#perl6-dev on 16 March 2018 - Raku Programming Language Log

Kaiepi

BeastieBot, build

01:04

BeastieBot, help

PufferBot, help

01:05

oh forgot to actually use the rakudo plugin with irc::client after the huge refactor

01:06

lol

BeastieBot, help

01:07

BeastieBot

Kaiepi, address me with "build", "test", or "spectest" to test building Rakudo, running tests, and running Roast's suite respectively on freebsd. Address me with "all" to attempt to run all three sequentially.

Kaiepi

BeastieBot, build

01:10

BeastieBot, help

BeastieBot

Kaiepi

BeastieBot, test

01:11

odd

01:12

PufferBot, build

Zoffix

Kaiepi: were those returns always there?

01:59

yoleaux

15 Mar 2018 20:22Z <cfa> Zoffix: /perl6/roast/issues/402

Zoffix

m: sub foo { my $p = start { return 42 }; await $p; $p }; say foo

camelia

MoarVM panic: Internal error: Unwound entire stack and missed handler

Zoffix

I don't think you can return from a method from within a promise and if you meant to return from the .protected bit only, then you need to make that a sub, not a block.

02:00

Also, what's sending output to IRC?

Kaiepi

the methods within the promises

Zoffix

FWIW returning a Promise from IRC::Client method sends return value to IRC when promise completes

Kaiepi

is there a way to make it so it doesn't address the user that called the command when you return a string from the method?

02:02

Zoffix

Ah. No

BeastieBot, source

02:03

BeastieBot, help

Where's the full source at?

Kaiepi

github.com/Kaiepi/p6-RakudoBot/blo...Rakudo.pm6

Zoffix

"No such method 'irc' for invocant of type 'IRC::Client::Plugin::Rakudo'"

02:09

"Attempt to return outside of immediatelly-enclosing Routine (i.e. `return` execution is outside the dynamic scope of the Routine where `return` was used)"

02:10

Kaiepi

Zoffix

Kaiepi: so you need to add ` does IRC::Client::Plugin;` to the class IRC::Client::Plugin::Rakudo line and, as I mentioned, change those blocks given to .protect() into sub {}

Kaiepi

02:11

gimme a sec, i gotta set up my vim plugins on my vps before i can test

Zoffix

Kaiepi: and I found that out by sticking CATCH { default { say "Error: $_" } } inside those `start` blocks. Also, IRC::Client.new can take a `:debug` flag that prints what it's doing (you can also install Terminal::ANSIColor to get preety colors)

02:12

Kaiepi

yeah, i'm adding a debug option to config.json for that

02:13

Zoffix

Kaiepi: here's the modifications I made. Seems to do things now and no errors show up in `build` routine. Still need to change blocks to subs in all the other protected blocks that try to return: gist.github.com/zoffixznet/3453f53...b3498b55a4

Zoffix &

Kaiepi

thanks

02:14

PufferBot, build

02:26

PufferBot, source

03:17

PufferBot

Kaiepi, github.com/Kaiepi/p6-RakudoBot.git

Kaiepi

PufferBot, build

03:18

BeastieBot, test

03:51

BeastieBot

[freebsd] Testing Rakudo...

[freebsd] | t/04-nativecall/02-simple-args.t (Wstat: 256 Tests: 14 Failed: 1)

03:53

[freebsd] Tests failed. See the output at fpaste.scsys.co.uk/575920

Kaiepi, done!

Geth

rakudo: Kaiepi++ created pull request #1620:
Skip NativeCall TakeUint16 test on FreeBSD

06:53

nine_

.tell jnthn On the issue of `composalizable` and `inheritalizable` I'm firmly in the "brilliant use of English derivational morphology" camp. But then, I'm native German speaker, so I'm used to torturing words ;)

06:58

yoleaux

nine_: I'll pass your message to jnthn.

Geth

rakudo: 51c4d4d89c | (Stefan Seifert)++ | lib/NativeCall.pm6
Fix precomp files of NativeCall users containing absolute paths

By storing the compiler object in a variable scoped to the Native role, it gets serialized into precomp files. This compiler object contains a lot of information including the path to the source file of the currently compiling module (the NativeCall user) and the path of the target bytecode file.
... (5 more lines)

08:52

nqp: 00b1f84f35 | (Stefan Seifert)++ | src/QRegex/NFA.nqp
Remove partially commented out timing debug code

Since this code went in in 2014 we have gained a profiler, rendering custom timing code pretty much obsolete. The global variable initialized to the current time however prevents reproducible builds.

09:31

[TuxCM]

Rakudo version 2018.02.1-166-g51c4d4d89 - MoarVM version 2018.02-171-geee5be412

csv-ip5xs	0.800	-	0.828
csv-ip5xs-20	7.557	-	7.772
csv-parser	36.197	-	37.209
csv-test-xs-20	0.437	-	0.445
test	8.811	-	8.984
test-t	2.542	-	2.869
test-t --race	1.041	-	1.054
test-t-20	45.303	-	45.760
test-t-20 --race	15.525	-	15.836

09:57

nine_

.tell jnthn What do you thing about this patch giving us reproducible nqp builds? With it a simple make clean && make install without changes won't break an installed rakudo anymore. gist.github.com/niner/c82676810ce4...e0ef086ef6

10:47

yoleaux

nine_: I'll pass your message to jnthn.

lizmat

Files=1236, Tests=76268, 315 wallclock secs (15.01 usr 5.36 sys + 2170.44 cusr 215.50 csys = 2406.31 CPU)

11:32

AlexDaniel

PufferBot: all

11:53

BeastieBot: all

BeastieBot

[freebsd] The current branch has uncommitted changes. Please tell Kaiepi to commit or reset any changes made before running your command again.

AlexDaniel, done!

AlexDaniel

Kaiepi: please commit or reset any changes

Kaiepi

BeastieBot, all

11:54

BeastieBot

[freebsd] Configuring Rakudo...

[freebsd] Building Rakudo...

11:55

[freebsd] Build failed. See the output at fpaste.scsys.co.uk/575922

11:58

Kaiepi, done!

Kaiepi

BeastieBot, build

11:59

BeastieBot

[freebsd] Configuring Rakudo...

[freebsd] Building Rakudo...

Kaiepi

PufferBot, all

12:00

PufferBot, build

12:01

ok weird, it should be using the same code beastiebot is

BeastieBot

[freebsd] Build failed. See the output at fpaste.scsys.co.uk/575924

12:15

Kaiepi, done!

Kaiepi

BeastieBot, build

13:01

BeastieBot

[freebsd] Configuring Rakudo...

[freebsd] Building Rakudo...

Kaiepi

PufferBot, all

13:03

PufferBot

[openbsd] Configuring Rakudo...

[openbsd] Building Rakudo...

AlexDaniel

Kaiepi: so what's up with these failures?

does it mean that rakudo is broken?

13:04

how do I interpret this result correctly?

nine

I'd guess it's running out of memory

yoleaux

12:13Z <jnthn> nine: Goodness, that's a tiny patch to achieve such a thing! Can't think of an objection :)

Kaiepi

the failures don't seem to happen when i run make manually

13:05

PufferBot

[openbsd] Build successful!

[openbsd] Testing Rakudo...

BeastieBot

[freebsd] Build failed. See the output at fpaste.scsys.co.uk/575932

Kaiepi, done!

PufferBot

[openbsd] Tests passed!

13:06

[openbsd] Running Roast test suite... (this will take a while)

Kaiepi

i'm either running out of memory or running out of swap space