jnthn sleep time; 'night 00:05
timotimo inlining is currently prevented when natives are involved 00:18
travis-ci Rakudo build failed. Elizabeth Mattijsen 'Manually inline swapping logic 01:44
travis-ci.org/rakudo/rakudo/builds/193059515 github.com/rakudo/rakudo/compare/4...4aeb32ddb2
buggable [travis build above] ☠ Did not recognize some failures. Check results manually. 01:45
MasterDuke .tell jnthn sorry, i guess i wasn't clear. to me, what you said may be useful sounded very like what you said the ops i added actually did. i don't mind ditching the PR (it was mostly done as a learning experience), but i just don't quite understand your comments 02:12
yoleaux2 MasterDuke: I'll pass your message to jnthn.
travis-ci Rakudo build passed. Elizabeth Mattijsen 'Remove destruction tests of doubtful value 03:31
travis-ci.org/rakudo/rakudo/builds/193060808 github.com/rakudo/rakudo/compare/c...a05c260329
Rakudo build errored. Elizabeth Mattijsen 'Make R:It.Batch handle Inf * large ints better 04:48
travis-ci.org/rakudo/rakudo/builds/193071395 github.com/rakudo/rakudo/compare/d...a66b7fa7fd
Rakudo build failed. Elizabeth Mattijsen 'Fix RT #130436 05:50
travis-ci.org/rakudo/rakudo/builds/193084913 github.com/rakudo/rakudo/compare/6...dc5f7cea73
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=130436
travis-ci Rakudo build failed. Zoffix Znet 'Log all commits to date 06:33
travis-ci.org/rakudo/rakudo/builds/193108134 github.com/rakudo/rakudo/compare/7...36f8f753b1
Rakudo build failed. Zoffix Znet 'Fix typo MasterDuke++' 07:15
travis-ci.org/rakudo/rakudo/builds/193118048 github.com/rakudo/rakudo/compare/f...a3c08fd12e
[Tux] This is Rakudo version 2016.12-388-g1aeea1062 built on MoarVM version 2016.12-115-ged2df9ac 07:16
csv-ip5xs 2.870
test 11.955
test-t 4.873
csv-parser 13.120
lizmat Files=1162, Tests=56138, 186 wallclock secs (10.94 usr 4.58 sys + 1101.42 cusr 117.27 csys = 1234.21 CPU) 08:58
travis-ci Rakudo build failed. Elizabeth Mattijsen 'Fix for RT #130439 09:19
travis-ci.org/rakudo/rakudo/builds/193194773 github.com/rakudo/rakudo/compare/9...f14a5c1f55
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=130439
lizmat hmmm... is Geth down ? 09:40
brokenchicken: ^^^ ?
timotimo looks kind of like
lizmat afk for a few hours& 09:41
timotimo it doesn't run on p6c, so i can't just jump in and restart it or anything
lizmat github.com/rakudo/rakudo/commit/c9...855660812a # just committed
afk&
timotimo ooooooh 09:43
brokenchicken hm. I see it reporting that one in the terminal log... some weird netsplit? 11:33
also, buggable is at 2.5GB of RAM and my box is swapping :'( 11:34
AlexDaniel brokenchicken: run it on whateverable server then? 11:35
brokenchicken AlexDaniel: so it'd eat all the RAM there? 11:36
AlexDaniel yea
brokenchicken Geth: ver 432 11:37
Geth: ver github.com/rakudo/rakudo/commit/c9...db9f9f42ed
Geth brokenchicken, version bump brought in these changes: github.com/perl6/nqp/compare/2016....7-g7b6fff9
AlexDaniel there are 10GB free atm, so it will handle one more bot easily
brokenchicken It'd just use all the 10GB 11:38
brokenchicken doesn't see the point
timotimo did you already get jnthn's latest fixes?
AlexDaniel it will take some time to get to this mark :)
brokenchicken Yes.
AlexDaniel: it depends on how often travis reports stuff
Took it 12 hours to get from 300MB to 2.5GB 11:39
jnthn I doubt you've got the latest round of fixes, which are still in PRs 11:44
yoleaux2 02:12Z <MasterDuke> jnthn: sorry, i guess i wasn't clear. to me, what you said may be useful sounded very like what you said the ops i added actually did. i don't mind ditching the PR (it was mostly done as a learning experience), but i just don't quite understand your comments
jnthn I will merge one of them pre-relesae
The other will wait until after
Because it introduces a few new sanity checks that will VM-panic
brokenchicken Ah, no, then no
Actually it's 36 hours. not 12
jnthn And I'm pretty comfortable they won't be tripped over by code out there, but I don't want to find out the day after the release about this ;) 11:45
timotimo oh, of course 11:48
maybe i should take the chance to finally merge in line_based_coverage after the release 11:50
brokenchicken \o/
timotimo that'll prevent all the conflicts all the damn time just because the oplist changes
jnthn timotimo: Can you PR it, and I'll give it a quick review before we merge it? 11:51
timotimo sure
i'm not 100% happy with the branch, because it still has a noticable false-negative rate for covered lines, and that's super annoying
lizmat wonders what masak would have to say about stackoverflow.com/questions/4173064...g-anything 12:37
arnsholt It's not entirely clear to me what he wants 12:43
timotimo i know what he wants 12:44
create a new Match object that gets put into the tree that has its own .orig and completely made up .from and .to and such 12:45
arnsholt "Doctor, it hurts when I do this."
Geth kudo/nom: 017c6cf523 | (Elizabeth Mattijsen)++ | src/core/List.pm
Make List/Array.sum about 30% faster

Especially for smaller lists/arrays		 12:48
arnsholt added a comment 12:50
jnthn I think the answer is "don't do that, use make/.made to hang extra data off the match object" 12:51
arnsholt Yeah, that's more or less what I tried to say 12:52
jnthn I'm also very curious about the scope of %Market_Shortcode in that code 12:53
If it's constant or so it's fine
Yeah, agree wiht your answer, arnsholt++ 12:54
*with
Geth kudo/nom: 22e589acce | (Elizabeth Mattijsen)++ | src/core/List.pm
Make List/Array.fmt() about 60x faster

Because we don't actually have to do a sprintf internally for the default format '%s', but we can do just a .Str. Apparently sprintf uses a regular expression somewhere under the hood, and that makes things a *lot* slower in this case.		 13:11
timotimo i wonder how often you actually .fmt()? 13:13
lizmat timotimo: I don't know either 13:14
timotimo how expensive is a check for "it's just '%s'" for when you want a different separator?
lizmat timotimo: ah, good point 13:16
timotimo we could perhaps also make a cheap check to see if it's just a single %s inside the string and do a regex-less thing? 13:17
like "(%s)" could be super fast that way
does simplematch still exist? or did all the optimizations from that get put into match itself? 13:20
lizmat yeah, it did 13:24
afaik
brokenchicken sprintf got a whole grammar for it: github.com/perl6/nqp/blob/master/s...printf.nqp 13:46
lizmat arnsholt: I think brian needs more guidance :-) 13:57
jnthn lizmat: I think it boils down to, "if you want to do this, you should probably think about your problem differently" :) 13:58
lizmat: And the different thinking is action methods, which two people already pointed out.
Or not even action methods, you can `make` in an inline block inside the token too 13:59
lizmat feels to me he's not grokking action methods, or the power of 'make'
yet
jnthn Aye, that can take a bit of time :) 14:01
Once he does, though, I suspect he'll be able to write a nice explanation of them :)
lizmat see it as an investment :-) 14:02
fwiw, I'm not fully grokking this either :-)
like I understand how an internal combustion engine works, but don't let me fix your car :-) 14:03
brokenchicken hehe 14:05
travis-ci Rakudo build passed. Zoffix Znet 'Fix typo timotimo++ ; add missing commit MasterDuke++'
travis-ci.org/rakudo/rakudo/builds/193226966 github.com/rakudo/rakudo/compare/b...eea1062f80
brokenchicken wow, it just finished doing it. 14:06
Geth kudo/nom: 7ef3682cf0 | (Elizabeth Mattijsen)++ | src/core/List.pm
Make List/Array.fmt("...") also about 60x faster

Provided the format string only has a single '%s' in it.
  timotimo++ for inspiration		 14:08
brokenchicken :o 14:09
timotimo <3 14:11
brokenchicken ♥ 14:12
perlpilot re bdf's SO question ... I dunno, I think he's just experimenting and looking for a way to manipulate the Match objects no matter what actually matched as a way to test the boundary of what's possible (rather than using this as an actual technique for anything) 14:16
DrForr You can probably do that from within nqp, but really from the user perspective just building a proper AST would be the way to go. 14:17
perlpilot aye 14:19
Perhaps I'm being to naive, but I look at it like he's exploring a territory by going beyond the bounds of that territory so that he knows exactly where the boundary is 14:20
s/to/too/
awwaiid One of my favorite learning techniques 14:21
lizmat commute to Niederrhein.pm meeting& 14:24
brokenchicken .seen FROGGS 14:30
yoleaux2 I saw FROGGS 30 Dec 2016 12:44Z in #moarvm: <FROGGS> samcv++
timotimo :(
brokenchicken .ask FROGGS how come no test for rt.perl.org/Ticket/Display.html?id...et-history ?
yoleaux2 brokenchicken: I'll pass your message to FROGGS.
arnsholt There. Tripled size of answer 14:49
lizmat arnsholt++ 15:06
arnsholt (As an aside, one of the reasons I backed his new book was specifically to see how he treated grammars while it was happening) 15:07
travis-ci Rakudo build passed. Elizabeth Mattijsen 'Make reified List.Array about 3x faster 15:11
travis-ci.org/rakudo/rakudo/builds/193323658 github.com/rakudo/rakudo/compare/1...a9bc855660
Geth kudo/nom: ed482ec9b1 | (Elizabeth Mattijsen)++ | src/core/List.pm
Make List/Array.join about 20% faster
 15:19
brokenchicken seems jvm nqp doesn't even build anymore? `make` is failing with gist.github.com/zoffixznet/ce21b1e...8288976480 15:40
re-run of make fuller error: gist.github.com/zoffixznet/f586489...b79f5eac85 15:41
Geth ast: 890ca310e7 | (Zoffix Znet)++ | S32-array/create.t
Test array elements get writable containers

RT#130583: rt.perl.org/Ticket/Display.html?id=130583		 15:56
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=130583
Geth ast: 1c61263ca4 | (Zoffix Znet)++ | S02-types/array-shapes.t
Test .List on uninited shaped arrays does not die

RT#130513: rt.perl.org/Ticket/Display.html?id=130513		 16:01
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=130513
brokenchicken :/ 16:07
Geth ast: 1f05be281e | (Zoffix Znet)++ | S02-types/array-shapes.t
Test @c[some shape] accepts a .new: :shape(same shape)...

RT#130510: rt.perl.org/Ticket/Display.html?id=130510
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=130510
brokenchicken \o/
lizmat waves from Cologne 17:41
brokenchicken \o 17:42
bartolin brokenchicken: wrt nqp-j not compiling: it seems to be 994d2159fc. with that commit revertet nqp-j builds 17:46
*reverted 17:47
brokenchicken .ask TimToady any idea for fixing this commit's breakage of JVM build? github.com/perl6/nqp/commit/994d2159fc `make` (in nqp) now fails with gist.github.com/zoffixznet/f586489...b79f5eac85
yoleaux2 brokenchicken: I'll pass your message to TimToady.
TimToady oh, gee, I'm always forgetting that end of things 17:52
yoleaux2 17:47Z <brokenchicken> TimToady: any idea for fixing this commit's breakage of JVM build? github.com/perl6/nqp/commit/994d2159fc `make` (in nqp) now fails with gist.github.com/zoffixznet/f586489...b79f5eac85
TimToady should just be the same thing as the MoarVM patch, decode fates[0] into fbeg and fend (32 bits each) and then offset the fate numbers returned from the actual NFA runner 17:53
I have to commute to norcal today, but will do it tonight if no one else gets to it 17:54
jnthn: and yes, separate compilation is an issue there, once we start trying to make use of the uniqueness
bartolin ah, this moarvm commit: github.com/MoarVM/MoarVM/commit/e167934e83 17:57
TimToady yes, that one 18:01
afk & 18:02
brokenchicken m: (1...*).perl 18:46
camelia rakudo-moar ed482e: OUTPUT«Cannot .elems a lazy list␤ in block <unit> at <tmp> line 1␤␤Actually thrown at:␤ in block <unit> at <tmp> line 1␤␤»
brokenchicken I guess that's normal...
m: (1, {*+*} ...*).perl # 'cause it could be weird stuff like this too 18:47
camelia rakudo-moar ed482e: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Malformed double closure; WhateverCode is already a closure without curlies, so either remove the curlies or use valid parameter syntax instead of *␤at <tmp>:1␤------> 3(1, {*+*}7⏏5 ...*).perl # 'ca…»
lizmat brokenchicken: one could argue that the error is LTA, and that it should say "cannot .perl a lazy list" ? 19:00
brokenchicken Maybe 19:01
[Coke] I think it's always LTA when the error refers to something the user didn't type 19:05
jnthn It's OK if they can see the stack trace showing how they ended up there, but yeah, better to put it straight on .perl 19:43
lizmat decommute& 19:56
Geth kudo/nom: a423edaf30 | (Elizabeth Mattijsen)++ | src/core/Any-iterable-methods.pm
Replace a die with an applicable Exception
 22:38
kudo/nom: dd911efeb9 | (Elizabeth Mattijsen)++ | src/core/List.pm
Remove unnecessary initialization

Probably a copy-pasto		 22:40
lizmat good night, #perl6-dev! 22:48
brokenchicken night

Powered by Cro and the Raku® Programming Language.

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

Thank you!