sena_kun MasterDuke, o/ 07:56
Can you please create a ticket for the regression, stating what you discovered? 07:57
I think it's a blocker for next release, assigned to vrurg.
MasterDuke sure 09:10
Xliff MasterDuke: What regression is this? 09:15
MasterDuke a performance regression
after vrurg++'s new coercion work, inlining is broken in some cases, and since that's a really big win for moarvm optimization, there have been some slowdowns seen 09:17
Xliff Ah. 09:18
Thanks for the explanation.
MasterDuke np 09:19
Xliff Was this regression in v2020.11-23-g551b67957?
MasterDuke 1f090e04dcd4ed
linkable6 (2020-11-15) github.com/rakudo/rakudo/commit/1f090e04dc Merge pull request #3891 from vrurg/raku_1285
Xliff Hmmm... definitely seeing higher compile times from 11/16 to now 09:20
MasterDuke oh, but yeah, it's still present in rakudo
Xliff Interesting I didn't see it last week.... 09:21
MasterDuke i haven't noticed any *new* slowdown in the past week, but i wasn't looking that closely 09:22
Xliff drive.google.com/file/d/1Po4QXSYFo...sp=sharing 09:23
See "Regular Timings" sheet 09:24
MasterDuke huh. so it got faster between 11/9 and 11/16, but then slower again 09:25
Xliff Yup
Geth_ ¦ rakudo: MasterDuke17 assigned to vrurg Issue Performance regression after merge of new coercion semantics github.com/rakudo/rakudo/issues/4056 09:47
MasterDuke Altai-man: github.com/rakudo/rakudo/issues/4056
lizmat Files=1345, Tests=117140, 231 wallclock secs (29.97 usr 8.75 sys + 3199.06 cusr 299.75 csys = 3537.53 CPU) 09:55
Altai-man MasterDuke, thank you. 10:09
MasterDuke welcome 10:10
lizmat notable6: weekly 11:07
notable6 lizmat, No notes for “weekly”
MasterDuke huh, i thought Xliff left a note over in #raku 11:09
AlexDaniel` didn't use the bot? colabti.org/irclogger/irclogger_lo...-11-23#l74 11:12
yes, arguable the bot could've caught that :)
arguably*
Altai-man lizmat, o/
lizmat, can I PM?
Xliff Actually, didn't want to be rude and use the bot. 12:02
If that's the preferred way, will do so from now on. 12:03
lizmat Xliff: yes, please, that makes life easier for me... and wouldn't be rude at all 12:28
still no guarantee it will be in the weekly, though :-)
Xliff What are the requirements? I've been writing gists for years. 12:29
And I thought that little script might be a bit useful.
lizmat well, a gist is just that: a gist, a fleeting moment in time, a fleet almost :-)
we don't have a place to store useful scripts 12:30
we do have a place to store useful code packaged in modules, though :-)
Xliff lizmat: You could say the same for "comments about raku", yet those are included. 12:31
lizmat the URLs for the comments most likely will continue to work for many years to come 12:32
Xliff I use gists as my sort-of code blog.
lizmat the URLs to gists are... not so stable ?
Xliff No. They are stable.
I have yet to have a gist change URL for years.
lizmat they won't be removed after X months ?
Xliff No. 12:34
MasterDuke nope, i have some from 5 years ago still accessible
Xliff I have gists going back 5 years
gist.github.com/Xliff/2907106c84e9eb14d4c0
lizmat ok, I was living under the impression that they would be removed...
Ah... maybe I'm confusing it with anonymous gists.... which we don't have anymore, right ? 12:35
Xliff There's a wealth of gist posts I've made there.
All about Raku
lizmat Xliff: so a section New Useful Code Examples in the weekly maybe ? 12:39
Xliff Sure!
lizmat just before the "New Modules" section ?
Xliff I might even be a regular contributor to that. 12:40
s/be/become/
lizmat left a commment :-) 12:42
ok, would be nice if there would be more than a single gist to kick off.... :-) 12:43
nudge nudge wink wink
I'll be afk for a few hours, will be picking up on the RWN after that 12:45
Xliff lizmat: If I can think of something quick to write up, then I'll do si. 13:27
s/si/so/
Geth_ nqp: 00e6237a37 | (Stefan Seifert)++ | src/how/NQPCurriedRoleHOW.nqp
Add missing debug type name for easier debuggin
 15:15
Geth_ rakudo: lizmat self-assigned REPL complaints about "useless" actions github.com/rakudo/rakudo/issues/4057
a68b8ab486 | (Stefan Seifert)++ | 5 files
 15:30
nine MasterDuke: how did you reproduce that Physics::Measure segfault? 15:50
[Coke] nine: what does that opcode do? (seems very niche?) 16:14
(nqp::setdebugtypename($type, 'Curried ' ~ $curried_role.HOW.name($curried_role));) 16:15
MasterDuke nine: checked out raku-Physics-Measure, raku-Physics-Unit, and raku-Physics-UnitAffix. in raku-Physics-UnitAffix, run `raku -I ../raku-Physics-Measure/lib -I ../raku-Physics-Unit/lib -I lib -e 'use Physics::Measure; use Physics::UnitAffixQ; my $l = 1nm; say ~$l;'` 16:19
s/checkout out/cloned from [email@hidden.address] 16:20
nine [Coke]: it sets the debug_name field on the STable (the MoarVM internal low level representation of a type). That's a C string that's really just there for debugging, i.e. being able to find out somewhat easily what type an object is when you're in gdb or the Raku debugger. Oh, and for error messages. 16:28
MasterDuke: I only ever see "No such symbol 'Physics::Measure::Length'" 16:30
[Tux] Rakudo v2020.11-27-ga68b8ab48 (v6.d) on MoarVM 2020.11-16-gaaa1af772
csv-ip5xs0.816 - 0.825
csv-ip5xs-208.735 - 8.796
csv-parser24.774 - 25.754
csv-test-xs-200.381 - 0.388
test7.592 - 7.678
test-t1.912 - 1.946
test-t --race0.848 - 0.858
test-t-2033.018 - 34.140
test-t-20 --race9.639 - 9.715
 16:43
lizmat m: dd (42;666;999) # weird 16:45
camelia (42, 666, 999)
MasterDuke nine: i'm at 9fa9ae30dcc24b489f322a96944158056cca4f32, cb40bd254c8ca219e89120b8df3d7e98757f933c, and 6d3738c4fe8af5a6052a984f0a9944ee92cea9da of Measure, Unit, and UnitAffix respectively 16:53
linkable6 (2020-11-08) github.com/p6steve/raku-Physics-Un...6d3738c4fe Merge branch 'master' of github.com/p6steve/raku-Physics-UnitAffix into master
Geth_ rakudo: eae309a49f | (Elizabeth Mattijsen)++ | src/core.c/REPL.pm6
Stop REPL complaining about "useless use"

As discovered by Horst H. von Brand in #4057, was introduced with
  github.com/rakudo/rakudo/commit/7c...501d8956c5
The solution is to wrap each command given in a ( ).sink, to stop the useless use warning.		 17:19
linkable6 RAKUDO#4057 [open]: github.com/rakudo/rakudo/issues/4057 REPL complaints about "useless" actions
rba May I ask for a commit bit on github.com/rakudo/rakudo.org or do we go with pr for everything? 17:38
lizmat rba: should have access now 17:41
rba lizmat: thank you :-) 17:42
lizmat yw, and sorry if you had it before... that repo got moved around a bit (by me)
rba lizmat: Yes and yes. And no problem. 17:44
patrickb ping rba 18:11
rba patrickb: pong
patrickb rba: Can you do another round of updates? (rakubrew v14 and rakudo.org) 18:12
wait, rakudo.org is already done :-D
rba rakudo.org is pulled. Please check. 18:13
patrickb see above :-)
rba V14 I do later this evening on rakubrew.
patrickb Thank you!
rba yw
lizmat and another Rakudo Weekly News hits the Net: rakudoweekly.blog/2020/11/23/2020-...t-release/ 19:37
Altai-man lizmat++ 19:39
lizmat afk for a few hours again& 19:42
rba patrickb: Updated rakubrew. Argh, was a pain again. Need to change the setup. 22:19
tellable6 rba, I'll pass your message to patrickb

Powered by Cro and the Raku® Programming Language.

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

Thank you!