MasterDuke samcv: did you see this? github.com/MoarVM/MoarVM/pull/501#...w-16889758 00:16
samcv MasterDuke, uhm not all the types have that pointer type? 00:17
but if you make that change does it supress all warnings?
there's MVMUnicodeGraphemeNameRegistry too 00:18
MasterDuke it surpresses the one warning during a moarvm compile
samcv yeah
I wouldn't make this change. i mean
better to fix the actual source than just suppress the warning
or maybe have one MVMUnicodeXXX and cast both the other ones to that 00:19
than to cast to a type which it is not
or maybe there's some way to do it to have the macro HASH_ADD_KEYPTR wants to be something that accepts both of them, i'm not sure 00:21
let me look for a sec
MasterDuke well, UnicodeNameRegistry and UnicodeGraphemeNameRegistry are pretty much identical
samcv they aren't the same though 00:22
don't they have different types in the containing items?
yeah they do
MasterDuke yeah, but i think the one is just a typedef of the other
samcv no they are not
MasterDuke (the differing ones)
samcv see unicode.h 00:23
here github.com/MoarVM/MoarVM/blob/mast...code.h#L14
one holds a codepoint value and one holds an index of another struct
so I really don't like the idea of making both the same type 00:24
even though the size of the values inside the struct is the same
MasterDuke typedef MVMint32 MVMGrapheme32;
src/6model/reprs/MVMString.h:19: 00:25
i have no strong feelings about what the correct solution is, i just made a (what seemed reasonable at the time without knowing anything about the code) change that silenced the warning 00:27
samcv your other changes seem fine. but I don't like changing that one unless we have a better solution
MasterDuke so feel free to do or suggest something else/better
won't hurt my feelings 00:28
samcv we could always use a union tbh 00:29
i think. i am thinking
or maybe this is the opposite of what i want. i'll get back to when I figure out what will be best :) 00:30
MasterDuke sure 00:32
samcv yeah we can use a union 00:33
but that may require more changes else where. so not a good change to just make quickly 00:34
i'm gonna look at the macro 00:35
MasterDuke do you mind replaying in the comment thread? 00:36
*replying
samcv ye
MasterDuke (no need to replay the whole conversation)
thanks
samcv MasterDuke, can you try casting as (keyptr) 00:39
see if that works 00:40
er wait that's wrong. ah 00:41
Geth ast: 5d837a7a26 | (Jonathan Scott Duff)++ | S03-operators/range-basic.t
Add tests for RT#130554
 05:44
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=130554
[Tux] This is Rakudo version 2016.12-337-g7f970357b built on MoarVM version 2016.12-115-ged2df9ac 07:09
csv-ip5xs 2.889
test 12.366
test-t 5.024
csv-parser 13.224
lizmat Files=1163, Tests=56135, 186 wallclock secs (11.01 usr 4.49 sys + 1102.06 cusr 118.87 csys = 1236.43 CPU) 08:25
samcv: the reason I repeated the string in the text, is that some browser may actually not display the emoji correctly 08:26
samcv kk
though it seemed redundant. "\c[charname]" (charname) 08:29
that was all
lizmat yeah, I know, but having the emoji only would have been lost on many 08:35
and just the text after it would look strange to me without the parens
hmmm... is there a reason why the @cycle on a .rotor() cannot be lazy (apart from implementation complexity?) 09:14
|Tux| I had some git mayhem, so I am unsure if this last timing was caused by that. Just to make sure, here is a new one ā€¦ 09:30
This is Rakudo version 2016.12-351-g0cd921e35 built on MoarVM version 2016.12-115-ged2df9ac
csv-ip5xs 2.803
test 12.007
test-t 4.878
csv-parser 13.332
337 ā†’ 351 09:31
lizmat 337 ā†’ 351 ?
ah, git commits
a new low! 09:32
:-)
DrForr I'd better update :) 09:34
lizmat should take more tea: the cycle on a rotor *can* be lazy, and *is* made lazy internally anyway :-) 09:35
afk for a few hours& 09:41
brokenchicken lizmat: there's this bug tho: rt.perl.org/Ticket/Display.html?id...et-history 11:57
lizmat ok, will take care of that
:-)
well, of the rotor case, not necessarily the slurpy arg case 11:58
hmmm... or maybe I must :-(
m: sub a(*@a) { say @a.iterator.is-lazy }; a(1...*) # underlying issue 12:01
camelia rakudo-moar 0cd921: OUTPUTĀ«Falseā¤Ā»
lizmat that should have said True
jnthn m: sub a(+@a) { say @a.iterator.is-lazy }; a(1...*) # curious 12:03
camelia rakudo-moar 0cd921: OUTPUTĀ«Trueā¤Ā»
ilmari what's the difference betwen *@a and +@a? 12:04
ah, docs.perl6.org/type/Signature#Sing...ule_Slurpy 12:05
shouldn't that be s/(engage)d/$0s/? 12:07
lizmat ilmari: there's github.com/perl6/doc and PR's :-) 12:09
stackoverflow.com/questions/4169657...-in-perl-6 # why do we need the inner eager ??? 12:25
that feels like a bug to me ?
timotimo is that just "last statement not being sunk correctly"? 12:26
i.e. if we put a "say 'finished'" at the end, will it work?
lizmat trying 12:27
yup 12:28
that's it
timotimo i'm glad my intuition was right :)
lizmat yeah, adding another statement will make the inner map sink 12:29
stackoverflow.com/questions/4169657...in-perl-6/ # answered 12:39
timotimo: basically, all is as it should be, it's just that a Seq of .Seqs getting sunk, doesn't iterate over the inner .Seqs 12:40
.oO( I'm so glad I don't have to read that aloud :-)
 12:41
timotimo OK!
lizmat and I don't think it should
jnthn looks but suspects not a bug
Yeah, no bug 12:42
lizmat yeah, the outer Seq is sunk, and that doesn't run the inners Seqs
jnthn But if you're just using map for the side-effects, why not write it with `for`?
timotimo lizmat: just pronounce it like "seek"
jnthn The point of map is you're mapping something to something else :P
lizmat well, there's that :-)
timotimo because after all it's short for Sequence
jnthn
.oO( I just mapped Seqs, and it felt so good... )
 12:43
lizmat well, I'm not sure I like being chased by men in turbans :-) 12:44
with their kirpans
.oO( stackoverflow.com/questions/4169657...5#41697345 )
 12:45
oops, wrong paste
.oO( en.wikipedia.org/wiki/Sikh )
timotimo jnthn: an array let me put my items inside it? 12:46
jnthn :)
brokenchicken On the topic of slurpies there's also this ticket: rt.perl.org/Ticket/Display.html?id...et-history 13:16
buggable: tag regression
buggable brokenchicken, There are 22 tickets tagged with REGRESSION; See perl6.fail/t/REGRESSION for details
brokenchicken NeuralAnomaly: status 13:17
NeuralAnomaly brokenchicken, [āœ˜] Next release will be in 3 days and 16 hours. Since last release, there are 89 new still-open tickets (89 unreviewed and 0 blockers) and 329 unreviewed commits. See perl6.fail/release/stats for details
samcv the newest atom beta that came out a week ago is so much faster brokenchicken 13:29
lizmat brokenchicken: RT #130566 is fixed 13:30
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=130566
samcv though your computer is probably faster than mine. I can actually open big files and it won't go all funny, and opening any file is like 2x faster
lizmat brokenchicken: also fixed: RT #130513
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=130513
lizmat and RT #130510 13:31
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=130510
brokenchicken cool 13:32
timotimo that sounds great! (and much needed)
samcv what I said? or?
the slurpy
timotimo no, atom performance 13:33
samcv ah yeah
jnthn Yeah, I moved to Vim largely because I got tired of <open large file> <hang> 13:37
Doubt I'll move back, alas.
Really liking how decluttered things are in Vim. Me. The code I'm working on. That's all. 13:38
timotimo vim has *big* trouble when a file contains a line like # ----------------------------------- 13:39
it seems like
haven't golfed it yet
but at least you can ctrl-c and it'll stop the syntax highlighter and you can just scroll away
brokenchicken .ask geekosaur you know a lot about this topic. Mind looking at rt.perl.org/Ticket/Display.html?id...et-history ? Is that even a bug? 13:40
yoleaux2 brokenchicken: I'll pass your message to geekosaur.
jnthn brokenchicken: I looked at that and thought "is it really reasonable to expect Perl 6 to do that" too, tbh. But yeah, geekosaur++ will likely know better on it :) 13:42
timotimo ah, yeah, pty support would be good indeed 13:43
nine The pty command sounds useful for that. Maybe one should just push that in between Perl 6 and the program it runs? 13:47
yoleaux2 6 Jan 2017 17:20Z <notviki> nine: is this meant to work? gist.github.com/zoffixznet/b6cb79c...ad08deb68d Seems to only use the sub from first loaded version and keeps using it that version even if I reverse the loading order, unless I also nuke the .precomp
6 Jan 2017 17:26Z <notviki> nine: here are the contents of the files in that tree: gist.github.com/zoffixznet/ccdd17f...c06be8afbe
6 Jan 2017 17:30Z <notviki> nine: err, `cat v2/Foo.pm6` has { say "two"; } instead of { say "one" } (and yeah, I then changed it to say "ones" and the code kept saying "one" until I nuked precomp
9 Jan 2017 04:11Z <japhb> nine: Which phasers are expected to happen *during* precompilation, which ones are expected to happen at need time, and which ones during import time?
nine .tell notviki the FileSystem repository is meant for development mostly and therefore ignores all version, auth and api requirements and just matches the name of the module. 13:56
yoleaux2 nine: I'll pass your message to notviki.
lizmat nine: I guess you want to leave that for brokenchicken :-) 13:57
Geth ast: ronaldxs++ created pull request #223:
Test for RT #130562		 13:58
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=130562
Geth ast: 89c51acfb9 | (Ronald Schmidt)++ | S05-metasyntax/longest-alternative.t
Test for RT #130562, ltm alternation only sees first sequential alternative.
 14:02
ast: e5efaec7ba | lizmat++ | S05-metasyntax/longest-alternative.t
Merge pull request #223 from ronaldxs/mixed-ltm-sequential-alternation

Test for RT #130562
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=130562
nine brokenchicken: the FileSystem repository is meant for development mostly and therefore ignores all version, auth and api requirements and just matches the name of the module. 14:05
lizmat wonders whether synopsebot6 should be taught not to mention the same ticket more than once a minute 14:11
brokenchicken hmm 14:17
$ perl6 -e 'use lib </tmp/>; use B; use A'
===SORRY!===
Failed to open dir: 13
mc: use lib </home/bisectable/test>; use B; use A 14:19
committable6 brokenchicken, Ā¦Ā«2015.12Ā»: Ā«exit signal = SIGSEGV (11)Ā»
brokenchicken bisect: use lib </home/bisectable/test>; use B; use A
bisectable6 brokenchicken, On both starting points (old=2015.12 new=0cd921e) the exit code is 0, exit signal is 11 (SIGSEGV) and the output is identical as well
brokenchicken, Output on both points:
brokenchicken k
That was rt.perl.org/Ticket/Display.html?id...et-history seems we had the issue for ages 14:20
nine 13 is permission denied
brokenchicken Ah
nine always loves the scavenger hunt for ERRNO constants in /usr/include 14:21
errno.h -> bits/errno.h -> linux/errno.h -> asm/errno.h -> asm-generic/errno.h -> asm-generic/errno-base.h 14:22
ilmari moreutils has errno(1) 14:34
Geth kudo/nom: 7d5bbefe7f | (Zoffix Znet)++ | docs/ChangeLog
Fix formatting
 14:35
ilmari $ errno ENOSPC
ENOSPC 28 No space left on device
$ errno 37
ENOLCK 37 No locks available
etc
brokenchicken lol 14:39
$ errno 25
ENOTTY 25 /* Not a typewriter */
lizmat afk& 14:41
ilmari brokenchicken: the modern wording for that one is Ā«Inappropriate ioctl for deviceĀ» 14:42
[Coke] samcv: any plans to add a routine to provide a runtime version of "\c[]" ? 14:48
samcv hmm? 14:49
oh.
that that could be nice
nine ilmari: wow, I'd have probably needed every single program contained in moreutils at one point or another. Thanks for pointing out that gem!
ilmari nine: yeah, it's very handy
samcv moreutils is great 14:50
ilmari I use sponge the most
samcv same
[Coke] samcv: without a routine, you end up with: github.com/coke/p6-unicode-mangler...gle.p6#L82 14:51
samcv [Coke], how does unifromname sound? though idk maybe makes it sound like it returnrs a Uni object
but uniname works only on strings or codes not on Uni so
there's that
[Coke] I mean, uniname is the right name, but it got stolen for the other direction. :) 14:53
I don't like unifromname, but don't have a great alternative. from-unicode-name?
samcv sure 14:56
idk. we need a poll
[Coke] I mean, they all suck. 14:57
samcv yeah
but one of them sucks the least
also i need to sleep
i think from-unicode-name could be ok i guess 14:59
brokenchicken Seems this month will have the biggest changelog that we had for a while... 16:33
[Coke] should we have a Uni.Range to show the range of codepoints? 16:42
timotimo btw do we expose our supported unicode version somewhere? 16:43
$*VM.uni-version or something could be used
[Coke] timotimo: been asking for that since christmas, nope. 16:47
timotimo oh
[Coke] even better: make it settable!
what's the highest uni ord? 16:48
timotimo mirror mirror on the wall .. 16:50
brokenchicken m: 0x10FFFF.chr.say 16:53
camelia rakudo-moar 7d5bbe: OUTPUTĀ«\x{10FFFF}ā¤Ā»
nine timotimo++ # reminding me of Blind Guardian which I haven't listened to in ages.
ilmari 'Supplementary Private Use Area-B' ends at U+10FFFF
brokenchicken m: 0x10FFFFFFFFFF.chr.say 16:54
camelia rakudo-moar 7d5bbe: OUTPUTĀ«chr codepoint cannot be negativeā¤ in block <unit> at <tmp> line 1ā¤ā¤Ā»
brokenchicken heh
timotimo nine: when i think "The Mirror", i'm reminded of Dream Theater 16:56
nine timotimo: :) 17:05
TimToady samcv: I'd just call it "uni", since it'll usually be obvious from the name what the input is :) 17:14
s/name/argument
and it's perfectly fine to name something by what it produces, generally 17:15
hmm, if she's on california time, looks like she coded all night and then collapsed in a heap 17:17
TimToady fondly remembers being young enough to do that...
[Coke] TimToady: I miss those days myself.
geekosaur re rt.perl.org/Ticket/Display.html?id...et-history they are expecting Proc::Async to be an Expect implementation, essentially. although it might be nice if it did provide the option to use a pty instead of a pipe (this should not be default as it's higher overhead and ptys are a limited resource compared to pipes) 17:45
yoleaux2 13:40Z <brokenchicken> geekosaur: you know a lot about this topic. Mind looking at rt.perl.org/Ticket/Display.html?id...et-history ? Is that even a bug?
b2gills .o( I wonder if Proc::Asyc.new could have a :tty flag added ) 17:55
geekosaur the real question is likely whether libuv provides it, since iirc Proc::Async defers to libuv for most stuff 18:05
timotimo it should 18:09
jnthn Seems not 18:10
timotimo huh, seems like 18:11
jnthn At least, the searches I did for libuv pty suggested not 18:12
timotimo it mean: seems like not 18:16
geekosaur actually, looks like it does but it doesn't connect the dots for you. docs.libuv.org/en/v1.x/tty.html 18:20
nor does it appear to have an API specifically for getting a pty; it just lets you specify one. (uv_process_t *does* support bidirectional "pipes", which only makes sense with a pty or on freebsd) 18:22
Geth kudo/nom: e041b0f827 | (Pawel Murias)++ | src/Perl6/World.nqp
Fix f67df8ad9d4f920bb63ea40501f3b1db9d8e661d.
 20:25
lizmat pmurias: is that a revert ? 20:26
Ah, it is... 20:27
too bad.. :-)
pmurias lizmat: it's not a revert
lizmat ah, good ! 20:28
b2gills That looks like it is an infinite loop github.com/rakudo/rakudo/commit/e0...330fcR1528 20:34
pmurias sorry 20:36
b2gills You can't learn from mistakes that you don't make 20:38
lizmat jnthn: would you mind terribly if I added a .perl method to IterationBuffer ? 20:39
m: my $l := IterationBuffer.new; $i.push(42); dd $i
camelia rakudo-moar 7d5bbe: OUTPUTĀ«5===SORRY!5=== Error while compiling <tmp>ā¤Variable '$i' is not declaredā¤at <tmp>:1ā¤------> 3my $l := IterationBuffer.new; 7ā5$i.push(42); dd $iā¤Ā»
lizmat m: my $l := IterationBuffer.new; $l.push(42); dd $l 20:40
camelia rakudo-moar 7d5bbe: OUTPUTĀ«IterationBuffer.newā¤Ā»
lizmat :-(
pmurias do we loose anything by adding a .perl method to IterationBuffer? 20:49
Geth kudo: usev6++ created pull request #996:
[JVM] Don't use --nqp-lib=blib for evalserver		 20:50
lizmat pmurias: well, I'm not 100% sure: jnthn was pretty adamant about keeping it as simple as possible 20:52
multi method perl(IterationBuffer:D:) { nqp::p6bindattrinvres(nqp::create(List),List,'$!reified',self) } 20:53
should do the trick
jnthn A .perl is probably only going to help people debug stuff 21:18
So long as we don't change the REPR or add any validation to the methods it should be fine. 21:19
I don't see any negatives to adding a .perl method :)
I guess the "should do the trick" example is missing a .perl though, and that's not really correct even 21:20
Because it doesn't actually create the IterationBuffe
*IterationBuffer
It'd create a List
lizmat multi method perl(IterationBuffer:D:) { nqp::p6bindattrinvres(nqp::create(List),List,'$!reified',self).perl } 21:29
it would convert the IterationBuffer to a List and then .perl it
that should work, no? 21:30
timotimo doesn't round-trip :) 21:31
Geth kudo/nom: 39bf63f574 | usev6++ | tools/build/create-jvm-runner.pl
[JVM] Don't use --nqp-lib=blib for evalserver

  ... and perl6-debug-j. That makes them work after
  'make install'. (They still fail to work otherwise.)		 21:32
kudo/nom: 71fd4178bf | lizmat++ | tools/build/create-jvm-runner.pl
Merge pull request #996 from usev6/jvm_build_blib

  [JVM] Don't use --nqp-lib=blib for evalserver
lizmat timotimo: don't care
actually, it would, as the reified normally *is* an IterationBuffer
well, in a way :-) 21:33
timotimo hm
jnthn :P 21:36
Maybe it should .perl to `do { given IterationBuffer.new { .push(1); .push(2); ... $_ } } :P 21:37
But yeah, if we add a better .perl it should try to round-trip 21:38
lizmat FLATTENABLE_LIST ? :-) 21:39
anyways, we can't build until pmurias has fixed World.nqp, right ? 21:43
bartolin lizmat: btw, it looks like r-j has the same problem with nqp::null/Mu with $!next in R::It.CrossIterables and R::It.CrossIterablesMap as it had in R::It.OneValue :-/ 21:48
lizmat: I'll try a spectest with a similiar workaround tomorrow
lizmat I feared it would and hoped the nqp::null / Mu issue could be solved at a deeper level
the thing is that checking for nqp::null is about 2x as fast as checking for Mu 21:49
and these are always hot paths, so I would like to keep that :-)
bartolin yeah, it definitely should -- but I don't know how to fix it
lizmat ok, from now on I will add jvm specific code in the new stuff I'm writing 21:50
bartolin will keep searching for a fix in nqp-j 21:51
lizmat maybe jnthn has some ideas 21:52
jnthn lizmat: No, that's an internal detail that may go away some day :)
(to FLATTENABLE_LIST)
lizmat yeah... and it even doesn't do what I wanted it to do (on a List) 21:58
jdv79 so... my laptop froze 2 times in a row running make install 22:11
wondering if i should try again 22:12
as in kernel crash
mst make *install* crashed it 22:16
how bizarre
bartolin .ask [Coke] do you have a suggestion whom I could contact about my CLA (cmp irclog.perlgeek.de/perl6-dev/2017-...i_13920361 ) 22:17
yoleaux2 bartolin: I'll pass your message to [Coke].
jdv79 actually no
lizmat jdv79: maybe it ates all the memory ?
jdv79: also, do you have pmurias' last commit as well ?
jdv79 that seems more plausible. box was dead but if i wait a few minutes it comes back
yes
lizmat ah, then you need to revert that locally for now
jdv79 oh cool. thanks! 22:18
sortiz Commit e041b0f broke rakudo's build.
lizmat pmurias: are you ok with a revert of e041b0f, or are you close to a solution ? 22:19
travis-ci Rakudo build errored. Pawel Murias 'Fix f67df8ad9d4f920bb63ea40501f3b1db9d8e661d.' 22:21
travis-ci.org/rakudo/rakudo/builds/192811365 github.com/rakudo/rakudo/compare/7...41b0f8276b
buggable [travis build above] ā˜  Did not recognize some failures. Check results manually.
geekosaur even travis piling on :p 22:24
lizmat m: say [3] ~~ Cool # huh ??? 22:25
camelia rakudo-moar 7d5bbe: OUTPUTĀ«Trueā¤Ā»
lizmat m: say [Nil] ~~ Cool # huh ???
camelia rakudo-moar 7d5bbe: OUTPUTĀ«Trueā¤Ā»
sortiz The problem with the commit is that the new build_container_add_to_sc method (line 1527) calls itself, a typo imo 22:26
Geth kudo/nom: 5 commits pushed by lizmat++
kudo/nom: c6662d14be | (Elizabeth Mattijsen)++ | src/Perl6/World.nqp
Revert "Fix f67df8ad9d4f920bb63ea40501f3b1db9d8e661d."

This reverts commit e041b0f8276b2abfbeefc3746c6fd26be33ef598.		 22:27
lizmat hmmm... apparently List is Cool ??? 22:29
that is a bit of a WAT to me
Geth kudo/nom: 92e704b4e7 | (Elizabeth Mattijsen)++ | src/core/Rakudo/Iterator.pm
Introducing R:It.Rotor, the new .rotor workhorse
 22:40
sortiz lizmat, the simple change in gist.github.com/salortiz/ceccd0c18...9d3b04ef89 fix pmurias commit _and_ fix DBIish also. 22:44
b2gills .oO( Maybe I should go and get a commit bit, because if I had one I would have done ^ over two hours ago when I pointed the problem out ) 22:49
brokenchicken Nil and Failure being Cool are more of a wat to me :/ 22:51
sortiz b2gills, sure. As pmurias acknowledged the problem I was only waiting for the fix...
brokenchicken But yeah, lists are. Recall that debate on stringy methods doing weird things with Lists, like .contains() almost works. 22:52
Geth kudo/nom: d7b82149d3 | (Elizabeth Mattijsen)++ | src/core/List.pm
Make List.rotor between 15x and 20x faster

By either using the new R:It.Batch or R:It.Rotor iterators		 22:57
lizmat is testing sortiz' patch 22:59
pmurias sorry, forgot I didn't push the commit :/ 23:00
lizmat *sigh* 23:03
b2gills lizmat: I tested it ( or rather I copied it by hand, and tested that ) Spec test passes
lizmat well, I've learned not to push unless I have done a "make spectest" the hard way
Geth kudo/nom: 703799ec24 | (Elizabeth Mattijsen)++ | src/Perl6/World.nqp
Restore pmurias++ work, sortiz++, b2gills++
 23:07
lizmat and all of a sudden,. t/spec/S12-construction/destruction.t has 2 TODOd tests passing
that file hasn't basically been touched since 2009!
not sure whether that is a good thing or not 23:08
brokenchicken lizmat++ # dat speed gain :o 23:11
pmurias do we have a 'make spectests' like thing that tests that the important modules in the ecosystem aren't broken? 23:12
brokenchicken none that I know of 23:13
sortiz Well, I can confirm DBIish is now fixed 23:14
lizmat whee!
brokenchicken pmurias: but that's a good idea. At the very least should be part of pre-release checks 23:15
lizmat well, that's called "cpan-testers" more or less :-)
maybe mst has some ideas about that :-)
pmurias cpan-testers is awesome 23:27

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!