ggoebel read the recent reddit thread... thought the following lessons learned from Audrey Tang might be worth reviewing: medium.com/@audrey.tang/lessons-i-...f5d8107e34 01:04
TimToady +1 01:18
samcv » doesn't use threading currently right? 04:36
so that shouldn't be causing any crashes?
[Tux] This is Rakudo version 2017.06-136-gd5d3bd276 built on MoarVM version 2017.06-30-g389e9732 06:04
csv-ip5xs 2.671
test 12.645
test-t 4.099 - 4.145
csv-parser 12.681
hgusa as a lurker outsider of the community, reading Zoffix lose an argument with chromatic is much more fun than reading a cheesy blog from a minister without portfolio of a fictive country 06:50
nine ggoebel: thanks for sharing Audrey's lessions! 06:51
llfourn +1 I hadn't seen that one. 07:53
Zoffix hgusa, glad you enjoyed yourself! 08:35
samcv: right, not yet.
samcv thanks 08:36
that's what i thought too
just wanted to double check
Zoffix I don't get Audrey's lessons. 08:41
Though there's one new one I can pitch in: people love drama. Hell, that reddit convo happend like a week ago, but people still can't stfu about it. Even someone lurking in the logs. 08:47
But no one bothered to correct me on my proposal to use uint64 for rat denominators and say that'd might end up paying too much for boxing all the time, because numerator can be an Int supposedly! 08:48
nine Zoffix: just to let you know, I'm pretty hurt right now for not even getting any kind of response for my apology. 08:53
jnthn Zoffix: The numerator is defined as an Int, but that doesn't mean there's no saving possible, it does mean that you might have to introduce a division op that takes a Int and uint64 08:57
Zoffix jnthn: OK thanks. 08:58
jnthn Also I don't see any discussion of that in the backlog here? 08:59
Zoffix Exactly! But you see discussion about Audrey lessons and the reddit thread!
jnthn Yeah, can't people do them on 09:00
#perl6 so we can work here? :P
Zoffix Unless you mean my original "road map" plan: irclog.perlgeek.de/perl6-dev/2017-...i_14820738
jnthn huh, dns fails to resolve on that :S
And that's probably just out of local history 09:02
llfourn also sometimes has DNS issues with perlgeek.de
Zoffix nine: sorry for storming out yesterday when you were giving your help. I was just annoyed the low-value crap from reddit from a week ago was getting dragged in to #perl6-dev. I don't have any goal or plans with chromatic or anyone else—whether they're positive or negative. What I did on reddit is just pass time on a boring bus ride, so all the advice and lessons about how to win arguments and bring peace 09:03
to world will fall on my deaf ears.
nine Fair enough. 09:04
Zoffix Here's a copy-paste of that log link: gist.github.com/zoffixznet/c658efa...3c057e7621
jnthn heh, finally it loaded 09:09
samcv Zoffix, did you see this regression. rt.perl.org/Ticket/Display.html?id=131703 your commit is linked that introduced it 09:24
Zoffix samcv: just saw the mention in #perl6 ... I guess ZofBot doesn't recognize "zoffix's" as a mention of me :/ 09:27
m: say "zoffix's" ~~ / «zoffix» / 09:28
camelia 「zoffix」
Zoffix weird
Oh, it's there in my Twitter log, I just only looked at the last message :} 09:29
c: f539a624043f5796fcefcca83412ee1ab66e0c4d~1 @*ARGS = 'meow'; get.say 09:32
committable6 Zoffix, ¦f539a624043f5796fcefcca83412ee1ab66e0c4d~1: «(HANDLED) Unable to open file 'meow'␤ in block <unit> at /tmp/pSNune1vHE line 1␤»
Zoffix samcv: how do you figure out the regression commit? Seems the same issue exists prior to it?
mc: @*ARGS = 'meow'; get.say
committable6 Zoffix, ¦2015.12: «(HANDLED) Unable to open file 'meow'␤ in block <unit> at /tmp/HwL39czJzp line 1␤»
Zoffix And in 2017.05 release
mc: @*ARGS = 'meow'; sub MAIN ($x) { get.say } 09:34
committable6 Zoffix, ¦2015.12: «(HANDLED) Unable to open file 'meow'␤ in sub MAIN at /tmp/NH9yzaD84J line 1␤ in block <unit> at /tmp/NH9yzaD84J line 1␤»
Zoffix m: @*ARGS = 'meow'; sub MAIN ($x) { get.say } 09:35
camelia Failed to open file /home/camelia/meow: No such file or directory
in sub MAIN at <tmp> line 1
in block <unit> at <tmp> line 1
Zoffix Ahhhh
The failure ends up handled in pre-commit
So I'd say that script was working for the wrong reason and the CatHandle swap fixed a bug that used to silence a failure. 09:36
samcv: wrote out explanation and closed the ticket: rt.perl.org/Ticket/Display.html?id...xn-1471842 09:42
"<samcv> but regardless shouldn't be using the arg as a file since it should be taken by MAIN" 09:46
Dunno. That get() would unambiguously mean $*IN inside MAIN then, because the only way to get inside is to "handle" all the args, so IMO that code should just use $*IN or just do $*ARGFILES = $*IN if it wants to to use argless IO routines. 09:48
samcv Zoffix, so the argument doesn't get eaten by MAIN?
Zoffix No
samcv i thought $*IN should be stdin
hmm
Zoffix it is.
samcv that... is very confusing
oh
get doesn't use STDIN it uses argfilesL
Zoffix get() lines() slurp() and words() operate on $*ARGFILES. $*ARGFILES (by default) is a CatHandle of @*ARGS, unless @*ARGS are empty, in which case it uses $*IN 09:49
Right.
samcv ah ok
why is argfiles set when MAIN accepts a variable though?
that is the odd functionality i think. i could understand if there were no args to main. but since it gets set as a variable in the signature 09:50
hm
Zoffix They're just unrelated. The "$*ARGFILES (by default) is a CatHandle of @*ARGS" behaviour exists, regardless of whether you used MAIN or not
The alternative is that $*ARGFILES would become $*IN the moment you use MAIN (or inside MAIN), so that begs the question why not just leave it as @*ARGS and write code inside as using $*IN 09:53
Maybe it should, dunno.
samcv yeah
worth considering
Zoffix 'cause it being @*ARGS isn't very useful. The second you have any named args, you no longer can use $*ARGFILES 09:54
samcv as i think it would be the least suprising to people that way. since MAIN is supposed to be all nice and good
Zoffix huggable: 6.d
huggable Zoffix, Proposals for 6.d language: github.com/perl6/specs/blob/master/v6d.pod
Zoffix Yeah. $*IN sounds goo. Wrote it up for 6.d: github.com/perl6/specs/blob/master...nside-main 10:00
samcv nice :) 10:06
lizmat .ask masak do you have any thoughts on deprecating (<+) in favour of baggy semantics of (<) and (<=) ? 11:39
yoleaux lizmat: I'll pass your message to masak.
lizmat .ask labster do you have any thoughts on deprecating (<+) in favour of baggy semantics of (<) and (<=) ?
yoleaux lizmat: I'll pass your message to labster.
lizmat .ask TimToady do you have any thoughts on deprecating (<+) in favour of baggy semantics of (<) and (<=) ?
yoleaux lizmat: I'll pass your message to TimToady.
lizmat m: dd @^10 # not what I expected, but correct nonetheless :-) 11:44
camelia one([], 10)
timotimo yup, earliest match wins 11:49
well, that and @^a would be valid, but @^10 wouldn't be
so it goes down the path of "is this a placeholder signature parameter?"
Zoffix huggable: cpan testers api :is: CPAN testers API: www.reddit.com/r/perl/comments/6lb...as_an_api/ 11:55
huggable Zoffix, Added cpan testers api as CPAN testers API: www.reddit.com/r/perl/comments/6lb...as_an_api/
Zoffix prolly something that can be baked into modules.perl6.org
Gonna do a few passes of straightening up &SEQUENCE, while I tease out all that it does, for my article 12:46
buggable: toast 12:47
buggable Zoffix, Between 2017.06-46-g43c1767 and 2017.06: 6 (0.73%) modules got burnt; 11 (1.33%) got unsucced; 199 (24.09%) out of 826 modules appear unusable. See toast.perl6.party/ for details.
Zoffix makes another toast
timotimo i'm hungry now 12:48
Geth rakudo/nom: 2504798464 | (Elizabeth Mattijsen)++ | 2 files
Make Setty eqv Setty between 2x and 3.5x faster

  - 2x in case of different number of elements
  - 3.5x in case of equivalence on a 26-element Set
  - also move it from set_operators.pm to Setty.pm
12:51
Element80 So what's happening with mix.t file? 13:01
t/spec/S03-operators/mix.t (Wstat: 1024 Tests: 66 Failed: 4)
Failed tests: 27, 30, 32, 65
ZOFVM: Files=1259, Tests=141105, 113 wallclock secs (18.85 usr 3.26 sys + 2138.17 cusr 205.49 csys = 2365.77 CPU)
Geth rakudo/nom: 001135ca2a | (Zoffix Znet)++ | src/core/operators.pm
Straighten up &SEQUENCE a bit
13:02
lizmat Element80: it's under investigation :-) 13:10
Element80 lizmat: is it the (<+) thing? 13:12
lizmat yeah, related to (<+) 13:13
Element80 Ah
lizmat I would like to hear masak labster TimToady's thoughts on deprecating (<+) before doing anything 13:14
Element80 buggable: draw 13:30
buggable Element80, The next Accidental /win Lottery draw will happen in -1 weeks, 2 days, 10 hours, 29 minutes, and 47 seconds. Currently have 0 ballots submitted by 0 users!
Element80 wonders who won
Hehe, jnthn did :) 13:31
timotimo yeah
i remember celebrating
Element80 buggable: testers f878354a-6183-11e7-89a5-c5f577a92919 13:36
buggable Element80, Proc::Q:ver(1.001003) test result PASS. See more at api.cpantesters.org/v3/report/f8783...f577a92919
Element80 hehe neat :)
Element80 hasn't figured out how to fetch info for Perl 6 dists by name yet 13:37
Hm. Only now do I realize the "see more" URL isn't human-friendly if you're not using Firefox :) 13:38
Geth rakudo/nom: 49b1b03b02 | (Elizabeth Mattijsen)++ | src/core/Baggy.pm
Make Baggy eqv Baggy upto 3.5x faster

  - 3.5x in case of equivalence on a 26-element Bag
  - no improvement on different number of elements, as that was already optimized
13:39
dogbert17 lizmat: if I were to write tests for RT #131303, where do you want me to put them? 13:42
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=131303
lizmat dogbert17: I think that's covered in S03-operators/set_difference.t 13:43
check out if there are any cases missing there :-) 13:44
dogbert17 lizmat: will do 13:47
Element80 .tell masak you have a bunch of PRs on this module: github.com/masak/data-pretty/pulls 13:51
yoleaux Element80: I'll pass your message to masak.
Element80 buggable: author masak
buggable Element80, Found 12 dists for masak. See temp.perl6.party/buggable/84904347...62717.html
Element80 .tell masak Actually a whole ton of PRs are waiting for merge in virtually all of your dists: temp.perl6.party/buggable/84904347...62717.html 13:53
yoleaux Element80: I'll pass your message to masak.
Geth rakudo/nom: 93d81d61a6 | (Elizabeth Mattijsen)++ | src/core/Setty.pm
Make Setty.ACCEPTS(Setty) about 3.5x faster

  - the difference between eqv and ACCEPTS is the strict type check only
  - moved most of logic from eqv to ACCEPTS
14:02
lizmat afk& 14:22
Element80 ugh 14:33
It doesn't take 1 hour to dump data as text -_-
Data::Dump::Tree-- over-engineered 14:34
dogbert17 Element80: would it be possible for you to close RT #131561, I see that tests have been added to roast
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=131561
dogbert17 Element80 == Mercury ? 14:35
Element80 dogbert17: www.youtube.com/watch?v=2LTxgx3KR30 14:38
It's a band that uses Perl 6. That song is about Promises
:trollface:
dogbert17: how come you don't close it? 14:39
dogbert17 Element80: I was afraid you should ask that :) it doesn't work, can only close my own bugs 14:40
Element80 dogbert17: what's your rt username?
dogbert17 dogbert17 14:41
sry, no my email address
Element80 What's the email address?
dogbert17 jan-olof.hendig(at)bredband.net 14:42
Element80 dogbert17: when was the last time you tried to close tickets? You're listed as admin... 14:43
dogbert17 I closed one just a few minutes ago 131303 14:44
Element80 dogbert17: and did you try closing 131561 ? 14:45
dogbert17 wtf, it worked, Element80++ 14:47
Element80 ok then :)
dogbert17 now I can go on a closing spree :) 14:48
Element80: suggestion for where I should put tests for RT #131529 14:51
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=131529
Element80 dogbert17: I'm guessing add `byte` to @int-types at the top of S02-types/int-uint.t and fix up the rest of the file to handle it, if needed. 14:56
dogbert17 Element80: thx 14:58
Element80 buggable: testers f878354a-6183-11e7-89a5-c5f577a92919 14:59
buggable Element80, Proc::Q:ver(1.001003) test result PASS. See more at temp.perl6.party/buggable/45441390...266756.txt
Element80 That's bettah
Geth roast: dogbert17++ created pull request #275:
Add test for RT #131529
15:11
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=131529
dogbert17 Element80: should I merge or do you see anything bogus? 15:14
Element80 dogbert17: looks good 15:16
dogbert17 thx
Element80 Though not what I meant when I add `byte` to @int-types :) It's a tangent point though, so I guess it don't matter for the ticket 15:17
Geth roast: 0331647a61 | (Jan-Olof Hendig)++ | S02-types/int-uint.t
Add test for RT #131529
roast: 3bc8e6df53 | dogbert17++ (committed using GitHub Web editor) | S02-types/int-uint.t
Merge pull request #275 from dogbert17/test-rt-131529

Add test for RT #131529
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=131529
Element80 Filed an Issue: github.com/perl6/roast/issues/276 15:19
dogbert17 do we really have more than one byte type, e.g. byte1, byte8 etc? 15:21
Element80 Don't think so 15:22
Geth roast: dogbert17++ created pull request #277:
Add tests for RT #131273
15:54
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=131273
Geth roast: ed74b5b449 | (Jan-Olof Hendig)++ | 2 files
Add tests for RT #131273
15:57
roast: d05653ace7 | dogbert17++ (committed using GitHub Web editor) | 2 files
Merge pull request #277 from dogbert17/test-rt-131273

Add tests for RT #131273
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=131273
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=131273
Element80 this is weird. The toast run for HEAD appearntly completed fine without my having to kill anything 16:17
m: $*PERL.compiler.version.say 16:19
camelia v2017.06.140.g.93.d.81.d.61.a
Element80 Yup. All the data seems to be there: toaster.perl6.party/
Unless I stroked off... but no, pretty sure I didn't need to kill anything.
buggable: toast 16:20
buggable Element80, Between 2017.06-140-g93d81d6 and 2017.06: 9 (1.08%) modules got burnt; 12 (1.45%) got unsucced; 201 (24.22%) out of 830 modules appear unusable. See toast.perl6.party/ for details.
Element80 or maybe I did kill it. I recall killing UNIX::Daemonize, but it says "Succ" for 2017.06 :| 16:21
Oh, "Dubious, test returned 1 (wstat 256, 0x100)". So it's not Succ :/ 16:22
And it got "Ambiguous use of »; use >> instead to mean hyper" warning up in it :/ 16:23
"daemonize(«rm $pidlockfile», :pid-file($pidlockfile))"
.ask TimToady is the `Ambiguous use of »` supposed to warn in cases like `run «rm $pidlockfile»` ? That's really annoying. I got it while running normal code and I see some ecosystem usage that now has the same warning pop out from their code. 16:24
jnthn
.oO( Unless you...what?! )
yoleaux Element80: I'll pass your message to TimToady.
Element80 jnthn: stroked off == had a stroke 16:25
jnthn Ah...it's...worse...in my dialect. :P
Element80 heh
greppable6: »
greppable6 Element80, Sorry, can't do that
jnthn The hyper check does seem a bit overzealous
It'd be nice if it only triggered when the thing actually successfully parsed as a hyper 16:26
Time for some break, dinner, etc. 16:30
Geth nqp: bc1b7361e6 | (Douglas Schrag)++ | t/hll/06-sprintf.t
Add tests for sprintf octal with precision

Test octal precision with and without the '#' flag. Ensure conformance with POSIX sprintf.
16:32
nqp: cd8d6ae046 | (Douglas Schrag)++ | src/HLL/sprintf.nqp
Do not add condition leading '0' if already present

If precision is sufficient to pad octal with a leading '0', then no additional prefix is needed with the '#' flag.
nqp: 1a90373c0f | (Zoffix Znet)++ (committed using GitHub Web editor) | 2 files
Merge pull request #364 from dmaestro/issue/353-sprintf-octal-padding

Issue/353 sprintf octal padding
nqp: 17c795cf6b | (Zoffix Znet)++ | tools/build/MOAR_REVISION
Bump MoarVM

MoarVM bump brought changes:
  github.com/MoarVM/MoarVM/compare/2...7-g4e29e4c
4e29e4c Fix an accidental nested redeclaration in proc. 33e46e2 Eliminate duplicate code in utf8 decoder. ... (6 more lines)
16:43
¦ nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2...7-g4e29e4c
rakudo/nom: d7e1046678 | (Zoffix Znet)++ | tools/build/NQP_REVISION
Bump NQP

NQP bump brought changes:
  github.com/perl6/nqp/compare/2017....5-g17c795c
17c795c Bump MoarVM 1a90373 Merge pull request #364 from dmaestro/issue/353-sprintf-octal-padding ... (32 more lines)
rakudo/nom: version bump brought these changes: github.com/perl6/nqp/compare/2017....5-g17c795c
99236d4009 | (Zoffix Znet)++ | docs/release_guide.pod
rakudo/nom: ff4039ca93 | (Zoffix Znet)++ (committed using GitHub Web editor) | docs/release_guide.pod
Move cautionary note

to be *before* the potentially dangerous instruction
16:52
nine huggable: 6.d 16:54
huggable nine, Proposals for 6.d language: github.com/perl6/specs/blob/master/v6d.pod
Geth rakudo/nom: b3c14c6d8f | (Zoffix Znet)++ (committed using GitHub Web editor) | docs/release_guide.pod
Remove "developmental" from releases list heading
16:55
Element80 pmurias: [Coke]++ was working on updated docs website. That's all the replacement work I'm aware of. 17:04
TimToady jnthn: but that's the point, it didn't successfully parse as a hyper, but gave a different error downstream 17:17
yoleaux 11:39Z <lizmat> TimToady: do you have any thoughts on deprecating (<+) in favour of baggy semantics of (<) and (<=) ?
16:24Z <Element80> TimToady: is the `Ambiguous use of »` supposed to warn in cases like `run «rm $pidlockfile»` ? That's really annoying. I got it while running normal code and I see some ecosystem usage that now has the same warning pop out from their code.
TimToady jnthn: we could perhaps make an exception list of things that are unlikely to be a postfix, but it's not impossible to have a postfix that starts with almost anything 17:20
Geth roast: dogbert17++ created pull request #278:
Added tests for RT #131272
17:21
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=131272
TimToady though I suppose if people start defining things like postfix:<;)> they deserve what they get ;) 17:22
Geth roast: e4302e5d7d | (Jan-Olof Hendig)++ | 2 files
Added tests for RT #131272
roast: 68d34de4b8 | dogbert17++ (committed using GitHub Web editor) | 2 files
Merge pull request #278 from dogbert17/test-rt-131272

Added tests for RT #131272
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=131272
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=131272
TimToady lizmat: on the (<+) thing, my preferences lean toward getting rid of it if we can, and making Baggy types more contagious with respect to Setty types, on the theory that sets are mostly just constrained bags 17:24
however, I missed the carpage earlier that caused the reversion, due to the drains braking in my house rather badly... 17:25
er, *breaking
but I don't like the idea of set operations coercing bags to sets and throwing away information 17:26
and of course the drains break when we have kids and grandkids visiting, so we've been dodging plumbers in hazmat suits and spending far too much time driving out to restrooms... 17:28
Element80 TimToady: what if we don't detect any postfixes but merely warn if there's stuff after » that's being parsed while still inside the `«»` quote constuct? 17:43
i.e. the `»` is not the closing delimiter
m: my $x = 42; my %h = :42a; say %h<<42 $x>>; 17:45
camelia Potential difficulties:
Ambiguous use of >>; use » instead to mean hyper, or insert whitespace before >> to mean a quote terminator (or use different delimiters?)
at <tmp>:1
------> 3my $x = 42; my %h = :42a; say %h<<42 $x7⏏5>>;…
Element80 It warns even here.
TimToady the » is already being parsed as a hyper at that point
Element80 I see. 17:46
TimToady it's relying on the benign failure of parsing a postfixes after the », and that's a problem with extensible postfix notation
we can guess that ; and ) are unlikely to be the beginnings of a postfix, but we can't be sure 17:47
guessing may be good enough, but then there will just be other places that people expect $foo>> to dwim 17:48
Element80 m: say [[<x>],]
camelia 5===SORRY!5=== Error while compiling <tmp>
Unable to parse expression in bracketed infix; couldn't find final ']'
at <tmp>:1
------> 3say [[<7⏏5x>],]
Element80 ^ is that the same sort of issue? The solution is to put the space rather than fixing ocde?
*codew
m: say [[ <x>],] 17:49
camelia [[x]]
TimToady reductions are one of the two places where we explicitly backtrack
and it commits to < being an infix before it ever thinks about circumfix:<[ ]> 17:50
the dwim in that rule is that if there are any spaces inside, it assumed it cannot be a reduction
Element80 ok, then I'll just close rt.perl.org/Ticket/Display.html?id...et-history 17:51
TimToady m: say [[x]] 17:52
camelia No zero-arg meaning for infix:<x>
in block <unit> at <tmp> line 1
TimToady m: say [[x ]] 17:53
camelia 5===SORRY!5=== Error while compiling <tmp>
Undeclared routine:
x used at line 1
TimToady and the rationale for the dwim is that most lists have multiple members, which most people separate with a comma and a space 17:54
so another spot where we optimize against code golf...
Element80 Well, I was writing [[<foo bar ber>], [<meow moo>]] 17:55
TimToady most people never even realize there's a formal ambiguity between reduction operators and array composers
Element80 m: with [[<foo bar ber>], [<meow moo>]] {}
camelia ( no output )
Element80 Oh maybe I wasn't
TimToady but this is also part of the reason for forbidding whitespace in metaops in general 17:57
Element80 Ah, OK, I wrote the above, but then shortened <foo bar ber> to just <foo> to get 1-el array. Though this was test code, not real program :) 17:58
TimToady P5 had many more dwims of that sort, which is why we have fewer :) 17:59
Element80 :)
Geth star: tbrowder++ created pull request #90:
add specific instructions for a system installation
18:01
roast: 94ba24dc09 | (Elizabeth Mattijsen)++ | 2 files
Test for baggy semantics on (<) and (<=) again

This reverts 3bab9ac2a397a2728 and f8d8bff1cc159db4b3fe
18:25
jnthn TimToady: In the original program I'm pretty sure $foo>>[0] was being parsed as a valid hyper 18:27
(The one the RT was originally about)
Element80 Yeah, and in `run «foo $bar»` the `»` also gets parsed as a hyper, and when it fails to find an infix that's when it knows it's the closing delim instead. That's how understand the problem at least. 18:29
m: sub postfix:<♥> { $^a.say }; say <<foo bar»♥()>> 18:31
camelia (foo bar»♥())
Element80 shrugs
m: sub postfix:<♥> { $^a.say }; my $x = "foo"; say <<foo $x»♥>> 18:32
camelia Potential difficulties:
Ambiguous use of >>; use » instead to mean hyper, or insert whitespace before >> to mean a quote terminator (or use different delimiters?)
at <tmp>:1
------> 3$^a.say }; my $x = "foo"; say <<foo $x»♥7⏏5>…
jnthn Yes, but "parsed as a valid hyper" I meant that it could something that completed it afterwards
*by
*could parse
I mean, if it fails to parse it as a hyper + thing after then it just falls back to being the quote terminator that was probably intended 18:33
So if we only complain when we parsed a >> as a hyper and it actually was finally interpreted as one, and it could conflict, then we'd complain 18:34
But if we just saw the >> part and the thing after didn't work out as a hyper and we fell back to treating it as a quote terminator anyway, we could just silently let it pass 18:35
Which would unbust the new warnings we've got
Element80 cool 18:36
Geth rakudo/nom: 75797af30d | (Elizabeth Mattijsen)++ | 3 files
Re-instate baggy semantics of (<) and (<=)

  - as indicated by irclog.perlgeek.de/perl6-dev/2017-...i_14830389
  - this reverts 724cbca71046a1858a8d398 and d4436e18ce1e49eb74733
18:41
Element80 Seems to do the trick 19:04
Element80 stresstests
ZOFFLOP: t/spec/S17-procasync/kill.t 19:09
ZOFFLOP: t/spec/S17-supply/syntax.t
ZOFVM: Files=1259, Tests=141127, 120 wallclock secs (20.44 usr 3.23 sys + 2261.25 cusr 217.42 csys = 2502.34 CPU)
hm, that's a high count of secs :| 19:10
Element80 stresstests again just to be sure
ZOFVM: Files=1259, Tests=141136, 113 wallclock secs (20.14 usr 3.66 sys + 2211.79 cusr 194.50 csys = 2430.09 CPU) 19:12
Geth rakudo/nom: d39f7b9aff | (Zoffix Znet)++ | 2 files
Make ambiguous » warning less zealous

Only warn when we sucessfully parsed a hyper and are about to make an AST for it instead of right when we spot » or >>
Fixes unwanted warnings in stuff like `run «foo $bar»` while still keeping them for cases like the problematic one in the ticket[^1] the original fix[^2] fixed.
  [1] rt.perl.org/Ticket/Display.html?id=131695
  [2] github.com/rakudo/rakudo/commit/d5...3516c25d6b
19:17
Element80 That still warns for stuff like `my $x = 42; say «foo $x»[0]»` but not for `my $x = 42; say «foo $x»` 19:18
dogbert17 does anyone have suggestions as to in which file I should add tests for RT #131386 ? 19:43
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=131386
Element80 S02-types/baggy.t ? 19:44
If I were writing tests for that ticket, I'd pre-emptively test all Set/SetHash, Bag/BagHash, and Mix/MixHash 19:45
m: for <Set Bag Mix> X ('', 'Hash') { my %x := ()."$_"; %x<a> } 19:46
camelia 5===SORRY!5=== Error while compiling <tmp>
Quoted method name requires parenthesized arguments. If you meant to concatenate two strings, use '~'.
at <tmp>:1
------> 3g Mix> X ('', 'Hash') { my %x := ()."$_"7⏏5; %x<a> }
Element80 m: for <Set Bag Mix> X ('', 'Hash') { my %x := ()."$_"(); %x<a> }
camelia No such method 'Set ' for invocant of type 'List'. Did you mean any of these?
Seq
Set
Str

in block <unit> at <tmp> line 1
dogbert17 heh
Element80 m: use Test; plan 1; subtest 'can access key of empty list coerced to type' => { my @tests = <Set SetHash Bag BagHash Mix MixHash Map Hash>; plan +@tests { lives-ok { my %x := ()."$_"(); %x<a> }, $_ } } 19:48
camelia 5===SORRY!5=== Error while compiling <tmp>
Unexpected block in infix position (missing statement control word before the expression?)
at <tmp>:1
------> 3sh Mix MixHash Map Hash>; plan +@tests7⏏5 { lives-ok { my %x := ()."$_"(); %x<a> …
Element80 m: use Test; plan 1; subtest 'can access key of empty list coerced to type' => { my @tests = <Set SetHash Bag BagHash Mix MixHash Map Hash>; plan +@tests; for @tests { lives-ok { my %x := ()."$_"(); %x<a> }, $_ } }
camelia 1..1
1..8
ok 1 - Set
ok 2 - SetHash
ok 3 - Bag
ok 4 - BagHash
ok 5 - Mix
ok 6 - MixHash
ok 7 - Map
ok 8 - Hash
ok 1 - can access key of empty list coerced to type
Element80 c: 87d0e0a39e~1 use Test; plan 1; subtest 'can access key of empty list coerced to type' => { my @tests = <Set SetHash Bag BagHash Mix MixHash Map Hash>; plan +@tests; for @tests { lives-ok { my %x := ()."$_"(); %x<a> }, $_ } }
committable6 Element80, gist.github.com/3cb40dd42e91b00c3d...bb421ea358
Element80 Cool. It was also broken on Mix :) 19:49
dogbert17 perhaps you should add this now that you've done all the work :)
Element80 Nope
Element80 decommutes
dogbert17 can I steal your code ? 19:50
Element80 Sure
dogbert17 then it will sonn be fixed :)
*soon
Geth roast: dogbert17++ created pull request #279:
Added tests for RT #131386. Zoffix++
20:03
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=131386
roast: d2d60f5a1c | (Jan-Olof Hendig)++ | S02-types/baggy.t
Added tests for RT #131386. Zoffix++
20:04
roast: e156d2a03d | dogbert17++ (committed using GitHub Web editor) | S02-types/baggy.t
Merge pull request #279 from dogbert17/test-rt-131386

Added tests for RT #131386. Zoffix++
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=131386
synopsebot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=131386
star: 7ed02c57e2 | (Tom Browder)++ (committed by Zoffix Znet) | README
add specific instructions for a system installation (#90)

  * add specific instructions for a system installation
  * tweak wording per Zoffix's comment
20:37
Element80 "Should probably have <!after '='> to agree w/spec, but after NYI." 21:29
Looks like this comment (code?) is outdated: github.com/rakudo/rakudo/blob/nom/....nqp#L4471 21:30
ugexe `my $proc = Proc::Async.new("ls"); $proc.stdout; await $proc.start;` # should this await indefinitely? 21:57
geekosaur yes? you told it to capture stdout but nothing is reading it, so it could well lead to a deadlock 21:59
ugexe something nice with Proc vs Proc::Async is Proc can do silent processes easier ala run("test-command", :out, :err).exitcode == 0 vs my $proc = Proc::Async.new("test-command") andthen {.stdout.tap; .stderr.tap;}; $proc.exitcode == 0 22:00
geekosaur and no, I don't think perl 6 or anything else can possibly make IPC behave the way people always think it should magically behave. if something is not consuming data on a pipe, the other end WILL block. you can't fix that on your (the perl 6) end except by reading the pipe 22:01
Element80 .exitcode == 0 can also be written as simply .so 22:08
ugexe ah yeah. im in the habit of using .exitcode because I do a lot of probing-for-existence where you get stuff like run("prove", "--help").exitcode == 1 22:20
jnthn ugexe: If you don't .tap it then it won't start reading; that's 'cus you can also do $proc2.bind-stdin($proc1.stdout) which will plumb the fds together, in which case you'd really not like a read to happen 22:26
Also, uh, does the thing you wrote with Proc really work? 22:30
[Coke] FYI, segfault for me in doc test: github.com/perl6/doc/issues/1411 22:31
ugexe the Proc one yeah. the Proc::Async needs a $proc.start in there
jnthn `run("test-command", :out, :err).exitcode == 0` will at the very least leak handles 22:32
I'm a tad surprised it doesn't deadlock if the commands produce sufficient output
What'd be nicer is if we could provide some simple methods on Proc::Async that handle the common use cases 22:34
ugexe i thought silent procs via :out/:err worked without leaking handles because of github.com/rakudo/rakudo/commit/e4...c69c52a4a0 22:40
if it *does* leak handles then I have some code to change heh 22:43
Element80 That just waits for the proc to be done. 22:45
to set status 22:46
run("test-command", :!out, :!err).so should work, I think, 22:47
yeah, it just taps to silence it without doing anything with data
ugexe ah, I like that better visually too 22:48
jnthn Yeah was just looking at the code
Those two will do the right thing.
(:!out and :!err that is) 22:49
ugexe i wonder if thats what is causing some people issues on travis-ci with big dependency chains 22:50
jnthn Quite possibly, if the issue is they run out of handles
ugexe no one has gotten a good error. it dies abruptly or segfaults
jnthn Also today I fixed an issue where error reporting when a process was launched could end up occasionally SEGVing due to a GC rooting accident 22:51
github.com/MoarVM/MoarVM/commit/4e29e4c717 fwiw 22:52
ugexe in the cases i've seen there should not be any errors to report, unless the error reporting you're talking about is at a lower level 22:53
jnthn No, this was in a case where for example launching a process failed 22:57
ugexe ah, i guess it could be trying to report the process failed but dies abruptly because of that 22:58
[Coke] trying to use samcv's perl6-lldb-m on mac; run the command, type "run" I get the 1..309 of the test plan, and then it just hangs. :| 23:18
samcv hmm hangs? 23:19
can you show output?
like hangs how?
[Coke] "stops emitting any output" 23:20
samcv hm
odd
did you edit the library folders? 23:21
in the script?
[Coke] so perl6-lldb-m xt/space-after-comma.t ; type "run" inside lldb; get 1..309, then nothing.
samcv: yes.
it emits the text from the first chunk as expected.
samcv i get the same thing 23:22
it works with everything else i've run though
Process 9007 stopped and restarted: thread 2 received signal: SIGCHLD
no it's still running but hm
odd
i see what you're seeing though 23:23
i get the same thing in gdb though
[Coke] ... guess I woin't bother installing gdb, then. :) 23:25
samcv heh 23:26
i was gonna add lldb script to rakudo
i should do that
working on adding that [Coke] should be almost done as lng as no hitches in the makefile 23:47
[Coke], i've found the apple site has pretty good info on using lldb developer.apple.com/library/conten...17-CH7-SW1 23:48
well organized
[Coke] does 'use strict' or 'no strict' actually do anything in rakudo other than set $*STRICT ? 23:55
ah, no strict allows you to auto declare vars. 23:57