#perl6-dev on 23 August 2019 - Raku Programming Language Log

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.

jnthn

vrurg: Ah, I see.

00:04

Yeah, that makes sense.

00:05

I'd probably do it by a different meta-object...I mean, GrammarHOW is very tiny :)

A v6.e GrammarHOW too, configured with the correct version of the parent, is probably a way to go

00:06

vrurg

Metamodel is not language-dependent, actually.

jnthn

Trying to figure out "which Grammar" in DefaultParent feels fragile

It ain't, but OTOH I think %*HOW is initially populated from CORE.setting

00:07

And so can be versioned that way

vrurg

Not really. A compunit always have a core attached. So, one can always find what's needed.

jnthn

I forget precisely how it happens, perhaps an EXPORTHOW

vrurg

EXPORTHOW, yes.

jnthn

Could even declare the new GrammarHOW in the CORE.e.setting

00:08

vrurg

There is something with how it's what made me doubt if it worth touching.

jnthn

Then it effectively is versioned :)

vrurg

But, anyway, if the code works, it would probably worth making an op of it. There will be more and more situations when one would need to find a symbol from client's lexical.

00:09

jnthn

Maybe it can be figured out down in the metamodel, I'm just not sure I'd do it that way. It "feels" wrong somehow.

vrurg

OTOH, I don't feel like multiplying HOWs is a good idea for as long as we don't really need it.

00:11

jnthn

Dunno, I'd much sooner have a new metaclass for new metabehavior.

00:17

Time to sleep o/

00:18

vrurg

00:19

timotimo

gn jn

Geth

roast: ZhongnianTao++ created pull request #569:
Add test for GB18030 encoding

00:57

01:16 MasterDuke left 01:28 lucasb left

vrurg

m: class Foo { has int $.a; }; use nqp; my $f = Foo.new; nqp::bindattr($f, Foo, q<$!a>, 42); say $f.a

02:27

camelia

vrurg

class Foo { has int $.a; }; use nqp; my $f := Foo.new; nqp::bindattr($f, Foo, q<$!a>, 42); say $f.a

m: class Foo { has int $.a; }; use nqp; my $f := Foo.new; nqp::bindattr($f, Foo, q<$!a>, 42); say $f.a

camelia

P6opaque: representation mismatch when storing value (of type Int) to attribute (of type int)
in block <unit> at <tmp> line 1

02:28

vrurg

^ anybody with ideas as to why is it happening?

m: class Foo { has int $.a; }; use nqp; my $f := Foo.new; nqp::bindattr($f, Foo, q<$!a>, nqp::unbox_i(42)); say $f.a

camelia

P6opaque: representation mismatch when storing value (of type Int) to attribute (of type int)
in block <unit> at <tmp> line 1

02:42 evalable6 left 02:47 evalable6 joined

nine

AlexDaniel: please don't assume that just because someone's logged onto a channel, they will actually see all the messages. I certainly don't.

03:22

tellable6

2019-08-21T10:22:21Z #perl6-dev <tbrowder> nine ^^^ THANKS!

AlexDaniel

nine: so-o-o… you want me to implement automatic .tell if the person didn't talk for a few hours?

03:23

I can do that

to clarify, the discussion was about this behavior:

03:24

MasterDuke: hello world! Just testing the bot

tellable6

AlexDaniel, I'll pass your message to MasterDuke

AlexDaniel

which currently only works if the person is not on the channel

you can still use .tell

03:25

to pass messages even to those who are still online

nine

AlexDaniel: the old bot didn't try to be smart at all and just delivered the message. That sometimes led to some superfluous deliveries, but it was absolutely reliable.

03:29

AlexDaniel

nine: this bot is even more reliable because it won't let you pass to non-existent nicknames

03:32

.tell ninez hello

tellable6

AlexDaniel, I haven't seen ninez around, did you mean nine?

nine

Just because someone talked within the last few hours doesn't actually tell you if they are still there. The last message could very well be a "good night" followed by a busy day at work with loads of backlog

AlexDaniel

yeah, you're right, so I guess there's no way to implement that

03:33

nine

I guess there are a few specific cases where you can be reasonably sure. Like the .tell being followed by an answer of the addressed person on the same channel with no other messages in between (and not too much time or disconnects). At least if the telling person is being addressed in the response

03:35

AlexDaniel

nine: that works, yeah

03:38

but that's not related to automatic .tell

03:39

03:47 evalable6 left 03:51 evalable6 joined

Geth

nqp: 051cbec2bc | usev6++ | src/vm/jvm/runtime/org/perl6/nqp/sixmodel/reprs/MultiCacheInstance.java
[JVM] Factor out some duplicated code

05:05

roast: 3301d4f1c7 | usev6++ | 3 files
[JVM] Unfudge some passing tests

05:54

|Tux|

Rakudo version 2019.07.1-193-g71646ead9 - MoarVM version 2019.07.1-118-g3e715b424

csv-ip5xs	0.661	-	0.668
csv-ip5xs-20	4.911	-	5.415
csv-parser	20.732	-	20.819
csv-test-xs-20	0.432	-	0.441
test	6.350	-	7.056
test-t	1.694	-	1.754
test-t --race	0.773	-	0.778
test-t-20	28.323	-	28.704
test-t-20 --race	8.884	-	9.146

08:37

10:46 Xliff joined 12:10 robertle joined 12:12 lucasb joined 15:12 a3f left 15:19 a3f joined 16:42 MasterDuke joined 16:44 robertle left 16:51 pamplemousse joined 17:50 robertle joined 19:01 pamplemousse left 19:02 pamplemousse joined 19:13 Kaiepi left, Kaiepi joined

Geth

roast: 1f7877acf8 | ZhongnianTao++ | 4 files
Add test for GB18030 encoding

19:22

roast: a5f03b1ae0 | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | 4 files
Merge pull request #569 from ZhongnianTao/master

Add test for GB18030 encoding

19:27 Ven``_ joined 19:43 pamplemousse left 19:44 Ven``_ left 19:45 Ven``_ joined

vrurg

I'm considering renaming directory core into core.c to conform with two other cores. Are there any objections to this?

19:47

AlexDaniel

lizmat: that GB18030 roast merge was a bit too early, moarvm pull request is not merged yet

19:51

tellable6

AlexDaniel, I'll pass your message to lizmat

AlexDaniel

vrurg: not really

19:52

vrurg

AlexDaniel: it's actually gonna be more than that. I'm opening a problem solving on this.

AlexDaniel

tellable6: jnthn please do something about this: github.com/perl6/problem-solving/pull/87

19:53

tellable6

AlexDaniel, I'll pass your message to jnthn

Geth

problem-solving: AlexDaniel assigned to jnthn Issue Reconsider Rakudo build structure github.com/perl6/problem-solving/issues/94
AlexDaniel assigned to jnthn Issue Support for custom knowhows isn't generic enough github.com/perl6/problem-solving/issues/90

Add one more template
If this looks alright, I'll add one for every label.

19:54

AlexDaniel

tellable6: jnthn correct me if I'm wrong, but to me it looks like vrurg is stumbling upon problems, files them as tickets, figures out what needs to be changed, and changes stuff

19:55

tellable6

AlexDaniel, I'll pass your message to jnthn

AlexDaniel

tellable6: jnthn so if they're able to drive the development by themselves, why not set them as a responsible dev for both `language` and `rakudo`, right next to you?

19:57

tellable6

AlexDaniel, I'll pass your message to jnthn

AlexDaniel

hmmm

20:06

vrurg

AlexDaniel: I just realized you gave me permissions on problem-solving. Strangely, I got no notification from github on this.

20:07

AlexDaniel

vrurg: I think that happened through `core` team

20:08

vrurg: when I locked the threads, I gave write permissions to the core team so that actual devs can leave comments

20:09

I guess it doesn't notify in that case?

vrurg

AlexDaniel: possibly, I have a feeling that I wasn't able to set labels pretty much recently.

AlexDaniel

vrurg: we should set up issue templates

20:10

vrurg

Ah, ok. Then this is what happened.

AlexDaniel

then labeling will happen automatically

vrurg

For well-designed templates it is necessary to analyse what tickets we have, what are the most common subjects they have.

20:11

AlexDaniel

one template per label will do for now

20:12

vrurg

Like, it makes no sense for a template for the build process because mine is probably gonna be a single one for a long time.

I'm off for lunch. Will try to think out something useful.