»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, std:, or /msg camelia p6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by masak on 12 May 2015.
cognominal m: sub a($a) { }; try { a() } # very unintuitive error message 00:01
camelia rakudo-moar 8957fc: OUTPUT«===SORRY!===␤Cannot find method 'orig'␤»
00:03 khw joined 00:04 nys joined 00:05 bjz left
cognominal sounds like somewhere an Match or Cursor is expected but something got in its place due to the try! 00:05
00:06 Ven left, captain-adequate left 00:08 Possum joined 00:16 mohij joined 00:18 telex left 00:20 kaare_ joined, telex joined, mohij_ left 00:34 pat_js left 00:35 BenGoldberg joined 00:50 atroxaper joined 00:51 vendethiel left 00:52 integral left 00:53 integral joined, integral left, integral joined 00:55 atroxaper left 01:02 yeahnoob joined
japhb seen flussence 01:06
.seen flussence
yoleaux I saw flussence 13 Jul 2015 17:20Z in #perl6: <flussence> wasn't writing it "do given" supposed to be a temporary workaround?
TimToady yay, plutohazblug :)
well, microblug, anyway 01:08
japhb .tell flussence Text-Tabs-Wrap seems rather unhappy: I can't install it with panda, and it seems to be failing quite a few tests. 01:09
yoleaux japhb: I'll pass your message to flussence.
01:15 yeahnoob left 01:17 rmgk_ joined, rmgk is now known as Guest79273, Guest79273 left, rmgk_ is now known as rmgk, kaare_ left 01:18 vendethiel joined 01:20 bowtie joined
dalek kudo-star-daily: 952076f | coke++ | log/ (9 files):
today (automated commit)
 01:20
01:20 bowtie is now known as Guest90936
dalek rl6-roast-data: eca0883 | coke++ | log/ (4 files):
Move old test files to archive/
 01:24
01:25 gfldex left 01:39 vendethiel left 01:46 ilbot3 joined 01:48 dayangkun joined 01:49 dayangkun left 01:50 colomon joined, dayangkun joined 01:51 dayangkun left, dayangkun joined 01:52 dayangkun left, dayangkun joined 01:53 dayangkun left 01:54 dayangkun joined, dayangkun left 01:56 dayangkun joined 01:57 prammer joined, dayangkun left 02:00 dayangkun joined 02:01 dayangkun left 02:07 dayangkun joined, dayangkun left 02:08 BenGoldberg left, dayangkun joined 02:09 dayangkun left 02:10 noganex_ joined 02:12 dayangkun joined 02:13 noganex left, dayangkun left, dayangkun joined
ugexe `use Test; BEGIN plan 1; use ThisDoesntExist; ok 1;` !== 0 exitcode. -- `use Test; plan 1; use ThisDoesntExist; ok 1` == 0 exitcode 02:15
alternatively putting 'use ThisDoesntExist" before 'use Test' will correctly exit with non-zero without the BEGIN 02:16
02:19 dayangkun left 02:20 dayangkun joined, vendethiel joined 02:21 yqt left
ugexe if you remove the END { } block in Test.pm it gives correct exit codes in all of those examples 02:25
02:37 aborazmeh joined, aborazmeh left, aborazmeh joined 02:41 vendethiel left
ugexe i had hoped BEGIN my $started; $started = 0; ... END { unless $started.DEFINITE { exit 255 } }; in Test.pm would get around that, but alas i was wrong 02:47
02:50 pyro- left 02:51 vendethiel joined 02:53 kaare_ joined 03:01 AlexDaniel left 03:03 nys left 03:07 [Sno] left 03:12 vendethiel left 03:36 aborazmeh left 03:39 atroxaper joined 03:41 laouji left 03:44 atroxaper left 03:52 vendethiel joined 03:58 atroxaper joined, davido___ left 03:59 davido___ joined 04:13 vendethiel left 04:17 kurahaupo joined 04:19 geekosaur joined 04:29 prevost joined 04:33 kurahaupo left, kurahaupo joined 04:34 ggoebel2 joined, ggoebel left 04:37 laouji joined
skids appends a cherrypickable fix for the above backtrace issue to PR#470. It was introduced in d8ce4cd47 in May. 04:38
04:43 skids left 04:47 prevost left 04:48 laouji left, laouji joined 04:52 kurahaupo left 04:53 kurahaupo joined 05:00 [Sno] joined
dalek kudo/nom: 22f7192 | lizmat++ | src/core/ (52 files):
Use colon syntax for named parameters consistently
 05:02
05:02 khw left
dalek ast: ed09f71 | lizmat++ | S09-typed-arrays/native- (2 files):
Tweak typed error checks
 05:05
05:06 kurahaupo left 05:09 mohij left 05:12 mohij joined, vendethiel joined, mr-foobar left
lizmat Looking at #125614 now: 05:14
m: my %xy{Mu}; %xy{1&2} = 3; say %xy.perl # wrong 05:15
camelia rakudo-moar 8957fc: OUTPUT«Hash[Any,Mu].new(1 => 3, 2 => 3)␤»
lizmat m: my %xy{Mu}; %xy.AT-KEY(1&2) = 3; say %xy.perl # is this right?
camelia rakudo-moar 8957fc: OUTPUT«Hash[Any,Mu].new(all(1 => 3, 2 => 3))␤»
lizmat m: my %xy{Mu}; %xy.ASSIGN-KEY(1&2,3); say %xy.perl # is this right?
camelia rakudo-moar 8957fc: OUTPUT«Hash[Any,Mu].new(all(1 => 3, 2 => 3))␤»
lizmat m: my %xy{Mu}; %xy.BIND-KEY(1&2,3); say %xy.perl # and this? 05:16
camelia rakudo-moar 8957fc: OUTPUT«Hash[Any,Mu].new(all(1 => 3, 2 => 3))␤»
ShimmerFairy Hm. Unless there's something in the specs stating otherwise, I'd think %xy{1&2} = 3 would make 1 => 3, 2 => 3 , since it is a Junction 05:17
lizmat m: my %xy{Any}; %xy{1&2} = 3; say %xy.perl # and this?
camelia rakudo-moar 8957fc: OUTPUT«Hash[Any,Any].new(1 => 3, 2 => 3)␤»
ShimmerFairy m: sub foo($a) { say $a }; foo(1&2)
camelia rakudo-moar 8957fc: OUTPUT«1␤2␤»
ShimmerFairy ^ that's my reason for thinking {1&2} is the one doing the right thing. 05:18
lizmat not if it is a typed hash with Mu: then the autothreading should not work (is what conception of the problem is_
) 05:19
ShimmerFairy Oh, I see. Well, I think in that case it should be any(1, 2) => 3 , since the Junction is the key (I don't think all(1 => 3, 2 => 3) is right, and I don't see how it could be immediately) 05:20
05:25 kurahaupo joined
lizmat added this discussion and some other observations to the ticket 05:32
05:35 vendethiel left 05:42 vendethiel joined
dalek kudo/nom: 52458b5 | lizmat++ | src/core/JSON/Pretty.pm:
Small string escape optimization
 05:54
05:55 kurahaupo left 06:01 coffee` left, ribasushi left 06:02 ribasushi joined, yeahnoob joined 06:04 vendethiel left
dalek kudo/nom: 81d569f | lizmat++ | src/core/ (3 files):
Remove trailing whitespace

This dif is git diff --ignore-space-at-eol clean.		 06:06
06:07 vendethiel joined
dalek kudo/nom: 8e5525e | lizmat++ | src/core/ (3 files):
DeTABify core settings
 06:10
06:11 ribasushi left 06:12 ribasushi joined 06:22 atroxaper left 06:29 mr-foobar joined 06:30 xinming joined 06:34 atroxaper joined
masak goooood morning, #perl6! 06:36
06:36 xinming left 06:37 xinming joined
moritz good morning, masak-san 06:45
lizmat masak moritz o/ 06:47
masak is there a mechanism in Rakudo to read a keystroke without blocking? 06:49
lizmat getc on $*IN ?
hmmm... guess not
06:49 domidumont joined
lizmat that blocks on newline 06:50
masak yeah, just noticed the same.
moritz sounds like a bug
masak would this be a case where Inline::Perl5 would help?
06:51 _mg_ joined, domidumont left
lizmat moritz: we don't have autoflush yet, afaik 06:51
that would probably help
06:51 domidumont joined
moritz lizmat: uhm, autoflush is for output, no? 06:51
lizmat hmmm... you're right 06:52
masak is by no means an expert, but suspects that even if `getc` returned immediately after the first character it got, it (a) would return a *character*, not a keystroke, and (b) would be blocking
lizmat guess it's getting time for me to go to bed...
masak lizmat: sleep tight 06:53
lizmat not there yet :-) 06:54
masak heh :) 06:55
06:55 mkz joined 06:57 RabidGravy joined 06:58 FROGGS joined
ugexe is there a way to do something crazy like pass a line of code like -e, except used like -M? easier to explain this way: perl6 -M'module :: { END { say "end" } };' script.pl 07:00
07:02 espadrine joined 07:08 laouji left
moritz ugexe: not that I'm aware of 07:11
07:12 darutoko joined 07:15 virtualsue joined 07:17 vendethiel left 07:18 atroxaper left 07:20 atroxaper joined 07:21 laouji joined
masak I'm alone at the office today. the motion-sensor activated overhead lights keep switching off, about once every 20 minutes. 07:24
kind of nice. reminds me to get up and stretch or move about. :P 07:25
ugexe: I suspect you could achieve much the same effect by writing a small wrapper script that evals some input file and then runs script.pl 07:26
dalek kudo/nom: 0493008 | lizmat++ | src/core/Temporal.pm:
Sanitize %UNITS out of global namespace
lizmat m: say %UNITS # soon to be gone
camelia rakudo-moar 8e5525: OUTPUT«day => 1, days => 1, hour => 1, hours => 1, minute => 1, minutes => 1, month => 1, months => 1, second => 1, seconds => 1, week => 1, weeks => 1, year => 1, years => 1␤»
masak TimToady: what if a macro calls a macro? would $COMPILING::x go one level up, or all the way? in the former case, would $COMPILING::COMPILING::x find $x in the mainline? 07:28
also, after reading the feedback from PerlJam and jnthn in yesterday's hygiene discussion, plus (especially) vendethiel's links, I'm back to thinking we have the right default with everything being hygienic by default. 07:32
"strong opinions, weakly held" :P
07:32 TimToady joined
masak especially since -- I realized -- it's not a very big step to opt into almost complete unhygiene. 07:32
`quasi :COMPILING { ... }` -- done.
DrForr Looks like I poked my proverbial nose back in at the right time :) 07:34
masak "The compiler says sorry for _your_ moronic typo!" -- serve, smash, and pun by jnthn++ in jnthn.net/papers/2015-fosdem-static-dynamic.pdf 07:35
lizmat m: say Date.today.is-leap-year(2012) # unrelated, but does it make sense to allow class method behaviour on instances ? 07:36
camelia rakudo-moar 8e5525: OUTPUT«True␤»
lizmat m: say Date.today.is-leap-year # as an instance method
camelia rakudo-moar 8e5525: OUTPUT«False␤»
lizmat m: say Date.is-leap-year(2012) # as a class method
camelia rakudo-moar 8e5525: OUTPUT«True␤»
ShimmerFairy I would say so. I don't know of a language that specifically disallows them. 07:37
lizmat m: say Date.is-leap-year # so we take this as collatetal damage ? 07:39
camelia rakudo-moar 8e5525: OUTPUT«Cannot look up attributes in a type object␤ in block <unit> at /tmp/8rD18_PnlM:1␤␤»
lizmat *collateral
07:42 diana_olhovik joined
lizmat perhaps this is a better error message: 07:44
$ 6 'say Date.is-leap-year'
Cannot call is-leap-year(Date: ); none of these signatures match:
(Dateish:D $: *%_)
(Dateish $: $y, *%_)
07:46 vendethiel joined
ShimmerFairy Agreed, that'd help explain it better. (And would help any attempt to call an instance method on a type object) 07:46
07:47 zakharyas joined
masak lizmat: I almost think `is-leap-year` is a *sub*. and should be called with Date::is-leap-year or something. 07:52
lizmat: it doesn't need an invocant, neither an instance nor a class one. 07:53
lizmat well, Date.foo is one character less than Date::foo
in P5 there wouldn't have been any difference in number of chars 07:54
personally, I don't like the Module::foo syntax for getting at what basically is a sub
especially since it would need to be an "our" to be visible outside 07:55
07:56 amurf left 07:57 amurf joined 07:58 abraxxa joined 08:00 ohb joined 08:01 amurf left 08:03 dakkar joined, dakkar left 08:04 dakkar joined
RabidGravy it appears that IO::Socket::INET.recv hides "Socket reset by peer" (and probably others) and instead just doesn't return anything 08:06
08:08 vendethiel left
lizmat that would explain the spectest failures 08:09
RabidGravy it manifested in an issue against HTTP::UserAgent 08:17
dalek kudo/nom: 0d660bb | lizmat++ | src/core/Temporal.pm:
Streamline is-leap-year and days-in-month

Don't use junctions and give better error message when trying to use the instance mode of the method on a class.		 08:22
lizmat and now it *is* time for me to go to bed :-) 08:23
RabidGravy :)
lizmat so good night, #perl6!
RabidGravy nighty night 08:28
08:30 dakkar left, dakkar joined 08:31 abraxxa left, abraxxa joined 08:32 Woodi left 08:35 rindolf joined 08:37 atroxaper left, espadrine left, Woodi joined 08:50 yeahnoob left 08:52 vendethiel joined
jnthn morning, #perl6 08:52
08:52 rindolf left
cdc Hello #perl6 08:57
kanl m: class Foo is List { method load( $f ) { self.push: $f } }; my $f = Foo.new.load: 0; $f.WHAT.say; my @f = ( 1 .. 2 ).map: { Foo.new.load: $_ }; @f.map: { .WHAT.say };
camelia rakudo-moar 0d660b: OUTPUT«(Foo)␤(Int)␤(Int)␤»
jnthn lizmat: Oh great, 22f7192269 is pointless and makes it harder to git blame stuff, and no, reverting won't help, just make things worse. :/
kanl m: class Foo is List { method load( $f ) { self.push: $f } }; my $f = Foo.new.load: 0; $f.WHAT.say; my @f = ( 1 .. 2 ).map: { $ = Foo.new.load: $_ }; @f.map: { .WHAT.say }; 08:58
camelia rakudo-moar 0d660b: OUTPUT«(Foo)␤(Foo)␤(Foo)␤»
kanl gurus, why do i need '$ =' ?
am i not inheriting in the right fashion? 08:59
cdc For some (stupid) reasons I can't connect to my Github account, so is there someone here that wishes to apply this patch to perl6/doc for me: gist.github.com/anonymous/660b9f7e778b9e7b0c82 (thanks in advance!) 09:01
09:01 atroxaper joined
jnthn lizmat: And looking at the result, it only makes me certain you're wrong in wanting "foo => 'bar'" not to be a named arg... 09:02
It's not really a net readability improvement. 09:03
09:05 espadrine joined 09:09 yeahnoob joined
dalek c: 0bcf8fc | RabidGravy++ | lib/Language/nativecall.pod:
Apply "Initial documentation for CUnion + embedded structures." from cdc
 09:09
RabidGravy cdc, ^^ done
cdc RabidGravy: thanks! 09:10
RabidGravy all good fun 09:11
FROGGS jnthn: I also prefer foo => bar when I can lineup the fat arrows for example 09:12
jnthn Yeah, me too
FROGGS when I write args in a single line I might prefer colonpairs
but, well, that's just style 09:13
jnthn I like them sometimes too, but I like having the choice.
FROGGS - fail X::Cannot::Infinite.new(:action<.push>, :what(self.^name)) 09:14
+ fail X::Cannot::Infinite.new(:action<push>, :what(self.^name))
was that intentional?
jnthn Yeah, that's the other thing - the patch changed more than it said in the subject line. But I thought I'd already beat lizmat++ up enough with my morning oh-shit-my-eyes-sting grumpiness. :) 09:15
FROGGS I'm just pasting that here so she can review... I can't tell if the said change is okay or not, be she will know and can act if needed 09:16
jnthn *nod*
I'm going to do Perl 6 things today :)
Or at least, most of the day 09:17
cdc jnthn: \o/
jnthn Dammit, back up to 1,036 tickets. :( 09:18
FROGGS :o(
ups, the changes to Temporal.pm change meaning too 09:25
- range => (0..^60),
+ :range('0..^60'),
and in Variable.pm: 09:28
- expected => <TypeObject default dynamic>,
+ :expected('TypeObject default dynamic'),
this also changes meaning but probably fixes a thinko: 09:29
- :feature<is default(*) on natives> )
+ :feature('is default(*) on natives'))
jnthn std: Yeah 09:30
camelia std 28329a7: OUTPUT«5===SORRY!5===␤Undeclared name:␤ 'Yeah' used at line 1␤Check failed␤FAILED 00:00 134m␤»
jnthn oops
FROGGS: Yeah
std: say 1 if;
FROGGS :P
camelia std 28329a7: OUTPUT«5===SORRY!5===␤Confused at /tmp/CoIjZ9XKQK line 1:␤------> 3say 1 7⏏5if;␤ expecting feed_separator␤Parse failed␤FAILED 00:00 135m␤»
jnthn std: repeat { } while; 09:31
camelia std 28329a7: OUTPUT«5===SORRY!5===␤Whitespace required after keyword at /tmp/0WWfGJEcts line 1:␤------> 3repeat { } while7⏏5;␤Parse failed␤FAILED 00:00 135m␤»
FROGGS m: say (:feature("Variable trait 'will final {...}'"))
camelia rakudo-moar 0d660b: OUTPUT«Unhandled exception: Stub code executed␤ at <unknown>:1 (/home/camelia/rakudo-inst-1/share/perl6/runtime/CORE.setting.moarvm:throw:4294967295)␤ from src/gen/m-CORE.setting:17346 (/home/camelia/rakudo-inst-1/share/perl6/runtime/CORE.setting.moarvm:t…»
09:31 salva joined
FROGGS there is a bug in there 09:31
09:32 atroxaper left
jnthn Yes :) 09:33
09:34 atroxaper joined, vendethiel left 09:35 yeahnoob left
FROGGS there is also the "foo => <bar baz>" => ":foo<'bar baz'>" change in traits.pm 09:36
when thinking about these not intentional changes I guess we have to revert? 09:38
and then fix the bug about the stub code that existed before?
jnthn *nod*
Yes, that one means something difference
Would have to be :foo<<'bar baz'>> to work
FROGGS okay
09:39 TEttinger left
FROGGS .tell lizmat I'm going to revert your formatting patch, it introduces changed meaning (some bugs, one fix) mentioned here on irc 09:40
yoleaux FROGGS: I'll pass your message to lizmat.
dalek kudo/nom: 270d179 | jnthn++ | src/Perl6/Grammar.nqp:
Improve error for malformed loop statement.

Brings us in line with STD.
09:40 vendethiel joined
dalek ast: e4eb44a | jnthn++ | S32-exceptions/misc.t:
Test for RT #125595.
 09:41
ast: c3f0861 | jnthn++ | S (6 files):
Merge branch 'master' of github.com:/perl6/roast
FROGGS ahh, dang, I've got unstaged changes... will revert later
09:42 Hor|zon left
FROGGS m: .perl.say for (foo => <bar baz>), (:foo<'bar baz'>), (:foo<<'bar baz'>>) 09:42
camelia rakudo-moar 0d660b: OUTPUT«:foo($("bar", "baz"))␤:foo($("'bar", "baz'"))␤:foo("bar baz")␤»
jnthn FROGGS: if you have time you could "git revert -p ..."
FROGGS: That will let you yes/no to the bits you want to revert.
And so can just pick out those that are behavior-changey 09:43
FROGGS ahh, good idea
will do that
jnthn loves git's -p
FROGGS .oO( so jnthn loves p of a git ) 09:44
FROGGS hides
09:44 diana_olhovik left
arnsholt I love git in general 09:45
09:46 telex left
arnsholt Today even more so, since I got to use git to handle an SVN branch merge 09:46
(New git branch, rebase onto master, then merge. Soooo much easier)
09:48 telex joined
jnthn discovers that "perl6 does-not-exist" has been reporting is error on stdout, not stderr 09:49
09:50 CIAvash joined
RabidGravy well the RPi2 has exposed a bug in Linux::Fuser (both P6 and P5 versions) 09:50
09:53 lizmat left
jnthn m: sub foo($!a) { } 09:54
camelia rakudo-moar 0d660b: OUTPUT«5===SORRY!5=== Error while compiling /tmp/gcHwOGjanS␤cannot use a $! parameter in a signature where no $?CLASS is available␤at /tmp/gcHwOGjanS:1␤------> 3sub foo($!a7⏏5) { }␤ expecting any of:␤ shape declaration␤»
jnthn m: class A { sub foo($!a) { } } 09:55
camelia rakudo-moar 0d660b: OUTPUT«5===SORRY!5=== Error while compiling /tmp/2wI4QkdHyr␤Attribute $!a not declared in class A␤at /tmp/2wI4QkdHyr:1␤------> 3class A { sub foo($!a) { } }7⏏5<EOL>␤»
jnthn m: class A { has $!a; sub foo($!a) { } }
camelia ( no output )
jnthn m: class A { has $!a; sub foo($!a) { }; foo(1) }
camelia rakudo-moar 0d660b: OUTPUT«Cannot look up attributes in a type object␤ in sub foo at /tmp/O4uOwHMTmn:1␤ in block <unit> at /tmp/O4uOwHMTmn:1␤␤»
dalek p: 35f6d68 | jnthn++ | src/HLL/Compiler.nqp:
Improve failure to run script errors.

Add a special case (and clear message) for directory, avoid spewing a pointless backtrace, and always report on stderr.		 09:57
09:58 Hor|zon joined 10:01 vendethiel left, gfldex joined
jnthn m: use Test; throws-like 'my class C { has $.x; submethod BUILD(:$.x) {} }', X::Syntax::VirtualCall; 10:03
camelia rakudo-moar 0d660b: OUTPUT« 1..2␤ not ok 1 - 'my class C { has $.x; submethod BUILD(:$.x) {} }' died␤ ␤# Failed test ''my class C { has $.x; submethod BUILD(:$.x) {} }' died'␤# at /tmp/6SvqO46Ab3 line 1␤ ok 2 - # SKIP Code did not die, can not check exception…»
cognominal I like the comment following Larry interview: "his tangents [go] back at the maim (sic) point". Beware, TimToady in not a straight shooter with his elliptic tangents that strike back in most expected ways :) 10:07
Probably not intended, but a very perlish pun. 10:09
itz is there anyway of specifying a list of required params in a sub signature as one and only one of these list of parameters are required?
moritz itz: no 10:12
10:13 vendethiel joined
jnthn You could hack it with a where clause though if you're willing to list them another time 10:14
10:15 atroxaper left
RabidGravy brane is struggling, "subset RawEncode of Array where (*.elems == 2 ) && (*[0] ~~ CArray[uint8]);" doesn't seem to do what I expect 10:19
m: use NativeCall; subset RawEncode of Array where (*.elems == 2 ) && (*[0] ~~ CArray[uint8]); sub foo() returns RawEncode { [CArray[uint8].new,1] }; foo(); 10:20
camelia rakudo-moar 0d660b: OUTPUT«Type check failed for return value; expected 'RawEncode' but got 'Array'␤ in sub foo at /tmp/Do_cPDUcTD:1␤ in block <unit> at /tmp/Do_cPDUcTD:1␤␤»
jnthn YOu don't need the * there, the thing to check against is already in $_
RabidGravy ah
perfecto!
10:30 skids joined
vendethiel m: say .[0] for 3; 10:38
camelia rakudo-moar 270d17: OUTPUT«3␤»
10:38 atroxaper joined 10:43 dayangkun left 10:45 zakharyas left 10:48 atroxaper left, mr-foobar left 10:49 mr-foobar joined 10:52 laouji left
dalek kudo/nom: 63ebb34 | jnthn++ | src/Perl6/ (2 files):
Add errors for bougs $!x and $.x parameter use.

Catch $.x usage in a submethod, plus catch sub ($!x) { } even when the sub is inside of a class. This took a bit more doing than first expected, since 'has ($.x, $.y)' is also parsed using a signature.		 10:55
ast: f21cb19 | jnthn++ | S12-methods/attribute-params.t:
Tests for RT #125591.
 10:56
arnsholt jnthn: "bougs"? I thought you moved to Prague, not Paris? ^_^
jnthn :P
nwc10 maybe he did. Paris isn't know for its beer, after all.
if so, the Balmer Peak was <--- thataway :-) 10:57
jnthn The only French city I've been to that I remember having decent beer is Strasbourg :)
arnsholt Then again, Alsace is Germany, 'cept it's in France
(To the extent that there are certain laws that only apply in Alsace) 10:58
nwc10 I wasn't aware of that special case 10:59
arnsholt It's basically due to the whole "is it French or is it German" thing. When the region was last handed over from Germany to France (late 1800s, IIRC), they were basically allowed to keep some bits of German law that differed from French law 11:00
Makes Alsace a beneficial place for incorporation, I've been told 11:01
jnthn
.oO( "le reinheitsgebot" )
moritz "le rein'eitsgebot"
11:02 g4 joined, g4 left, g4 joined 11:03 mr-foobar left
jnthn I've a patch for RT #125608. It actually fails one of the NQP test files, but goes on to build Rakudo fine anyway and fixes the ticket. 11:04
make test is alright. Guess I'll spectest and see the fallout.
11:04 proudzhu1 joined 11:05 proudzhu1 left 11:10 pdcawley joined
jnthn hm, it seems OK 11:12
dalek p: 1dd3a77 | jnthn++ | src/NQP/Grammar.nqp:
LTM-terminate error catching case for POD.

The existing grammar here relied on the incorrect semantics of LTM and sequential alternations, which is about to be fixed.		 11:14
jnthn Unfortunately, the fix does seem to slow down parsing somewhat. 11:15
11:18 atroxaper joined 11:19 atroxaper left 11:22 virtualsue left
dalek p: 5834191 | jnthn++ | src/QRegex/NFA.nqp:
Fix sequential alternations and LTM interaction.

The first branch of the alternation should be factored into the longest prefix.		 11:24
FROGGS jnthn: what if we do this instead in line 99? ] || [ \s || <.panic: 'Illegal pod directive'> ] \N*
here I mean: github.com/perl6/nqp/blob/1dd3a771...ar.nqp#L99
jnthn FROGGS: Likely also fixes it
FROGGS: I just went for least invasive change :) 11:25
FROGGS ahh, and the slowdown comes from 5834191?
jnthn FROGGS: Oh, yes, very much so
The first isn't a slowdown
FROGGS I see, first I was suspecting it comes from 1dd3a77 :o) 11:26
11:26 mr-foobar joined
FROGGS k 11:26
jnthn Sorry, wasn't clear
FROGGS np
dalek kudo/nom: d378509 | jnthn++ | tools/build/NQP_REVISION:
Bump NQP_REVISION for LTM and error fixes.
ast: 9f936e7 | jnthn++ | S19-command-line/arguments.t:
Test error for running a directory.

Plus tweak another test to expect the error on STDERR, not STDOUT.		 11:27
11:29 avalenn_ joined 11:31 atroxaper joined, yqt joined, mkz left, krakan_ joined, rindolf joined 11:32 jonathan__ joined, haroldwu_ joined, mrf_ joined, sunnavy_ joined, Brock joined 11:33 silug_ joined, skids1 joined
dalek ast: 44e1719 | jnthn++ | S05-metasyntax/longest-alternative.t:
Tests for RT #125608.
 11:34
11:34 amurf joined 11:35 rmgk_ joined, rmgk left, rmgk_ is now known as rmgk
jnthn Down to 1,032... 11:35
nwc10 open? 11:36
11:37 skids left, RabidGravy left, sftp left, sftp_ joined
nwc10 aha yes: rt.perl.org/Public/ 11:37
11:38 sftp_ is now known as sftp
nwc10 Perl 5 only has 52% more open 11:38
jnthn new/open I guess
But yeah, ones that need dealing with 11:39
11:39 amurf left 11:40 mkz joined 11:50 FROGGS left 11:58 abraxxa left 12:01 atroxaper left
dalek kudo/nom: e4077e3 | jnthn++ | src/ (3 files):
Good error on parameterizing a non-parametric.

Now it shows what went wrong and where, rather than just a method not found error.		 12:02
ast: 7a7dd28 | jnthn++ | S32-exceptions/misc.t:
Tests for RT #115398 and RT #115400.
jnthn Two RTs for the price of one patch :) 12:03
12:05 rurban joined
jnthn m: say grammar Gram { regex TOP { ('XX')+ %% $<delim>=<[a..z]>* }; }.parse('XXXXXX'); 12:06
camelia rakudo-moar d37850: OUTPUT«｢XXXXXX｣␤ 0 => ｢XX｣␤ delim => ｢｣␤ 0 => ｢XX｣␤ delim => ｢｣␤ 0 => ｢XX｣␤ delim => ｢｣␤»
dalek ast: 0a3a69d | jnthn++ | S05-capture/caps.t:
Test for RT #125391.
 12:11
jnthn tosses out rt.perl.org/Ticket/Display.html?id=125486 for anyone who fancies writing a test to close an already-fixed ticket 12:14
lunch &
12:14 mr-foobar left 12:19 AlexDaniel joined
[Coke] (cannot splice a fixed dimension array) - what if you're replacing a chunk with a chunk of the same size? 12:24
12:27 zostay left 12:28 zostay joined
hoelzro morning #perl6 12:31
12:35 muraiki joined
cdc m: my Buf $a = run('bash', '-c', 'echo $\'\xff\xff\'', :out, :bin).out.slurp 12:39
camelia rakudo-moar e4077e: OUTPUT«run is disallowed in restricted setting␤ in sub restricted at src/RESTRICTED.setting:1␤ in sub run at src/RESTRICTED.setting:14␤ in block <unit> at /tmp/QICpGp3I5n:1␤␤»
cdc well, my point was: .slurp fails because of "Malformed UTF-8"
I thought that specifying :bin was enough to slurp it as a Bug instead of a Str 12:40
JimmyZ www.cnbeta.com/articles/411259.htm # chinese news about www.linuxvoice.com/interview-larry-wall/
cdc How should I proceed to read the output of an external command in a binary form? Is it ^^ NYI? 12:42
oh never mind ... 12:43
my Buf $a = run('bash', '-c', 'echo $\'\xff\xff\'', :out).out.slurp(:bin)
12:46 yqt left 12:48 andreoss joined 12:52 g4 left 12:54 avalenn_ left, avalenn joined 12:57 yqt joined 13:00 g4 joined, g4 left, g4 joined 13:01 atroxaper joined 13:06 atroxaper left 13:13 Hor|zon left
masak I've got a clean revert of 22f7192269. should I push it? 13:21
jnthn [Coke]: (splice) we can relax that case later if we need to, I guess 13:23
masak: Define "clean revert"?
masak jnthn: I tried reverting the whole patch on the tip of nom, and it worked without conflicts.
jnthn masak: Oh...we were only going to revert the semantic-changing accidents, I thought 13:24
masak I see. hm, maybe it was only I who thought that the whole thing was a net impoverishment in terms of readability. 13:25
ok, I won't push it, then. someone else can -p painstakingly.
dalek p: 37a0d5f | jnthn++ | src/QRegex/P6Regex/Actions.nqp:
Fix capture handling with regard to &.
jnthn masak: I thought that too :) 13:26
But didn't want to make the git blame situation even worse. 13:27
ShimmerFairy I personally don't see what the big deal is with that commit
masak ShimmerFairy: well, for one, it pushes one side of a non-consensus/debated #perl6 topic into the code base.
you know that something's not right when a patch is made and people exclaim "now I know why I like the other way" :P 13:28
ShimmerFairy That's fair. It's just that I found the 'git blame' explanation to be nonsense, simply put. 13:29
13:30 VinceDee left
masak kebab case. arg pair syntax. these are things that aren't broken and don't need fixing. they have a very skewed risk/importance ratio. 13:32
ShimmerFairy But then why has kebab case been allowed to be pushed through, and arg pair (apparently) isn't? 13:33
jnthn ShimmerFairy: It doesn't matter what you think. I'm one of the top committers/debuggers on the project. If making git blame messier is a problem for me, it's a problem. End of story.
13:33 raiph joined
ShimmerFairy jnthn: then where are the complaints for github.com/rakudo/rakudo/commit/8e5525eedf and github.com/rakudo/rakudo/commit/81d569f915 ? 13:33
jnthn ShimmerFairy: They were rather smaller. 13:34
But yeah, also obscure things
jnthn uses the version history a lot.
13:35 diana_olhovik joined
masak do I understand correctly that 22f7192269 introduces errors into nom? 13:35
jnthn masak: Quite possibly, though not spectest failures.
masak then (a) why isn't it better to revert the whole one rather than wait for someone to have tuits to revert -p it?
(b) if there are no spectest failures, isn't that "interesting"?
jnthn On (b), yes. 13:36
Though I suspect many of them fall into the "we avoid testing exact messages"
masak ok.
ShimmerFairy jnthn: Alright, it's just that you appeared to be using 'git blame' as an excuse to rail against a commit doing cleanup on an issue you disagree with, since I didn't see you complain about the other cleanup commits.
jnthn ShimmerFairy: I've various times refused PRs because they did incidental cleanups along with what they claimed to... 13:37
masak I should ask it like this: if there's a rough consensus that a patch was a net negative for the nom branch, why shouldn't we immediately revert it?
especially if the patch wasn't actually meant to change any semantics.
jnthn masak: Well, there's not a rough consensus, given lizmat thought it was the right thing ;)
ShimmerFairy jnthn: OK. Like I said, it just seemed suspicious that you only complained about one of those commits :)
masak: I also agree with lizmat on the separation of Pairs and named params, fwiw :) 13:38
masak drops the issue and hopes someone with tuits picks it up
ShimmerFairy: as you're free to.
ShimmerFairy: but (orthogonally to that), do you see how it can be a bad idea to push those opinions into Rakudo? 13:39
ShimmerFairy briefly considers using this disagreeable reaction to argue against all this kebab-case nonsense :)
jnthn masak: Well, I thought FROGGS++ was working on the -p approach already 13:40
ShimmerFairy masak: sure, I definitely think it was a poor choice on lizmat's part to just do that. My opposition here is just in response to what I saw was a needless (and inconsistent) attack on the matter, to help balance the matter a bit.
masak I don't mind if people always use the :foo(bar) form in their own code.
jnthn: well, that was 4 hours ago.
ShimmerFairy masak: Basically, I disagree with that commit on a "you should know this would cause problems" basis, not on a "I disagree with always-colonpair" basis :) 13:41
13:42 lucasb joined
jnthn One more useful matters, can anybody find a bit in S05 that talks about whether /<Some::Grammar>/ should be equivalent to /<Some::Grammar::TOP>/ ? 13:42
*on 13:43
masak has a look
jnthn RT #67128 seems to think it should
It's one of our oldest RTs :P
Hm, I don't immediately find the bit that the latter should work... 13:44
13:45 aborazmeh joined, aborazmeh left, aborazmeh joined
jnthn ShimmerFairy: btw, sorry for being harsh to you back there. Bad sleep + allergies getting me down :/ 13:45
ShimmerFairy No worries, I can understand that. 13:46
13:46 yqt left
jnthn hopes lizmat++ won't be upset over the reactions to her surely well-intended commit also 13:46
13:46 raiph left
ShimmerFairy jnthn: I _think_ using a different grammar is supposed to use the :lang adverb, but I'm of course not sure about that :) 13:47
masak jnthn: note that that ticket doesn't claim that it *should* mean ::TOP, only that it oughtn't NPA
jnthn masak: Godo point.
masak jnthn: though I personally think (as I seem to have in 2009) that the syntax kind of makes sense... :)
jnthn: but the ticket is technically closeable 13:48
jnthn masak: Agree, but it feels like a cop-out :)
masak *nod* -- it's a good opportunity to think about what it would do (and what should do that) 13:49
13:49 zakharyas joined
jnthn ShimmerFairy: Interesting point :) 13:49
ShimmerFairy Unsurprisingly, :lang is NYI :)
masak .oO( what was that name of that feature? it's on the tip of my :lang... ) 13:50
jnthn I love how the example in S05 doesn't talk about what kind of thingy is meant to be in $funnylang too :P
masak well, obviously INTERCAL. duh. 13:51
jnthn Or how the method call on it is meant to magically work out
ShimmerFairy Or what the unbalanced method does exactly :P
jnthn And get the current position
masak waves his hands frantically
jnthn But yeah, I don't find any spec justification for <Grammar::Rule> working 13:52
Even though I'm pretty sure if I kill it off folks won't be happy
ShimmerFairy Though I do get the feeling <Grammar::foo> should work too, regardless of :lang . After all, you can call subs from regexes (as in <&foo>), so why not from external grammars too?
masak jnthn: I didn't find it either. 13:53
DrForr I'd support a COMEFROM operator :)
jnthn std: /<Some::Grammar>/
camelia std 28329a7: OUTPUT«ok 00:00 136m␤»
ShimmerFairy I tried looking for :: , TOP, and grammar, none of which yielded anything
masak jnthn: how about installing an error message? "Unsupported syntax for calling another grammar. Please use ... instead."
jnthn masak: Well, for one because folks may well be relying on /<FooGrammar::rule>/ working today, for two 'cus we don't have a ... to use instead, and three because if you just mean to do it for /<Some::Grammar>/ then it's less work to actually make it call TOP on that grammar than it is to implement the error ;) 13:56
masak :)
well, you've convinced me. :) go for it!
ShimmerFairy I'd still be interested in seeing how :lang would work, if <Some::Grammar> doesn't obsolete it :) 13:58
jnthn Yeah, I think TimToady will have to weigh in on :lang
masak .oO( everybody wants some :lang action )
jnthn I think I'll note the semantics of <Some::Grammar> and <Some::Grammar::rule> in S05 13:59
And then make the former work out
masak ++jnthn
jnthn And make sure the tests are alright
And hopefully everyone'll be happy enough :)
timotimo .o( quoteharder )
lucasb Just saying that I wish there was code review *before* commiting changes. Other projects use tools like Phabricator and Gerrit. But I understand this project runs on the EAFP model. And also heavy-weight tools could slow down the process. 14:01
masak lucasb: yeah, I don't wish for code review before.
lucasb: not all the time.
timotimo we could make a super lightweight code review thing that utilises IRC 14:02
jnthn We do pretty well out of post-review
masak what jnthn said.
lucasb Everybody reviews every commit?
14:02 raiph joined
ShimmerFairy It would be nice if git blame had a --skip-commits option, so you could list the useless commits :) (I'm guessing it doesn't have that kind of feature already) 14:02
lucasb GitHub doesn't even show some big commits, but that is very rare
masak I prefer if people police themselves and have a low threshold for thinking "hm, this one isn't 100% non-controversial. let's put it in a branch and point people to it." 14:03
timotimo someone pushes a commit, it has five minutes of grace time where you can say "aye" or "nay" to the bot and if nobody says anything, or there are many more ayes than nays, it gets pushed on to github
jnthn j: say 'alive?'
camelia rakudo-jvm e4077e: OUTPUT«alive?␤»
jnthn j: say 'rule1 foo rule2 bar' ~~ /^ ( 'rule1' || 'rule2' )* %% (.+?) $/
camelia rakudo-jvm e4077e: OUTPUT«｢rule1 foo rule2 bar｣␤ 0 => ｢rule1｣␤ 1 => ｢ ｣␤ 1 => ｢ f｣␤ 1 => ｢ fo｣␤ 1 => ｢ foo｣␤ 1 => ｢ foo ｣␤ 0 => ｢rule2｣␤ 1 => ｢ ｣␤ 1 => ｢ b｣␤ 1 => ｢ ba｣␤ 1 => ｢ bar｣␤»
jnthn m: say 'rule1 foo rule2 bar' ~~ /^ ( 'rule1' || 'rule2' )* %% (.+?) $/ 14:04
camelia rakudo-moar e4077e: OUTPUT«｢rule1 foo rule2 bar｣␤ 0 => ｢rule1｣␤ 1 => ｢ ｣␤ 1 => ｢ f｣␤ 1 => ｢ fo｣␤ 1 => ｢ foo｣␤ 1 => ｢ foo ｣␤ 0 => ｢rule2｣␤ 1 => ｢ ｣␤ 1 => ｢ b｣␤ 1 => ｢ ba｣␤ 1 => ｢ bar｣␤»
ShimmerFairy (Or better yet, if git had some concept of "cleanup" commits, à la Wikimedia's minor edits)
jnthn grmbl...that oen is probably going to be a code-gen bug that has to be fixed on both backends...
*one
And it's probably going to be hard to figure out
AlexDaniel uhm, so what's the problem with reviewing every commit yourself? 14:06
timotimo i bet jnthn already knows about -w ("ignore whitespace") for git blame, and the commit made formatting changes that go beyond white space? 14:07
timotimo hasn't looked closely 14:08
masak timotimo: it's a cute idea, but even it introduces overhead and brittleness worse than the original problem.
dalek ast: b95280f | hoelzro++ | S17-procasync/no-runaway-file-limit.t:
Add test for RT #125616 (Proc::Async not closing tapped handles)
 14:09
jnthn hoelzro: That needs the bump to work, yes? :)
ShimmerFairy m: say 'rule1 foo rule2 bar' ~~ /^ ( 'rule1' || 'rule2' )* %% (.+) $/ # jnthn: causes similar problems with greedy separator
camelia rakudo-moar e4077e: OUTPUT«｢rule1 foo rule2 bar｣␤ 0 => ｢rule1｣␤ 1 => ｢ foo ｣␤ 1 => ｢ foo r｣␤ 1 => ｢ foo ru｣␤ 1 => ｢ foo rul｣␤ 1 => ｢ foo rule｣␤ 1 => ｢ foo rule2｣␤ 1 => ｢ foo rule2 ｣␤ 1 => ｢ foo rule2 b｣␤ 1 => ｢ foo rule2 …»
hoelzro jnthn: correct
14:09 diana_olhovik left, Hor|zon joined
hoelzro once you bump, would you mind adding the new file to spectest.data? 14:09
dalek p: ae07ba3 | jnthn++ | tools/build/MOAR_REVISION:
Bump MOAR_REVISION for async proc handle leak fix.
 14:10
jnthn Oh, it's a new file :)
Yeah, can do
timotimo masak: my main worth in life is coming up with ideas that sound kinda-sorta-reasonable, but ultimately don't work out :) 14:11
dalek kudo/nom: 0d47f78 | jnthn++ | tools/build/NQP_REVISION:
Bump NQP_REVISION for regex and proc async fixes.
ast: 76676ad | jnthn++ | S05-capture/caps.t:
Unfudge test for RT #117955.
 14:12
ShimmerFairy r: say "ab" ~~ /^ a +% (b) $/; say "ab" ~~ /^ a *% (b) $/; say "ab" ~~ /^ a ?% (b) $/; 14:13
camelia rakudo-jvm e4077e: OUTPUT«===SORRY!===␤setcodeobj can only be used with a CodeRef␤»
..rakudo-moar e4077e: OUTPUT«===SORRY!===␤QAST::Block with cuid cuid_5_1436969626.75453 has not appeared␤»
ShimmerFairy r: say "ab" ~~ /^ a ?% (b) $/; # this is the one causing the issue 14:14
camelia rakudo-moar e4077e: OUTPUT«===SORRY!===␤QAST::Block with cuid cuid_1_1436969646.9145 has not appeared␤»
..rakudo-jvm e4077e: OUTPUT«===SORRY!===␤setcodeobj can only be used with a CodeRef␤»
ShimmerFairy r: say "ab" ~~ /^ a +% (b) $/; say "ab" ~~ /^ a *% (b) $/;
14:14 aborazmeh left
camelia rakudo-{moar,jvm} e4077e: OUTPUT«Nil␤Nil␤» 14:14
jnthn ShimmerFairy: I think that may be a spearate issue
ShimmerFairy jnthn: me too, I just found it while messing around with the bug you showed :)
I'll file that as a bug then, if necessary 14:16
jnthn ShimmerFairy: It looks familiar, so have a check if it ain't already in RT
ShimmerFairy Yeah, I'm making sure to search 14:17
jnthn hoelzro: Aww, the test fails on Windows :(
ah, cat 14:18
I'll fix it in a bit
14:18 Hor|zon left 14:20 FROGGS joined
masak m: my %h = foo => 42; say %h<foo> :exists 14:21
camelia rakudo-moar e4077e: OUTPUT«True␤»
jnthn wonders what masak was curious about there :) 14:22
vendethiel spacing?
masak no, I usually space it out like that.
I wanted to check that it actually returns True/False, not the value.
(which is quite obvious in retrospect that it doesn't, because then :exists wouldn't be better than not using :exists) 14:23
then I decided (in the code I'm writing) to not use :exists, because I'm actually interested in the value as well as checking for existence, and I have full control over the %h, so I know I won't have semipredicate troubles. 14:24
vendethiel
.oO( the worse trouble is not being aware you're in trouble )
ShimmerFairy jnthn: I'm not seeing it, so I'll submit a bug to be sure it's there :)
masak goes searching for the Ultimate Metasyntactic Symbol instead, something that will never collide with anything else, ever 14:25
jnthn ShimmerFairy: OK, thanks
ShimmerFairy Legends say it's somewhere past U+10FFFF... 14:26
jnthn OK, bad news on <Grammar::Foo>
masak :(
jnthn /<Grammar>/ probably also should work
So I put in an "is there a type Grammar" check
Then <num> gets rather upset 'cus we have a type "num"...
masak hah! 14:27
mebbe rules should take precedence?
they *are* more proximal...
jnthn Rules are late-bound
masak oh argh
jnthn But we should decide this at compile time.
masak ok, well. then that's a good reason not to do this, methinks. 'cus it's impossible :P 14:28
jnthn I could of course look at the type and see if it ~~ Grammar
masak hm.
jnthn But that feels like ratcheting the clever level up a bit more than we tend to like.
masak yeah. 14:29
ShimmerFairy So perhaps :lang isn't obsolete after all :)
masak it feels fragile for possibly-unintended reasons.
jnthn OK, going to stash the patches and delegate to TimToady++.
masak narrowly evades using ^Z as a metasyntactic symbol 14:30
14:31 g4 left 14:33 khw joined
dalek ast: 6ffdc85 | jnthn++ | S17-procasync/no-runaway-file-limit.t:
Pick a program that exists for Windows.
 14:35
lucasb m: my %h; say 'yes' if %h<foo> :exists :!p 14:36
camelia rakudo-moar e4077e: OUTPUT«yes␤»
lucasb The thing I find funny is that :exists sounds like a predicate, but when mixed with other adverbs, one must remember that it is not a predicate anymore and becomes an utility function.
dalek kudo/nom: 3ae9970 | jnthn++ | t/spectest.data:
Add Proc::Async handle closing test.
masak lucasb: nice.
m: my $a = "a"; say so "fooAbar" ~~ /foo <{$a.uc}> bar/ 14:37
camelia rakudo-moar e4077e: OUTPUT«True␤»
timotimo m: my %h = foo => 42; say %h<foo> :exists:value
camelia rakudo-moar e4077e: OUTPUT«Unexpected adverb 'value' and unsupported combination of adverbs (exists) passed to slice on %h␤ in block <unit> at /tmp/RXHgu7XtQA:1␤␤»
timotimo m: my %h = foo => 42; say %h<foo> :exists:v
camelia rakudo-moar e4077e: OUTPUT«Unsupported combination of adverbs (exists v) passed to slice on %h␤ in block <unit> at /tmp/agQ_Jf7nP_:1␤␤»
masak m: my $a = "a"; say so "fooAbar" ~~ /foo $a.uc() bar/
camelia rakudo-moar e4077e: OUTPUT«5===SORRY!5=== Error while compiling /tmp/gtBRCfJy7T␤Null regex not allowed␤at /tmp/gtBRCfJy7T:1␤------> 3a = "a"; say so "fooAbar" ~~ /foo $a.uc(7⏏5) bar/␤»
timotimo wasn't there a way ...
m: my %h = foo => 42; say %h<foo>:kv:exists
camelia rakudo-moar e4077e: OUTPUT«foo True␤»
timotimo m: my %h = foo => 42; say %h<foo>:v:exists
camelia rakudo-moar e4077e: OUTPUT«Unsupported combination of adverbs (exists v) passed to slice on %h␤ in block <unit> at /tmp/RXLKIoBexk:1␤␤»
masak should I expect variable interpolation in regexes to work like in qq strings, inclusing the ability to put method calls at the end? 14:38
jnthn masak: Not according to my memory of the grammar
14:39 vytas left
PerlJam Greetings #perl6 people 14:39
masak PerlJam: greetings, earthling 14:40
jnthn o/ PerlJam
masak: Though if you use a double-quoted string in the regex... ;) 14:41
m: my $a = "a"; say so "fooAbar" ~~ /foo "$a.uc()" bar/
camelia rakudo-moar e4077e: OUTPUT«True␤»
masak ooh
jnthn Then you can have your braid and eat it.
masak yeah, that looks better. jnthn++
PerlJam Better add that to the FAQ for P5 people 14:42
(do we have a faq?) 14:43
14:43 vytas joined
masak .oO( Q: Do we have a FAQ? A: Yes. ) 14:44
ShimmerFairy Q: Do we have an FAQ? A: Nope. 14:45
timotimo .. when the question "do we have a faq?" becomes a faq itself, the faq needs to be made more obvious to find ..
DrForr My translator does "foo\u$x" --> "foo{uc $x}" :)
masak DrForr: cute!
jnthn If only it was at faq.perl6.org! 14:46
PerlJam :) 14:47
ShimmerFairy I like the sound of faq.perl6.org :)
14:47 Hor|zon joined
lucasb Q: What are all those $@%* signs for? A: Watch your $@*$!% mouth, buddy! 14:49
timotimo jnthn: if only the bugtracker were at bugs.perl6.org; if only you could find contact information at contact.perl6.org, if only you could download a perl6 implementation at download.perl6.org ;)
lucasb I always laugh when I remember that one
ShimmerFairy subdomain-help.perl6.org 14:50
14:50 atroxaper joined
[Coke] bugs.perl6.org could be a redirect to rt.perl.org 14:53
or a placeholder page with a list of all the various places to submit bugs would actually be better. 14:54
14:54 atroxaper left
hoelzro really needs to set up a Windows test machine 14:55
14:56 JimmyZ_ joined 15:00 _mg_ left 15:01 FROGGS left
hoelzro speaking of Windows, would one of you Windows users out there mind testing out Linenoise on Windows for me? 15:02
15:03 n0tjack joined
timotimo [Coke]: especially since rt.perl.org shows you a front page with "sign in or register, please" and in the corner there's a little note of a "public interface", and even then you have bugs for perl5 and perl6 and it's not always very easy to tell apart depending on what piece of RT you're looking at 15:08
i think the best thing we could implement would be a "perl6 web ring"
ShimmerFairy In the case of RT, I think it'd be nice if there was a separate Perl 6 RT (though I imagine the migration would be painful) 15:09
15:10 domidumont left, raiph left
timotimo maybe rt has some way to get a second read-write "view" of the same underlying database? 15:10
i mean, we already have rakudobug@ which puts bugs into the perl6 "part" of the rt 15:11
maybe there could be something that leaves the "choose between perl5 and perl6" parts out completely
15:11 amurf joined
PerlJam some way to cleanly separate the p5 parts from the p6 parts would be nice. 15:13
There have been cases of a p5per commenting on a p6 bug as if it were a p5 bug 15:14
15:14 _mg_ joined
timotimo ah, yes, i remember seeing a mail about that 15:15
"your sub has an illegal prototype, dood"
15:15 lizmat joined
lizmat computer, messages 15:15
yoleaux 09:40Z <FROGGS> lizmat: I'm going to revert your formatting patch, it introduces changed meaning (some bugs, one fix) mentioned here on irc
15:16 amurf left
ShimmerFairy All the more reason for a separate RT :) 15:17
PerlJam ShimmerFairy: Let us know when you've got that working :) 15:18
timotimo :\ 15:19
15:20 AlexDaniel left
PerlJam ShimmerFairy: I imagine it would require a data dump of the existing RT to get the ticket numbers right. And then it's going to need someone to admin it. And the p6 queue in the "old" RT will need to be shutdown in some way so that people don't accidentally send tickets there any more. 15:21
not insurmountable problems, but it will be lots of work. 15:22
lizmat good *, #perl6!
PerlJam lizmat: o/
lizmat well, I guess I made a lot of people happy last night :-(
.tell FROGGS re irclog.perlgeek.de/perl6/2015-07-15#i_10898154 , yes that was intentional 15:23
yoleaux lizmat: I'll pass your message to FROGGS.
ShimmerFairy Sure, I'm not saying it'd be easy :) But it does seem weird that the two languages are still sharing a ticket system. (And fwiw, parrot managed to make the move away, as I recall)
15:23 Ven joined
skids1 lizmat: eh, everyone stays indecisive until something actualy happens. 15:23
PerlJam ShimmerFairy: isn't the plan for those same two languages to share a CPAN? Sharing a ticket system wouldn't seem that out of place. 15:24
(though, now that I think about it ... rt.cpan.org exists for modules on CPAN; how is it to differentiate between p5 and p6 modules that are named the same?) 15:25
lizmat .tell FROGGS since the error messages could also come from using the sub version, so the prefixing period could be misleading
yoleaux lizmat: I'll pass your message to FROGGS.
15:26 khw left
lizmat .tell jnthn How would reverting my patch make things worse? It reverts fine afaixs 15:26
yoleaux lizmat: I'll pass your message to jnthn.
ShimmerFairy PerlJam: yeah, but I think "bug reports for compilers" is a bit of a different case than "extensions for languages" :)
lizmat .tell FROGGS re irclog.perlgeek.de/perl6/2015-07-15#i_10898200 , the range info is just for display, so a string will do fine 15:27
yoleaux lizmat: I'll pass your message to FROGGS.
15:27 araujo left
jnthn lizmat: It's just that when using "git blame" to understand the history of a chunk of code - which I do a decent bit - it means there's an extra level to dig through 15:27
yoleaux 15:26Z <lizmat> jnthn: How would reverting my patch make things worse? It reverts fine afaixs
jnthn Not the end of the world.
I just didn't want to rush to a full revert if it wasn't needed. 15:28
TimToady would prefer a revert
TimToady thinks readability was negatively affected 15:29
dalek kudo/nom: cadee02 | lizmat++ | src/core/ (52 files):
Revert "Use colon syntax for named parameters consistently"

This reverts commit 22f719226914bd6b913e706e0321d027449cca7a.
nine PerlJam: Perl 6 modules have a long name that should be used for the name of the RT queue anyway. 15:30
PerlJam nine: ah, good point.
masak lizmat++ # lots of useful commits overall 15:31
15:31 _mg_ left 15:32 rindolf left
vendethiel lizmat++ :) 15:32
jnthn If you commit enough things, somebody won't like some of them :) 15:33
masak that's an Umeshism, at least if it's stated as "if people like all of your commits, you're not making enough commits" :) 15:34
lizmat jnthn: that's true 15:38
and then I guess there are the people that have committed to doing patches, and then don't commit
timotimo yeah, and lizmat has committed so many things that were great!
lizmat I don't like that either :-(
masak jnthn: sometimes I wish for a "deep blame", tracking all changes all the way back... 15:39
lizmat masak: +1
jnthn masak: git log -L can sorta kinda ish help
TimToady feels deeply blamed :)
PerlJam TimToady: then you should also feel deeply praised :) 15:40
TimToady gets too much of that already 15:41
TimToady wonders if he's going to talk about himself in 3rd person all day...
nine jnthn: since I discovered vim's git support (:Gblame), I don't mind a few extra blame steps anymore.
lizmat thinks it helps creating perspective
timotimo nine: yup, fugitive is great; you may be interested in gitv, too 15:42
emacs also has a git mode that has some super cool name, iirc
"Magit"
TimToady agrees that it can function as a kind of distancing method 15:43
TimToady gets tired of typing /me though
PerlJam nine++, timotimo++ TIL about fugitive :) 15:44
15:44 andreoss left
skids1
.oO(talk like an illeist day)
 15:45
vendethiel loves magit
.oO( come to the emacs side, we have cookies... and evil-mode )
PerlJam vendethiel: vi(m) fits my brain better. (just like Perl fits my brain better than most other programming languages) 15:46
masak I enjoy vim for coding and Emacs for prose. 15:47
timotimo that's why ven recommends evil-mode
muraiki but there's still the "learn all the other emacs stuff that isn't covered in evil-mode" problem 15:48
PerlJam exactly 15:49
vendethiel PerlJam: evil-mode is great.
PerlJam (or ... "Why does my editor speak with a lisp?")
ShimmerFairy I like my editors not requiring keystrokes to begin editing :P 15:50
muraiki well, I think a vim vs emacs debate will not be a very useful discussion. because clearly acme is the best editor.
JimmyZ_ Is evil-mode better than vim?
vendethiel PerlJam: I'll admit to liking lisp ;-)
JimmyZ: I'm not sure it's "better", but I've never missed anything 15:53
15:53 zakharyas left
skids1 Like just about every bit of software these days, with emacs there is a growing list of defaults package maintainers think are cool that simply must be killed. Die electric indent. 15:53
JimmyZ_ always use nano for committing git message. 15:54
skids1
.oO(Don't get me started on bash-completions)
 15:55
JimmyZ_ uses zsh 😊
geekosaur you do realize zsh comes with a completion nightmare that puts bash-completions to shame? 😊 16:01
16:01 smls joined
geekosaur (at least you get to choose whether to activate it on first run...) 16:01
16:01 yqt joined
skids1 It's really an art form: how can we make the CLI as WAT as moderm UIs 16:01
timotimo skids1: electric indent is when you write something inside a line that changes the indentation of that line you're in? 16:03
skids1 Something like that. All I remember is frantically googling to find out how to turn it off and then saying "i really need to start keeping a custom .emacsrc on a flash drive" 16:04
ShimmerFairy timotimo: A while ago a new version of emacs changed things around so <Enter> did newline+indent, and C-j did just a newline, the exact opposite of the previous behavior.
timotimo oh? 16:07
ShimmerFairy Needless to say, I corrected this bug pretty quickly :P 16:08
16:09 Ven left
skids1 I do like the new CTRL-L behavior though. 16:10
ShimmerFairy I don't think I've ever used that.
arnsholt skids1: Having looked at the introduction to the Unix Hater's Handbook earlier today, I'm reminded of this quote from it: "What is this horrible fascination with Unix? The operating system of the 1960s, still gaining in popularity in the 1990s." 16:11
To be fair, it follows on with "A horrible system, except that all the other commercial offerings are even worse."
skids1 Lately my pet peeve is closed source products implementing space-completion with no way to turn it off. Trend obviously motivated by users who don't have tab keys on their phone softkeys. 16:12
arnsholt Mal a propos Unix hating, there's been some hilarious comments on HN when people see Rob Pike and Ken Thompson's comments on how modern Unix is kind of terrible in various places 16:15
Or realizing that Rob Pike's preferred editor is *mouse-driven* =D 16:17
moritz
.oO( mine is visual )
 16:19
16:20 Spot__ left, timotimo left, yogan left, retupmoca left, hoelzro_trying_w left, ugexe left, yogan joined, hoelzro_trying_w joined, timotimo joined, retupmoca joined, ugexe joined, spider-mario joined 16:21 telex left 16:22 telex joined, Spot__ joined
arnsholt moritz: I think GUI editors are (grudgingly) tolerated. But an editor that does not strive towards being 100% keyboard-driven is heresy, pure and simple =) 16:23
lizmat is moving away from the keyboard for a bit 16:25
sjn thought voice-based programmng was all the rage in rage these days... "Siri, make me an app that pays my bills"
16:38 atroxaper joined 16:40 cognominal left
TimToady "Okay, you are now an app that pays your bills." 16:41
16:41 cognominal joined 16:42 atroxaper left 16:47 diana_olhovik joined
DrForr m: say q{{}} 16:47
camelia rakudo-moar cadee0: OUTPUT«␤»
16:48 JimmyZ_ left
smls Why doesn't Perl 5 allow non-literal values in regex quantifiers, like Perl 6's "** {$x}" :( 16:54
Also, now that I've been spoiled by Perl 6, I'm really missing the ability to get at the capture values of recursed subpatterns... 16:55
PerlJam smls: I think you want #perl ;) 16:56
smls OTOH, I can't use Perl 6 for what I'm doing because it involves matching against binary blobs
16:56 virtualsue joined
smls (which won't be supported in 6.0) 16:56
PerlJam: Nah, that would come across as trolling :) 16:57
17:01 amurf joined 17:02 salv0 joined, dakkar left 17:04 jeffreykegler joined
jeffreykegler TimToady: re the Eco _Search for a Perfect Language_ book -- I thought it was excellent and am glad that you did ... 17:05
it's biggest fault is that it is very Euro-centric.
17:06 amurf left
jeffreykegler How can you write a book about attempts at a perfect language, and omit Sanskrit? 17:06
smls heh, I guess (?{ $i = 0 }) (?<loop> ... (?(?{ $i < $x })(?&loop)) ) works as a dynamic quantifier in P5 regexes, if there's no recursion
it's hideous though, and requires externally declared vars
17:10 vendethiel left 17:12 vendethiel joined, felher joined 17:13 araujo joined, araujo left, araujo joined
smls So yeah, TimToady++ for boldly redesigning the regex language :) 17:16
17:17 araujo_ joined, araujo_ left 17:18 araujo left 17:20 Hor|zon left
geekosaur was tempted to mention that about how to do it with p5, but would prefer to quietly forget it exists... 17:20
dalek rl6-roast-data: 7c94661 | coke++ | / (9 files):
today (automated commit)
 17:31
17:33 vendethiel left
[Coke] (parrot made the move away from RT) what a hellish waste of time that was. 17:34
17:34 vendethiel joined 17:35 domidumont joined 17:36 espadrine left
[Coke] Moving ticketing systems is a bunch of infrastructure work that 1) takes time away from other infrastructure work, 2) has little practical benefit, 3) ends up happening more than once (cvf. parrot). Our best bet for perl6 is to document all the places bugs can be opened and make it easy for people to find existing bugs or create new ones. 17:37
I would be not quite opposed to saying "all new tickets go in the new system" if we were to do that, but we would have to LOVE whatever the new system was. 17:38
there would have to be clear technical advantages. If you think moving ticketing systems is something you want to work on, ping me and I will be happy to suggest other infrastructurey things. :)
17:39 zostay left, zostay joined 17:42 nys joined 17:51 lucasb left 17:54 jeffreykegler left 17:56 yqt left 17:59 skids1 left
dalek ast: 9ace330 | lizmat++ | S09-typed-arrays/native- (2 files):
Revert "Tweak typed error checks"

This reverts commit ed09f71f7bab673948a65b7c787296a08f109d20.		 18:03
lizmat since we also reverted 22f719226914bd6b913 18:04
18:05 domidumont left 18:13 zakharyas joined 18:16 VinceDee joined, Hor|zon joined, yqt joined 18:17 spider-mario left 18:18 spider-mario joined 18:21 Hor|zon left 18:22 [Sno] left, [Sno] joined
PerlJam [Coke]++ That was a very good summary. 18:22
18:22 bin_005 joined 18:27 [Sno] left, ][Sno][ joined
PerlJam lizmat: IMHO (others may differ), the one thing I liked best about your earlier patch that was reverted was the removal of those dots. They can be removed without loss of information and possibly remove confusion when the user uses a sub form rather than a method form. 18:29
lizmat patches welcome :-) 18:30
18:30 rurban left
PerlJam As soon as my time + energy + circumstances align, perhaps (If I don't forget by then :) 18:31
18:42 virtualsue left 18:44 virtualsue joined 18:46 zakharyas left
dalek kudo/nom: 648a585 | lizmat++ | src/core/ (6 files):
Remove confusing . from error messages, PerlJam++
 18:46
ast: 91bd47e | lizmat++ | S09-typed-arrays/native- (2 files):
Revert "Revert "Tweak typed error checks""

This reverts commit 9ace33081f9862771bc400ae04839b11b6ca62d6.
lizmat PerlJam: vi `ack -l '<\.' src/core` is your friend :-) 18:47
off to see Butte Creek Falls &
18:50 xfix joined, xfix left, xfix joined 18:52 domidumont joined
masak [Coke]++ # "happy to suggest other infrastructurey things" 18:53
smls: if you want to privmsg at some point and discuss regex matching of binary blobs, we might make something happen pretty quickly in module space. 18:54
18:57 Hor|zon joined 19:09 domidumont left
moritz m: say $?FILE 19:12
camelia rakudo-moar cadee0: OUTPUT«/tmp/bbmXANMUMx␤»
dalek kudo/nom: 89847f8 | moritz++ | src/core/IO/Notification.pm:
Give IO::Notification::Change a gist method
 19:23
19:27 atroxaper joined 19:31 diana_olhovik left, atroxaper left, n0tjack left
dalek c: 89e66f8 | moritz++ | lib/Type/IO/Notification.pod:
Document IO::Notification
 19:34
19:38 CIAvash left
hoelzro is there a minimum version of Windows that Rakudo will support? 19:39
I'm thinking of building a Windows VM (since I so often break things on that OS), and I'm wondering which version I should install
jnthn hoelzro: Well, if you try it on Win95 it's probably not going to end well... :) 19:40
hoelzro =P
jnthn I can't imagine we don't work on anything Vista or later
moritz hoelzro: I wouldn't recommend anything older than XP, though I don't know if that's actually new enough
jnthn is still on Win7, 'cus Win8 was horrid 19:41
And yeah, for your own sake, skip Vista too :P
hoelzro I have licensed copies of XP and 8.1, but I have 8.1 installed on the desktop
so I don't know what I'm legally allowed to do even
I'm guessing XP would be fine, since it's not installed anywhere
19:41 Peter_R joined
moritz doesn't get the windows 8 jumps between tiles and traditional desktop 19:42
hobbs it wants to be all tiles and new and ...
moritz looks like "big boss wants tiles, users are used to traditional desktop, so both get both, and everybody's confused equally" 19:43
hobbs but a million existing apps don't play nice with that and it can't break them without becoming irrelevant
19:49 vendethiel left 19:51 darutoko left 19:56 smls left, smls_ joined
hoelzro apparently Microsoft makes VM images available for testing \o/ 20:02
20:05 salv0 left
b2gills Perl 5 is still tested on Win2k SP4 20:10
20:13 colomon left
hoelzro Perl 5 is also still tested on VMS, I believe =) 20:16
nwc10 VMS is still supported by its vendor
Win2k, not so much -) 20:17
hoelzro touche
nwc10 no, I don't know why George Greer likes smoking on Win2k
but we don't argue with useful donated work.
20:32 xfix left, colomon joined 20:34 muraiki left 20:35 spider-mario left 20:38 spider-mario joined, amurf joined 20:43 amurf left 20:50 kaare_ left 20:51 colomon left 20:54 colomon joined 21:01 synbot6 joined 21:04 bin_005_l joined, synbot6 left, bin_005 left
masak 'night, #perl6 21:06
21:12 captain-adequate joined 21:14 colomon left 21:15 atroxaper joined 21:20 TEttinger joined 21:21 atroxaper left 21:29 smls_ left 21:38 colomon joined 21:53 nightfrog joined
jnthn Another week, another progress report: 6guts.wordpress.com/2015/07/15/thi...ood-stuff/ 22:05
'night, #perl6 22:13
labster jnthn++ good work last wek 22:14
week 22:15
22:18 virtualsue left 22:23 bin_005_l left 22:31 spider-mario left 22:36 jonathan__ left 22:40 coffee` joined 22:53 VinceDee left 22:54 telex left 22:56 telex joined 23:10 pdcawley left 23:16 nebuchadnezzar left, nebuchadnezzar joined 23:24 amurf joined 23:44 VinceDee joined 23:47 Celelibi joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!