MasterDuke they both could use a lot of work 00:00
00:55 Kaiepi left 00:57 Kaiepi joined 01:03 BenGoldberg joined 01:07 PufferBot joined 01:45 ilbot3 left 01:58 ilbot3 joined, ChanServ sets mode: +v ilbot3 03:18 j3nnn1 joined
AlexDaniel weekly: reportable: 04:55
notable6 AlexDaniel, Noted!
samcv jnthn: i think those optimizations to nqp::atpos make it 4% faster 04:56
well it makes this nqp script which does 5 nqp::atpos's 10000000 times. about half the cpu usage is from atpos so i'm estimating it makes atpos 4% faster even though the full script itself is only 2% faster 04:59
05:27 j3nnn1 left
tyil samcv: I've done some work on vim-perl6, but I'm not a maintainer on the project 05:54
I can take a look tomorrow if there's an issue, I have an all-day privacy event today 05:55
Geth rakudo: 740671273e | (Samantha McVey)++ | tools/build/NQP_REVISION
Bump NQP/MoarVM for .words speedup and string bitop fixes

Changes: 2018.05-12-gc00142108..2018.05-17-g628d70fa5 628d70fa5 [JVM] Use peer/socket host/port in async sockets 67c8b96dd Bump MoarVM for str bitwise op fixes and .words speed impv. 41a0d017f [js] Implement nqp::getuniname for private-use and surrogate characters d5f13c037 [js] Fix bug f08a58269 [js] Handle _ and unicode unary minus in string to number coercions
¦ rakudo: version bump brought these changes:
nine ~~ 06:07
yoleaux 27 May 2018 22:08Z <lizmat> nine: perhaps "my class Rakudo::Internals::IterationSet is repr('VMHash') { }" could be of help?
27 May 2018 22:09Z <lizmat> nine: or perhaps we need to generalize this a bit like IterationBuffer
nine .tell lizmat close, but I'd need that in Inline::Perl5. I have some patches for making the meta model code deal with Perl 6 hashes correctly. But it's kinda unelegant. Also it'd mean that Inline::Perl5 would depend on a not even released rakudo. 06:10
yoleaux nine: I'll pass your message to lizmat.
06:10 BenGoldberg left
Geth roast: 0a7bd6fe68 | (Samantha McVey)++ | S03-operators/bit.t
Add tests for string bitwise ops with combining characters

These are tests for issue MoarVM/MoarVM#867 and ensuring that bitwise ops are doing the operations by codepoint.
07:10 dct joined 07:16 shareable6 joined
[Tux] Rakudo version 2018.05-26-g740671273 - MoarVM version 2018.05-28-gfede5cf50
csv-ip5xs0.888 - 0.917
csv-ip5xs-208.741 - 9.561
csv-parser25.986 - 26.287
csv-test-xs-200.440 - 0.450
test8.820 - 9.280
test-t2.348 - 2.389
test-t --race0.983 - 1.018
test-t-2043.761 - 44.361
test-t-20 --race15.480 - 16.464
07:36 AlexDani` joined 07:37 AlexDaniel left 08:23 shareable6 left
lizmat Files=1239, Tests=76355, 318 wallclock secs (15.51 usr 5.63 sys + 2193.64 cusr 218.08 csys = 2432.86 CPU) 08:45
yoleaux 06:10Z <nine> lizmat: close, but I'd need that in Inline::Perl5. I have some patches for making the meta model code deal with Perl 6 hashes correctly. But it's kinda unelegant. Also it'd mean that Inline::Perl5 would depend on a not even released rakudo.
09:16 dct left
lizmat .tell zoffix I can't seem to find that module on, suggestions? 09:19
yoleaux lizmat: I'll pass your message to zoffix.
lizmat .tell Zoffix I can't seem to find that module on, suggestions? 09:24
yoleaux lizmat: I'll pass your message to Zoffix.
09:36 robertle joined 09:38 shareable6 joined, ChanServ sets mode: +v shareable6 09:53 zostay_ joined 10:01 shareable6 left, [Tux] left, Undercover left, SourceBaby left, zostay_ is now known as zostay, squashable6 joined, ChanServ sets mode: +v squashable6 10:07 [Tux] joined 10:17 pmurias joined 11:00 yoleaux joined 11:20 brrt joined
brrt \o 11:21
samcv \o
brrt samcv: do you know how the website is generated? 11:22
my exact question is: could we add a docs/ section to that website
samcv brrt: yes we could 11:23
also the releases page is generated
brrt that would be fairly cool
samcv some documents are in markdown, others are in asciidoc (which is nice because it gives us nice table of contents and more advanced features) 11:24
brrt we can maybe convert it all to one format? 11:27
also, I like org-mode :-) 11:28
samcv yeah
brrt but, i want auto-generated public documentation more than I want to keep using org-mode only 11:32
12:36 j3nnn1 joined
lizmat hmmm.. not sure what to make of 13:45
brrt let a thousand flowers bloom? 13:50
El_Che lizmat: we're back to hashes :) 13:51
pmurias lizmat: it feels very javascripty
El_Che remembers Ouside-in objects fondly
(actually not really, the code was unreadable) 13:52
I picked Inside Out object because it was in the PBP book
lizmat yeah, InsideOut objects always felt like an even bigger hack to me than just blessed hashes 13:53
brrt it's a performance hack 13:54
and not one that is really applicable to perl
El_Che was it not about protecting the inside of the objects? 13:55
brrt in C/C++ it works well, because you'd maybe do some transformation on 100s of the same attributes of the same objects, and inside-out objects make sure that these attributes are all in the same part of memory
you could use it for that, but that makes little sense to me, tbh
perl being perl, what are you really going to do to 'protect' parts of the program from one another 14:00
actually writing good code is a much saner way to use your time
El_Che brrt: it's useful for poeple coming from other languages that have that 14:04
brrt no, it's not 14:05
(this is my opinion, though)
actually learning how to write perl in a sane way is a better way to spend your time than trying to fit perl into something that it is not
El_Che brrt: in an ideal world 14:06
brrt: in reality everyone takes some luggage from other languages
which can be good
brrt because - provided your program is actually useful and hence likely to be expanded by other people than yourself - *somebody* will break your house-of-cards mental model
much better to have a mental model that actually corresponds to what the code does 14:07
this is related to a 'baby perl' discussion that was held recently at some $workplace - bunch of folks arguing that no, we should only use the obvious parts of perl, and others saying no, you should learn how to use the language 14:10
doesn't help that perl (5) resources are still writing BS like 'perl array literals are delimited by parentheses'
El_Che brrt: I agree with you on principle, but I see where the need for some features comes from 14:12
14:23 MasterDuke left 14:26 MasterDuke joined 14:33 skids joined
lizmat m: sub a(Callable:U :$a! is rw) { $a = { say "foo" } }; a a => my &bar; bar() # never thought I would ever use Callable:U in a Signature 14:43
camelia foo
lizmat m: sub a(:&a! is rw) { &a = { say "foo" } }; a a => my &bar; bar() # cause this is not allowed, why? 14:50
camelia 5===SORRY!5=== Error while compiling <tmp>
Can only use 'is rw' on a scalar ('$' sigil) parameter, not '&a'
at <tmp>:1
geekosaur guesses someone assumed all the other sigils were already r/w containers 14:52
lizmat in case one wonders why I'm doing this: I'm working on porting the Memoize module from perl 5 14:54
15:43 brrt left 15:44 shareable6 joined 16:08 robertle left
El_Che lizmat: one can say that you're consistent :) 16:16
lizmat perhaps also stubborn?
actually, doing this I realized why we ripped out "is cached" 16:17
updating a simple in-memory cache from multiple threads needs locking :-(
or perhaps an approach where each thread has its own cache (perhaps optional) 16:19
and some flushing logic that would create a single hash for permanent backends
MasterDuke is 'is cached' gone for good? 16:23
lizmat well, I've not been able to find any code that does anything anymore 16:24
I guess we need to come up with some roles, and a default case that uses a built-in client of that role 16:25
16:28 shareable6 left
Geth nqp: 8385b98bd5 | (Tom Browder)++ (committed using GitHub Web editor) | docs/
update per current nqp

the =begin/=end identifier enforcement was added in commit ???.
16:55 j3nnn1 left 17:02 AlexDani` is now known as AlexDaniel 17:26 shareable6 joined 17:28 brrt joined 18:00 dct joined 18:12 dct left
AlexDaniel m: use experimental :cached; sub foo($x) is cached { say ‘hello’; $x }; say foo 42; say foo 42 18:18
camelia hello
AlexDaniel what's gone exactly?
18:19 stmuk_ joined 18:20 stmuk left
lizmat hmmm... 19:07
I was looking for the code that handles this the other day, and couldn't find it
aaaahhh... duh, it lives in lib/experimental.pm6, not in src/core 19:10
19:22 shareable6 left 19:32 Ven`` joined 19:35 Ven` joined, Ven`` left 19:37 brrt left 19:42 Ven` left
Geth roast: 8ae773213a | usev6++ | S03-operators/bit.t
[JVM] Fudge newly added tests
roast: 8d29945eb8 | usev6++ | 6 files
[JVM] Adjust fudging (mostly unfudge passers)
20:03 skids left 20:16 shareable6 joined 20:25 shareable6 left
Geth rakudo: 0666f6a864 | usev6++ | t/
[JVM] Run some more test files (all tests pass)
MasterDuke anybody have any thoughts/comments on and ? 20:43
20:46 dct joined 20:49 Kaiepi left
lizmat MasterDuke: just mulling over it 20:54
MasterDuke cool. no hurry, just wanted to throw the occasional mention out 20:56
lizmat and another Perl 6 Weekly hits the Net: 21:02
notable6: weekly reset 21:03
notable6 lizmat, Moved existing notes to “weekly_2018-05-28T21:03:57Z”
MasterDuke lizmat: i believe samcv's .lines and .words optimization is mostly (only?) visible for Str.(lines|words)' 21:08
timotimo oh, i think MasterDuke is right 21:09
lizmat ah, hmmm... 21:10
I guess you're right :)
fixed 21:12
MasterDuke++ timotimo++
MasterDuke lizmat++ a machine with the weeklies 21:13
AlexDaniel lizmat++ indeed 21:14
AlexDaniel loves weeklies
samcv MasterDuke: yeah that's probably right? though may not be right 21:45
anything that uses findcclass. so from what others have said IO.lines may not do that? 21:46
timotimo yeah, i'd expect IO.lines to use the separator functionality we have for handles 22:13
i'd go so far as to claim that that's The Truth 22:17
23:37 dct left 23:56 pmurias left 23:58 PufferBot left