00:52 sena_kun joined 00:53 Altai-man left 01:22 AlexDani` joined 01:24 AlexDaniel left 03:30 AlexDani` left, AlexDani` joined 04:51 Altai-man joined 04:53 sena_kun left 07:38 leont joined 08:08 nwc10 joined
nwc10 good *, #raku-dev 08:08
lizmat Files=1334, Tests=113514, 223 wallclock secs (29.22 usr 8.67 sys + 3113.80 cusr 303.79 csys = 3455.48 CPU) 08:32
nwc10 o/
08:51 sena_kun joined 08:53 Altai-man left 09:22 hungrydonkey joined 09:33 hungrydonkey left 09:35 hungrydonkey joined 10:45 [TuxCM] joined
[TuxCM] Rakudo version 2020.07-64-g9d6d8dd7a - MoarVM version 2020.07-19-g9a5203395
csv-ip5xs0.846 - 0.856
csv-ip5xs-207.932 - 8.251
csv-parser24.258 - 24.288
csv-test-xs-200.389 - 0.395
test7.465 - 7.654
test-t1.826 - 1.951
test-t --race0.832 - 0.935
test-t-2030.106 - 30.490
test-t-20 --race8.546 - 9.293
11:09 [TuxCM] left, [TuxCM] joined 11:39 [TuxCM] left 11:53 hungrydonkey left 12:08 hungrydonkey joined
lizmat notable6: weekly 12:14
notable6 lizmat, 2 notes: gist.github.com/c5236975850d2cc72a...5c0808729d
lizmat .tell demostanis still not sure why this wouldn't need to be in the ecosystem, for ease of installation 12:15
tellable6 lizmat, I'll pass your message to demostanis
lizmat notable6: weekly reset
notable6 lizmat, Moved existing notes to “weekly_2020-08-17T12:15:26Z”
nwc10 does NQP implement todo for tests? Or just skip? 12:35
12:38 hungrydonkey left
MasterDuke looks like it has todo 12:43
nwc10 OK, I missed that :-) How/where?
12:51 Altai-man joined 12:52 demostanis joined
demostanis Hello, world! 12:52
tellable6 2020-08-17T12:15:19Z #raku-dev <lizmat> demostanis still not sure why this wouldn't need to be in the ecosystem, for ease of installation
demostanis lizmat: It can't be in ecosystem as it requires root to install and to use. 12:53
12:53 sena_kun left
lizmat why does it require root ? 12:54
I'm not sure I want to promote a product that requires root for something that does not actually requires root access to work
MasterDuke nwc10: github.com/Raku/nqp/blob/master/sr...qp#L39-L42 12:56
demostanis lizmat: Its main purpose it to install software from Github, which might need root, for example to install executables in /usr/bin, or for many other reasons. As install instructions will be specified in user-provided rules, they must be root-owned, so that malware/untrusted users/anything else can't modify these rules and thus can't replace rules by something dangerous. It can't prompt for user credentials 12:58
once the software are getting installed as HubToDate may run inside of a crontab in the future.
nwc10 thanks 13:03
13:30 |Tux| joined 13:41 [TuxCM] joined
Geth nqp/more-hash-tests: dd6b6fee54 | (Nicholas Clark)++ | t/nqp/108-vmhash.t
More tests for NQP hash behaviour.

Tests for iterators, including deleting the hash key corresponding to the current iterator position. Tests for exceptions before and after starting iteration (which are not yet consistent between the three backends).
Test `.elems`. More thorough tests for `.kv`.
Also fix a test description.
nqp: nwc10++ created pull request #657:
More tests for NQP hash behaviour.
lizmat and another Rakudo Weekly News hits the Net: rakudoweekly.blog/2020/08/17/2020-...ion-types/ 14:03
[Coke] nwc++ 14:05
nwc10 this was why I wanted to build JS
Altai-man lizmat++
[Coke] are there instructions for doing the JS build? 14:07
I got to the point where I just use the node package because I'm lazy
nwc10 I found that I needed node v10 14:09
v11 and v12 didn't work, I think because soemting ("us") has pinned npm packages that are too old to deal with v11 or later 14:10
so I built the most recent v10 release from source and used that
on a machine with a new enough gcc
(That's linux)
seems that it only supports g++ on Linux, and (only?) clang on OS X (and *BSD?)
but this might be a constraint from whatever v8 demands. 14:11
anyway, it's fussy.
14:12 demostanis left 14:38 [TuxCM] left, [TuxCM] joined 14:39 [TuxCM] left 14:40 [TuxCM] joined 15:10 [TuxCM] left 15:11 [TuxCM] joined 15:40 [TuxCM] left 15:41 [TuxCM] joined 16:10 [TuxCM] left 16:11 [TuxCM] joined, [TuxCM] left 16:12 [TuxCM] joined
Geth rakudo/sorting-with-junctions: 410a428783 | (Elizabeth Mattijsen)++ | src/core.c/Rakudo/Sorting.pm6
Make .sort about 4.4x as fast on Int/Str/Num arrays

  AlexDaniel++ expressed concern about the Junction check causing a
slow down in sorting. And he was right. However, when looking at the profile, it occurred to me that most of the overhead of sorting was caused by using the default infix:<cmp> logic. With the multi sub "compare" now inbetween, it was possible to turn this overhead ... (9 more lines)
16:41 [TuxCM] left
16:42 [TuxCM] joined, [TuxCM] left, [TuxCM] joined 16:45 AlexDani` is now known as AlexDaniel, AlexDaniel left, AlexDaniel joined 16:52 sena_kun joined 16:53 Altai-man left 17:09 [TuxCM] left
AlexDaniel sourceable6: ‘foo’ cmp ‘bar’ 17:12
sourceable6 AlexDaniel, github.com/rakudo/rakudo/blob/9d6d....pm6#L3650
AlexDaniel sourceable6: <foo> cmp <bar> 17:17
sourceable6 AlexDaniel, github.com/rakudo/rakudo/blob/9d6d....pm6#L3650
AlexDaniel sourceable6: <42e1> cmp <42e1> 17:19
sourceable6 AlexDaniel, github.com/rakudo/rakudo/blob/9d6d...s.pm6#L184
AlexDaniel lizmat: I made a review, I hope it's not too long 17:33
lizmat: but TL;DR is that `compare` multis form their own cmp table just like this one, except that it's different: github.com/rakudo/rakudo/blob/9d6d...s.pm6#L172 17:34
so basically you're trying to recreate the same thing but fast… I mean… yeah, it's something the compiler should do 17:35
lizmat: if you care about my advice then I'd suggest to talk to jnthn to figure out why a call to `cmp` is not optimized into a simple nqp::cmp_* call 17:38
or whoever knows about this
[Coke] m: say e 17:53
camelia 2.718281828459045
18:47 JRaspass joined
releasable6 Next release in ≈4 days and ≈23 hours. 2 blockers. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft 19:00
19:44 vrurg left 19:45 vrurg joined
Geth rakudo/sorting-with-junctions: 3f60d0a70b | (Elizabeth Mattijsen)++ | src/core.c/Rakudo/Sorting.pm6
Emulate exact behaviour on lists with IntStr/NumStr

First check numerically, if they're the same, then compare as strings.
20:33 codesections left 20:51 Altai-man joined 20:53 sena_kun left
Altai-man lizmat, hi! I have moved your RFC article to 19th, hope it isn't wrong for me to move it. Thanks a lot for articles! 21:56
Oh, no, this is probably a very bad idea.
Ok, moving it back. 21:57
AlexDaniel lizmat: I'm sorry, but as I said, I don't think an optimization in this particular form can work without sort regressions 22:47
I think we'll have to figure out something else
22:54 [TuxCM] joined 23:37 leont left