Geth rakudo: MasterDuke17++ created pull request #1068:
Better error if punning role with required methods		 01:25
rakudo: MasterDuke17++ created pull request #1069:
Make errors in MultipleInheritance more awesome		 01:53
rakudo/nom: 84eb359903 | (Zoffix Znet)++ | src/core/IO/Handle.pm
Don't sink IO::Handle.close's value in .slurp

This lets users .slurp a ::Pipe of a failed proc without explosions and mirrors old .slurp-rest's behaviour.		 03:00
roast: 172e5dfd1d | (Zoffix Znet)++ | S32-io/io-handle.t
Test failed Proc can be .slurp(:close)'ed

Rakudo fix: github.com/rakudo/rakudo/commit/84eb359903		 03:01
Zoffix "MasterDuke+│ fwiw, slurp(:close) would cause `The spawned command '<...>' exited unsuccessfully (exit code: 2)`"
^ the above commit fixes that
MasterDuke_ Zoffix: cool 03:02
i can rebase my PR and add the :close back in 03:04
PR updated 03:08
Zoffix xkcd.com/1822/ 05:58
I'm gonna respond with that to all our tickets from now on.
Geth rakudo/nom: 6d28d7887e | MasterDuke17++ | src/Perl6/Metamodel/MultipleInheritance.nqp
Make errors in MultipleInheritance more awesome

Include relevant information where possible.		 06:11
rakudo/nom: c76a15d602 | niner++ (committed using GitHub Web editor) | src/Perl6/Metamodel/MultipleInheritance.nqp
Merge pull request #1069 from MasterDuke17/make_errors_in_MultipleInheritance_more_awesome

Make errors in MultipleInheritance more awesome
[Tux] This is Rakudo version 2017.04.3-33-gc76a15d60 built on MoarVM version 2017.04-44-gf0db8822 06:37
csv-ip5xs 3.035
test 13.567
test-t 5.116 - 5.176
csv-parser 13.074
Zoffix samcv: m// is not the same thing as rx// 07:10
samcv yeah
apparently
Zoffix It executes the match, whereas you need to give .match a regex. 07:11
samcv ah i see 07:12
is there a way we can throw?
Zoffix Yes. 07:16
Geth rakudo/nom: 1c21974d29 | (Zoffix Znet)++ | 3 files
Fix dispatch hang in Str.match: Nil
 07:20
roast: 0a67c4316e | (Zoffix Znet)++ | S05-substitution/match.t
Test Str.match: Nil does not hang
 07:21
samcv \o/ Zoffix 07:22
Geth rakudo/nom: cf733e963c | MasterDuke17++ | t/05-messages/01-errors.t
Add missing '.' to end of expected text
 11:42
rakudo/nom: 43f193ae59 | lizmat++ (committed using GitHub Web editor) | t/05-messages/01-errors.t
Merge pull request #1067 from MasterDuke17/fix_RT79288_tests

Add missing '.' to end of expected text for RT #79288 test
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...l?id=79288
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...l?id=79288
rakudo/nom: d87de586b6 | MasterDuke17++ | src/core/traits.pm
Better error if punning role with required methods

  `role R {method a{...};method b{say"b"};method c{...}}; class C is R {}`
used to say `Method 'a' must be implemented by R because it is required by roles: R.`, but now it says `Class C can't pun role R because it has required methods: a, c. Did you mean to use "does" instead?`.
Fixes RT #118263
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=118263
rakudo/nom: 32d4a36dd6 | lizmat++ (committed using GitHub Web editor) | src/core/traits.pm
Merge pull request #1068 from MasterDuke17/better_error_when_trying_to_is_a_role_with_required_methods

Better error if punning role with required methods
Zoffix . 12:11
yoleaux 07:33Z <eater> Zoffix: your "Thanks." looks very passive aggressive on PRs :')
jnthn Why do people think that?! :)
Zoffix Passive agressive thanks?
People are fucking weird.
jnthn Maybe some people are so attuned to lack of niceness online that when it happens they just look for an alternative explanation :P 12:12
brrt people have arms that are adapted to tree climbing and afterwards to tool holding. they are pretty weird, yes 12:13
stmuk it's a joke isn't it since it has a smiley on the end of it? 12:29
Zoffix I thought jokes were supposed to be funny 12:31
[Coke] Hello thought I'm dad. 12:46
(so, not always. :) 12:47
Zoffix m: with '/tmp/foo42'.IO { my $fh = .open(:w, :bin); $fh.spurt: Buf.new(200,200); $fh.close; dd .slurp(:bin) } 13:15
camelia Buf[uint8].new(200,200)
Zoffix Oh, nevermind.
Zoffix read that as `Buf[utf8]` first >_<
timotimo yeah, utf8 is already a buf-like (a blob, really) 13:16
Geth rakudo/nom: 6fa4bbcb6c | (Zoffix Znet)++ | src/core/io_operators.pm
[io grant] Straighten out &slurp/&spurt/&get/&getc/&close

  - Pass all args via captures, not just some
  - Fail, not throw
  - Add IO::Handle candidate for spurt, to maintain consistency
  - Change Cool:D slurp candidate to IO()
  - Add type constraints on subs that are missing it
   (IO::ArgFiles is IO::Handle)		 13:17
roast: 18d9c0441d | (Zoffix Znet)++ | S32-io/spurt.t
[io grant] Cover IO::Handle.spurt

With direct tests, rather than indirect tests via IO::Path.spurt.
MasterDuke_ anyone mind merging github.com/rakudo/rakudo/pull/1066 ? then i can close the RT 13:46
Geth rakudo/nom: d00de936d1 | MasterDuke17++ | t/05-messages/01-errors.t
Add tests for RT #113954

Spaces or quotes in literal params to MAIN() are quoted correctly in the useage statement.		 13:48
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=113954
rakudo/nom: a87d708e30 | MasterDuke17++ | t/05-messages/01-errors.t
Make sure to :close the output when slurp()ing it
rakudo/nom: 64c303297f | (Zoffix Znet)++ (committed using GitHub Web editor) | t/05-messages/01-errors.t
Merge pull request #1066 from MasterDuke17/tests_for_RT113954_fix

Add tests for RT #113954
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=113954
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=113954
MasterDuke_ Zoffix++ thanks 13:50
Geth nqp: cee4a7ed16 | (Pawel Murias)++ | t/nqp/083-math.t
Test nqp::cos with infs
 13:58
nqp: 485855ad5e | (Pawel Murias)++ | src/vm/js/nqp-runtime/runtime.js
[js] Less annoying js exception handling
nqp: d2eb8081d4 | (Pawel Murias)++ | src/vm/js/Compiler.nqp
[js] Fix using locals in pre_deserialize/post_deserialize
rakudo/nom: 34b58d1b5e | (Zoffix Znet)++ | 2 files
[io grant] Straighten out &lines/&words

  - Just pass all args via captures; don't do own processing
  - Make &words default to $*ARGFILES, just like the rest of the routines
   in this group.
  - Add rudimentary IO::ArgFiles.words impl; will be replaced by IO::CatHandle		 14:24
roast: 8f78ca604c | (Zoffix Znet)++ | S16-io/words.t
[io grant] Test &words with IO::ArgFiles
japhb My full-screened chat window is showing an almost solid wall of Geth. Must be a productive morning 'round these parts. 14:58
:-)
Zoffix buggable: speed 30
buggable Zoffix, ▃▃▁▃▁▂▇▅▂▁▃▃▃▃▃▃▄▃▃█▄▆▄▄▄▄▆▅▄▃ data for 2017-04-12–2017-04-24; range: 4.994s–5.479s; 1% faster
Zoffix buggable: speed 10
buggable Zoffix, ▂▄▃▃▃▂▄▃▂▁ data for 2017-04-21–2017-04-24; range: 5.116s–5.351s; 1% faster
[Coke] what do those mean? 14:59
is there a help link?
Zoffix Don't think so.
buggable: speed
buggable Zoffix, █▂▃▂▂▁▂▂▂▃▆▂▆▃▆▂▃▂▂▂▄▄▃▄▃▃▆▅▃▂▄▃▄▄▃▃▄▃▃▇▄▅▅▄▄▄▅▅▄▃ data for 2017-04-03–2017-04-24; range: 4.829s–5.646s; 10% faster
japhb [Coke]: It's a sparkline graph of the last N days of [Tux]'s benchmark.
Zoffix ^ that's the data from [Tux]'s bench. And if you give it N in the command, it uses only last N entries
japhb And I think the "faster" bit is last versus first within the sparkline 15:00
Zoffix Not N days, last N entries. There are at least two measurements per day
Yes, the faster is 1st vs last
japhb Zoffix: Ah, that's new. Fair enough.
Zoffix Gonna try avoiding all the IO() coercers and adding a multi for IO::Path:D to them; that's the only thing I can think of as the cause of recent slowage 15:01
[Coke] shorter is better? 15:03
Zoffix Yes 15:04
[Coke] Thanks. 15:05
MasterDuke_ buggable: bench 15:08
timotimo huggable: speed 15:09
huggable timotimo, tux.nl/Talks/CSV6/speed4.html
MasterDuke_ isn't there also a buggable command to show how to run the bench yourself? 15:10
oh, you used huggable 15:12
huggable: bench
huggable MasterDuke_, nothing found
Zoffix huggable: bench 15:14
huggable Zoffix, nothing found
Zoffix huggable: bench csv
huggable Zoffix, export PATH=`pwd`/install/bin:$PATH; cd CSV; for i in $(seq 1 10000); do echo 'hello,","," ",world,"!"'; done > /tmp/hello.csv; time perl6 -Ilib -MText::CSV test-t.pl </tmp/hello.csv
MasterDuke_ thanks
Geth roast: dc7d1a7dbb | (Pawel Murias)++ | S02-literals/radix.t
[js] Fudge unicode problem
 15:18
roast: 3569cfd1bf | (Pawel Murias)++ | S02-names/pseudo.t
Fix test

CALLER only looks in the immediate caller's scope
roast: 29c8a9da4a | (Pawel Murias)++ | S02-names/pseudo.t
Test CALLERS
Zoffix ZOFFLOP: t/spec/S11-modules/nested.t 15:23
ZOFFLOP: t/spec/S10-packages/use-with-class.t 15:24
MasterDuke_ use-with-class.t, i don't remember seeing that ever flop before?
Geth rakudo/nom: d0cd137294 | (Zoffix Znet)++ | src/core/io_operators.pm
[io grant] Make dir take any IO(), not just Cool

To maintain consistency with all the other routines
Zoffix Nope. IO() coercer isn't it. It's only really present on a few routines. 15:33
ZofBot: must be solar flares
ZofBot Zoffix, it's kind of hard to get a good result when it already switches directions at 4
Geth nqp/master: 4 commits pushed by (Pawel Murias)++ 15:52
Zoffix m: with "/tmp".IO.add(rand) { .spurt: "abc"; with .open { dd .getc; dd .tell; dd .getc; dd .tell } } 16:06
camelia "a"
2
"b"
3
Zoffix seems off, isn't it?
m: with "/tmp".IO.add(rand) { .spurt: "abc"; with .open(:bin) { dd .read(1).decode; dd .tell; dd .read(1).decode; dd .tell } } 16:07
camelia "a"
1
"b"
2
Zoffix m: with "/tmp".IO.add(rand) { .spurt: "abc"; with .open { dd .readchars(1); dd .tell; dd .readchars(1); dd .tell } }
camelia "a"
2
"b"
3
Zoffix Also bustified like .getc
geekosaur need to think about how the unget affects things 16:08
also is that exposed, such that someone could unget a different thing than was read (C stdio allows this)? which means the file offset is in fact distinct and should behave like that 16:09
Zoffix Don't think we have an unget 16:10
cygx Zoffix: binary and char-based routines don't mix well
ages ago, I started doing something about it: github.com/MoarVM/MoarVM/pull/319/files , but it did not go anywhere
Zoffix cygx: the implication being that .tell is binary? 16:11
geekosaur effectively yes
seeking and non-byte/record IO don't mix very well
in any language
Zoffix :/
geekosaur I mean, you can try, but contemplate for example the difference between seek/tell and setpos/getpos in POSIX 16:12
one might argue that tell should be left as is and a getpos/setpos type thing should be used to paper over the utf8 readahead the same way it is used to paper over newline differences 16:13
but it will be imperfect at best 16:14
cygx geekosaur: why? in case of random access files (and ignoring concurrent access), perfection is possible in principle 16:15
geekosaur is it? can you reliably get to the next position from the current one and the contents of a string you have read? 16:16
(tail is an example of a program that wants to do this)
Geth roast: ea137f64d3 | (Zoffix Znet)++ | S32-io/tell.t
[io grant] Cover IO::Handle.tell

Closes RT#123838: rt.perl.org/Ticket/Display.html?id=123838		 16:17
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=123838
cygx not sure I understand the scenario: the OS tracks the position in the file, and the decoding buffer knows how many bytes it has read ahead 16:20
geekosaur but does the user code know, or can it get at that information? 16:21
that's the specific difference between seek/tell and setpos/getpos
or does the lanbguage try to paper it over and can it do so with 100% reliability (i.e. knowing the user program didn't decide to seek by some different amount instead)? 16:22
cygx what I woud provide are byte-oriented seek/tell methods and the ability to skip (from current position) in any of the unites of interest (bytes, codepoints, graphemes) 16:29
Zoffix is burnt out 16:42
Gonna take a break until next weekend.
Ping me if need anything
Zoffix & 16:43
cygx have fun 16:44
cygx isn't burnt out, but needs to get groceries
MasterDuke_ if anyone is interested, here is a profile of test-t.pl: gist.github.com/MasterDuke17/0f579...463cd6c8b8 16:52
Geth nqp: 5c29d1840c | (Pawel Murias)++ | 2 files
[js] Implement nqp::setdispatcherfor
 18:02
nqp: 43c520138c | (Pawel Murias)++ | t/nqp/100-dispatcher.t
Test nqp::setdispatcherfor
nqp: 05b51d4564 | (Pawel Murias)++ | 4 files
[js] Run fixjsstyle
rakudo: MasterDuke17++ created pull request #1070:
Add test for RT #118263 fix		 19:02
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=118263
stmuk timotimo: is it possible to disable telemeh from being compiled? I ask since "x86intrin.h" isn't present on old GCC versions it seems 19:43
I mean is there a compile time option?
timotimo not yet 19:48
i can put a define in 19:49
i.e. put a -DNO_TELEMEH
in the cflags
nit sure how to work the configure.pl script
would be best if there wqs a probe in the configure script 19:50
but dont know how to
stmuk that might help on platforms such as OpenBSD with its ancient gcc 4.2.1
timotimo damn
stmuk I have a old note of rakudobrew build moar --configure-opts=--moar-option=--cc=clang36 19:56
I assume a -DNO_TELEMEH might be passed in a similar way .. at least as a start 19:57
I could probably work out how to add that into the Configure.pl 19:58
timotimo actually, no, not just like that 20:00
but i'm about to write a probe for this 20:01
stmuk ok
dogbert17 are the versions of 'say' called on lines 517 and 523 different or the same, i.e. is say != nqp::say ? 20:17
oops, github.com/perl6/nqp/blob/master/s...r.nqp#L515 20:18
MasterDuke_ i think nqp's say() is just sugar for nqp::say() 20:19
timotimo it's probably a sub in nqp's setting though 20:20
MasterDuke_ yeah, like note(). which is why i also think it doesn't work in some places, if it's too early in the nqp compilation 20:21
dogbert17 MasterDuke_, timotimo: thx, have just written this gist.github.com/dogbert17/f97842d6...95b0409155 20:22
if it doesn't look like junk it will be part of a PR for RT #130760 20:24
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=130760
timotimo yeah, if you don't have note, you need to use nqp::sayfh(nqp::getstderr, "foobar") 20:27
dogbert17 when I tried the code it definitely looked as if note was available
timotimo stmuk: can you review my attempt at making --no-telemeh/--telemeh and also "probe if rdtscp exists" exist? 20:39
Geth nqp: dogbert17++ created pull request #355:
Fix for RT #130760		 20:42
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=130760
timotimo stmuk: i pushed a branch to moarvm, i called it configure-telemeh-probe-rdtscp 20:43
stmuk: it's not much perl5 i wrote, but i did write a bit of perl5, and i'm not experienced with perl5 whatsoever 20:47
Geth rakudo/nom: 956a880556 | MasterDuke17++ | t/05-messages/01-errors.t
Add test for RT #118263 fix

The message when trying to pun a role with required methods should have the names of the child, parent, required methods, and suggest "does".		 20:48
rakudo/nom: c0c9a7d7a1 | MasterDuke17++ | t/05-messages/01-errors.t
Check for desired parts of message, not exact text
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=118263
rakudo/nom: f0414c4681 | lizmat++ (committed using GitHub Web editor) | t/05-messages/01-errors.t
Merge pull request #1070 from MasterDuke17/add_test_for_RT118263_fix

Add test for RT #118263 fix
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=118263
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=118263
lizmat and another Perl 6 Weekly hits the Net: p6weekly.wordpress.com/2017/04/24/...ing-times/ 21:46
timotimo hm. can we get %?RESOURCES not break apart? 21:52
lizmat tries 21:55
alas 21:58
suggestions ?
timotimo no clue
lizmat .u question
yoleaux U+003F QUESTION MARK [Po] (?)
U+00BF INVERTED QUESTION MARK [Po] (¿)
U+037E GREEK QUESTION MARK [Po] (;)
lizmat rephrased the line so it should come out better (usually) 22:00
fwiw, in my preview, it did not break up :-(
timotimo yeah, i tend to view websites at some amount of zoomage 22:01
lizmat good night, #perl6-dev! 22:04
timotimo gnite lizmat 22:05
lizmat++ #weekly
Zoffix lizmat++ good weekly 22:09
yoleaux 20:29Z <El_Che> Zoffix: I'll release updated 2017.03 packages tomorrow with Windows 10 bash support. I'll wait with 2014.04 until npq get released.

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!