lizmat boarding& 00:27
Kaiepi boarding& 00:35
?
MasterDuke i believe she's flying home from orlando 00:37
Kaiepi ah 00:46
Geth 6.d-prep: 433f7d7d1f | (Zoffix Znet)++ (committed using GitHub Web editor) | TODO/FEATURES.md
List remaining tasks of R#1523 as part of 6.d TODO
 01:08
synopsebot R#1523 [open]: github.com/rakudo/rakudo/issues/1523 [testneeded] .subst/.subst-mutate should always return a List when :g is specified
Geth rakudo: 829ea8e37e | (Zoffix Znet)++ | src/core/Str.pm
Make Str.match(:x) constistently return a non-Slip List

All the other multi-match-object return variants (e.g. :g) return an empty List, but `:x` returns an `Empty`. Make that consistent.
This patch makes `:x` return an empty List as well. Unsure if
  `Empty` is a better choice here—I just went for the consistent
option with least impact.		 02:07
rakudo: e0d39321e1 | (Zoffix Znet)++ | src/core/Str.pm
Fix return of multi-match options of Str.subst-mutate

Currently only `:g` option returns a list of matches, however other combinations (when used with `.match`) also return multiple Match objects, such as `:x` or the `:nth`, `:th`, `:st`, `:nd`, and `:rd` options set to Iterable values.
Fix by just returning what `.match` gives us.
Also addreses R#1523 github.com/rakudo/rakudo/issues/1523		 02:10
synopsebot R#1523 [open]: github.com/rakudo/rakudo/issues/1523 [testneeded] .subst/.subst-mutate should always return a List when :g is specified
roast: c537763529 | (Zoffix Znet)++ | S05-substitution/subst.t
Test multi-match .subst-mutate return

Rakudo fix: github.com/rakudo/rakudo/commit/e0d39321e1		 02:11
roast: 8c358cfc87 | (Zoffix Znet)++ | S05-substitution/subst.t
Test non-matching .subst-mutate returns empty List…

  …when used with multi-match-producing options.
Closes github.com/rakudo/rakudo/issues/1523 Rakudo fix: github.com/rakudo/rakudo/commit/77794a28aa
   github.com/rakudo/rakudo/commit/e0d39321e1		 02:13
tbrowder what’s the protocol here for adding tests that are offered for review. how long should i wait before merging them myself? 02:16
AlexDaniel bisectable6: uptime 03:20
bisectable6 AlexDaniel, 8 hours, 55 minutes, and 1 second, 212.214844MiB maxrss. This is Rakudo version 2018.01-191-g555db42a3 built on MoarVM version 2018.01-89-ga01cdb449 implementing Perl 6.c.
AlexDaniel dogbert11: 236→212 MiB in 5 hours, that looks surprisingly right. Yes, it's so much better now. 03:22
tbrowder: it depends on how obvious the change is 03:24
there's no protocol besides “patience is a virtue” :) 03:25
tbrowder: if you're talking about NQP#412, then according to CI the test failed on rakudo-jvm 03:26
synopsebot NQP#412 [open]: github.com/perl6/nqp/pull/412 add a test for stderr in spawnprocasync
AlexDaniel and therefore can't be merged 03:27
(yet)
bisectable6: uptime 08:16
bisectable6 AlexDaniel, 13 hours, 50 minutes, and 55 seconds, 212.214844MiB maxrss. This is Rakudo version 2018.01-191-g555db42a3 built on MoarVM version 2018.01-89-ga01cdb449 implementing Perl 6.c.
AlexDaniel this is so awesome
still 6G of free RAM after 13 hours, swap is empty 08:17
OK that may not sound that impressive, but I've been struggling with this issue for… um… a few months I think?
jnthn AlexDaniel`: ooh, so leak gone? :) 10:08
tbrowder AlexDaniel: sounds like congrats are in order! 10:34
ok, the days of ignoring failing tests on the jvm are over, so i’ll try to get a jvm lab working... 10:48
DrForr jnthn: Is .cro.yml meant to be a place to store application-level settings, or is it just for the `cro' binary? 10:52
I.E. if I had something like a "supported list of languages" should that go in .cro.yml or am I better off reading it from some other file? 10:55
If you haven't thought about it, that's fine too, I can always put it one place and move it elsewhere when the decision is made. 10:56
jnthn DrForr: .cro.yml is meant to be a purely development-time artefact; when we start creating a Dockerfile in the project stubs, we'll actually put that into the .dockerignore 10:57
(e.g. it's not intended to go off to production)
DrForr Thanks. Were you intending something like config/{dev,stg,prod}.yml for this? 10:58
jnthn I'd encourage taking configuration in via the environment 10:59
DrForr (what Dancer does, for comparison)
jnthn To the degree there's a suggested way, that's the Cro one
You *can* put stuff to be injected into the environment: github.com/croservices/cro/blob/ma...nvironment
uh, cro.services/docs/cro-yml#Environment is nicer :) 11:00
DrForr I see there's an env: [] block there; I just wasn't sure of the intent.
jnthn Again, the intention being that you put your development settings into .cro.yml
And arrange something to put them there for production use
For example, if deploying services with Kubernetes it has a whole bunch of stuff for putting config into the environment, from secrets stores for example 11:01
But anything can set up env-vars, including a startup script :)
DrForr What I see o that page is stuff that you don't necessarily want in a filesystem anyone can snoop. 11:04
jnthn The k8s secret store? Yes, the secrets store is only for, well, secret stuff. There's other simpler mechanisms for non-secrets. :) 11:05
If you do want a configuration file, that's fine, but I'd still suggest injecting the path to it using an env var
DrForr Ah, the *path* to it, gotcha. 11:06
jnthn btw, there's also a #cro, since this maybe isn't entirely on topic for #perl6-dev :)
Geth roast: c4211535c6 | (Zoffix Znet)++ | S32-num/real-bridge.t
Remove trailing whitespace
 11:13
DrForr TT 11:16
Geth roast: b5425a0a74 | (Zoffix Znet)++ | S32-num/real-bridge.t
Cover .Bridge on core types
 11:18
DrForr jnthn: Over on #perl6 someone wants to sponsor a hackathon. 11:43
timotimo .o( and since our env vars are utf8-c8, you can even inject binary data via env vars ) 12:12
tbrowder hm, compiling nqp for jvnis real slow, prob need to bump X* env var 12:22
anyone have suggestion for min memory settng for jvm compile? i’m currently at 500Mb 12:24
on a 4Gb host
i have JAVA_ARGS=-Xms512m -Xmx512m 12:27
dogbert11 tbrowder: try increasing -Xmx 13:00
tbrowder dogbert11: i bumped all up to 3Gb and that helped a lot. thanks% 13:04
any hints on debugging tests that pass moar but fail jvm? 13:06
dogbert11 tbrowder: sorry, I don't I'm afraid 13:22
|Tux| Rakudo version 2018.01-196-ge0d39321e - MoarVM version 2018.01-91-g1c79ca35d
csv-ip5xs0.806 - 0.818
csv-ip5xs-207.520 - 7.959
csv-parser11.293 - 11.712
csv-test-xs-200.456 - 0.469
test9.135 - 9.395
test-t2.537 - 2.544
test-t --race1.055 - 1.076
test-t-2045.126 - 45.295
test-t-20 --race15.585 - 16.144
 14:27
tbrowder .ask Zoffix i am having jvm test failures (but not moarvm) when i test spawnprocasync to capture stderr. did you try it on yr jvm run_command or do you know if it should work without special handlng? 14:39
yoleaux tbrowder: I'll pass your message to Zoffix.
Zoffix . 15:02
yoleaux 14:39Z <tbrowder> Zoffix: i am having jvm test failures (but not moarvm) when i test spawnprocasync to capture stderr. did you try it on yr jvm run_command or do you know if it should work without special handlng?
Zoffix tbrowder: I don't know. I don't use JVM backend, but IIRC Proc::Async works on it, so I'd expect swanprocasync to work just fine as well. 15:03
tbrowder i just noted that was js, not jvm, sorry
stdout works fine, prob is stderr 15:04
thanks
AlexDaniel bisectable6: uptime 15:11
bisectable6 AlexDaniel, 20 hours, 46 minutes, and 14 seconds, 216.15625MiB maxrss. This is Rakudo version 2018.01-191-g555db42a3 built on MoarVM version 2018.01-89-ga01cdb449 implementing Perl 6.c.
AlexDaniel dogbert11: ↑ I'm so happy :)
Zoffix haha 15:16
bisectable6: uptime
bisectable6 Zoffix, 12 seconds, 191.972656MiB maxrss. This is Rakudo version 2018.01-191-g555db42a3 built on MoarVM version 2018.01-89-ga01cdb449 implementing Perl 6.c.
Zoffix I was trying to give it some work to do, to see if memory would grow, but it crashed on the second attempt to bisect `bisectable6: $*PERL.compiler.version > 2017.12 and die ` 15:17
dogbert11 Zoffix: already bored with Fallout 4?
Zoffix dogbert11: no, I'm currently away from the gaming box :) 15:19
dogbert11 AlexDaniel: as for the SEGV's, you could try running your bots with the JIT disabled, at least for the time being, until the bug is fixed
AlexDaniel` Zoffix: yeah, segv-ed I think :( 15:27
Geth nqp: 1c211d3bd8 | (Tom Browder)++ (committed using GitHub Web editor) | docs/ops.markdown
remove description of 'shell' and 'spawn' subs which are no longer available in nqp
 16:49
travis-ci NQP build failed. Tom Browder 'remove description of 'shell' and 'spawn' subs which are no longer available in nqp' 17:07
travis-ci.org/perl6/nqp/builds/341524453 github.com/perl6/nqp/compare/70d13...211d3bd848
Geth rakudo: 1dc1f038f1 | (Zoffix Znet)++ | 3 files
Fix .Numeric/.Real on :U numerics (and .Real on :D allomorphs)

We go through Mu.Real/.Numeric which default to Int type.
Fix by implementing methods that return types that match the Numeric the methods are called on (or the numeric portion of the allomorphs, for allomorphs).		 18:18
roast: f2d745f1a8 | (Zoffix Znet)++ | S02-literals/allomorphic.t
Spec .Numeric.Real on :U numerics (and :D allomorphs)

Rakudo fix: github.com/rakudo/rakudo/commit/1dc1f038f1		 18:19
pmurias the semaphore test seems to be failing on travis 18:39
Geth nqp/master: 4 commits pushed by pmurias++
geekosaur that'll make timotimo feel a bit better, maybe. (tried to speed up testing by using a second core, triggered the semaphore failure, backed off in confusion) 18:43
travis-ci NQP build failed. pmurias 'Test nqp::iterate with lowlevel lists' 18:56
travis-ci.org/perl6/nqp/builds/341568814 github.com/perl6/nqp/compare/1c211...1ce4f2bbd3
Rakudo build failed. Zoffix Znet 'Fix .Numeric/.Real on :U numerics (and .Real on :D allomorphs) 19:05
travis-ci.org/rakudo/rakudo/builds/341560593 github.com/rakudo/rakudo/compare/e...c1f038f1df
buggable [travis build above] ☠ All failures are due to: failed make test (1 failure). Across all jobs, only t/02-rakudo/repl.t test file failed.
AlexDaniel ev 19:12
evalable6: uptime
evalable6 AlexDaniel, 4 hours, 18 minutes, and 12 seconds, 240.285156MiB maxrss. This is Rakudo version 2018.01-191-g555db42a3 built on MoarVM version 2018.01-89-ga01cdb449 implementing Perl 6.c.
AlexDaniel squashable6: uptime
squashable6 AlexDaniel, 19 hours, 13 minutes, and 30 seconds, 229.796875MiB maxrss. This is Rakudo version 2018.01-191-g555db42a3 built on MoarVM version 2018.01-89-ga01cdb449 implementing Perl 6.c.
AlexDaniel jnthn: yes! Definitely 19:13
jnthn++
well, next problem is this issue with segvs… but maybe disabling the JIT is a reasonable workaround for now 19:14
Geth nqp: tbrowder++ created pull request #413:
add a test for stderr in spawnprocasync		 19:35
lucasb looks like github.com/perl6/nqp/commit/1c211d3bd8 removed the wrong heading. removed the file/directory one instead of the external commands 20:22
Geth nqp: b52d2000d0 | (Tom Browder)++ (committed using GitHub Web editor) | t/nqp/059-nqpop.t
correct no. of tests
 20:29
tbrowder lucasb: i made the mistake and will fix it, ok? 20:36
lucasb ok :) 20:38
Geth nqp: 6b12bab53e | (Tom Browder)++ (committed using GitHub Web editor) | docs/ops.markdown
restore correct comment for the opcode group
 20:44
nqp: tbrowder++ created pull request #414:
add a test for stderr in spawnprocasync		 20:48
travis-ci NQP build passed. Tom Browder 'correct no. of tests'
travis-ci.org/perl6/nqp/builds/341610848 github.com/perl6/nqp/compare/2e1ce...2d2000d0e9
tbrowder lucasb 20:49
lucasb: i think that commit 6b12bab53e fixed it 20:50
lucasb tbrowder: thanks! 20:51
tbrowder sorry 'bout that! 20:52
lucasb no problem
travis-ci NQP build passed. Tom Browder 'restore correct comment for the opcode group' 21:00
travis-ci.org/perl6/nqp/builds/341616554 github.com/perl6/nqp/compare/b52d2...12bab53ee1
lucasb I added a new comment to github.com/perl6/6.d-prep/issues/4 21:58
I tried to make some points. Let me know if none of them makes sense :) I just don't know how to argue...
AlexDaniel m: my str $x; dd $x 22:04
camelia ""
AlexDaniel m: my num @x; dd @x[5] 22:06
camelia 0e0
AlexDaniel hehehe
Geth rakudo: e930e2537f | (Timo Paulssen)++ | 2 files
teach X::Syntax::Number::LiteralType about natives

until now `my num $foo = 1` was a somewhat cryptic low-level run-time error ("cannot unbox"). Now it's a compile-time error of type X::Syntax::Number::LiteralType just like
  `my Num $foo = 1` has already been.
If changing this from a runtime to a compile-time error is seen as a problem, I can insert a run-time throw of this specific exception instead before release.		 23:31
timotimo AlexDaniel: you may have an opinion about this 23:32
AlexDaniel timotimo: I think this is totally OK, as long as it doesn't break any errata tests 23:34
if it does I can think again :)
timotimo i haven't run spectest yet, and also not the errata suite 23:36
Geth rakudo: 63a1d25ac6 | (Timo Paulssen)++ | src/Perl6/Optimizer.nqp
missed the case where Rat was assigned to int
 23:43
roast: d3a3957ab1 | (Timo Paulssen)++ | S02-types/type.t
test X::Syntax::Number::LiteralType for natives

used to be a low-level run-time exception on rakudo. now it's a compile-time object exception.		 23:44
timotimo hm, is there a make target for the errata tests? if not, there should be ... 23:48
regular spectest succeeds at least 23:54
running 6.c-errata now 23:55

Powered by Cro and the Raku® Programming Language.

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

Thank you!