Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6-dev, 2016-08-25

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

All times shown according to UTC.

Time Nick Message
00:10 dalek roast/6.c-errata: ab52151 | (Zoffix Znet)++ | S06-advanced/wrap.t:
00:10 dalek roast/6.c-errata: Avoid assuming too much about internals of throws-like
00:10 dalek roast/6.c-errata:
00:10 dalek roast/6.c-errata: The exception thrown depends on internal implementation
00:10 dalek roast/6.c-errata: of the testing function if one is used to contain the exception.
00:10 dalek roast/6.c-errata:
00:10 dalek roast/6.c-errata: See https://github.com/rakudo/rakudo/pull/743#issuecomment-210489723
00:10 dalek roast/6.c-errata: for more information.
00:10 dalek roast/6.c-errata: review: https://github.com/perl6/roast/commit/ab521517a4
00:12 skids joined #perl6-dev
00:21 Zoffix I got t/spec/S32-exceptions/misc.t failing in 6.c-errata
00:21 Zoffix https://gist.github.com/zoffixznet/3b8edc1744e8ad1e5b567dd5673d929a
00:22 Zoffix Ah, that just looks like the recent addition of Exception::JSOn or whatever it was
00:26 dalek rakudo/nom: 547fc51 | (Zoffix Znet)++ | lib/Test.pm6:
00:26 dalek rakudo/nom: Revert to using multi-dispatch for subtest
00:26 dalek rakudo/nom:
00:26 dalek rakudo/nom: Reverts commit 9e4760256647336d8e5b2a3dc71d01c0f0485a3e That commit was
00:26 dalek rakudo/nom: introduced to prevent a spectest failure in 6.c-errata, however, the
00:26 dalek rakudo/nom: failure was due to the test itself assuming too much about the internal
00:26 dalek rakudo/nom: implementation of the throws-like test that was used to contain the
00:26 dalek rakudo/nom: exception.
00:26 dalek rakudo/nom:
00:26 dalek rakudo/nom: The test has now been fixed in
00:26 dalek rakudo/nom: https://github.com/perl6/roast/commit/ab521517a4
00:26 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/547fc5154c
00:30 dalek roast/6.c-errata: 9c1ba24 | (Zoffix Znet)++ | S32-exceptions/misc.t:
00:30 dalek roast/6.c-errata: Port 2ab7a1f67d3 fix to 6.c-errata
00:30 dalek roast/6.c-errata:
00:30 dalek roast/6.c-errata: We now have Exceptions:: class, which the test naively assumes
00:30 dalek roast/6.c-errata: we don't. Fix test to avoid such assumption.
00:30 dalek roast/6.c-errata: review: https://github.com/perl6/roast/commit/9c1ba24f0b
00:32 timotimo kind of strange to have it named Exception::JSON?
00:32 timotimo like, the test shouldn't expect we don't, it's just those are usually called X::Something?
00:33 Zoffix The class is not *an* Exception but rather a way to output exceptions
00:33 timotimo oh!
00:33 timotimo that new thing! yes, indeed
00:33 Zoffix Like, it changes the regular output to being output as JSON
00:33 timotimo yes, i remember it now
00:33 timotimo it's fantastic!
00:33 Zoffix \o/
00:35 Zoffix m: my $sig = :(| ($a) = 2);
00:35 camelia rakudo-moar 547fc5: OUTPUT«===SORRY!=== Error while compiling <tmp>␤Cannot put default on anonymous slurpy parameter␤at <tmp>:1␤------> my $sig = :(| ($a) = 2⏏);␤    expecting any of:␤        constraint␤»
00:35 Zoffix Is it really anonymous? Is it really slurpy?
00:36 Zoffix I'm not exactly sure what that sig is meant to be, but a side glance at https://github.com/perl6/roast/commit/f096ae50f4944926c752c52661416a7462fa65f7 seems to suggest it's muting tests that indicate a commit that's breaking something.
00:36 Zoffix (they're also failing in 6.c-errata now)
00:37 timotimo the slurpy parameter there is anonymous
00:37 Zoffix So the tests should be removed?
00:37 timotimo yeah, i believe it was a recent fix to make that error happen
00:37 timotimo for all slurpy parameters
00:37 Zoffix cool
00:37 timotimo they are not allowed to have defaults now
00:37 timotimo shouldn't have been allowed to in the first place
00:44 * timotimo BBL
00:56 Zoffix committable6, HEAD dd [ $*PERL.compiler.version ]
00:56 committable6 Zoffix, ¦«HEAD»: [v2016.07.1.220.g.6.e.84.a.5.e]
00:57 Zoffix .tell AlexDaniel looks like committable6 never recovered from my tag fiasco
00:57 yoleaux2 Zoffix: I'll pass your message to AlexDaniel.
00:57 AlexDaniel .
00:57 yoleaux2 00:57Z <Zoffix> AlexDaniel: looks like committable6 never recovered from my tag fiasco
00:57 Zoffix m: dd [ $*PERL.compiler.version ]
00:57 camelia rakudo-moar 547fc5: OUTPUT«[v2016.08.1.21.g.547.fc.51]␤»
00:57 AlexDaniel Zoffix: no, it's because we were working on some things
00:57 Zoffix Ah, OK :)
00:57 AlexDaniel mostly compression
01:00 cognominal joined #perl6-dev
01:01 dalek roast/6.c-errata: fb53102 | (Zoffix Znet)++ | S06-signature/introspection.t:
01:01 dalek roast/6.c-errata: Remove bogus tests
01:01 dalek roast/6.c-errata:
01:01 dalek roast/6.c-errata: The tests assumed defaults on slurpies are a thing, but they were
01:01 dalek roast/6.c-errata: giving MoarVM errors when used. In
01:01 dalek roast/6.c-errata: https://github.com/rakudo/rakudo/commit/560170f31ef22a2dbc7cb1fe44df464b2082626c
01:01 dalek roast/6.c-errata: those errors were caught earlier and now cause Rakudo Exceptions to be
01:01 dalek roast/6.c-errata: thrown, which causes these tests to fail.
01:01 dalek roast/6.c-errata: review: https://github.com/perl6/roast/commit/fb53102538
01:02 dalek roast: 8627307 | (Zoffix Znet)++ | S06-signature/introspection.t:
01:02 dalek roast: Remove bogus tests
01:02 dalek roast:
01:02 dalek roast: The tests assumed defaults on slurpies are a thing, but they were
01:02 dalek roast: giving MoarVM errors when used. In
01:02 dalek roast: https://github.com/rakudo/rakudo/commit/560170f31ef22a2dbc7cb1fe44df464b2082626c
01:02 dalek roast: those errors were caught earlier and now cause Rakudo Exceptions to be
01:02 dalek roast: thrown, which causes these tests to fail.
01:02 dalek roast: review: https://github.com/perl6/roast/commit/8627307ba7
01:02 [Coke] zoff; Changing errata tests should not be done litely make sure you have buyin from somebody
01:03 [Coke] Which you may have already doneI'm just going off the commit message
01:03 Zoffix Yeah, I've confirmed with timotimo: <Zoffix> So the tests should be removed?
01:03 Zoffix <timotimo> yeah, i believe it was a recent fix to make that error happen
01:04 Zoffix Basically, even though the test passed, when someone actually tried to use that sig in a subroutine, they'd get a MoarVM error. The test never went that deep and merely used the Signature object, calling .perl on it, so the test was passing, even though that signature was impossible to be actually used on a routine
01:04 [Coke] Okay it's a big enough thing that you should check with more than just one personIt's basically changing the contract we have with our users about what six CS
01:04 [Coke] what 6 c is
01:05 Zoffix OK
01:05 [Coke] ut I am going to bed ZZ :)
01:05 [Coke] joined #perl6-dev
01:11 Zoffix ....and now, instead of waiting for it to get all the way down to MoarVM, we throw in Rakudo.
01:11 * Zoffix loves async text chat in that you can finish a thought you stared 7m ago and it won't seem weird :D
01:34 bisectable6 joined #perl6-dev
01:35 bisectable6 joined #perl6-dev
01:35 committable6 joined #perl6-dev
01:45 MasterDuke lizmat: some interesting performance swings for .comb here: https://gist.github.com/Whateverable/0956fb70fd2e5cff0ddbaf4340525bca
01:48 ilbot3 joined #perl6-dev
01:48 Topic for #perl6-dev is now Perl 6 language and compiler development 2.0 | Logs at http://irclog.perlgeek.de/perl6-dev/today
01:57 AlexDaniel lizmat: to make it clearer: this commit made things much slower https://github.com/rakudo/rakudo/commit/977797f
02:03 AlexDaniel lizmat: if I see it correctly, then this .comb is in sink context, yet no warning is thrown. Weird synthetic thing, not sure if there is any actual code suffering from this slowdown, but perhaps it's worth taking a look
02:09 mst joined #perl6-dev
03:10 committable6 joined #perl6-dev
03:10 bisectable6 joined #perl6-dev
04:20 cognominal joined #perl6-dev
04:38 [Tux] joined #perl6-dev
05:00 dalek rakudo/nom: 597052f | lizmat++ | lib/TAP.pm6:
05:00 dalek rakudo/nom: Lose to more returns in TAP.pm6
05:00 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/597052f5ca
05:20 brrt joined #perl6-dev
05:30 brrt good * #perl6
05:30 yoleaux2 19 Aug 2016 10:55Z <Zoffix> brrt: might be worth mentioning to jnthn++ as he's been working on making stuff like that unhang. My test runs stopped hanging after that work. Never had anything flap or hang for quite a while.
05:30 yoleaux2 19 Aug 2016 11:18Z <jnthn> brrt: About said hanging test, can you give me more info? I had that running a couple of hundred times in a row on a loaded system without issue after my last round of fixes...
05:30 brrt early, in european zone
05:30 brrt ohhai, yeah, i'll try it out
05:30 brrt it may be that my moar is far behind
05:41 travis-ci joined #perl6-dev
05:41 travis-ci Rakudo build errored. Elizabeth Mattijsen 'Remove unneeded returns from TAP.pm6
05:41 travis-ci https://travis-ci.org/rakudo/rakudo/builds/154801258 https://github.com/rakudo/rakudo/compare/76e1ede7a6bb...33be22fd2e9d
05:41 travis-ci left #perl6-dev
05:41 buggable ☢ [travis build above] Build log missing from at least one job. Check results manually.
05:49 [TuxCM] joined #perl6-dev
05:50 lizmat joined #perl6-dev
05:55 zengargoyle joined #perl6-dev
06:00 [TuxCM] This is Rakudo version 2016.08.1-22-g597052f built on MoarVM version 2016.08
06:00 [TuxCM] csv-ip5xs       10.064
06:00 [TuxCM] test            15.281
06:00 [TuxCM] test-t           7.297
06:00 [TuxCM] csv-parser      16.045
06:12 brrt joined #perl6-dev
06:18 * brrt is currently retesting with latest moar
06:21 lizmat joined #perl6-dev
06:32 pyrimidi_ joined #perl6-dev
06:41 travis-ci joined #perl6-dev
06:41 travis-ci Rakudo build errored. Elizabeth Mattijsen 'Oops, wrong order of checks, Brad Gilbert++'
06:41 travis-ci https://travis-ci.org/rakudo/rakudo/builds/154815880 https://github.com/rakudo/rakudo/compare/33be22fd2e9d...a4cc1ca41e47
06:41 travis-ci left #perl6-dev
06:42 buggable ☢ [travis build above] Build log missing from at least one job. Check results manually.
06:44 brrt joined #perl6-dev
06:44 brrt .tell jnthn https://gist.github.com/anonymous/6c4b5eece51500252ad57fab423316b6
06:44 yoleaux2 brrt: I'll pass your message to jnthn.
06:44 brrt sorry, i broked it again
07:24 RabidGravy joined #perl6-dev
07:28 dalek rakudo/nom: 945b916 | lizmat++ | t/harness6:
07:28 dalek rakudo/nom: Remove superfluous returns from t/harness6
07:28 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/945b916821
07:40 cognominal joined #perl6-dev
07:52 [TuxCM] https://gist.github.com/anonymous/e62ee2c340d105edb020c3e1147803d8
07:52 [TuxCM] I need a bit of NativeCall help there
07:53 [TuxCM] Cannot invoke object with invocation handler in this context
07:53 [TuxCM] in method CALL-ME at /pro/3gl/CPAN/rakudobrew/moar-blead-nom/install/share/perl6/sources/24DD121B5B4774C04A7084827BFAD92199756E03 (NativeCall) line 328
07:53 [TuxCM] in block <unit> at nc-c.pl line 67
07:53 [TuxCM] which in the gitst is line 51
07:53 * [TuxCM] wants to see if I can get libcsv3 working with NativeCall in perl6
08:04 [TuxCM] joined #perl6-dev
08:15 nine One can now install Inline::Perl5 with just git clone git@github.com:niner/Inline-Perl5.git && cd Inline-Perl5 && perl6 configure.pl6 --test --install
08:15 nine Without needing panda or zef or any other dependencies :)
08:16 nine See mst's talk on Friday for why this is useful ;)
08:20 JimmyZ [TuxCM]: as per
08:20 JimmyZ [TuxCM]: https://github.com/rakudo/rakudo/blob/nom/t/04-nativecall/04-pointers.t
08:20 JimmyZ [TuxCM]: you may want Pointer
08:26 lizmat joined #perl6-dev
08:34 [TuxCM] JimmyZ, that caused an interesting change :)
08:35 [TuxCM] > ===SORRY!=== Error while compiling:
08:35 [TuxCM] Undeclared routines:
08:35 [TuxCM] hello used at line 1. Did you mean 'shell'?
08:35 [TuxCM] world used at line 1. Did you mean 'words'?
08:35 [TuxCM] those are *runtime* errors (repeating 50000 times)
08:35 lizmat joined #perl6-dev
09:06 [TuxCM] joined #perl6-dev
09:22 dalek rakudo/nom: 8a95de1 | leont++ | / (3 files):
09:22 dalek rakudo/nom: Restore "Convert create-moar-runner to perl 6"
09:22 dalek rakudo/nom:
09:22 dalek rakudo/nom: This reverts commit 66883a40e396c23253a14708f039317beb33301c.
09:22 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/8a95de14bf
09:22 dalek rakudo/nom: 9ccd848 | lizmat++ | / (3 files):
09:22 dalek rakudo/nom: Merge pull request #846 from Leont/less5
09:22 dalek rakudo/nom:
09:22 dalek rakudo/nom: Restore "Convert create-moar-runner to perl 6"
09:22 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/9ccd84801f
09:37 mst DOH
09:37 mst if you add a generated file, remember to add it to 'make clean' or confusion may ensue
09:40 mst nine: um
09:40 mst p5helper.c:237:5: warning: value computed is not used [-Wunused-value]
09:40 mst /usr/bin/ld: /home/matthewt/local/perl-5.24.0/lib/5.24.0/x86_64-linux/CORE/libperl.a(op.o): relocation R_X86_64_32S against `PL_opargs' can not be used when making a shared object; recompile with -fPIC
09:40 mst /home/matthewt/local/perl-5.24.0/lib/5.24.0/x86_64-linux/CORE/libperl.a: could not read symbols: Bad value
09:40 mst argh, hang on, did I forget to build multiplicity or whatever
09:41 mst aha, -Dusershrplib
09:41 mst as documented that I missed
10:03 [TuxCM] joined #perl6-dev
10:05 geekosaur it's kinda broken that it even tries to treat a static library as a shared library
10:05 geekosaur then again I should probably expect that; on 32-bit linux you could actually get away with that, and it's probably wired into dlopen or something and was never disabled even though it can never work on 64-bit
10:14 mst yeah, I'm not sure why it's using libperl.a rather than libperl.so
10:14 mst but I'm trying to just crowbar everything into working right now, not asking why ;)
10:17 lizmat joined #perl6-dev
10:45 mst argh
10:45 mst nine: you about?
10:51 lizmat joined #perl6-dev
10:52 [TuxCM] joined #perl6-dev
10:54 nine mst: now
10:55 mst nine: found it
10:56 mst RepositoryRegistry uses nqp::backend unless RKAUDO_PREFIX is set
10:56 mst so also looks in the fucking moar dirs
10:57 * mst tries with the env var for the moment
10:59 cognominal joined #perl6-dev
11:00 mst oh, wait, the env var isn't actually the prefix
11:00 mst because that would be helpful
11:04 [TuxCM] lizmat, here it is ...
11:04 [TuxCM] This is Rakudo version 2016.08.1-25-g9ccd848 built on MoarVM version 2016.08
11:04 [TuxCM] csv-ip5xs       10.061
11:04 [TuxCM] test            15.519
11:04 [TuxCM] test-t           7.338
11:04 [TuxCM] csv-parser      16.431
11:04 lizmat yeah, so no difference reallly: if any a negative difference  :-(
11:04 nine lizmat: nadim's happy now :)
11:05 lizmat nine: cool  :-)
11:12 ilmari[m] joined #perl6-dev
11:49 lizmat joined #perl6-dev
12:05 [TuxCM] joined #perl6-dev
12:13 lizmat joined #perl6-dev
12:17 [Coke]_ joined #perl6-dev
12:17 [TuxCM] joined #perl6-dev
12:20 japhb joined #perl6-dev
12:24 TheLemonMan joined #perl6-dev
13:11 [TuxCM] joined #perl6-dev
13:26 [TuxCM] joined #perl6-dev
13:30 skids joined #perl6-dev
13:35 travis-ci joined #perl6-dev
13:35 travis-ci Rakudo build errored. Zoffix Znet 'Revert to using multi-dispatch for subtest
13:35 travis-ci https://travis-ci.org/rakudo/rakudo/builds/154913908 https://github.com/rakudo/rakudo/compare/a4cc1ca41e47...547fc5154cd2
13:35 travis-ci left #perl6-dev
13:36 buggable ☠ [travis build above] 2 builds failed. NONE due to the timeout
13:38 [TuxCM] joined #perl6-dev
13:38 lizmat jnthn: any idea why this doesn't compile in the setting, but outside it does?
13:38 lizmat multi sub postcircumfix:<[ ]>(array[int] \SELF, Int $pos) is raw {
13:38 lizmat nqp::atposref_i(SELF,nqp::unbox_i($pos))
13:38 lizmat }
13:40 jnthn Did you place it after class array?
13:40 yoleaux2 06:44Z <brrt> jnthn: https://gist.github.com/anonymous/6c4b5eece51500252ad57fab423316b6
13:41 jnthn (It depends on the ^parameterize method inside of class array, so just pre-declaring it ain't enough)
13:47 lizmat jnthn: I did
13:48 lizmat ok, lemme check that later, now watching rakudo running on a SailfishOS phone  :-)
13:48 jnthn :D
13:55 lizmat joined #perl6-dev
13:57 lizmat joined #perl6-dev
14:00 TheLemonMan lizmat, can you give a look at #127803 ? it looks like something you'd fix in the blink of an eye :)
14:00 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=127803
14:03 [Coke] RT: 1365; @LARRY: 18; BOOTSTRAP: 4; BUG: 478; BUILD: 10; CONC: 47; DOCS: 1; GLR: 4; IO: 15; JVM: 68; LTA: 116; MATH: 5; NATIVECALL: 15; NEW: 857; NYI: 89; OO: 7; OSX: 5; PERF: 26; POD: 18; PRECOMP: 12; REGEX: 29; REPL: 4; RFC: 50; SEGV: 37; SPESH: 1; STAR: 5; TESTNEEDED: 30; TODO: 10; UNI: 30; UNTAGGED: 317; WEIRD: 3
14:04 lucasb_ joined #perl6-dev
14:06 lizmat joined #perl6-dev
14:06 dalek rakudo/nom: 6ae6ecb | niner++ | / (2 files):
14:06 dalek rakudo/nom: Make Distribution::Path handle native libraries correctly
14:06 dalek rakudo/nom:
14:06 dalek rakudo/nom: With this change we can now use it in install-dist.pl and get rid of the
14:06 dalek rakudo/nom: prototype.
14:06 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/6ae6ecbfca
14:19 cognominal joined #perl6-dev
14:24 lizmat joined #perl6-dev
14:25 lizmat TheLemonMan: looking at it (#127803)
14:25 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=127803
14:29 [Coke] RT: 1365; @LARRY: 18; BOOTSTRAP: 4; BUG: 485; BUILD: 11; CONC: 49; DOCS: 1; GLR: 4; IO: 15; JVM: 68; LTA: 117; MATH: 6; NATIVECALL: 15; NEW: 850; NYI: 90; OO: 7; OSX: 6; PERF: 26; POD: 18; PRECOMP: 12; REGEX: 34; REPL: 4; RFC: 53; SEGV: 37; SPESH: 1; STAR: 6; TESTNEEDED: 32; TODO: 10; UNI: 30; UNTAGGED: 296; WEIRD: 3
14:29 [Coke] ^^ untagged under 300
14:29 timotimo so many tags :S
14:30 timotimo note my much improved errors for the p6opaque error!
14:30 timotimo i made that! *proud*
14:30 TheLemonMan clear error messages for p6opaques *giggles*
14:31 timotimo :D
14:32 timotimo lizmat: my first guess is it's trying to access the result of the code block classify gets and assumes it's a List for some reason, but really it gets a junction that doesn't autothread (for some reason)
14:32 [Coke] timotimo++
14:32 timotimo yay
14:39 arnsholt joined #perl6-dev
14:45 tomboy65 joined #perl6-dev
14:59 [TuxCM] joined #perl6-dev
15:01 pmichaud good morning, #perl6
15:01 pmichaud er, #perl6-dev
15:01 timotimo hey pm!
15:10 dalek rakudo/nom: 7fa2ba7 | lizmat++ | src/core/Hash.pm:
15:10 dalek rakudo/nom: Fix RT #127803
15:10 dalek rakudo/nom:
15:10 dalek rakudo/nom: Definitely a case of premature optimization  :-)
15:10 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/7fa2ba77d4
15:10 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=127803
15:15 TheLemonMan lizmat++ :D
15:37 MetaZoffix joined #perl6-dev
15:39 MetaZoffix pmichaud: would you say this can be implemented right now, without waiting for 6.d too? (Providing it doesn't break anything in roast) https://github.com/perl6/specs/blob/master/v6d.pod#signatureattribute-defaults-based-on-definedness
15:39 MetaZoffix basically adding a //= for signature defaults that would assign the default if the parameter is absent or is not defined
15:41 * MetaZoffix is now a proud owner of http://Perl6.Fail :D
15:43 MetaZoffix To host my bug app at http://i.imgur.com/lqKvZpi.png
15:45 lizmat MetaZoffix++
15:50 jdv79 MetaZoffix++++
15:58 hoelzro MetaZoffix++ looks fancy!
16:16 lizmat joined #perl6-dev
17:03 [TuxCM] joined #perl6-dev
17:06 brrt joined #perl6-dev
17:39 cognominal joined #perl6-dev
18:38 gfldex joined #perl6-dev
19:12 pmichaud MetaZoffix:  I will look
19:13 pmichaud [Coke]: ping
19:14 [Coke] pong
19:14 pmichaud have a moment?  I feel like I need a little recalibration on the current state of things
19:15 pmichaud Looking on your latest comment on b14828b (about the /b /B stuff), you say "I've no issues with it going into 6.c".  What exactly does that mean?
19:15 pmichaud I mean, how does something "go into 6.c"?  As I understand it, 6.c was released in December.
19:15 jnthn It doesn't.
19:15 jnthn Everything that's new now will go into either a 6.c.1 or a 6.d
19:16 jnthn Aside from stuff we classify as errata, 6.c is "immutable"
19:16 pmichaud Right, that's the perspective I've had.
19:16 pmichaud And that's part of what's been confusing me about this particular set of tickets/threads.
19:17 jnthn I suspect what it really means is "can we put this in *now* as opposed to waiting for us to have the machinary to make it only happen in the presence of an explicit `use 6.d.a` and not happen with a `use 6.c`"
19:17 pmichaud Hmm.
19:17 jnthn And I suspect the "go into 6.c" wording is being used as shorthand for that rather more unwieldy thing.
19:18 pmichaud Well, does "use 6.c" mean that untested things are guaranteed to work the same as they did with Rakudo's corresponding release?  I would hope not.
19:18 jnthn No
19:18 jnthn That's near-impossible to achieve.
19:19 pmichaud Agreed.
19:19 jnthn I attempted to document this stuff clearly at some point. :)
19:19 pmichaud So, even in the presence of "use 6.c", Rakudo should be able to have features that are outside of the literal 6.c specification, yes?
19:19 [Coke] I specifically meant "add a spectest in 6.c"
19:19 [Coke] which was incorrect; should have been 6.c.1
19:20 [Coke] but it was in response to your comment about it not having to wait until 6.d
19:20 MetaZoffix joined #perl6-dev
19:20 pmichaud Hmm.
19:20 [Coke] my mental setup for this is based on jnthn++'s doc.
19:20 MetaZoffix If there aren't any use 6.c or use 6.d, what does the user get? The latest and greatest that they have installed?
19:21 pmichaud What's the url for jnthn++'s doc?  I should read that.
19:21 jnthn pmichaud: https://gist.github.com/jnthn/c10742f9d51da80226fa is what I wrote up fwiw
19:21 pmichaud danke
19:21 jnthn And we really could do with it living somewhere better than a gist
19:21 MetaZoffix huggable: perl6 versions :is: https://gist.github.com/jnthn/c10742f9d51da80226fa
19:21 huggable MetaZoffix, Added perl6 versions as https://gist.github.com/jnthn/c10742f9d51da80226fa
19:22 pmichaud MetaZoffix: In answer to your earlier question of "can we implement signature attributes based on defaults now?", my response would be that there's nothing to prevent it from being implemented.
19:23 pmichaud in my mental model (which exists before reading jnthn++'s document), it's normal and even expected for compilers to lead the language specification.  Language specification is retrospective, not prospective.
19:23 pmichaud whether it should be part of the language (or compiler) is a slightly different topic... but a compiler doesn't have to wait for a language feature to be approved before implementing it.
19:23 MetaZoffix pmichaud: ok. Do I need someone else's review of that feature (from the language design perspective) or are you OKing it and I should start working on it?
19:24 pmichaud MetaZoffix: working code and rough consensus is the general model, I think.
19:24 [TuxCM] joined #perl6-dev
19:24 MetaZoffix Cool. I'll make a PR for review when I get a chance.
19:24 pmichaud it's more ideal if it can be done as a module first, though.
19:24 MetaZoffix Yeah, I'll do that, since that was my original plan anyway :) (and a blog post to go with it too)
19:27 pmichaud After studying a bit about language specification at a meta level, and looking at the history, I became convinced that language spec should be retrospective.  As a more relevant example -- Internet standards normally ratify what has already been done and proven by implementations, they don't map out the features yet-to-be-tried.  Or, when they are mapping out a possible future, they're termed "RFC" and
19:27 timotimo all we have is the "errata"
19:27 pmichaud not "standards" or "specifications".
19:27 timotimo ooes, didn't scroll
19:28 pmichaud I particularly wouldn't want to have "6.d" be a list of proposed features that are adopted prior to implementation.  That heads down the (wrong) path taken by C99, for example.
19:28 jnthn pmichaud: My doc was in many ways influenced by your language specs talk fwiw :)
19:28 pmichaud jnthn: I suspect that to be the case, but you may have added insights or choices that I hadn't reached yet.  :)
19:28 jnthn pmichaud: I see releases as snapshots in time for passing spectests :)
19:29 jnthn s/for/of/
19:29 pmichaud jnthn: we're in complete agreement there :)
19:29 jnthn Dinner time here, back later. :)
19:30 pmichaud Now that I'm having a chance to look into it more deeply -- that's the dissonance I feel in reading the v6d.pod in general.
19:31 pmichaud The whole "plans for 6.d" approach feels wrong.
19:31 pmichaud it should be "plans for Perl 6", and how strongly we feel about them.
19:31 pmichaud if there's a consensus that a particular feature would be good for Perl 6 (or Rakudo), it can proceed to implementation in Rakudo or other compilers.
19:32 pmichaud But we shouldn't tie it to prospective future language specifications *until* it's had a chance to be vetted in implementation somewhere.
19:33 timotimo we'll still have a 6.d "beta" phase
19:33 timotimo where you're not supposed to rely on how exactly the 6.d features behave
19:33 timotimo but if you don't "use 6.d" you won't get any of them at all
19:33 timotimo is how i understood it, anyway
19:34 MetaZoffix We have 54 RFC tickets: https://jsfiddle.net/2bsj7ecr/1/
19:34 pmichaud timotimo: that feels completely backwards to me
19:35 pmichaud if language ("6.d") is retrospective, I don't see how one can do a "use 6.d" for something that doesn't exist yet.
19:36 MetaZoffix (to answer my earlier question: "Any Perl 6 program or module that does not declare a language version will be interpreted as expecting to run on the latest non-draft version supported by the implementation")
19:36 timotimo i bet someone else can explain my thought better than me in any case
19:37 pmichaud MetaZoffix: I don't understand "latest non-draft version"
19:37 pmichaud it doesn't fit with my mental model
19:37 MetaZoffix pmichaud: "(for example, it could get 6.c, 6.c.1, 6.d, or 6.d.1 - but never 6.d.a, 6.e.b, etc.) "
19:37 pmichaud what exactly is a "non-draft version" of the language?
19:37 pmichaud oops
19:37 MetaZoffix I'm guessing 6.d.a would be the "beta" version that timotimo mentioned
19:37 timotimo ah! yes, letters
19:37 pmichaud I mean, what exactly is a "draft version" of the language?
19:38 timotimo "dude we should totally disallow \b in regexes, k?"
19:38 gfldex 6.d.a reads "six . dee . alpha" for me
19:39 pmichaud if the language is defined as a snapshot of spectests... what's a "draft version" of that?
19:39 timotimo a rakudo would ship with a 6.c spec test snapshot and a 6.d spec test snapshot that we're not yet committed to support in the future
19:39 MetaZoffix pmichaud: a version that can be used in the wild by implementation to see if it's fine to use, before actually being finalized and released is my guess
19:40 pmichaud let me finish reading jnthn++'s document.
19:40 pmichaud it talks about draft versions.
19:40 * MetaZoffix does same
19:40 MetaZoffix Yeah, my quote was from it
19:41 pmichaud my first guess:  "Any Perl 6 program or module that does not declare a language version will be interpreted as expecting to run on the implementation's default version."
19:42 pmichaud To me, it could be draft.  I shouldn't have to explicitly declare a draft version to use new features of an implementation.
19:42 MetaZoffix eh, won't have time to finish it. Need to pack things up before commutting
19:42 MetaZoffix \o
19:45 [Coke] pmichaud: in other news, I assigned an RT about rakudo.org to you that I have no idea how to fix.
19:45 pmichaud [Coke]: just now?
19:45 [Coke] today
19:45 [Coke] one sec.
19:46 pmichaud #128664
19:46 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=128664
19:46 [Coke] https://rt.perl.org/Ticket/Display.html?id=128664
19:46 [Coke] yes.
19:50 pmichaud looking.
19:55 [Coke] danke!
19:55 [Coke] I couldn't find it on disk or in wp
19:55 pmichaud Now fixed.
19:56 [Coke] woot!
19:56 pmichaud Dashboard -> Appearance -> Header, Select Image, (re)load the image from the media library.
19:56 pmichaud Apparently the url was originally set when the site was still "rakudoperl.org" and had never been updated.
19:57 [Coke] oh wow, wouldn't ahve figured that ,yah
19:57 pmichaud It took me a bit to find it as well.
19:57 pmichaud I just happened to notice "Header" in the appearance menu and said "hey, what's that?"
19:57 pmichaud otherwise I was randoming looking around and grepping odd things as well.
19:57 pmichaud anyway, now fixed.  :)
20:02 [Coke] down to 1364 tickets! ;)
20:03 pmichaud I have to go do other things for a while -- I'll re-read the jnthn++ versioning doc and make my comments from there.  Thanks everyone for your patience.  :)
20:04 [Coke] pmichaud: thanks for yours. :)
20:04 gfldex m: sub foo(:$where = -> $p { $p // 'default'}){};
20:04 camelia rakudo-moar 7fa2ba: ( no output )
20:34 lizmat joined #perl6-dev
20:59 cognominal joined #perl6-dev
21:31 cognominal joined #perl6-dev

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