Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6-dev, 2017-07-31

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

All times shown according to UTC.

Time Nick Message
01:15 deep-book-gk joined #perl6-dev
01:18 deep-book-gk left #perl6-dev
01:26 BenGoldberg joined #perl6-dev
01:52 ilbot3 joined #perl6-dev
01:52 Topic for #perl6-dev is now Perl 6 language and compiler development 2.0 | Logs at http://irclog.perlgeek.de/perl6-dev/today | For toolchain/installation stuff see #perl6-toolchain | For MoarVM see #moarvm
01:56 stmuk joined #perl6-dev
03:08 pharv joined #perl6-dev
03:23 geekosaur joined #perl6-dev
03:25 geekosaur joined #perl6-dev
03:29 geekosaur joined #perl6-dev
05:38 GK1wmSU joined #perl6-dev
05:39 GK1wmSU left #perl6-dev
05:52 _GK1wmSU joined #perl6-dev
05:54 _GK1wmSU left #perl6-dev
06:47 [Tux] This is Rakudo version 2017.07-97-ga256c26d8 built on MoarVM version 2017.07-266-ged84a632
06:47 [Tux] csv-ip5xs        2.705
06:47 [Tux] test            12.616
06:47 [Tux] test-t           4.362 - 4.371
06:47 [Tux] csv-parser      12.909
08:18 vendethiel- joined #perl6-dev
08:52 robertle joined #perl6-dev
09:14 robertle joined #perl6-dev
09:17 lizmat Files=1217, Tests=67446, 256 wallclock secs (13.75 usr  5.01 sys + 1675.72 cusr 162.44 csys = 1856.92 CPU)
09:21 vendethiel joined #perl6-dev
10:01 vendethiel- joined #perl6-dev
10:09 Geth ¦ rakudo/nom: 3e078d4d47 | (Timo Paulssen)++ | src/Perl6/Actions.nqp
10:09 Geth ¦ rakudo/nom: space after import modulename shall not explode the compiler
10:09 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/3e078d4d47
10:34 vendethiel joined #perl6-dev
11:13 stmuk hmm I suppose appveyor can be used as a Windows MSI build servic
11:14 stmuk assuming the free/OSS service allows it
12:00 Geth ¦ rakudo/nom: a7d2ad1db8 | (Elizabeth Mattijsen)++ | src/core/traits.pm
12:00 Geth ¦ rakudo/nom: Make Attribute is default() functional
12:00 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/a7d2ad1db8
12:43 Geth ¦ rakudo/nom: 148ba7f2e4 | (Elizabeth Mattijsen)++ | src/core/traits.pm
12:43 Geth ¦ rakudo/nom: Use proper check for Scalarness of attribute
12:43 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/148ba7f2e4
12:51 lizmat m: 42
12:51 camelia rakudo-moar 148ba7: OUTPUT: «WARNINGS for <tmp>:?Useless use of constant integer 42 in sink context (line 1)?»
12:52 lizmat m: class A { has $.a = 42 }; for ^100000 { my $a = A.new }; say now - INIT now
12:52 camelia rakudo-moar 148ba7: OUTPUT: «0.26820687?»
12:52 lizmat m: class A { has $.a is default(42) }; for ^100000 { my $a = A.new }; say now - INIT now
12:52 camelia rakudo-moar 148ba7: OUTPUT: «0.20388547?»
12:52 lizmat m: say 26820687 / 20388547
12:52 camelia rakudo-moar 148ba7: OUTPUT: «1.315478097?»
12:53 lizmat not bad for a one line change  :-)
12:53 lizmat jnthn++ for pointers
12:55 travis-ci joined #perl6-dev
12:55 travis-ci Rakudo build failed. Elizabeth Mattijsen 'Make Attribute is default() functional'
12:55 travis-ci https://travis-ci.org/rakudo/rakudo/builds/259338464 https://github.com/rakudo/rakudo/compare/3e078d4d47e8...a7d2ad1db850
12:55 travis-ci left #perl6-dev
12:55 buggable [travis build above] ? Did not recognize some failures. Check results manually.
12:57 lizmat appears to be a flapper  :-(
13:00 Zoffix But the two aren't equivalent?
13:01 jnthn No, they ain't
13:01 jnthn the is default(...) form comes with the usual mass of caveats
13:01 jnthn Which aren't going to go away
13:01 jnthn my $a is default([])
13:01 jnthn For example, is one array shared between all the variables
13:02 jnthn Same will happen with attrs
13:02 Zoffix m: class A { has $.a = 42 }; A.new(:a(Nil)).a.say
13:02 camelia rakudo-moar 148ba7: OUTPUT: «(Any)?»
13:02 Zoffix m: class A { has $.a is default(42) }; A.new(:a(Nil)).a.say
13:02 camelia rakudo-moar 148ba7: OUTPUT: «42?»
13:02 lizmat indeed, that is the difference
13:02 jnthn Also that :)
13:02 jnthn Also you can't refer to an earlier attribute with is default
13:02 lizmat which you could consider a feature, fwiw  :-)
13:02 jnthn Well, it does what `is default` does
13:03 lizmat indeed
13:03 jnthn Which is good in that it should work on attributes
13:04 ggoebel joined #perl6-dev
13:15 Geth ¦ rakudo/nom: c63c57e9a8 | (Elizabeth Mattijsen)++ | src/core/IO/Handle.pm
13:15 Geth ¦ rakudo/nom: Use "is default" for IO::Handle atttributes
13:15 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/c63c57e9a8
13:15 lizmat should help in file intensive operations
13:15 lizmat although the big problem really still is the number of named parameters on IO::Handle.open
13:16 Zoffix let's toss most of them in 6.d \o/
13:17 jnthn heh
13:18 jnthn There's one more needed too, I think (:translate-nl)
13:18 Zoffix + one more (:chmod)
13:19 lizmat if we could only have a sig for "no named params"
13:20 jnthn where !%_ :)
13:22 lizmat jnthn: are you implying we have such a thing already ?
13:23 lizmat ah, using "where"
13:23 lizmat that will kill performance
13:24 lizmat although, in the case of IO::Handle.open, that may still be better  :-)
13:24 lizmat hmmm
13:24 jnthn Did you measure it against the status quo? :)
13:24 jnthn Right :)
13:24 Zoffix lizmat: reading the code, the last commit has a bug in it, of sorts. If you set $fh.nl-in = Nil; it'll set the attribute to default, but that default won't be set in the decoder. I think the $nl-in in this piece needs to be replaced with $!nl-in on these two lines: https://github.com/rakudo/rakudo/blob/nom/src/core/IO/Handle.pm#L194-L195
13:25 lizmat Zoffix: agree
13:25 lizmat will do
13:27 Geth ¦ rakudo/nom: 08f705b9f7 | (Elizabeth Mattijsen)++ | src/core/IO/CatHandle.pm
13:27 Geth ¦ rakudo/nom: Remove now superfluous assignment
13:27 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/08f705b9f7
13:27 Geth ¦ rakudo/nom: d0419a0f2a | (Elizabeth Mattijsen)++ | src/core/Rakudo/Iterator.pm
13:27 Geth ¦ rakudo/nom: Remove superfluous assigments
13:27 Geth ¦ rakudo/nom:
13:27 Geth ¦ rakudo/nom: Native ints default to 0
13:27 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/d0419a0f2a
13:28 nicq20 joined #perl6-dev
13:28 nicq20 Hello! o/
13:28 Zoffix \O\
13:32 nicq20 Do core types like Hash need to be defined in NQP?
13:34 timotimo interesting question
13:36 lizmat they're not defined in NQP, they get defined in the Rakudo metamodel bootstrap
13:37 jnthn To get a hash in NQP do nqp::hash()
13:37 nicq20 The reason I ask is because I think I have a (probably poor) working version of `Hash ~~ Hash` but written in regular p6.
13:38 jnthn That'd need to go into CORE.setting
13:38 jnthn Which is written in Perl 6 :)
13:38 jnthn So no NQP re-writing required :)
13:38 Geth ¦ rakudo/nom: 9f5686ec81 | (Elizabeth Mattijsen)++ | src/core/IO/Handle.pm
13:38 Geth ¦ rakudo/nom: Make sure nl-in changes are passed to decoder
13:38 Geth ¦ rakudo/nom:
13:38 Geth ¦ rakudo/nom: Spotted by Zoffix++
13:38 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/9f5686ec81
13:38 nicq20 Woo! \o/
13:38 Zoffix nicq20: what does it do?
13:39 nicq20 Really it just checks to see if each key exists, then checks that the value is eqv.
13:40 nicq20 It's like 5 lines with the sig.
13:40 Zoffix nicq20: how come not just self eqv $got ?
13:40 Zoffix From S03:
13:40 Zoffix Hash        Hash      hash mapping equivalent $_ eqv X
13:40 Zoffix Associative Hash      force hash comparison   $_.Hash eqv X
13:41 nicq20 That's a good point...
13:45 nicq20 I think I did that because I thought eqv still uses `.perl`.
13:45 Zoffix s: &infix:<eqv>, \(%(:42a), %(:42a))
13:45 SourceBaby Zoffix, Sauce is at https://github.com/rakudo/rakudo/blob/c63c57e/src/core/Map.pm#L423
13:46 Zoffix nicq20: it does what you described (checking key exists and eqv'ing the value) except it also takes shortcuts when the two objects are the same address or have different number of elements
13:48 Zoffix m: .signature.say for &infix:<eqv>.candidates
13:48 camelia rakudo-moar d0419a: OUTPUT: «($?)?(Any:U \a, Any:U \b)?(Any:D \a, Any:U \b)?(Any:U \a, Any:D \b)?(Any:D \a, Any:D \b)?(Iterable:D \a, Iterable:D \b)?(Stringy:D \a, Stringy:D \b)?(Numeric:D \a, Numeric:D \b)?(Blob:D \a, Blob:D \b)?(Str:D \a, Str:D \b)?(Capture:D \…»
13:48 Zoffix tons of candidates that don't use .perl
13:48 nicq20 `eqv` it is then.
13:50 dogbert17 lizmat: where is RabidGravy, is he trying to stop Brexit by himself?
13:50 lizmat .oO( that would be quite a Boom! :-)
13:51 dogbert17 he seemed quite invested in that particular question if memory serves
13:51 lizmat then more power to him!
13:57 Geth ¦ rakudo/nom: c8c5c7129f | (Elizabeth Mattijsen)++ | src/core/IO/Socket/INET.pm
13:57 Geth ¦ rakudo/nom: Use is default in favour of BUILDALL base init
13:57 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/c8c5c7129f
14:16 |Tux| This is Rakudo version 2017.07-104-g9f5686ec8 built on MoarVM version 2017.07-266-ged84a632
14:16 |Tux| csv-ip5xs        2.481
14:16 |Tux| test            12.380
14:16 |Tux| test-t           4.140 - 4.201
14:16 |Tux| csv-parser      12.610
14:16 travis-ci joined #perl6-dev
14:16 travis-ci Rakudo build passed. Elizabeth Mattijsen 'Use "is default" for IO::Handle atttributes'
14:16 travis-ci https://travis-ci.org/rakudo/rakudo/builds/259362506 https://github.com/rakudo/rakudo/compare/148ba7f2e4de...c63c57e9a823
14:16 travis-ci left #perl6-dev
14:16 |Tux| (after Liz' change to default attribute values)
14:20 Zoffix \o/
14:20 Zoffix lizmat++
14:26 Geth ¦ rakudo/nom: 40d5b27d52 | (Elizabeth Mattijsen)++ | src/core/IO/Socket.pm
14:26 Geth ¦ rakudo/nom: Use is default instead of BUILDALL based init
14:26 Geth ¦ rakudo/nom:
14:26 Geth ¦ rakudo/nom: In the same manner as IO::Handle.
14:26 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/40d5b27d52
14:39 tadzik joined #perl6-dev
14:43 Zoffix FWIW, in IO::Handle a Nil encoding means we're in binary mode; now in IO::Socket it means utf8. It doesn't look like it's doing anything with it when it's Nil, but still inconsistent.
14:44 Zoffix Also, IO::Socket has $.encoding, and IO::Socket::INET does IO::Socket, but IO::Socket::Async doesn't do IO::Socket and has $.enc for the same purpose :/
14:44 lizmat yeah, something for 6.d I'm afraid
14:45 * Zoffix fears more like 6.e
14:45 jnthn I'd pondered just :enc(:$encoding) everywhere :)
14:45 jnthn IO::Socket::Async can't reasonably do IO::Scoket
14:46 Zoffix I see
14:47 jnthn The APIs are completely difference
14:47 jnthn The point of the IO::Socket role is for when IO::Socket::INET is joined by IO::Socket::Domain or so
14:48 * Zoffix starts inconsistencies.txt
14:48 Zoffix In case there'll be a point where they could be ironed out.
14:54 ilmari[m] joined #perl6-dev
14:54 CIAvash[m] joined #perl6-dev
14:56 Geth ¦ rakudo/nom: fb5db59220 | (Elizabeth Mattijsen)++ | src/core/Duration.pm
14:56 Geth ¦ rakudo/nom: Streamline Duration creation
14:56 Geth ¦ rakudo/nom:
14:56 Geth ¦ rakudo/nom: - set default using "is default"
14:56 Geth ¦ rakudo/nom: - shortcut new() by bypassing Mu.bless
14:56 Geth ¦ rakudo/nom: - makes it about 1.5x faster
14:56 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/fb5db59220
14:57 Zoffix Logged 4 so far, off memory: https://temp.perl6.party/inconsistencies.txt
14:57 Geth ¦ rakudo: nicqrocks++ created pull request #1126: Add `ACCEPTS` method to Hash for defined hashes.
14:57 Geth ¦ rakudo: review: https://github.com/rakudo/rakudo/pull/1126
15:01 travis-ci joined #perl6-dev
15:01 travis-ci Rakudo build failed. Elizabeth Mattijsen 'Remove now superfluous assignment'
15:01 travis-ci https://travis-ci.org/rakudo/rakudo/builds/259367025 https://github.com/rakudo/rakudo/compare/c63c57e9a823...08f705b9f788
15:01 travis-ci left #perl6-dev
15:01 buggable [travis build above] ? Did not recognize some failures. Check results manually.
15:03 Zoffix make: *** [gen/moar/stage1/nqpmo.moarvm] Segmentation fault: 11
15:05 lizmat wow
15:05 timotimo Zoffix: did you see the latest moarvm commit?
15:07 Zoffix timotimo: why?
15:07 Zoffix I'm just copy-pasting why travis failed.
15:08 Zoffix That commit is over my head.
15:08 jnthn Apparently the one it was fixing was over mine too :)
15:09 timotimo oh, ok
15:09 timotimo how's about we teach travis how to do a "thread apply all bt full" when it finds a crash like that ...
15:16 Geth ¦ rakudo/nom: 413b71163b | (Elizabeth Mattijsen)++ | src/core/Proc/Async.pm
15:16 Geth ¦ rakudo/nom: Use is default instead of BUILDALL based init
15:16 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/413b71163b
15:28 Geth ¦ rakudo/nom: fbdbe6fb55 | (Elizabeth Mattijsen)++ | src/core/IO/Pipe.pm
15:28 Geth ¦ rakudo/nom: Use is default instead of BUILDALL based init
15:28 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/fbdbe6fb55
15:29 Zoffix nicq20: how crummy is your crummy machine? 35m sounds like you forgot to set TEST_JOBS and you're running a 1-core spectest
15:30 Zoffix Related: would be a good idea to default TEST_JOBS to, say, 8 instead of 1
15:31 ugexe i've often wished I could use a default in signatures, so something like `sub foo(:$a, :$b) { bar(:a, :$b) }; sub bar(:$a is default(1), :$b is default(True)) { say $a; say $b }; foo()` would print 1 and True instead of (Any)(Any)
15:31 MasterDuke joined #perl6-dev
15:31 lizmat ugexe: is next on my list
15:32 jnthn Huh, that's just =
15:32 ugexe otherwise Ive been having to do like sub foo(*%args [:$a, :$b]) { bar(|%args) }; bar(:$a, :$b) { say $a; say $b };
15:32 Zoffix lizmat: with that behaviour? :/
15:32 jnthn Also Nil defaulting is about assignment, but it's signature *binding*
15:33 ugexe m: sub foo(:$a, :$b) { bar(:$a, :$b) }; sub bar(:$a = 1, :$b = True) { say $a; say $b }; foo()
15:33 camelia rakudo-moar 413b71: OUTPUT: «(Any)?(Any)?»
15:33 nicq20 2 core. I checked that var though, I had set it to 2 and forgot to revert it back to a higher number.
15:33 nicq20 Oops.
15:33 jnthn The only place I think `is default` can apply on a parameter is when it's `is copy`
15:33 Zoffix ugexe: there was a proposal to add `(:$a //= 1, :$b //= True)` syntax that'd assign the default based on definedness.
15:33 ugexe that sounds like it would work too
15:34 Zoffix I was meant to make a module to test the feature, but I never managed to do it
15:34 Zoffix By managed, I mean I never wrote anything, not that I had any challenges.
15:35 Zoffix nicq20: ah, if it's 2 core then yeah, likely higher values won't do too much good.
15:35 nicq20 Zoffix: Spectest just finished at ~25 min.
15:35 Zoffix yikes.
15:36 Zoffix And I was complaining about *stress* test running in 140 seconds instead of the usual 113 :)
15:37 nicq20 This is my work machine too.
15:42 ugexe jnthn: = doesn't work well if you want do do both 1) use a named argument in a function B) pass that named argument to another function that wants to set a default for that parameter
15:44 ugexe well, it works well. but there is no way around the boilerplate to do the //= bit Zoffix mentioned
15:45 Zoffix IMO all these changes from `=` to `is default` are introducing a bunch of inconsistencies. We're making a semantic change for the sake of performance. We now have different behaviours for when Nil is given to .new, vs. when it's passed as arguments.
15:46 jnthn ugexe: Yeah, I sometimes hit that. There's the destructuring approach you mentioned.
15:46 jnthn I don't think `is default` is the solution
15:46 jnthn `is default` is really a way of saying "what does this default to if unitialized or assinged to Nil"
15:46 jnthn And it's a property of a scalar container
15:46 jnthn But 99% of parameters don't involve one
15:47 lizmat Zoffix jnthn: want me to revert ?
15:48 jnthn I'm not passing judgement on using is default on attributes, just saying it's about useless on parameters
15:49 Zoffix e.g. Proc::Async.new: :enc(Nil) vs Proc::Async.Supply: :enc(Nil). Or even more dangerous: $.translate-nl as Nil that when passed in .new becomes True, but when passed into methods gets ignored.
15:50 lizmat FWIW, all is default changes where spectested before I pushed them
15:51 lizmat otherwise I wouldn't have pushed
15:52 Zoffix lizmat: I think it's introducing a lot of inconsistencies that will make it more difficult to learn the behaviour of code as there's a difference between instantiating/assigning to attribute and using some value for it in a method call (like .translate-nl and .enc). Do I want it reverted... no opinion, but I think we'll see these inconsistencies bite someone in the future.
15:53 travis-ci joined #perl6-dev
15:53 travis-ci Rakudo build failed. Elizabeth Mattijsen 'Remove superfluous assigments
15:53 travis-ci https://travis-ci.org/rakudo/rakudo/builds/259367154 https://github.com/rakudo/rakudo/compare/08f705b9f788...d0419a0f2acf
15:53 travis-ci left #perl6-dev
15:53 buggable [travis build above] ? All failures are due to timeout (0), missing build log (0), GitHub connectivity (1), or failed make test (0).
15:54 ugexe the only problem with the destructuring approach is it does not work for less basic functions where you might need to pass the same parameter to two functions which only share that single parameter (e.g. you want to destructure :$a, :$b, :$c into both %first [$a, :$b] and %second [:$b, :$c])
15:54 Zoffix Especially with .enc/.encoding parameters.  IO::Handle.new: :enc(Nil) means "binary mode". Proc::Async.new:  :enc(Nil) means "utf8"
15:55 * Zoffix & lunch
15:56 Geth ¦ rakudo/nom: a8016278cd | (Elizabeth Mattijsen)++ | 5 files
15:56 Geth ¦ rakudo/nom: Revert potentially confusing "is default" changes
15:56 Geth ¦ rakudo/nom:
15:56 Geth ¦ rakudo/nom: See https://irclog.perlgeek.de/perl6-dev/2017-07-31#i_14947645
15:56 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/a8016278cd
15:59 * Zoffix waits for [Tux]'s tomorrow's bench numbers before feeling bad :}
15:59 |Tux| want a run now?
16:01 lizmat I deston't think my changes will show up in test-t
16:01 lizmat there aren't that many handles opened
16:04 perlpilot joined #perl6-dev
16:04 Zoffix buggable: speed 4
16:04 buggable Zoffix, ???? data for 2017-07-31–2017-07-31; range: 4.140s–4.371s; 6% faster
16:04 Zoffix I thought that was result of is default..
16:10 lizmat yeah, that looks about right
16:13 Zoffix actually, I lied. IO::Handle.new: :enc(Nil) means utf8 too, but IO::Handle.new.encoding = Nil means utf8 :/
16:13 Zoffix I mean: actually, I lied. IO::Handle.new: :enc(Nil) means utf8 too, but IO::Handle.new.encoding = Nil means binary :/
16:16 Geth ¦ rakudo/nom: 45ca084e7c | (Nic Q)++ | src/core/Map.pm
16:16 Geth ¦ rakudo/nom: Add `ACCEPTS` method to Map to compare Maps.
16:16 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/45ca084e7c
16:16 Geth ¦ rakudo/nom: 6e001867f2 | (Zoffix Znet)++ (committed using GitHub Web editor) | src/core/Map.pm
16:16 Geth ¦ rakudo/nom: Merge pull request #1126 from nicqrocks/nom
16:16 Geth ¦ rakudo/nom:
16:16 Geth ¦ rakudo/nom: Add `ACCEPTS` method to Hash for defined hashes.
16:16 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/6e001867f2
16:17 Zoffix lizmat: 6% for something without many handles sounds like a lot.
16:17 Zoffix Now I feel bad :)
16:45 Zoffix I'm gonna revert it back. 'cause taking a closer look at the affected attributes, looks like I'm mostly talking shit. IO::Handle :enc(Nil) doesn't mean bin; IO::H.chomp takes a Nil, but doesn't work with it in later code (.e.g. from .lines), nl-out don't take it, ; IO::Pipe $!eof/$!closed are private attrs, and Proc::Async/IO::Socket .enc/.encoding attributes aren't `is rw` so they don't got the "Nil is
16:45 Zoffix binary" meaning as in IO::Handle. That leaves Proc::Async.translate-nl, but it's not even documented (and not tested?) for methods and if you think of it, passing Nil to methods kinda does work as a "default"; the default being whatever $!translate-nl is
16:46 AlexWithFootgun GitHub: “Major service outage”
16:48 Zoffix ZOFVM: Files=1267, Tests=144378, 131 wallclock secs (21.28 usr  3.93 sys + 2601.97 cusr 247.95 csys = 2875.13 CPU)
16:48 Zoffix ZOFFLOP: t/spec/S11-modules/require.t
16:48 Zoffix ZOFFLOP: t/spec/S11-modules/nested.t
16:48 Geth ¦ rakudo/nom: 7fdbb49ccc | (Zoffix Znet)++ | 5 files
16:48 Geth ¦ rakudo/nom: Revert "Revert potentially confusing "is default" changes"
16:48 Geth ¦ rakudo/nom:
16:48 Geth ¦ rakudo/nom: This reverts commit a8016278cd1cb8836f49e52de0f9bf6cf3982404.
16:48 Geth ¦ rakudo/nom:
16:48 Geth ¦ rakudo/nom: On closer examination of affected attributes, it seems Nil don't
16:48 Geth ¦ rakudo/nom: usually work anyway, so the potential confusion is a lot lesser,
16:48 Geth ¦ rakudo/nom: and the perf gain of ~6% on our canary bench justifies it.
16:48 Geth ¦ rakudo/nom: https://irclog.perlgeek.de/perl6-dev/2017-07-31#i_14947791
16:48 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/7fdbb49ccc
16:49 Geth ¦ roast: cbb7302388 | (Zoffix Znet)++ | S03-smartmatch/hash-hash.t
16:49 Geth ¦ roast: Unfudge now-passing tests
16:49 Geth ¦ roast:
16:49 Geth ¦ roast: Fixed by https://github.com/rakudo/rakudo/commit/6e001867f2
16:49 Geth ¦ roast: review: https://github.com/perl6/roast/commit/cbb7302388
16:51 Zoffix a clue for the above floopers. The roast repo file shows an uncommited change in packages/RT128156/Needed.pm6 from `class Needed {};` to `class Needed { method version() { 2 } }` so whatever changes it either dies before changing back or maybe that's why the tests are failing
16:51 synopsebot6 Link:  https://rt.perl.org/rt3/Public/Bug/Display.html?id=128156
16:53 Zoffix .tell lizmat I put the `is default` revert back in 'cause at least half of my original concerns turned out to be invalid.
16:53 yoleaux Zoffix: I'll pass your message to lizmat.
16:53 * Zoffix &
17:06 robertle joined #perl6-dev
17:06 lizmat .
17:06 yoleaux 16:53Z <Zoffix> lizmat: I put the `is default` revert back in 'cause at least half of my original concerns turned out to be invalid.
17:07 lizmat Zoffix: ok, fine by me  :-)
17:07 lizmat fwiw, I think it's a feature being able to pass Nil to get the default
17:07 lizmat without having to know what the default is
17:29 travis-ci joined #perl6-dev
17:29 travis-ci Rakudo build passed. Elizabeth Mattijsen 'Use is default in favour of BUILDALL base init'
17:29 travis-ci https://travis-ci.org/rakudo/rakudo/builds/259378198 https://github.com/rakudo/rakudo/compare/9f5686ec81f9...c8c5c7129f55
17:29 travis-ci left #perl6-dev
17:46 lizmat hmnmm... what to think of this test:
17:46 lizmat ok  (\@a eqv \@a), "eqv on array references (1)";
17:47 lizmat the problems in my view:
17:47 lizmat 1. array references, WTF ?
17:47 lizmat 2. it's not checking for True, it's checking for thruthiness
17:48 lizmat unsurprisingly, these tests date from 2008
18:04 b2gills Hmm what do *I* think of that test? … .oO( **EXTERMINATE** )
18:06 * lizmat hopes moritz has some thoughts, being the original author of said tests :-)
18:23 moritz if we have capture equvialence tests somewhere else, sure
18:23 lizmat moritz: well, the comments state that they are array reference tests  :-)
18:23 lizmat that'd be a bit confusing, no ?
18:23 moritz lizmat: it is; but a comment is easily adjusted, no?
18:24 lizmat true
18:25 lizmat I was more thinking of a treatment like I've done in t/spec/S03-operators/set_subset.t
18:25 lizmat would that be ok with you, moritz?
18:27 moritz lizmat: I claim no ownership
18:27 lizmat ok, consider it up for demolition and rebuild  :-)
18:27 lizmat but after I've written this week's Perl 6 Weekly
18:28 moritz in other news: https://pbs.twimg.com/media/DGFWOCjXkAApImR.jpg
18:29 lizmat whee!
18:29 * lizmat is still waiting for hers  :-(
18:30 moritz there's got to be at least some perks to writing this stuff :-)
18:31 Zoffix moritz++
18:37 [Coke] the shrink wrap makes it look like a ring is under the plastic. ;)
18:37 [Coke] congrats!
19:42 travis-ci joined #perl6-dev
19:42 travis-ci Rakudo build passed. Elizabeth Mattijsen 'Use is default instead of BUILDALL based init'
19:42 travis-ci https://travis-ci.org/rakudo/rakudo/builds/259410970 https://github.com/rakudo/rakudo/compare/fb5db59220d3...413b71163bee
19:42 travis-ci left #perl6-dev
19:57 ugexe what is the empty folder https://github.com/rakudo/rakudo/tree/nom/dynext needed for?
19:59 moritz ugexe: that's where rakudo's .o and .so files go
20:27 travis-ci joined #perl6-dev
20:27 travis-ci Rakudo build passed. Elizabeth Mattijsen 'Use is default instead of BUILDALL based init'
20:27 travis-ci https://travis-ci.org/rakudo/rakudo/builds/259415233 https://github.com/rakudo/rakudo/compare/413b71163bee...fbdbe6fb550f
20:27 travis-ci left #perl6-dev
21:10 pharv joined #perl6-dev
21:22 lizmat and another Perl 6 Weekly hits the Net: https://p6weekly.wordpress.com/2017/07/31/2017-31-moar-smaller/
21:37 dct joined #perl6-dev
21:40 Geth ¦ 6.d-prep: 2cc614bde4 | (Zoffix Znet)++ (committed using GitHub Web editor) | TODO/FEATURES.md
21:40 Geth ¦ 6.d-prep: List deprecation of .flatmap in 6.d
21:40 Geth ¦ 6.d-prep: review: https://github.com/perl6/6.d-prep/commit/2cc614bde4
21:54 Zoffix moritz: is "Mark Powers" someone you know?
21:54 Zoffix They sent me a "review copy of Perl 6 Fundamentals" in a PDF I don't really wanna open :P
21:56 Zoffix And there's a leanpub's version from 28ths. Are leanpub's versions still current?
22:07 Zoffix Oh, Mark is asking for Amazon review. Wouldn't it show that I didn't buy it? Can you even review without buying?
22:10 Zoffix OK, I opened their PDF. Book looks sweet. Too bad there ain't no code formatting. I guess that's so they could run it B&W and save some money?
22:10 travis-ci joined #perl6-dev
22:10 travis-ci Rakudo build failed. Zoffix Znet 'Merge pull request #1126 from nicqrocks/nom
22:10 travis-ci https://travis-ci.org/rakudo/rakudo/builds/259433563 https://github.com/rakudo/rakudo/compare/a8016278cd1c...6e001867f215
22:10 travis-ci left #perl6-dev
22:10 buggable [travis build above] ? All failures are due to timeout (0), missing build log (0), GitHub connectivity (1), or failed make test (0).
22:13 Zoffix I've no idea what my amazon login is so, I gave up with the review stuff :)
22:16 Zoffix "This item has not been released yet and is not eligible to be reviewed"
22:16 Zoffix I've no idea idea what I'm doing.
23:04 travis-ci joined #perl6-dev
23:04 travis-ci Rakudo build failed. Zoffix Znet 'Revert "Revert potentially confusing "is default" changes"
23:04 travis-ci https://travis-ci.org/rakudo/rakudo/builds/259441532 https://github.com/rakudo/rakudo/compare/6e001867f215...7fdbb49ccc4b
23:04 travis-ci left #perl6-dev
23:04 buggable [travis build above] ? All failures are due to timeout (0), missing build log (0), GitHub connectivity (0), or failed make test (1). Across all jobs, only t/04-nativecall/12-sizeof.t test file failed.
23:09 BenGoldberg joined #perl6-dev
23:17 lizmat hmmm... do we have a value that will always return False when compared with eqv ?
23:21 geekosaur m: say (0/0) eqv (0/0)
23:21 camelia rakudo-moar 7fdbb4: OUTPUT: «False?»
23:22 lizmat hmmm...  :-)
23:23 geekosaur or maybe not
23:23 geekosaur m: my $a = (0/0); say $a eqv $a
23:23 camelia rakudo-moar 7fdbb4: OUTPUT: «True?»
23:23 lizmat meh
23:23 lizmat m: dd NaN eqv NaN
23:23 camelia rakudo-moar 7fdbb4: OUTPUT: «Bool::True?»
23:29 bisectable6 joined #perl6-dev
23:29 quotable6 joined #perl6-dev
23:29 committable6 joined #perl6-dev
23:29 evalable6 joined #perl6-dev
23:29 coverable6 joined #perl6-dev
23:29 greppable6 joined #perl6-dev
23:29 unicodable6 joined #perl6-dev
23:29 bloatable6 joined #perl6-dev
23:29 benchable6 joined #perl6-dev
23:29 statisfiable6 joined #perl6-dev
23:31 timotimo i've started reading my review copy of perl6 FUNdamentals but i've gotta go sleep
23:31 timotimo it's looking real good so far, though
23:31 timotimo moritz++
23:32 timotimo also, someone has to write a little pamphlet "Perl 6 Pundamentals"
23:32 timotimo maybe we can get Warry Lall to write a fun little foreword
23:32 lizmat woreford you mean ?
23:33 timotimo Of Course!
23:33 timotimo m: say "larry wall".trans(["r", "l"] => ["l", "r"])
23:33 camelia rakudo-moar 7fdbb4: OUTPUT: «rally warr?»
23:33 timotimo warr, huh, what is it good for?
23:34 lizmat uh!
23:34 timotimo nothing!
23:34 timotimo say it again y'all
23:35 timotimo hmpf. the cat is occupying a good 3/5ths of the bed again
23:35 AlexDaniel O_O what kind of a cat is that
23:36 timotimo well, the problem is less with the cat
23:36 timotimo you see, she's positioned herself quite close to the center
23:36 timotimo stretched out quite a bit
23:36 timotimo and when i get too close she tends to jump and run away
23:37 AlexDaniel ok, the problem is with the bed then? Should be bigger?
23:38 https_GK1wmSU joined #perl6-dev
23:39 AlexDaniel statisfiable6: bed size
23:39 statisfiable6 AlexDaniel, I cannot recognize this command. See wiki for some examples: https://github.com/perl6/whateverable/wiki/Statisfiable
23:39 AlexDaniel statisfiable6: libmoar0
23:39 statisfiable6 AlexDaniel, OK! Working on it…
23:40 timotimo haha
23:40 https_GK1wmSU left #perl6-dev
23:40 timotimo it's already a rather wide bed. before i moved in here i considered going for a smaller bed because my room is also smaller
23:40 statisfiable6 AlexDaniel, https://gist.github.com/80bac571cfb9db0e3cb40fdfeac11674
23:41 timotimo but i decided against buying a new bedframe + lath floor (??) + mattress
23:42 AlexDaniel hm it's nice to have these graph but nobody really needs them :)
23:43 geekosaur bed sizing: a human requires 2x their body size, a dog 3x, a cat 9x :p
23:43 timotimo i'll look at the graph again the next time we do something to make moar noticably smaller
23:43 timotimo geekosaur: that seems accurate
23:44 timotimo now she's looking at me and making some freaky noises
23:44 AlexDaniel sooo… humans are significantly more efficient?
23:44 timotimo nah, they just don't have standards
23:45 AlexDaniel .oO( then why would we need bigger toilets )
23:45 timotimo a cat knows she's worth a full king-sized bed
23:45 timotimo anyway, this being #perl6-dev, that's rather off-topic :)
23:51 sivoais joined #perl6-dev

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