/nom: We will have all Emoji properties except Emoji_All. With that change to
06:00 dalek rakudo/nom: MoarVM and this change to Rakudo we now pass the roast Emoji property
06:00 dalek rakudo/nom: tests (aside from Emoji_All).
06:00 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/3baffe71a8
06:16 pyrimidine joined #perl6-dev
06:33 lizmat .
06:33 yoleaux2 03:44Z <notviki> lizmat: I see comment "canary until 2014". Does it mean it can be removed now? https://github.com/rakudo/rakudo/commit/66cd275b018722057b58da796a9e314b3492155e
06:33 yoleaux2 04:17Z <notviki> lizmat: for RT#130418 (chdir), the commit says "so we don't lose failures", so maybe if we `return` them instead of throwing that'd work out? https://github.com/rakudo/rakudo/commit/7dec9a9540a8a4b265e70a67d741f70c83492b12
06:33 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=130418
06:33 lizmat .tell notviki the canary can go
06:33 yoleaux2 lizmat: I'll pass your message to notviki.
06:34 lizmat .tell notviki re #130418, judging from 7dec9a9540a8a4b2 I think the throw was added because it is very dangerous to "miss" a chdir() failing
06:34 yoleaux2 lizmat: I'll pass your message to notviki.
06:34 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=130418
06:35 lizmat .tell notviki so I'm leaning towards changing the documentation / tests rather than the code
06:35 yoleaux2 lizmat: I'll pass your message to notviki.
06:35 lizmat breakfast&
06:36 samcv morning lizmat
06:41 * geekosaur had that thought re chdir as well
06:46 travis-ci joined #perl6-dev
06:46 travis-ci Rakudo build passed. Samantha McVey 'uniprop: return Bool's for Emoji properties to prepare for MoarVM adding
06:46 travis-ci https://travis-ci.org/rakudo/rakudo/builds/187137264 https://github.com/rakudo/rakudo/compare/ca919d96c97e...3baffe71a8e1
06:46 travis-ci left #perl6-dev
07:14 samcv m: '1'.uniprop('Numeric_Value_Denominator').say
07:14 camelia rakudo-moar 3baffe: OUTPUT«1␤»
07:14 samcv m: Rakudo::Internals.PROPCODE('Numeric_Value_Denominator').say
07:14 camelia rakudo-moar 3baffe: OUTPUT«18␤»
07:15 samcv weird. that's the same value i have atm on the branch i'm working on. coincidence prolly
07:15 samcv m: Rakudo::Internals.PROPCODE('Numeric_Value_Denominator').say
07:15 camelia rakudo-moar 3baffe: OUTPUT«18␤»
07:15 samcv m: Rakudo::Internals.PROPCODE('Numeric_Value_Numerator').say
07:15 camelia rakudo-moar 3baffe: OUTPUT«11␤»
07:15 samcv bisectable6, Rakudo::Internals.PROPCODE('Numeric_Value_Numerator').say
07:15 bisectable6 samcv, Bisecting by output (old=2015.12 new=3baffe7) because on both starting points the exit code is 0
07:16 bisectable6 samcv, bisect log: https://gist.github.com/2bb2552cbc3d3b817c1f6523fb608b07
07:16 bisectable6 samcv, (2016-09-28) https://github.com/rakudo/rakudo/commit/2cad3d214f6f151df537a535760e341838332c21
07:20 pyrimidi_ joined #perl6-dev
07:26 pyrimidine joined #perl6-dev
07:26 samcv anybody know how unival works? this is the one unicode property i can't seem to get to work :\
07:27 samcv oh wait forgot to say
07:28 samcv unicode.c says at the top This file is generated by ucd2c.pl from the Unicode database.
07:28 samcv but ucd2c.pl has no references to this file
07:28 samcv and it has never changed when i have rebuild the unicode database
07:29 samcv .ask jnthn This file is generated by ucd2c.pl from the Unicode database says unicode.c, but I have never had the file change when running ucd2c.pl before
07:29 yoleaux2 samcv: I'll pass your message to jnthn.
07:29 samcv .ask jnthn unival is the one property I can't get to work after working hard on the script file. the other ones seem alright
07:29 yoleaux2 samcv: I'll pass your message to jnthn.
07:35 samcv oh interestingly '1'.uniprop('Numeric_Value_Denominator') on my test build gives me a 1
07:35 samcv but .unival just returns 0
07:36 samcv and Numeric_Value_Numerator returns 1 too
07:38 samcv m: say Rakudo::Internals.PROPCODE("Numeric_Value_Denominator")
07:38 camelia rakudo-moar 3baffe: OUTPUT«18␤»
07:38 samcv cool :)
07:39 samcv I think i can fix this !!!!
07:40 samcv \\o//
07:41 samcv WOOO! fixed!
07:41 * samcv does a dance
07:52 dalek rakudo/nom: dbbf9dd | samcv++ | src/ (2 files):
07:52 dalek rakudo/nom: Use full Unicode name for Numeric_Value_* for Cool.pm and Actions.nqp
07:52 dalek rakudo/nom:
07:52 dalek rakudo/nom: Also switch from nqp::unipropcode in Cool.pm to
07:52 dalek rakudo/nom: Rakudo::Internals.PROPCODE which does some caching on the Rakudo side
07:52 dalek rakudo/nom: which should probably have a bit of speed improvement for lookups of the
07:52 dalek rakudo/nom: same character.
07:52 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/dbbf9dd169
08:02 lizmat joined #perl6-dev
08:06 moritz good morning
08:07 lizmat moritz o/
08:12 moritz samcv: welcoem to the Rakudo core dev team. It's always great to see new, enthusiastic contributors!
08:12 samcv :)
08:12 samcv happy to be here!
08:14 tadzik yay samcv! \o/
08:14 samcv \o/
08:14 samcv i'm rich!
08:15 samcv perl 6 is the most exciting programming language I know and i'm glad to be enriched by being able to contribute
08:15 tadzik you enrich us all, everyone wins :)
08:16 [Tux] This is Rakudo version 2016.12-119-gdbbf9dd16 built on MoarVM version 2016.12-21-g3dc56470
08:16 [Tux] csv-ip5xs        3.070
08:16 [Tux] test            13.721
08:16 [Tux] test-t           5.196
08:16 [Tux] csv-parser      13.984
08:18 lizmat whee!
08:18 lizmat samcv++  :-)
08:48 RabidGravy joined #perl6-dev
08:50 dalek rakudo/nom: 434bf75 | lizmat++ | src/core/ (2 files):
08:50 dalek rakudo/nom: Make .sort always return a .Seq
08:50 dalek rakudo/nom:
08:50 dalek rakudo/nom: This is not really about performance but really about (possible)
08:50 dalek rakudo/nom: correctness.  This makes a few spectests fail that test for the return
08:50 dalek rakudo/nom: type of .sort being a List (which one could consider to be faulty).
08:50 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/434bf75b9d
08:53 samcv m: '💩'.uniprop('Block').say
08:53 camelia rakudo-moar dbbf9d: OUTPUT«Miscellaneous Symbols and Pictographs␤»
08:53 samcv hm
08:53 samcv this test looks bad
08:54 samcv unless we're actually testing it with spaces and not underscores which is how roast says in the print
08:54 samcv hm
08:56 samcv no it's good
08:57 samcv not too bad if that's the only one failing in this test
08:57 samcv i was failing like down from 100 to only 3 (testing Miscellaneous_Symbols_And_Pictographs with different _ or capitalization)
08:59 dalek roast: 9b1a7cb | moritz++ | S32-list/grep.t:
08:59 dalek roast: RT #130365: test last in grep on infinite list
08:59 dalek roast: review: https://github.com/perl6/roast/commit/9b1a7cb528
08:59 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=130365
09:14 lizmat m: multi a(Callable:D &a, *@v) { "one" }; multi a(+v) { "two" }; say a { $^a cmp $^b }, 1,2,3  # MMD dispatch error ?
09:14 camelia rakudo-moar 434bf7: OUTPUT«two␤»
09:15 lizmat m: say { $^a cmp $^b } ~~ Callable:D
09:15 camelia rakudo-moar 434bf7: OUTPUT«True␤»
09:27 dalek roast: ab31dc4 | samcv++ | S05-mass/properties-general.t:
09:27 dalek roast: Add test for matching ' ' ~~ /<:White_Space>/
09:27 dalek roast:
09:27 dalek roast: Remove some extra whitespace in the file as well.
09:27 dalek roast: review: https://github.com/perl6/roast/commit/ab31dc41dd
09:37 pyrimidi_ joined #perl6-dev
09:58 ThorstenB joined #perl6-dev
09:59 ThorstenB Hi. I'm up too early and have a question about WSL. I'm not sure who to address this to, so figured this channel was as good a start as any.
10:00 ThorstenB WSL is microsoft's "Bash on Windows" effort, it runs Ubuntu natively on top of Windows. Useful for dev work on that platform, particularly if the target is a Linux server.
10:01 ThorstenB Rakudo * will build there, with one caveat: the build fails first, then you can disable execstack on libmoar.so, try again, then it works. That's using rakudobrew.
10:02 samcv but then it succeeds after that?
10:02 samcv moar compiles fine, nqp compiles fine, but rakudo fails?
10:02 ThorstenB This is because executable stack is a feature not implemented in WSL. I asked the MS devs when this feature would come, waiting on answer
10:02 samcv oh you mean all of rakudo? which part does it fail on
10:02 samcv ah
10:02 ThorstenB The build moar fails , then succeeds
10:03 ThorstenB I have a link to reproduce this, on iPhone might take me a moment. Let me ask
10:03 ThorstenB My question though:
10:03 ThorstenB Is it reasonable for the rakudo build environment to detect WSL and compile with -noexecstack in order to build more easily there?
10:04 ThorstenB Or do you prefer to just wait until MS has implemented execstack support ?
10:05 ThorstenB That promised link. The rakudo * / moarvm stuff is at the very end, the thread is about execstack in WSL in general: https://github.com/Microsoft/BashOnWindows/issues/286
10:06 samcv that would be a question for jnthn i think
10:10 ThorstenB Do they have a time they're usually on so I can ask? I have code to detect WSL from a perl 5 effort (which was not needed in the end, MS fixed a bug and perl 5 started working). I'd need pointing at where in the build environment such detection would
10:11 ThorstenB Live, but be happy to take a stab at it
10:13 samcv can you show me the code?
10:14 samcv also was curious how the windows experience was, without WSL for rakudo *?
10:16 ThorstenB Code: there's a patch attached to this: https://rt.perl.org/Public/Bug/Display.html?id=130258
10:18 ThorstenB There's also a code snippet in this github thread (yorickdowne is myself): https://github.com/Microsoft/BashOnWindows/issues/186
10:18 ThorstenB In a nutshell, check /proc/version to detect WSL
10:18 samcv "In short, Perl modules cannot be installed because, for some reason, the files aren't really there even though I can browse, view, edit, delete, etc. those files." :P
10:18 samcv a unix style file then?
10:19 ThorstenB Yeah that's fixed now. File::Find lost its mind because MS would return link count 2 for a directory, always. They changed that to 0 and now File::Find knows not to rely on that count
10:20 ThorstenB Yes, WSL runs Ubuntu, with a kernel
10:20 ThorstenB Shim to intercept system calls
10:20 ThorstenB It's a clever piece of engineering
10:21 samcv yeah
10:21 ThorstenB You asked about the rakudo * experience on Windows without WSL. I'm assuming you mean the windows binary provided ? If so, I shamefully admit I haven't even tried that
10:21 samcv ah
10:21 ThorstenB I don't target windows server, so while my laptop runs windows, ultimately code would run on Linux
10:22 samcv ah so you want the closest environment then?
10:22 ThorstenB WSL is just a convenient way not to have to launch an Ubuntu vm
10:22 samcv to deployment
10:22 samcv yeah
10:22 ThorstenB Yeah
10:22 samcv i don't see execstack mentioned on the RT or on the github post?
10:24 samcv ThorstenB, so moarvm works fine? but building rakudo or nqp fails?
10:24 samcv need to know which one
10:24 ThorstenB We're talking about different issues. Let me try to untangle
10:24 samcv oh
10:24 samcv there's the file issue and the other one is building
10:24 * samcv is more concerned about the building issue atm
10:24 ThorstenB The last two links were about code to detect WSL
10:25 ThorstenB Because you asked for it :)
10:25 samcv yes. ok but
10:25 ThorstenB Nothing more. That file issue is fixed
10:25 samcv so we know We are on WSL
10:25 ThorstenB Exactly
10:25 samcv and so we need to change how Rakudo builds? or is it moarvm or nqp that fails
10:25 ThorstenB Let me find the link again that talks about building moarvm and execstack
10:25 samcv build process goes. MoarVM -> nqp -> rakudo
10:25 samcv kk
10:26 ThorstenB https://github.com/Microsoft/BashOnWindows/issues/286
10:26 ThorstenB Very end of that thread, yorickdowne again
10:26 ThorstenB Moarvm build fails
10:26 ThorstenB It's because it tries to use execstack in libmoar.so and yeah, no go on WSL
10:26 ThorstenB So the way to fix that is:
10:27 ThorstenB Rakudobrew build moar
10:27 ThorstenB Failure
10:27 samcv ok so MoarVM fails
10:27 ThorstenB Execstack -c path/to/libmoar.so
10:28 Ven joined #perl6-dev
10:28 samcv so if MoarVM builds fine. then we are in the clear?
10:28 ThorstenB Rakudobrew build moar
10:28 ThorstenB Success
10:28 ThorstenB Yes the rest of the build works after that
10:28 samcv ok but if we have a clean install
10:28 samcv that won't help
10:28 samcv hm
10:28 samcv well unless we do it right after we make the .so maybe hm
10:29 samcv if you can have it fail and provide an error log
10:29 ThorstenB So if rakudobrew compiled libmoar with -noexecstack (and a couple other switches), this would likely work
10:29 samcv and open a ticket https://github.com/MoarVM/MoarVM/issues
10:29 samcv also would be good
10:29 ThorstenB Okay let me hop over to pc for that, no ticket opening while on couch with phone :)
10:29 samcv can you download moar and then
10:30 samcv compile moar standalone without rakudobrew
10:30 samcv then provide output
10:30 ThorstenB Okay let me try that
10:31 ThorstenB To be clear Though, im not suggesting to detect the failure. I'm suggesting to detect WSL, compile without execstack (thats just an option to gcc), so that there is no failure
10:32 samcv yes
10:32 samcv i understand that :)
10:36 ThorstenB This is also emphatically not a bug in moar. It's reasonable to assume execstack is there when you detect an Ubuntu environment. That WSL is a partial implementation of Linux syscalls is an unexpected wrinkle. :)
10:37 samcv yeah i understand
10:37 dalek rakudo/nom: 5023bf3 | samcv++ | src/core/REPL.pm:
10:37 dalek rakudo/nom: REPL: only suggest installing Linenoise/Readline if not on Windows
10:37 dalek rakudo/nom:
10:37 dalek rakudo/nom: Readline and Linenoise packages don't work on Windows. We should not ask
10:37 dalek rakudo/nom: them to install something which is not compatible.
10:37 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/5023bf39f1
10:39 stmuk I'm sure I've seen linenoise work on Windows -- although I think it only works with gcc and not MSVC (?)
10:39 yoleaux2 16 Dec 2016 00:47Z <Zoffix> stmuk: any ideas for https://rt.perl.org/Ticket/Display.html?id=130192 ? I can't reproduce the issue with a built-from-source Rakudo, but it does exists in 2016.04 and 2016.11 Stars (the only two I tried)
10:39 samcv stmuk, ok so should we only ask to install Linenoise and not Readline if not on windows?
10:39 samcv hmm
10:39 samcv i mean
10:40 stmuk I'll check on a windows VM in the next few hours .. hassle me if I don't get back to you shortly
10:40 samcv kk :)
10:41 samcv ugh and https://github.com/drforr/perl6-readline/issues/12
10:41 samcv i had to make a symlink for it to even work
10:42 samcv and used to just enter an endless loop before some change to the REPL code
10:42 samcv (when missing readline 7)
10:42 samcv system updated readline and then perl 6 repl was very non plussed
10:42 samcv and Linenoise doesn't support UTF-8
10:43 samcv so I really want Readline to have a good user experience
10:51 kalkin- joined #perl6-dev
11:06 ThorstenB Samcv issue 470 is open for this
11:10 samcv nice thanks :)
11:10 FROGGS joined #perl6-dev
11:13 notviki
11:13 yoleaux2 06:33Z <lizmat> notviki: the canary can go
11:13 yoleaux2 06:34Z <lizmat> notviki: re #130418, judging from 7dec9a9540a8a4b2 I think the throw was added because it is very dangerous to "miss" a chdir() failing
11:13 yoleaux2 06:35Z <lizmat> notviki: so I'm leaning towards changing the documentation / tests rather than the code
11:13 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=130418
11:13 notviki OK
11:14 notviki gonna do all in ~2 hours
11:43 b2gills1 joined #perl6-dev
11:55 * notviki grumbles seeing even more spectest failures on nom -_-
11:57 notviki And about this: https://github.com/perl6/roast/commit/08aa3ea37128fa37255ef8cc4ec21d58498b81b4
11:58 notviki So now we're telling users that even if the stuff is in the roast, we still might change it if we feel the intent of the test is satisfied?
11:59 notviki feels like thist stuff should wait until 6.d
12:04 notviki results of spectest on 6.c-errata: https://gist.github.com/zoffixznet/fb682e087e02461405aa71e24168f872
12:04 notviki will run stresstest in ~1 hours
12:04 * notviki relocates
12:08 samcv whoever wants to revert that moar pull is free to
12:09 samcv would be a good idea…
12:10 samcv notviki, i didn't remove any of those tests from roast btw. just from that file
12:10 samcv there's another huge file devoted to them
12:11 samcv but I think we should revert moarvm. then once we have unicode 1 name support we can do Unicode names + unicode alias names, then if that fails we could fall back to check the unicode 1 name
12:12 samcv notviki, i will add that to the 6.d proposal file
12:17 notviki I don't have a commit bit for MoarVM
12:18 samcv i know :(
12:18 samcv i asked pmuritas in #perl6 maybe he can do it
12:19 notviki It's just we already have a lot of user affecting changes and "it's not in 6.c" is our excuse. But if we now start making changes that also include 6.c test suite changes....
12:19 samcv yep
12:23 samcv and i really hate how it breaks compiling too :\ that's really not good
12:28 samcv yeah
12:28 samcv oops
12:30 jnthn Did we bump MOAR_REVISION?
12:30 yoleaux2 07:29Z <samcv> jnthn: This file is generated by ucd2c.pl from the Unicode database says unicode.c, but I have never had the file change when running ucd2c.pl before
12:30 yoleaux2 07:29Z <samcv> jnthn: unival is the one property I can't get to work after working hard on the script file. the other ones seem alright
12:31 samcv oh hey
12:31 samcv yeah. can you revert that last pull
12:31 jnthn I guess...but doesn't your next PR lined up fix spectest?
12:32 samcv it's more complicated than that… sadly
12:32 samcv we have tests in 6.c-errata that rely on unicode 1 names
12:32 samcv wel. we have one test
12:32 jnthn Ah, ok
12:32 samcv and a bunch in roast currently. yeah
12:32 samcv that's the only problem with it :\
12:32 jnthn I figured you'd spectested it before submitting the PR ;)
12:33 jnthn Revert done
12:33 samcv ok cool
12:34 samcv i thought it passed but I must have made some mistake, i was working with a lot of branches
12:34 jnthn Ah, OK
12:34 * jnthn tends to merge locally and test PRs from folks he's never heard of, but trusts the regulars to have done the testing. :-)
12:35 samcv my mistake!
12:35 samcv oh but also
12:35 jnthn No worries, we all do them now and then.
12:35 samcv i am pretty close to getting this working
12:35 jnthn \o/
12:35 samcv not the unicode 1 names but
12:35 jnthn Got code-gen fixes in NQP also?
12:35 samcv code-gen?
12:36 samcv no
12:36 samcv i would, but. i don't know how :P
12:36 samcv you know much more what is wrong
12:36 samcv jnthn, now i just need to make sure all the alias names we spectest for work. but other than that
12:36 jnthn Well, my initial idea was to try and case-analyze the <:Foo>
12:36 samcv everything is working and no duplicate property names with different values
12:36 samcv and they are sorted in order
12:37 samcv this seems fine tbh
12:37 jnthn Since we know the set of things that are general category values at compile time
12:37 samcv yeah
12:37 jnthn So we can then spit out code to do a gen-cat check only
12:37 samcv for nqp?
12:37 jnthn And similar for boolean enumerated props
12:37 jnthn Well, in the QAST regex compiler, yes.
12:38 samcv well i'm most concerned about the space bug
12:38 samcv because it stil persists even with nothing wrong in the data structure
12:38 jnthn What I'd really like to do is to get rid of all the value names from the property names table.
12:38 samcv no "Space" and "space" conflicting
12:38 samcv just plain space
12:38 jnthn Hm
12:38 samcv no colliding keys
12:38 samcv White_Space still works though
12:39 jnthn What if it should have been White_Space all along, and we just got lucky that Space worked? :)
12:39 samcv you mean space?
12:39 jnthn Uh, yes
12:39 samcv ha
12:39 samcv welll!
12:39 samcv space has always been broken
12:39 samcv for at least 2 years
12:39 jnthn This is the code behind \s iirc
12:40 samcv space != White_Space
12:40 jnthn So the real question is "what property should \s be using"
12:40 samcv space === Line_Break=space
12:40 samcv White_space
12:40 samcv White_Space*
12:41 samcv it uses space?
12:41 samcv you sure?
12:41 jnthn Well, can check easy enough
12:42 samcv m: "\t".uniprop-bool('space').say
12:42 camelia rakudo-moar 5023bf: OUTPUT«False␤»
12:43 samcv space isn't space!
12:43 samcv terrible
12:43 MasterDuke jnthn: whenever you've got a moment, did you happen to see my question and conversation with JimmyZ here? https://irclog.perlgeek.de/perl6-dev/2016-12-27#i_13806796
12:44 jnthn Oh, I mighta been confused
12:44 jnthn https://github.com/MoarVM/MoarVM/blob/master/src/strings/ops.c#L1659 uses WHITE_SPACE
12:44 jnthn I somehow thought that was the line we were talking about before
12:45 jnthn Wait, I think that *was* the one you changed to _SPACE yesterday?
12:46 jnthn (To work around things)
12:46 jnthn Anyway, it looks like it's correct
12:46 samcv yep
12:46 samcv yeah it is correct there
12:47 jnthn OK, so the incorrect thing is... <:space> ?
12:47 samcv yep
12:47 jnthn OK. I was confused :)
12:47 jnthn MasterDuke: I can't easily guess off-hand, but I don't know a deep reason why that coudln't change. It may just need some code changes inside of Cursor.pm or so in core
12:48 jnthn (src/core that is)
12:48 jnthn Lunch time here; I'll bbiab
12:48 samcv kk
12:58 lucasb joined #perl6-dev
13:16 notviki Results of stresstest on 6.c-errata: https://gist.github.com/zoffixznet/8f99558441b923e0f4e616bcbfac00db
13:26 notviki IO::Path:U.chdir uses $*CWD :S
13:27 notviki seems like a wat
13:32 dalek nqp: 252d760 | (Zoffix Znet)++ | tools/build/MOAR_REVISION:
13:32 dalek nqp: (moar bump) Bring over revert of MoarVM's #469
13:32 dalek nqp:
13:32 dalek nqp: https://github.com/MoarVM/MoarVM/pull/469
13:32 dalek nqp:
13:32 dalek nqp: It was found the work breaks some of the tests. Revert until fix
13:32 dalek nqp: is found.
13:32 dalek nqp:
13:32 dalek nqp: Brings in: https://github.com/MoarVM/MoarVM/compare/3dc56470...3c5253c
13:32 dalek nqp: review: https://github.com/perl6/nqp/commit/252d7609fa
13:34 samcv thanks notviki
13:34 dalek rakudo/nom: d35efb6 | (Zoffix Znet)++ | tools/build/NQP_REVISION:
13:34 dalek rakudo/nom: (nqp bump) Bring over revert of MoarVM's #469
13:34 dalek rakudo/nom:
13:34 dalek rakudo/nom: https://github.com/MoarVM/MoarVM/pull/469
13:34 dalek rakudo/nom:
13:34 dalek rakudo/nom: It was found the work breaks some of the tests. Revert until fix
13:34 dalek rakudo/nom: is found.
13:34 dalek rakudo/nom:
13:34 dalek rakudo/nom: Brings in changes: https://github.com/perl6/nqp/compare/4a950e19...252d760
13:34 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/d35efb602f
13:34 notviki Stresstest results with latest MoarVM (seems still failure in module stuff): https://gist.github.com/zoffixznet/66586582da0c2bdf616b144825a0858a
13:35 notviki that's 6.c-errat
13:35 samcv hm
13:35 samcv oh that's with the revert
13:36 notviki yeah
13:49 notviki ZOFVM: Files=1208, Tests=130855, 153 wallclock secs (20.97 usr  3.25 sys + 2574.02 cusr 255.80 csys = 2854.04 CPU)
13:58 dalek roast: c9cf71d | (Zoffix Znet)++ | S32-list/sort.t:
13:58 dalek roast: Amend tests testing return type of .sort
13:58 dalek roast:
13:58 dalek roast: The tests were added[^1][^2] on July, 30, 2016 and are not part of 6.c.
13:58 dalek roast: The purpose of the tests is to check we're not returning a BOOTArray
13:58 dalek roast: per tickets[^3][^4], by checking the return type is a List. However,
13:58 dalek roast: Rakudo was changed[^5] to always return a Seq from .sort
13:58 dalek roast:
13:58 dalek roast: [1] https://github.com/perl6/roast/commit/9697a3b3453a97deb0987ad37558383b2a0
13:58 dalek roast: [2] https://github.com/perl6/roast/commit/94d5bc99c37896271d391c573be7958b0b6
13:58 dalek roast: [3] https://rt.perl.org/Ticket/Display.html?id=126921#ticket-history
13:58 dalek roast: [4] https://rt.perl.org/Ticket/Display.html?id=128779#ticket-history
13:58 JimmyZ samcv: linenoise works on windows, if there is a msvc
13:59 samcv alright. we can recommend Linenoise on windows and Linenoise or Readline on linux?
14:01 kalkin- joined #perl6-dev
14:06 notviki Seems some todos in t/spec/S15-unicode-information/uniprop.t are now passing https://gist.github.com/zoffixznet/e1316dd5112d6ea2a8eb48296d34e6ad
14:07 samcv notviki,
14:07 samcv sorta
14:07 samcv ok 146 - uniprop for Emoji_Modifier returns False for non modifier Emoji's# TODO Emoji properties NYI in MoarVM
14:08 samcv we can stil unfudge them though
14:09 samcv ugh. i'm considering like. maybe we should do a workaround in nqp for like maybe 2 property names
14:09 samcv because no matter how good everything else works, those ones don't
14:09 samcv well 'alpha' doesn't work but 'Alpha' does
14:09 samcv m: say Rakudo::Internals.PROPCODE('alpha')
14:09 camelia rakudo-moar d35efb: OUTPUT«29␤»
14:09 samcv m: say Rakudo::Internals.PROPCODE('Alpha')
14:09 camelia rakudo-moar d35efb: OUTPUT«29␤»
14:10 samcv argh
14:10 samcv at least consistency is good though
14:10 samcv instead of just rondom things breaking
14:17 notviki .ask lizmat what are your thought on removing this https://github.com/rakudo/rakudo/commit/3962ee51d9fab ? Feels like a footgun: if you end up with a unitialized IO::Path type object (say, from unset named param), the code will silently assume CWD. Since $*CWD has .chdir, doesn't seem like it's worth the rist of also having this behaviour. Thoughts?
14:17 yoleaux2 notviki: I'll pass your message to lizmat.
14:18 notviki s/rist/risk/;
14:18 notviki there are no tests for it
14:26 notviki /o\
14:26 * notviki is slipping into the IO rabbit hole
14:29 notviki A whole bunch of subs and methods have $foo // $foo.throw in them
14:30 notviki and the docs contradict taht
14:31 notviki and also, if we really needed that throwage, kinda makes our whole Failure mechanism questionable if we have to avoid it for IO AND for list returning things :/
14:33 notviki I guess part of the reason is the better IO branch that never got merged?
14:46 samcv jnthn, i'm trying to find the function in moar which resolves the property names to the values it spits back
14:48 samcv is it all in unicode_ops.c?
14:48 samcv MVM_OP_unipropcode
14:48 samcv but trying to see how to trace that to where it goes
14:57 samcv also if i want to add some prints. i don't seem to be getting any output to my terminal
14:59 notviki And we also have a PROCESS::chdir that returns True instead of the $*CWD? :S https://github.com/rakudo/rakudo/blob/nom/src/core/io_operators.pm#L148
15:00 * notviki notes not to visit IO land
15:01 samcv chdir False
15:01 notviki m: say WHAT "IOfsdfdsfdsfds".IO.chdir: "foo"
15:01 camelia rakudo-moar d35efb: OUTPUT«(Failure)␤»
15:01 notviki m: say WHAT chdir "IOfsdfdsfdsfds".IO
15:01 camelia rakudo-moar d35efb: OUTPUT«Failed to change the working directory to '/home/camelia/IOfsdfdsfdsfds': does not exist␤  in block <unit> at <tmp> line 1␤␤Actually thrown at:␤  in any  at gen/moar/Metamodel.nqp line 3072␤  in block <unit> at <tmp> line 1␤␤»
15:01 notviki sub throws but method Fails :S
15:01 samcv notviki, does roast say it has to return True?
15:02 samcv or just something Truey?
15:02 notviki no idea
15:02 * notviki types git reset --hard and gives up on this
15:03 * notviki hopes lizmat gives another go for better-io branch
15:04 samcv link to branch?
15:04 samcv what.
15:04 samcv # mkdir True; chdir True #> /home/samantha/git/rakudo/True.IO
15:04 samcv awesome
15:05 samcv well. that was sarcastic
15:05 notviki no idea. From what I understand lizmat hacked much better IO stuff than we have now but no one paid enough attention to it to merge it or something.
15:05 samcv and i can rmdir True
15:05 samcv sounds like uh. that isn't good
15:06 samcv if something else ends up returning true, you end up trying to cd into a directory named True or False…
15:06 MasterDuke think it is/was https://github.com/rakudo/rakudo/tree/newio
15:07 samcv Cannot resolve caller Str(Bool: ); none of these signatures match:
15:07 samcv (Bool:D $: *%_)
15:07 samcv hmm
15:07 samcv that's if i try to pass just Bool undefined
15:07 samcv why do we have a chdir multi for Bool?????????
15:10 timotimo that's not a chdir multi for Bool, that's trying to coerce a Bool to a Str
15:10 samcv SourceBaby, &chdir
15:10 timotimo samcv: ^
15:10 samcv yes that's what i thought
15:10 timotimo m: sub test(Str() $a) { say $a.perl }; test(True)
15:10 camelia rakudo-moar d35efb: OUTPUT«"True"␤»
15:11 timotimo i have no idea why it'd error out like that, though
15:11 samcv is there a way to make it not do that
15:11 timotimo m: sub test(Str() $a) { say $a.perl }; test(Bool)
15:11 camelia rakudo-moar d35efb: OUTPUT«Cannot resolve caller Str(Bool: ); none of these signatures match:␤    (Bool:D $: *%_)␤  in sub test at <tmp> line 1␤  in block <unit> at <tmp> line 1␤␤»
15:11 timotimo that's how
15:11 samcv well i passed an undefined bool
15:11 notviki s: &chdir
15:11 SourceBaby notviki, Sauce is at https://github.com/rakudo/rakudo/blob/d35efb6/src/core/io_operators.pm#L164
15:11 samcv m: Bool.Str
15:11 camelia rakudo-moar d35efb: OUTPUT«Cannot resolve caller Str(Bool: ); none of these signatures match:␤    (Bool:D $: *%_)␤  in block <unit> at <tmp> line 1␤␤»
15:11 timotimo well, what should Bool give as a string?
15:11 samcv i meant for chdir
15:11 timotimo well, chdir just has Str(), which means it'll coerce anything to Str that you give to it (as long as it's Cool)
15:12 samcv if you pass it a boolean it shouldn't create a directory named True or False. i would think…
15:12 samcv maybe unavoidableish
15:12 samcv yeah
15:12 timotimo that's "just" how Cool works
15:12 timotimo m: say "100".sin
15:12 camelia rakudo-moar d35efb: OUTPUT«-0.506365641109759␤»
15:12 samcv yea
15:12 notviki timotimo: why Cool? .Str is a Mu method
15:13 samcv m: chdir True
15:13 camelia rakudo-moar d35efb: OUTPUT«Failed to change the working directory to '/home/camelia/True': does not exist␤  in block <unit> at <tmp> line 1␤␤Actually thrown at:␤  in any  at gen/moar/Metamodel.nqp line 3072␤  in block <unit> at <tmp> line 1␤␤»
15:13 notviki And I'm guessing there's a proto in Bool.pm which is why Bool.Str craps out
15:13 timotimo yeah, but Str() is short for Str(Cool), no? or is it Str(Any)?
15:14 notviki Any
15:14 notviki m: say Str(class {})
15:14 camelia rakudo-moar d35efb: OUTPUT«Use of uninitialized value of type <anon|63097232> in string context.␤Methods .^name, .perl, .gist, or .say can be used to stringify it to something meaningful.␤  in block <unit> at <tmp> line 1␤␤»
15:14 notviki m: -> Str() $x { say $x }(class {})
15:14 camelia rakudo-moar d35efb: OUTPUT«Use of uninitialized value of type <anon|49929984> in string context.␤Methods .^name, .perl, .gist, or .say can be used to stringify it to something meaningful.␤  in block  at <tmp> line 1␤␤»
15:15 notviki m: quietly chdir class {}
15:15 camelia rakudo-moar d35efb: ( no output )
15:15 regnarg joined #perl6-dev
15:15 notviki tehe
15:15 regnarg left #perl6-dev
15:15 timotimo OK
15:15 timotimo so it's Any
15:16 samcv m: chdir &chdir
15:16 camelia rakudo-moar d35efb: OUTPUT«Sub object coerced to string (please use .gist or .perl to do that)␤  in block <unit> at <tmp> line 1␤Failed to change the working directory to '/home/camelia/chdir': does not exist␤  in block <unit> at <tmp> line 1␤␤Actually thrown at:␤  in an…»
15:16 notviki even Mu it seems
15:16 notviki m: say Str(Mu)
15:16 camelia rakudo-moar d35efb: OUTPUT«(Str(Mu))␤»
15:16 samcv this is more the type of warning i would want True to warn
15:16 samcv when trying to chdir into True or False
15:17 notviki m: -> Str() $x { say "all good" }(Mu)
15:17 camelia rakudo-moar d35efb: OUTPUT«Type check failed in binding to $x; expected Any but got Mu (Mu)␤  in block <unit> at <tmp> line 1␤␤»
15:17 samcv that actually coerced but it did print a  warning
15:17 notviki Ah, no, Any
15:17 samcv m: mkdir &mkdir
15:17 camelia rakudo-moar d35efb: OUTPUT«mkdir is disallowed in restricted setting␤  in sub restricted at src/RESTRICTED.setting line 1␤  in sub mkdir at src/RESTRICTED.setting line 8␤  in block <unit> at <tmp> line 1␤␤»
15:21 kalkin- joined #perl6-dev
15:25 samcv m: chdir True
15:25 camelia rakudo-moar d35efb: ( no output )
15:25 samcv heh
15:26 samcv m: say dir $*PWD
15:26 camelia rakudo-moar d35efb: OUTPUT«Dynamic variable $*PWD not found␤  in block <unit> at <tmp> line 1␤␤Actually thrown at:␤  in block <unit> at <tmp> line 1␤␤»
15:26 samcv m: say dir $*CWD
15:26 camelia rakudo-moar d35efb: OUTPUT«("/home/camelia/.cpanm".IO "/home/camelia/.local".IO "/home/camelia/.npm".IO "/home/camelia/.perl6".IO "/home/camelia/.perlbrew".IO "/home/camelia/.rcc".IO "/home/camelia/.ssh".IO "/home/camelia/Perlito".IO "/home/camelia/evalbot".IO "/home/camelia/log".IO…»
15:27 samcv m: $*CWD.dir.grep(/True/).say
15:27 camelia rakudo-moar d35efb: OUTPUT«("/home/camelia/True".IO)␤»
15:30 samcv m: $*CWD.dir.grep(/True/).say
15:30 camelia rakudo-moar d35efb: OUTPUT«()␤»
15:30 samcv ok. cool. don't want to leave that directory just lying around
15:31 notviki m: say $*CWD
15:31 camelia rakudo-moar d35efb: OUTPUT«"/home/camelia".IO␤»
15:31 notviki m: say chdir True
15:31 camelia rakudo-moar d35efb: OUTPUT«Failed to change the working directory to '/home/camelia/True': does not exist␤  in block <unit> at <tmp> line 1␤␤Actually thrown at:␤  in any  at gen/moar/Metamodel.nqp line 3072␤  in block <unit> at <tmp> line 1␤␤»
15:31 notviki m: chdir True
15:31 camelia rakudo-moar d35efb: OUTPUT«Failed to change the working directory to '/home/camelia/True': does not exist␤  in block <unit> at <tmp> line 1␤␤Actually thrown at:␤  in any  at gen/moar/Metamodel.nqp line 3072␤  in block <unit> at <tmp> line 1␤␤»
15:31 notviki why did it work when you did it?
15:31 samcv m: use nqp; nqp::mkdir('/home/camelia/True', 0o777)
15:31 camelia rakudo-moar d35efb: ( no output )
15:32 notviki Ah
15:32 samcv hahaha
15:32 samcv m: say chdir True
15:32 camelia rakudo-moar d35efb: OUTPUT«"/home/camelia/True".IO␤»
15:32 [Coke] m: my $a = 3; temp temp temp temp temp temp temp temp temp temp $a;
15:32 camelia rakudo-moar d35efb: ( no output )
15:32 samcv m: say chdir True; say dir $*CWD
15:32 camelia rakudo-moar d35efb: OUTPUT«"/home/camelia/True".IO␤()␤»
15:33 samcv m: say chdir True; m: use nqp; nqp::mkdir(False, 0o777)
15:33 camelia rakudo-moar d35efb: OUTPUT«"/home/camelia/True".IO␤This type cannot unbox to a native string: P6opaque, Bool␤  in block <unit> at <tmp> line 1␤␤»
15:34 samcv m: say chdir True; m: use nqp; nqp::mkdir('False', 0o777)
15:34 camelia rakudo-moar d35efb: OUTPUT«"/home/camelia/True".IO␤»
15:34 samcv m: say chdir True; m: use nqp; nqp::mkdir('False', 0o777)
15:34 camelia rakudo-moar d35efb: OUTPUT«"/home/camelia/True".IO␤»
15:34 samcv m: say chdir True; m: use nqp; say nqp::mkdir('False', 0o777)
15:34 camelia rakudo-moar d35efb: OUTPUT«"/home/camelia/True".IO␤False␤»
15:52 * notviki stares at incorrect test wondering how the hell it's still passing..
15:52 * notviki notices it's an `is` test -_-
15:53 notviki m: use Test; is (), ((),)
15:53 camelia rakudo-moar d35efb: OUTPUT«ok 1 - ␤»
15:53 [Coke] Slip-erry slope?
15:54 [Coke] m: say ().perl; say ((),).perl;
15:54 camelia rakudo-moar d35efb: OUTPUT«()␤((),)␤»
15:54 notviki They stringify the same
15:54 notviki `is` is a Bad Idea™, especially in roast
15:55 notviki m: use Test; is [1, [2, [3]]], "1 2 3", 'totally the same'
15:55 camelia rakudo-moar d35efb: OUTPUT«ok 1 - totally the same␤»
15:59 notviki m: use Test; is-deeply ().Seq, ()
15:59 camelia rakudo-moar d35efb: OUTPUT«ok 1 - ␤»
16:00 notviki m: use Test; is-deeply ().Seq, []
16:00 camelia rakudo-moar d35efb: OUTPUT«not ok 1 - ␤␤# Failed test at <tmp> line 1␤# expected: $[]␤#      got: $()␤»
16:03 notviki m: sub foo (@a, @b) { dd [@a,  @b] }(().Seq, ())
16:03 camelia rakudo-moar d35efb: OUTPUT«[(), ()]␤»
16:03 notviki m: my $a = ().Seq; my $b = (); dd [ $a, $b ]; sub foo (@a, @b) { dd [@a,  @b] }(().Seq, ())
16:03 camelia rakudo-moar d35efb: OUTPUT«[().Seq, ()]␤[(), ()]␤»
16:04 notviki looks like infix:<eqv>(@a, @b) is buggy https://github.com/rakudo/rakudo/blob/d35efb6/src/core/Mu.pm#L857
16:07 notviki Heh, turns out *I* added all those `is` tests... In July
16:07 notviki What a difference half a year makes :o
16:08 samcv :P
16:09 samcv yeah is isn't that great an idea unless you are doing it on purpose
16:09 dalek roast: 3ea6fa6 | (Zoffix Znet)++ | S32-list/combinations.t:
16:09 dalek roast: Use is-deeply instead of is for tests
16:09 dalek roast:
16:09 dalek roast: As they're sensitive to the datastructure returned
16:09 dalek roast: review: https://github.com/perl6/roast/commit/3ea6fa6c71
16:10 samcv i don't usually use `like` but i used it for testing the unicode version name returned, since like 1.1 and v1.1 and V1.1 are all valid aliases
16:11 samcv well assuming it didn't print out like 1.1.1.1 i guess haha
16:13 samcv i kind of want to figure out how the moarvm data structure actually works so I can actually just scrap a lot of the code in the unicode generation file
16:15 samcv then just eventually rewrite it or something. it's just not maintainable at all. like reading through tthe same file multiple times and not storing the data in a hash or something to process or an array
16:16 samcv at least totally scrap the code that generates the hash table for the database for sure. the rest may be ok enough
16:16 notviki m: dd combinations(0.5, 2),
16:16 camelia rakudo-moar d35efb: OUTPUT«((),).Seq␤»
16:17 ggoebel joined #perl6-dev
16:18 notviki m: dd combinations(-0.5,   -2)
16:18 camelia rakudo-moar d35efb: OUTPUT«()␤»
16:21 notviki t/spec/S15-unicode-information/uniname.rakudo.moar (Wstat: 512 Tests: 40 Failed: 2)
16:21 notviki Failed tests:  15, 18
16:21 notviki Non-zero exit status: 2
16:21 notviki samcv: I guess these were fixed by the MoarVM PR that we reverted?
16:21 notviki And now they're broken again?
16:21 samcv yeah
16:21 notviki k
16:22 samcv i need to sleep. see you on the other side
16:22 notviki bye
16:25 dalek rakudo/nom: d86c419 | (Zoffix Znet)++ | src/core/List.pm:
16:25 dalek rakudo/nom: Fix combinations() cases incorrectly returning 1-item list
16:25 dalek rakudo/nom:
16:25 dalek rakudo/nom: Currently, asking for, say, 5-item combinations out of a 3-item list
16:25 dalek rakudo/nom: returns a 1-item list containing an empty list. This is incorrect,
16:25 dalek rakudo/nom: as that result applies only when we ask for 0-item combination.
16:25 dalek rakudo/nom:
16:25 dalek rakudo/nom: Fix that by changing short-circuiting logic at the start of the
16:25 dalek rakudo/nom: routine:
16:26 dalek rakudo/nom:
16:26 dalek rakudo/nom: n < 1 → we have an empty list to pick from
16:26 dalek rakudo/nom: k = 0 → can pick just 1 combination (empty list); return ((),)
16:26 notviki rakudo/nom: n < k → we don't have enough items to pick a combination of k items; return ()
16:26 notviki rakudo/nom:
16:26 notviki rakudo/nom: Note that negative k is not the same as k=0; it's a conscious design
16:26 notviki rakudo/nom: decision: you can have a set of zero items, but not a set of -1 items:
16:26 notviki rakudo/nom: https://irclog.perlgeek.de/perl6/2016-12-28#i_13810513
16:26 notviki rakudo/nom: review: https://github.com/rakudo/rakudo/commit/d86c419623567a7b206008323f5
16:27 dalek roast: af8002e | (Zoffix Znet)++ | S32-list/combinations.t:
16:27 dalek roast: Fix incorrect combinations() tests
16:27 dalek roast:
16:27 dalek roast: The tests were added[^1][^2] on July 4, 2016 and are not part of 6.c.
16:27 dalek roast: Three of them assume the result of picking a 2-item combination from an empty list is
16:27 dalek roast: an 1-item list containing empty list. It was a now-fixed[^3] bug. Fourth one assumes
16:27 dalek roast: negative k is equivalent to k=0 and was passing only due to the test using `is` (it was
16:27 dalek roast: changed[^4] to `is-deeply` in a commit previous to this one).
16:27 dalek roast:
16:27 dalek roast: [1] https://github.com/perl6/roast/commit/2f8f0a4ceb5a3f7e6a2c6f3216a2a9e2c70cff3d
16:27 dalek roast: [2] https://github.com/perl6/roast/commit/0ff33b312bbaf63dae0c19fdfbbbfae97ef53883
16:27 dalek roast: [3] https://github.com/rakudo/rakudo/commit/d86c419623567a7b206008323f5
16:27 notviki https://github.com/perl6/roast/commit/af8002e3c94689585dd9fa2cb182775d6d6a00e9
16:27 notviki lazy ass robot
16:29 japhb You didn't use a hyphen, and yet I still thought of https://xkcd.com/37/
16:31 notviki heh
16:34 notviki mmm
16:35 samcv what i can buy a unicode 9.0 book
16:35 samcv i read the 5.0 was the last one they made
16:35 * samcv buys quickly
16:36 samcv two volumes each 500 pages
16:36 notviki yikes
16:36 notviki That was a quick sleep
16:36 samcv very
16:37 samcv so quick it did not exist
16:39 jnthn .tell samcv unipropcode is implemented in interp.c, which calls MVM_unicode_name_to_property_code which is in unicode_ops.c, which looks in the property_codes_by_names_aliases hash, which is constructed from unicode_property_keypairs in unicode_db.c generated from ucd2c.pl
16:39 yoleaux2 jnthn: I'll pass your message to samcv.
16:39 samcv hi
16:39 yoleaux2 16:39Z <jnthn> samcv: unipropcode is implemented in interp.c, which calls MVM_unicode_name_to_property_code which is in unicode_ops.c, which looks in the property_codes_by_names_aliases hash, which is constructed from unicode_property_keypairs in unicode_db.c generated from ucd2c.pl
16:39 jnthn Heh, I saw you'd gone for sleep in the backlog... :)
16:39 samcv i wish
16:40 samcv ok wrote that down
16:41 * jnthn seems to have some cold-y thing sapping all his energy and can do not that much more than sleep :/
16:41 moritz jnthn: get well soon!
16:42 jnthn Thanks...here's hoping :)
16:42 jnthn At least I got it *after* enjoying Christmas.
16:42 * [Coke] wishes jnthn some ... caffienated soup?
16:42 * moritz had a few days without illness but not interest in programming whatsoever
16:42 moritz it's slowly returning now
16:47 japhb Few things more annoying than wishing you had time and energy to hack for a few months, getting a nice calm vacation with lots of time and energy -- and not wanting to hack.  Because you just *know* that when you return to work, suddenly you'll be thinking "Oh man, why didn't I use that time to hack?  I want to hack SO BAD right now."
16:53 dalek rakudo/nom: b0e61f3 | (Zoffix Znet)++ | src/core/List.pm:
16:53 dalek rakudo/nom: Restore fast-path for negative k in combinations()
16:53 dalek rakudo/nom:
16:53 dalek rakudo/nom: It was accidentally lost in
16:53 dalek rakudo/nom: https://github.com/rakudo/rakudo/commit/d86c419623567a7b206008323f5
16:53 dalek rakudo/nom:
16:53 dalek rakudo/nom: No functional changes
16:53 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/b0e61f32b8
16:55 moritz japhb: could be worse; I can hack at work :-)
16:55 moritz also, work starts at 2017-01-10 for me, so no hurry
16:56 moritz if I'm able to write my next blog post for the p6 book in the next 1.5 weeks, I'm happy
17:02 notviki m: sub foo (Positional \a, Positional \b) { dd [a, b] }(().Seq, ())
17:02 camelia rakudo-moar d86c41: OUTPUT«[(), ()]␤»
17:02 notviki m: sub foo (\a, \b) { dd [a, b] }(().Seq, ())
17:02 camelia rakudo-moar d86c41: OUTPUT«[().Seq, ()]␤»
17:02 * notviki doesn't get it
17:07 dalek rakudo/nom: db1836a | (Zoffix Znet)++ | src/core/List.pm:
17:07 dalek rakudo/nom: Make combinations() always return a Seq on succcess
17:07 dalek rakudo/nom:
17:07 dalek rakudo/nom: Currently, it returns Seq for all cases except ones that return
17:07 dalek rakudo/nom: empty lists, in which case it returns a List.
17:07 dalek rakudo/nom:
17:07 dalek rakudo/nom: Fix by returning a Seq for those cases as well.
17:07 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/db1836a45d
17:10 jnthn notviki: If you pass Seq to a sub and it expects Positional, we automatically call .cache on the Seq. If we didn't, you couldn't write sub foo(@a) { }; foo(@b.grep(* < 5)) for example
17:10 jnthn Well you could write it, but it'd bind fail :P
17:11 notviki hmm...
17:11 notviki Well, the side effect of that is eqv equates a Seq with a List
17:12 notviki m: say (1, 2, 3).Seq eqv (1, 2, 3)
17:12 camelia rakudo-moar db1836: OUTPUT«True␤»
17:12 notviki How come on 32-bit boxes isbig_I() starts to return true after 2**28-1 ?
17:13 notviki m: use nqp; say nqp::isbig_I(2**32)
17:13 camelia rakudo-moar db1836: OUTPUT«1␤»
17:13 mst my guess would be that it uses the extra bits to stash flags of some sort?
17:13 notviki m: use nqp; say nqp::isbig_I(2**32-1)
17:13 camelia rakudo-moar db1836: OUTPUT«1␤»
17:13 mst fairly common approach to allow ints to be held inline in a slot that could also be a pointer
17:22 notviki Do we have maximum non-is-big_I stored in a constant somewhere? I thought I recalled seeing something or other
17:22 notviki m: dd +SOCK_MAX
17:22 camelia rakudo-moar db1836: OUTPUT«6␤»
17:22 notviki m: dd +PF_MAX
17:22 camelia rakudo-moar db1836: OUTPUT«4␤»
17:22 notviki heh
17:23 notviki m: dd $UINT64_UPPER
17:23 camelia rakudo-moar db1836: OUTPUT«Int $UINT64_UPPER = 18446744073709551616␤»
17:23 notviki That's the one I saw..
17:24 notviki m: dd IntLexRef
17:24 camelia rakudo-moar db1836: OUTPUT«IntLexRef␤»
17:24 [Coke] seems like we shouldn't have SOCK_MAX exposed like that.
17:25 notviki m: dd $in_deprecation
17:25 camelia rakudo-moar db1836: OUTPUT«Any $in_deprecation = Any␤»
17:25 notviki we got all kinds of crap lying around :)
17:26 notviki lol
17:26 notviki This is a good one:
17:26 notviki m: say DON'T_DIVIDE_NUMBERS 1, 2, 3, 4
17:26 camelia rakudo-moar db1836: OUTPUT«0.5␤»
18:02 japhb notviki: Um, how can the maximum value of a base-2 number be even?
18:07 notviki japhb: ? I'm not following
18:07 japhb From your dd above: Int $UINT64_UPPER = 18446744073709551616␤
18:08 japhb or is _UPPER not the maximum value?
18:08 notviki I don't even know what that value is.
18:08 geekosaur "brain-hurty"
18:08 japhb Ugh:
18:08 japhb src/core/UInt64.pm:my Int $UINT64_UPPER = nqp::pow_I(2, 64, Num, Int);
18:08 japhb src/core/UInt64.pm:subset UInt64 of Int where { 0 <= $_ < $UINT64_UPPER }
18:09 japhb So it's _MAX + 1
18:09 notviki Is it even max?
18:09 notviki based on my big_I questino above, seems some bits are used for something
18:10 japhb The threshold for big_I is used for different things (knowing when you can do
18:10 geekosaur you may be confusing representation of a notional Int in MoarVM with the correct range of a uint64_t
18:10 japhb "big integer" math using only small integers, for one.
18:11 japhb geekosaur: To whom are you speaking?
18:11 geekosaur to notviki, since I also saw the limit they saw (and with similar lack of context)
18:11 japhb ah
18:11 notviki I'm about to define the big_I max, so we could report correct limits in error messages in combinations() and permutations()
18:11 geekosaur but yes there are several possible reasons for a lower limit than just what the bits can hold
18:13 notviki Also, TIL we have UInt64 subset... I thought UInt was the only one
18:13 geekosaur ideally there are subsets for all of the sized ones, providing boxy targets for the unboxed native types
18:14 geekosaur otherwise using the unboxed natives with things that require boxed types is annoying to codegen for
18:15 * japhb wonders if UInt64 originally went in for NativeCall or Rat
18:21 notviki m: my UInt $x
18:21 camelia rakudo-moar db1836: ( no output )
18:21 notviki m: my UInt64 $x
18:21 camelia rakudo-moar db1836: ( no output )
18:21 notviki m: say 42 ~~ UInt
18:21 camelia rakudo-moar db1836: OUTPUT«True␤»
18:21 notviki m: say 42 ~~ UInt64
18:21 camelia rakudo-moar db1836: OUTPUT«True␤»
18:21 notviki :/
18:22 notviki m: say Int ~~ UInt
18:22 camelia rakudo-moar db1836: OUTPUT«True␤»
18:22 notviki m: say Int ~~ UInt64
18:22 camelia rakudo-moar db1836: OUTPUT«Invocant requires an instance of type Int, but a type object was passed.  Did you forget a .new?␤  in any accepts_type at gen/moar/Metamodel.nqp line 3444␤  in block <unit> at <tmp> line 1␤␤»
18:22 notviki Ah, there we go
18:24 pyrimidine joined #perl6-dev
18:29 notviki " notviki │ I'm about to define the big_I max, so we could report correct limits"
18:30 notviki never mind... $*KERNEL is not available as early as where I want to use it :/
18:31 TimToady .oO(Quoth the raven, "Nevermind!")
18:33 notviki TimToady: should IO methods throw or fail?
18:33 notviki like spurt, open, chdir
18:34 * geekosaur draws distinctions there
18:34 geekosaur spurt and open are file operations, fail is reasonable
18:35 geekosaur process-level state changes (chdir) should be a harder failure mode
18:35 notviki and pipe, slurp, indir, tmpdir, homedir, rename, copy, symlink, and link
18:35 geekosaur ...but I think chdir is the only such that gets exposed; most of them would be highly OS specific
18:35 notviki all of the listed currently throw... though for some reason move() appears to fail
18:36 notviki or rather throw on anything, but fail on X::IO::Copy|X::IO::Unlink :/
18:37 notviki hm...
18:38 notviki It's even worse: many subs throw, but methods fail :S
18:38 moritz can we just always throw exceptions please?
18:38 moritz fail was meant to make concurrent code cheaper
18:38 notviki moritz: for all IO routines?
18:38 moritz but start blocks catch exceptions anyway, and rethrow them when accessing Promise.result
18:39 moritz notviki: all IO routines, but I mean really everywhere
18:39 notviki :o
18:39 moritz I think fail() was a mistake, and it's time we admitted it
18:39 notviki I'd find that really annoying to write try {} everywhere
18:40 moritz my $fh = try open $filename
18:40 moritz not much worse than my $fh = open $filename
18:40 moritz and you check $fh anyway later on to make sure it doesn't throw when you use it
18:42 moritz I see none of the benefits that fail() was supposed to give us anywhere in real-world code
18:42 kalkin- joined #perl6-dev
18:43 notviki What were the benefits?
18:44 moritz making sure some massively parallel computations didn't die just because a single sub-computation failed
18:44 moritz and not burdening parallel code with error handling
18:44 notviki And that's it?
18:44 moritz (which we still need, because we can't know that stuff only fails and never will die)
18:45 moritz I'm not aware of other supposed benefits
18:45 notviki I always seen it as an easier way to throw that doesn't encoumber you with all the try {} ugliness
18:46 notviki Like from-json here throws, so I have that try block and because } ends statement I have to place // in an weird place: https://github.com/zoffixznet/perl6-buggable/blob/master/lib/Buggable/Plugin/Eco.pm6#L40
18:46 notviki And here $!r6.stats fails() and it's just an `or` statement: https://github.com/zoffixznet/na/blob/master/lib/NA/Plugin/Release.pm6#L15
18:49 notviki ugggh... so github now made it possible for submitters to request a specific person to review a PR....
18:49 notviki ... and now I'm asked to review a couple for no good reason? :/
18:50 TimToady moritz: we've gone over this many time, it also allows you to treat failure with logic instead of putting try blocks everywhere
18:51 TimToady and if you fail in sink context, it throws immediately anyway, just like you want
18:53 notviki TimToady: any idea for what to do for IO stuff? Fail or throw? We're now kinda half-way in between and I wanna know how to fix it.
18:54 TimToady fail was mostly invented for IO stuff in the first place
18:54 TimToady errand &
18:55 notviki alright
18:58 hankache joined #perl6-dev
19:00 japhb Also, with/orwith/without make more sense in a world with Failure in it.
19:01 japhb Looking at the doc for that is the first time I've seen the new doc code highlighting.  Not bad!
19:02 MasterDuke if anyone is interested https://gist.github.com/MasterDuke17/4b9dc25bbc9b45fc5ade7fac0b2556cd has 'perf report > compile.report' and 'perf report --call-graph=none --no-children > compile.report_no_call-graph_no_children'
19:02 japhb Although occasionally it highlights methods as keywords; see https://docs.perl6.org/language/concurrency for several examples
19:02 MasterDuke after doing 'perf record --call-graph dwarf /home/dan/Source/perl6/install/bin/moar --libpath="/home/dan/Source/perl6/install/share/nqp/lib" --libpath="/home/dan/Source/perl6/install/share/nqp/lib" perl6.moarvm --nqp-lib=blib --setting=NULL --ll-exception --optimize=0 --target=parse --stagestats --output=CORE.setting.moarvm gen/moar/CORE.setting'
19:03 japhb --optimize=3 was too heavy?
19:04 MasterDuke no, just timotimo suggested that the first time i profiled a compile
19:04 MasterDuke though it's actually a no-op, since i had --target=parse anyway
19:04 notviki japhb: please report those on https://github.com/perl6/atom-language-perl6
19:06 MasterDuke i'm running one with the usual --optimize=3 --target=mbc now, but parsing is the long pole in the tent
19:09 moritz TimToady: so have you observed benefits from fail() in real-world code?
19:10 notviki I have. No need to stick try {} all over the place.
19:11 MasterDuke and here are reports for the usual build https://gist.github.com/MasterDuke17/0fe1a713392bc960be8b1a17f7a867b4
19:11 moritz notviki: so do you actually do error reporting for each thing that could fail?
19:11 moritz or error handling, for that matter
19:12 notviki sometimes I let it just sink and explode.
19:13 notviki I mean, it's really close to just exceptions *without typing try {}* all over teh place
19:13 notviki or sometimes I do // "some value"
19:13 notviki instead of try {} // "some value"
19:24 japhb notviki: highlighting bug reported
19:24 japhb https://github.com/perl6/atom-language-perl6/issues/33
19:24 japhb I tend to use // a LOT.  One of my favorite operators.
19:25 japhb In fact I use it so often, I have to force myself to use || when I really do mean Truthiness and not just Definedness
19:58 dalek roast: 9eafc2c | ronaldxs++ | S04-blocks-and-statements/temp.t:
19:58 dalek roast: Add tests for temp array and hash (#206)
19:58 dalek roast: review: https://github.com/perl6/roast/commit/9eafc2c026
21:38 dalek rakudo/nom: 7193df1 | TimToady++ | src/Perl6/Actions.nqp:
21:38 dalek rakudo/nom: treat ++ and -- as mutators under hyper
21:38 dalek rakudo/nom:
21:38 dalek rakudo/nom: A ++<<@foo shouldn't complain about useless use in sink context.
21:38 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/7193df1d23
22:26 MasterDuke timotimo: have you seen those gists i linked earlier (perf reports of perf recording a rakudo compile)? anything useful in them?
22:26 timotimo oh, i haven't
22:27 timotimo i have some difficulty reading that format, unfortuantely
22:28 MasterDuke you have to look at the raw file, but at the bottom of the gists is the no-children format
22:30 timotimo oh!
22:30 timotimo um, which of those files is what?
22:31 MasterDuke https://gist.github.com/MasterDuke17/4b9dc25bbc9b45fc5ade7fac0b2556cd#file-compile-report_no_call-graph_no_children
22:31 timotimo i download this and open it with "perf report"?
22:31 MasterDuke that is the report
22:31 MasterDuke same with this https://gist.github.com/MasterDuke17/0fe1a713392bc960be8b1a17f7a867b4#file-compile-report_no_call-graph_no_children
22:32 timotimo interesting. a lot of hits in GC
22:32 MasterDuke those are the output of 'perf report --call-graph=none --no-children'
22:32 timotimo if only we could tell which gc_mark the top one is
22:32 MasterDuke first for --target-parse, second for --target=mbc
22:33 MasterDuke any way for me to get that info if i re-run it?
22:33 timotimo nope
22:34 timotimo i don't think so, at least
22:50 lizmat joined #perl6-dev
22:53 Ven joined #perl6-dev
22:55 dalek rakudo/nom: d1d0e46 | (Zoffix Znet)++ | src/core/List.pm:
22:55 dalek rakudo/nom: Improve error reporting for combinations()
22:55 dalek rakudo/nom:
22:55 dalek rakudo/nom: - Fix reported maximum $n: on 32-bit boxes it's smaller than 2147483647
22:55 dalek rakudo/nom: - Remove range check for $k: the shortcurcuits already take care of all
22:55 dalek rakudo/nom:     the cases where $k would be out of range (the $n < $k case and out
22:55 dalek rakudo/nom:     of range $n throwage).
22:55 dalek rakudo/nom: - For consistency, check args first, before shortcuircuting into fast
22:55 dalek rakudo/nom:     paths. This makes failure mode more defined, as previously we
22:55 dalek rakudo/nom:     allowed huge $n's as long as $k was zero or negative, now we throw
22:55 dalek rakudo/nom:     regardless of what the $k might be.
22:55 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/d1d0e46c36
23:01 MasterDuke fwiw, here's the top 20 $name's seen in MATCH() during a rakudo compile (nc=1 vs nc=2 is which branch in MATCH() it came from), https://gist.github.com/MasterDuke17/ad31d008085e7f10891ac243e02f67bb
23:12 dalek rakudo/nom: 347271d | lizmat++ | src/core/Rat.pm:
23:12 dalek rakudo/nom: Fix RT #130427
23:12 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/347271d214
23:12 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=130427
23:16 * lizmat has been in the car for 7+ hours and is tired
23:16 yoleaux2 14:17Z <notviki> lizmat: what are your thought on removing this https://github.com/rakudo/rakudo/commit/3962ee51d9fab ? Feels like a footgun: if you end up with a unitialized IO::Path type object (say, from unset named param), the code will silently assume CWD. Since $*CWD has .chdir, doesn't seem like it's worth the rist of also having this behaviour. Thoughts?
23:18 lizmat notviki: not sure what I was thinking then.  Feels like a footgun to me now as well
23:18 notviki alright
23:23 lizmat good night, #perl6-dev!
23:23 notviki night

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