jdv79 i tried using 6.d.preview but i still had to up the max threads to get my very async program to not bind up. i thought that shouldnt happen. 03:48
i basically kick off 30 or so start blocks that sleep and loop forever 03:51
ugexe actual sleep, or a Promise.in(...) type thing 04:02
?
jdv79 literal sleep() 04:03
geekosaur I thought 6.d's sleep was fixed to be multithread sane 04:40
Geth rakudo: ugexe++ created pull request #1125:
Add candidates method as CURI recommendation manager
05:19
[Tux] This is Rakudo version 2017.07-90-g51e59eeb5 built on MoarVM version 2017.07-263-gb161c851 08:34
csv-ip5xs 2.613
test 12.610
test-t 4.389 - 4.369
csv-parser 12.861
samcv bumping nqp now so we get the fixed string compare so cmp has deterministic comparing of synthetic codeponits. before it compared the synthetic codepoints as negative numbers instead of the actual codepoints that it held 10:12
the speed should be the same except if the codepoint which is different between the two strings is a synthetic, in which case it checks the codepoint inside the synthetic and compares that
for two synthetics with the same starting codepoints but one is shorter than the other, the grapheme with less codepoints in it will be 'Less' 10:13
also improvement to make .codes 350% faster
Geth rakudo/nom: 0564891e9c | (Samantha McVey)++ | tools/build/NQP_REVISION
Bump NQP for new nqp::codes op, string cmp fixes

Changes brought in by MoarVM bump: abc38137 Fix MVM_string_compare to support deterministic comparing of synthetics b161c851 be safe about marking NULL call graph nodes 1d3a139b Avoid range check on every SC object access. 5c67d732 calloc a tospace instead of memset old fromspace. ... (10 more lines)
10:16
rakudo/nom: version bump brought these changes: github.com/perl6/nqp/compare/2017....g393eb0165
3c6277c77e | (Samantha McVey)++ | src/core/Str.pm

Makes it much less costly than calling .NFC.codes. For long strings the times are not quite as different since the bulk of the time is actually iterating through the string and not the unneeded calls.
Ā¦ rakudo/nom: version bump brought these changes: github.com/perl6/nqp/compare/2017....g393eb0165
Zoffix There's nothing to "fix" with sleep. If you want non-blocking sleep, use Promise.in(42) 10:47
jdv79: 6.d.PREVIEW made `await` nonblocking. You're still blocking your threads with sleep().
m: use v6.d.PREVIEW; await ^100 .map: {start await Promise.in: 2} 10:52
camelia ( no output )
Zoffix m: use v6.d.PREVIEW; await ^100 .map: {start await Promise.in: 2}; say now - INIT now;
camelia 2.022665?
Zoffix m: await ^100 .map: {start await Promise.in: 2}; say now - INIT now;
camelia (timeout)
stmuk is there anything written for the 6.d release plans next year .. or is it just in backlog? 10:58
dogbert2 anyone seen this? 11:11
Stage optimize : 8.577
Stage mast : No registered operation handler for 'codes'
while building rakudo 11:12
travis-ci Rakudo build failed. Samantha McVey 'Have .codes use nqp::codes op. 350% faster for short strings 11:16
travis-ci.org/rakudo/rakudo/builds/259058339 github.com/rakudo/rakudo/compare/5...6277c77ed0
stmuk :)
buggable [travis build above] ? Did not recognize some failures. Check results manually.
dogbert2 sigh 11:20
same error on travis 11:21
samcv hmm what did i do wrong 11:55
oh crap 11:56
Geth nqp: 59421ffe1b | (Samantha McVey)++ | src/vm/moar/QAST/QASTOperationsMAST.nqp
Add nqp::codes to QASTOperationsMAST.nqp
11:57
timotimo heh. 11:58
Geth rakudo/nom: e051dd2def | (Samantha McVey)++ | tools/build/NQP_REVISION
Bump NQP to get nqp::codes op actually added
rakudo/nom: version bump brought these changes: github.com/perl6/nqp/compare/2017....g59421ffe1
4db230641e | (Elizabeth Mattijsen)++ | src/core/Rakudo/Iterator.pm
timotimo happens when you have three separate repositories
Zoffix stmuk: 6.d release is this year 12:00
stmuk: everything is here: github.com/perl6/6.d-prep 12:01
stmuk: aside from plans to cut a new language point release quarterly next year (basically for each R* release)
stmuk ah excellent I'll keep an eye on that repo 12:03
Zoffix: is that something like 6.d.1 .2 etc?
Zoffix Yeah 12:05
stmuk is Oct likely to be a "soft" release or (from marketing perspective) should we "big" it up a bit? 12:13
Zoffix has no idea what that means 12:21
stmuk a soft release is a quiet release made due to the possibility of problems rather than more agressive promotion 12:24
Zoffix It will be promoted once we know for certain when we'll be able to release. Wouldn't think we'd be doing anything "agressive" considering most of the new features were stuffed into 6.c compiler, so the actual list of things people will get with use 6.d is rather slim 12:26
and because overall performance is still in need of improvement. Those who weren't impressed with earlier releases would likely measure that first thing when they hear about hyped up 6.d release, which will likely end like this: www.reddit.com/r/perl6/comments/6h...o_gnu_yes/ 12:29
stmuk yes that's true maybe best to hold fire until speed improvements are more noticeable 12:32
I did notice the perl 6 search queries peaked around Larry's Perl 6 release annoucement in Oct 2015 rather than the actual release 12:33
trends.google.com/trends/explore?d...q=perl%206 12:34
jnthn I think we get ourselves something we'd be willing to ship and call 6.d, and *then* announce 13:02
NeuralAnomaly: status 13:03
NeuralAnomaly jnthn, [?] Next release will be in 2 weeks and 6 days. Since last release, there are 23 new still-open tickets (6 unreviewed and 0 blockers) and 97 unreviewed commits. See perl6.fail/release/stats for details
travis-ci Rakudo build passed. Samantha McVey 'Bump NQP to get nqp::codes op actually added' 13:12
travis-ci.org/rakudo/rakudo/builds/259073195 github.com/rakudo/rakudo/compare/3...51dd2def01
Zoffix Yeah, sounds good. 13:56
dogbert2 hmm, quite a few ZOFFLOP's all of a sudden ... 15:38
t/spec/S02-types/unicode.t, t/spec/S32-temporal/DateTime.t, t/spec/S32-str/length.t and t/spec/S32-temporal/Date.t 15:40
dogbert2 wonders if he's missing something obvious 15:41
m: say "\c[LATIN CAPITAL LETTER A WITH DOT ABOVE, COMBINING DOT BELOW]".codes 15:47
camelia 1?
dogbert2 hmm 15:48
timotimo m: say "\c[LATIN CAPITAL LETTER A WITH DOT ABOVE, COMBINING DOT BELOW]".NFC.perl 16:10
camelia Uni.new(0x1ea0, 0x0307).NFC?
timotimo m: say uninames(|"\c[LATIN CAPITAL LETTER A WITH DOT ABOVE, COMBINING DOT BELOW]".NFC.list)
camelia Cannot resolve caller uninames(Int, Int); none of these signatures match:? (Str:D $str)? in block <unit> at <tmp> line 1??
timotimo m: say uninames("\c[LATIN CAPITAL LETTER A WITH DOT ABOVE, COMBINING DOT BELOW]".NFC.list)
camelia Cannot resolve caller uninames(Seq); none of these signatures match:? (Str:D $str)? in block <unit> at <tmp> line 1??
timotimo m: .&uniname.say for "\c[LATIN CAPITAL LETTER A WITH DOT ABOVE, COMBINING DOT BELOW]".NFC.list
camelia LATIN CAPITAL LETTER A WITH DOT BELOW?COMBINING DOT ABOVE?
timotimo that's strange 16:11
er, huh 16:12
didn't sam just change the codes function
at the moment the implementation is the same as chars, i.e. it just reads the size property out of the string 16:14
dogbert2 timotimo: perhaps she'd hacked all night and was so tired that she forgot 16:26
timotimo i've certainly done that before
dogbert2 :) 16:27
did you see the bug which was uncovered yesterday?
gist.github.com/dogbert17/e23c3340...eda3feb18f
timotimo yeah, that was fascinating, but i haven't invested any time in it yet
dogbert2 should I report it or wait for you to fix it :) 16:29
timotimo not sure if it could end up in a roast test, so maybe a RT isn't suitable? 16:30
dogbert2 MoarVM issue perhaps 16:33
I wonder if it possible for valgrind to tell which variable is uninitialized 16:35
timotimo spread it out over multiple lines :P 16:41
yoleaux Zoffix: Zoffix testing: brson.github.io/2017/07/10/how-rust-is-tested 18:12
Zoffix Shoo, robot 19:00
mst: "??? _ :moon.freenode.net 465 ZofBot You are banned from this server- You have violated freenode's terms of service. Please contact [email@hidden.address] if you think this kline has been set in error. (2017/7/30 15.26)" ??? _ :ZofBot!zofbot-zof@perl6.party QUIT K-Lined 19:01
What sort of service did I violate :/
mst Zoffix: lemme guess, linode?
Zoffix mst: Yeah 19:02
mst Zoffix: yeah, I accidentally half of linode again earlier
Zoffix haha
mst they hand out single v6 addresses rather than normal sized blocks
Zoffix \o/
ZofBot: I missed you, buddy
ZofBot Zoffix, new
mst so what superficially appears to be a tight v6 mask that should only affect the person you're attempting to k-line 19:03
Zoffix Ah :)
mst I do try to remember this before dropping a v6 k-line, but I don't always succeed 19:04
lizmat . 19:21
hmmm... I see quite a lot of spectest breakage atm 19:28
they all seem related to unicode stuff 19:29
samcv checking now lizmat 19:36
timotimo i wonder if including WWW in star would be a good idea? 19:39
Zoffix IMO a good replacement for LWP::Simple, but not much more than that. 19:40
timotimo OK
Zoffix Though its design is a bit botched.
timotimo i haven't realized
should LWP::Simple see a bit more development? like, more methods than just get and post? 19:41
Zoffix post() takes headers as a positional arg and form params as named, but if you send raw body, then it takes headers as named
IMO LWP::Simple was botched by making it UNsimple. Why the hell do I need to type the module name twice to make a request :/
samcv lizmat, what tests do you see failing?
lizmat samcv: gist.github.com/lizmat/83df10669b2...ebced975bf 19:42
Zoffix Actually, scratch that. Don't pack WW into star. I may want to explode it and make a redesigned version and I rather it not get the current version too popular :) 19:43
samcv lizmat, fixing 19:47
i feel i was asleep yesterday
ahh i see what happened. i made a commit when i was detached... looking at what i did
and so it didn't get commited to master but i pushed and thought it went
Geth nqp: 6932192bbb | (Samantha McVey)++ | tools/build/MOAR_REVISION
Bump MoarVM to get nqp::codes working

Changes brought in: ed84a632 Have MVM_string_codes iterate the string with codepoint iterator b07acdfd disable jit when we're not allowed to make memory executable
19:50
Ā¦ nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2...-ged84a632
rakudo/nom: 4eff491951 | (Samantha McVey)++ | tools/build/NQP_REVISION
Bump NQP to get bumped MoarVM, with nqp::codes actually working
19:51
rakudo/nom: version bump brought these changes: github.com/perl6/nqp/compare/2017....g6932192bb
0c19f54904 | (Elizabeth Mattijsen)++ | src/core/Rakudo/Iterator.pm
samcv lizmat, ok it should be fixed
lizmat samcv: testing 19:52
samcv the old MVM_string_codes just returned .chars. and that had been commited on a detached head
and not my changes which made it actually iterate the string and return the codes 19:53
hopefully people who need .codes will be happy now that it's much faster than calling .NFC.codes
lizmat samcv: clean now :-) 20:05
travis-ci Rakudo build failed. Samantha McVey 'Bump NQP to get bumped MoarVM, with nqp::codes actually working' 21:17
travis-ci.org/rakudo/rakudo/builds/259159469 github.com/rakudo/rakudo/compare/4...ff49195188
buggable [travis build above] ? All failures are due to timeout (0), missing build log (0), GitHub connectivity (0), or failed make test (1). Across all jobs, only t/04-nativecall/01-argless.t test file failed.
jdv79 Zoffix: odd, switching to 6d await and Promise.in doesn't fix it 22:58
Geth rakudo/nom: a256c26d85 | (Elizabeth Mattijsen)++ | src/core/array_slice.pm
Make @a[Range] about 3x faster

  - for completely refied list/arrays only
  - without any adverbs
  - by adding a highly specialized postcircumfix:<[ ](List:D, Range:D) candidate
23:01
jdv79 is 6d await known to have issues? 23:03
timotimo can you show your code a little bit?
jdv79 its a little big
timotimo oh, hmm
jdv79 i'll try to golf it a bit
timotimo that'd be nice 23:05
Zoffix jdv79: you sure you got use v6.d.PREVIEW as first thing in the file? I think the mechanism that complains about it being in the wrong place isn't working yet 23:24
m: use v6.c; sub foo { use v6.d.PREVIEW; await start say 42 }()
camelia 42?
Geth 6.d-prep: 2d53c6fce7 | (Zoffix Znet)++ (committed using GitHub Web editor) | TODO/FEATURES.md
Add throwage on wrongly placed use pragma to TODO
23:26
jdv79 yeah, its first. must be something else.
cog_ jnthn++ Our hero. Not only he slays dragons, but he knows how to tell his heroic battles. 23:45
timotimo aye 23:48