Perl 6 language and compiler development | Logs at colabti.org/irclogger/irclogger_log/perl6-dev | For toolchain/installation stuff see #perl6-toolchain | For MoarVM see #moarvm
Set by Zoffix on 27 July 2018.
00:17 AlexDani` joined 00:18 p6bannerbot sets mode: +v AlexDani` 00:21 AlexDaniel left 00:22 AlexDani` is now known as AlexDaniel, fake_space_whale left 02:36 pyrimidine left 02:37 pyrimidine joined, p6bannerbot sets mode: +v pyrimidine 07:25 stmuk_ joined 07:26 p6bannerbot sets mode: +v stmuk_ 07:28 stmuk left 07:38 stmuk joined 07:39 p6bannerbot sets mode: +v stmuk 07:41 stmuk_ left 07:45 stmuk_ joined, p6bannerbot sets mode: +v stmuk_ 07:46 stmuk left
lizmat Files=1255, Tests=76374, 344 wallclock secs (15.77 usr 5.56 sys + 2425.56 cusr 227.24 csys = 2674.13 CPU) 09:00
still slower than weeks before, but not as bad as yesterday 09:01
09:34 robertle left 09:35 robertle joined 09:36 p6bannerbot sets mode: +v robertle
Geth rakudo: f06e8582fe | (Elizabeth Mattijsen)++ | src/core/metaops.pm6
Activate the Hyper refactor

  - in the simplest way, by taking over the HYPER proto
  - makes simple cases such @a <<+>> @a, @a <<+ 42 between 1.4x and 1.8x as fast
  - more complicated cases benefit more from needing to pass fewer parameters
  - no other code is removed yet: this is yet to come
09:59
lizmat afk&
10:16 committable6 left, committable6 joined, ChanServ sets mode: +v committable6 10:17 p6bannerbot sets mode: +v committable6 10:33 robertle left 10:46 robertle joined 10:47 p6bannerbot sets mode: +v robertle 12:37 pmurias joined, p6bannerbot sets mode: +v pmurias
Geth rakudo: 73eb297f6c | (Elizabeth Mattijsen)++ | src/core/Hyper.pm6
Fix some containerization errors in Hyper

  - some copy/pastos
12:45
pmurias nine: nqp::serializetobuf is broken when the serialization data is not stuffed into the bytecode file 12:53
(untested ops)---
Geth rakudo: 194daab717 | (Elizabeth Mattijsen)++ | src/core/metaops.pm6
Remove old HYPER implementation
13:07
13:07 pmurias left 13:18 pmurias joined, p6bannerbot sets mode: +v pmurias
lizmat does anyone have an idea why dispatch:<hyper> is calling HYPER(), while it only has to do with hypering in name? 13:25
it looks to me that it is a very convoluted way to respectively call nodemap or deepmap 13:26
13:27 pmurias left 13:28 pmurias joined, p6bannerbot sets mode: +v pmurias 13:30 pmurias left 13:33 pmurias joined, p6bannerbot sets mode: +v pmurias 13:36 pmurias left 13:37 pmurias joined, p6bannerbot sets mode: +v pmurias
lizmat bisectable6: dd :{ a => 42, b => 666 } <<+>> 4 13:38
bisectable6 lizmat, On both starting points (old=2015.12 new=194daab) the exit code is 0 and the output is identical as well
lizmat, Output on both points: «:{:a(46), :b(670)}␤»
lizmat bisectable6: dd :{ 42 => "a", 666 => "b" } <<~>> "c" 13:39
bisectable6 lizmat, On both starting points (old=2015.12 new=194daab) the exit code is 0 and the output is identical as well
lizmat, Output on both points: «:{42 => "ac", 666 => "bc"}␤»
lizmat bisectable6: old=2018.09 dd :{ 42 => "a", 666 => "b" } <<~>> "c"
bisectable6 lizmat, Bisecting by output (old=2018.09 new=194daab) because on both starting points the exit code is 0
lizmat, bisect log: gist.github.com/67281782ed5b72be6d...bafa13459e 13:40
lizmat, There are 45 candidates for the first “new” revision. See the log for more details
pmurias is just casting pointer from a a MVMString* to MVMObject* allowed? 13:48
s/pointer/a pointer/
lizmat perhaps someone on #moarvm knows ? /me doesn't 13:49
pmurias sorry meant to ask on #moarvm ;)
lizmat is there an (easy) way to see if an operator does assignment, aka to differentiate between &[+] and &[+=] *without* looking at the name 13:50
m: dd &[+] ~~ Sub 13:53
camelia Bool::True
lizmat m: dd &[+=] ~~ Sub
camelia Bool::False
AlexDaniel c: 2018.09,HEAD dd :{ 42 => "a", 666 => "b" } <<~>> "c" 14:01
yoleaux 04:00Z <jmerelo> AlexDaniel: bitrot again?
committable6 AlexDaniel, ¦2018.09: «Hash[Mu,Any] element = :{42 => "ac", 666 => "bc"}␤» ¦HEAD(194daab): «:{42 => "ac", 666 => "bc"}␤»
AlexDaniel lizmat: bisectable is probably right, it's one of the [js] commits 14:02
14:20 AlexDaniel left
|Tux| :( 15:17
Rakudo version 2018.10-101-g194daab71 - MoarVM version 2018.10-74-g2fdde4a21
csv-ip5xs0.954 - 0.955
csv-ip5xs-207.126 - 7.668
csv-parser22.902 - 23.036
csv-test-xs-200.432 - 0.432
test8.451 - 8.580
test-t1.890 - 2.001
test-t --race0.979 - 0.983
test-t-2033.140 - 34.113
test-t-20 --race12.820 - 13.254
lizmat not good indeed :-( 15:20
dogbert2_ lizmat, are you referring to |Tux| numbers 15:34
lizmat yes 15:35
15:36 pmurias left, pmurias joined, p6bannerbot sets mode: +v pmurias
dogbert2_ last good date is Oct. 31 15:37
test-t 1.671 - 1.699
pmurias how should I base64 a buffer in nqp? 16:10
16:21 pmurias left 16:27 pmurias joined, p6bannerbot sets mode: +v pmurias
timotimo which are the BSDs we run on? 16:41
someone on twitter is sad it's not packaged on FreeBSD 16:42
16:42 robertle left
timotimo Kaiepi: you're working on freebsd as well as openbsd at least, right? is there some kind of clue when it could land in ports perhaps? 16:47
wakelift.de/p/9c1c2260-aba5-40f4-a...3bebf9bb6/ - looking for proofreaders and criticism :) 17:25
lizmat any takers for www.reddit.com/r/perl6/comments/9v...unctional/ ? 17:27
timotimo: reading
timotimo don't see anything odd offhand 17:32
afk for a bit&
17:45 j3nnn1 joined, p6bannerbot sets mode: +v j3nnn1 17:47 AlexDaniel joined, p6bannerbot sets mode: +v AlexDaniel
pmurias github.com/MoarVM/MoarVM/pull/994 - I made a pull request to fix nqp::serializetobuffer so I can unbreak nqp-js 17:52
I have an uncommited test for that
18:23 ExtraCrispy joined, p6bannerbot sets mode: +v ExtraCrispy
Geth nqp/master: 4 commits pushed by (Paweł Murias)++ 20:38
timotimo wakelift.de/2018/11/09/where-did-i...-at-key-s/ - i published the latest work report for the moarvm profiler frontend 21:23
Geth nqp: 9c01ea679e | (Timo Paulssen)++ | src/vm/moar/HLL/Backend.nqp
profiler: generate highest_child_id in sql output

i had some SQL code inside moarperf that could add this in post-processing, but it slows to a crawl once the data sets get above a few hundred thousand rows, and you can reach that more easily than you think.
22:43
timotimo pmurias: since you just recently worked on serialization/deserialization, do you see a sensible way forward for "this attribute shall not have a value set for it when deserializing, and its value shall not be part of the serialization either"?
Kaiepi timotimo, no clue, freebsd is not my main language 23:12
timotimo language? :) 23:13
Kaiepi er
os
timotimo should i retract my claim that "we've got someone working on the BSDs :)"?
Kaiepi pushy aren't you? lol 23:14
i'll take a look at it
timotimo i kind of meant to be the opposite of pushy :| 23:15
Kaiepi ah 23:17
sorry i take things the wrong way a lot
timotimo i'm glad you pointed it out, though 23:18
23:23 MasterDuke joined, p6bannerbot sets mode: +v MasterDuke, MasterDuke left, MasterDuke joined, herbert.freenode.net sets mode: +v MasterDuke, p6bannerbot sets mode: +v MasterDuke
Kaiepi openbsd's port could use an update as well 23:25
they're on 2017.something atm iirc
Geth rakudo: a575effeaa | (Elizabeth Mattijsen)++ | src/core/Rakudo/Iterator.pm6
We don't need to .push an IterationBuffer

A nqp::push() will do fine. Makes @a <<+>> @b about 20% faster still. And @a <<+<< @b and @a >>+>> @b about 10% faster (on account they only use one DWIM iterator).
23:40
lizmat sleep& 23:41
Geth nqp: ad27192cc4 | (Timo Paulssen)++ | src/vm/moar/HLL/Backend.nqp
profiler: use right kind of id for highest_child_id
23:42