#perl6-dev on 24 January 2018 - Raku Programming Language Log

samcv

i've merged the UCD changes and going to bump nqp and rakudo now

01:01

TimToady wonders if there's any way to persuade rakudo to turn a socket coming in on fd 0 into an IO::Socket::Async automatically...

01:02

Geth

nqp: e2f3776b67 | (Samantha McVey)++ | tools/build/MOAR_REVISION
Bump MoarVM for Unicode property fixes

Changes: 2017.12.1-62-ga04d1099b..2017.12.1-65-gcd902eb58 4e5161542 ucd2c.pl: Fix White_Space/space Unicode Property +many other improvements 21267d80f Implement tile for OR operator

nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2...gcd902eb58
45d46a9ee4 | (Jeremy Studer)++ | src/QAST/Block.nqp

TimToady

rakudo should really not be assuming that stdin is always a file or a terminal

01:05

timotimo

agreed, rakudo should become much more flexible with regards to that

01:11

TimToady is trying to bring up a udp daemon under xinetd, and it suxors

01:18

geekosaur

udp and inetd have always been very strange bedfellows

01:25

tbrowder

lizmat: my two p5->p6 modules are now showing om

01:26

samcv

AlexDaniel: additionally i wrote a test script that tests that all aliases for propeties have the same property code, and it passed all of them. but before my change it fails

tbrowder

on modules.perl6.org/t/CPAN5

01:27

samcv

a few of them

cf, Case_Folding; sc, Script; STerm, Sentence_Terminal; Upper, Uppercase; WSpace, White_Space, space

01:30

thuogh likely only it's the shortened terms which were failing. cf, sc, STerm, Upper and space (plus the White_Space) but other than that only the shortened ones i believe

01:31

Geth

rakudo/post-release: 494ffb6a13 | (Zoffix Znet)++ | src/core/Any-iterable-methods.pm
Revert "Fix Any.skip: *-1 crash"

This reverts commit 44ff0a2a4c4fa6c250841a4f065136318be87383.
.tail(Callable) is not equivalent to .skip(Callable); spectest
from master branch catching this was not yet pulled into postrelease branch when this commit was applied, so nothing was caught

02:30

rakudo/post-release: 8efba3a897 | (Zoffix Znet)++ | src/core/Any-iterable-methods.pm
Alternate fix for Any.skip: Callable

Fixes github.com/rakudo/rakudo/issues/1384

03:07

roast/post-release: cca7b90403 | (Zoffix Znet)++ | S32-list/skip.t
Fix/expand Any.skip: Callable test

We're meant to have some elements since we're meant to leave
*-$n elems from the end.
github.com/rakudo/rakudo/issues/1384
Rakudo fix: github.com/rakudo/rakudo/commit/8efba3a897

03:09

rakudo: 49dce163e8 | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/build/NQP_REVISION
[NQP Bump] Brings 11 commits

NQP bump brought: github.com/perl6/nqp/compare/2017....g45d46a9ee 45d46a9ee Fix blocktype display in Block extra node info e2f3776b6 Bump MoarVM for Unicode property fixes 91e4a3e32 Merge pull request #396 from tbrowder/master 9a89c769c clarify what 'word' is being discussed ... (7 more lines)

05:48

¦ rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2017....g45d46a9ee

[Tux]

Rakudo version 2017.12-288-g49dce163e - MoarVM version 2017.12.1-65-gcd902eb58

csv-ip5xs	1.073	-	1.083
csv-ip5xs-20	11.706	-	12.062
csv-parser	11.800	-	11.820
csv-test-xs-20	0.464	-	0.488
test	11.158	-	11.261
test-t	2.754	-	2.829
test-t --race	1.147	-	1.204
test-t-20	49.284	-	49.324
test-t-20 --race	17.244	-	17.277

08:05

AlexDaniel

m: say `foo`

08:44

camelia

5===SORRY!5===
Argument to "say" seems to be malformed
at <tmp>:1
------> 3say7⏏5 `foo`
Bogus postfix
at <tmp>:1
------> 3say 7⏏5`foo`
expecting any of:
infix
infix stopper
postfix
…

AlexDaniel

m: `foo`

camelia

5===SORRY!5=== Error while compiling <tmp>
Bogus statement
at <tmp>:1
------> 3<BOL>7⏏5`foo`
expecting any of:
prefix
term

AlexDaniel

FWIW github.com/perl6/doc/issues/1745#i...-360056350

“===> Failed to find dependencies: DateTime::Parse” and what is that even about

09:04

AlexDaniel attempts to reproduce

looks fine here

09:10

weird

when installing DateTime::Parse itself it got “No candidates found matching identity: DateTime::Parse”

09:13

that's from toaster

looks like a fluke but I'm still interested to know what happened there

09:14

stmuk

could be a network issue (the iso standard excuse for everything)

09:55

jnthn

Comments welcome: news.perlfoundation.org/2018/01/gra...l-6-1.html

10:13

AlexDaniel

yey!

10:18

El_Che

"jnthn is an all round nice guy. Dogs and babies seem to like him. He devides his spare time in saving children in India and cleaning up plastic from the Atlantic Ocean"

10:21

jnthn

I...I think the comments are supposed to be on my proposal :P

10:22

El_Che

"If he ever proposes, I will be the happiest person on the planet"

10:23

too much?

AlexDaniel

“… Therefore I fully support this grant extension request”

jnthn

:D

El_Che

that's the magic

AlexDaniel

speaking about off-topic, now github has a convenient button to report spam, abuse and off-topic stuff

10:26

so not only you can close and edit tickets, you can freaking report them to github :o

10:27

El_Che

just commented, waiting moderation

10:31

(I think I didn't use much profanity)

lizmat

tbrowder++

10:52

Files=1233, Tests=76106, 339 wallclock secs (15.23 usr 5.16 sys + 2310.61 cusr 236.38 csys = 2567.38 CPU)

10:53

.tell tbrowder looks to me like your Net::IP::Lite is *not* a port of Net::IP::Lite on CPAN, am I correct ?

11:21

yoleaux

lizmat: I'll pass your message to tbrowder.

11:23

stmuk

is the release likely to escape today?

11:26

lizmat

AlexDaniel` would know :-)

11:37

so re: github.com/lizmat/P5wantarray/blob...tarray.pm6

it appears it bombs trying to tell me I'm using the wrong way to set a slang

note("Deprecated use of %*LANG\<$name> assignment detected in $tag; module should export syntax using \$*LANG.define_slang(\"$name\",<grammar>,<actions>) instead")

however, if I change my assignment to: %*LANG.define_slang('MAIN',Scalar::Grammar,Scalar::Actions);

11:38

AlexDaniel`

Yes, very likely

lizmat

I get: Cannot find method 'define_slang': no method cache and no .^find_method

AlexDaniel`

As soon as s.a.m.c.v wakes up and ready we will let it free :)

11:40

tbrowder

lizmat: that’s correct, NOT a port of CPAN5 Net::IP::Lite. given that it’s a subset of Net::IP, it should probably be renamed. in either event, i guess it doesn’t qualify as a port since it’s a subset. i will remove the tag since it’s misleading.

11:41

yoleaux

11:23Z <lizmat> tbrowder: looks to me like your Net::IP::Lite is *not* a port of Net::IP::Lite on CPAN, am I correct ?

lizmat

tbrowder: rename it to Net::IP, or at least make a copy to Net::IP and work on complete support ?

11:42

tbrowder

ok, but is there a way to do a rename other than deprecate the current name and publish under a new name?

11:47

lizmat

if there is, I don't know of it :-)

11:48

tbrowder

okey dokey

lizmat

.tell DrForr looks like Slang::Roman doesn't install anymore due to test failures

11:50

yoleaux

lizmat: I'll pass your message to DrForr.

lizmat

.tell DrForr investigating, will produce a PR when cause is found

yoleaux

lizmat: I'll pass your message to DrForr.

lizmat

hmmm... Slang::Kazu doesn't install for the same reason my P5wantarray fails:

11:58

New type for Perl6::Grammar must have a matching representation (P6opaque vs Uninstantiable)

I wonder if Slang::Tuxic still installs

11:59

cause if it doesn't, then Text::CSV won't install either

12:00

.tell DrForr PR sent

12:15

yoleaux

lizmat: I'll pass your message to DrForr.

DrForr

Merged - I've been watching in the background.

12:21

yoleaux

11:50Z <lizmat> DrForr: looks like Slang::Roman doesn't install anymore due to test failures

11:50Z <lizmat> DrForr: investigating, will produce a PR when cause is found

12:15Z <lizmat> DrForr: PR sent

Zoffix

lizmat: note in that slang error message, the sigil on LANG is different. Also, if you did not mix in core grammar/action, use .refine_slang method instead

12:25

huggable: slang

huggable

Zoffix, mouq.github.io/slangs.html Make an op that chains right (uses unsupported feature: slangs): BEGIN { $?LANG.refine_slang('MAIN', role { token infix:sym«smaller» { <sym> <O('prec', 'm=', 'assoc', 'left', 'dba', 'chaining', 'iffy', 1, 'diffy', 1, 'pasttype', 'chain')> } } )}; sub infix:<smaller> { $^a < $^b }; say 1 smaller 2 smaller 3

Zoffix

like that ^

lizmat

Zoffix: thanks, had gotten to that point already, looking at other fails now :-(

12:27

psch

thread cleanup in java is a mess :l

12:29

i'm hoping cleaning up circular references between a GlobalContext and its ThreadContexts is enough for the EvalServer, buuut

'cause yeah, apparently killing the threads spawned by the EvalServer doesn't necessarily clean up everything those threads allocated

12:30

humm, actually mark-and-sweep

12:35

guess i should instead look closely at the GC roots :s

lizmat

AlexDaniel`: I guess the failing test in S05-mass/properties-general.t is why you're waiting for samcv ?

13:10

AlexDaniel

failing test? :O

tbrowder

.tell lizmat just uploaded Net::IP with CPAN5 tag and another, obsolete version of Net::IP::Lite with NO CPAN5 tag.

yoleaux

tbrowder: I'll pass your message to lizmat.

lizmat

tbrowder++

yoleaux

13:10Z <tbrowder> lizmat: just uploaded Net::IP with CPAN5 tag and another, obsolete version of Net::IP::Lite with NO CPAN5 tag.

lizmat

AlexDaniel: test 615 fails for me consistently

13:11

m: say "9" ~~ /<:Numeric_Type<Digit>>/

camelia

Nil

tbrowder

AlexDaniel: i made the change to the nqp detabify PR, but it may need closing and reopening--your call, of course.

13:12

AlexDaniel

lizmat: ok, well, I was waiting because I cannot proceed without a moarvm release

13:13

lizmat

ah, ok :-)

AlexDaniel

lizmat: now… regarding that test… How come I only have 606

ah

I see…

lizmat

my roast is on master, afaics

13:14

AlexDaniel

ok that's a problem, thanks for letting me know

13:15

yeah, I definitely see a failure in t/spec/S05-mass/properties-general.t

13:43

.tell samcv S05-mass/properties-general.t file has a failing test

yoleaux

AlexDaniel: I'll pass your message to samcv.

lizmat

m: $_ = 42; dd $_.VAR.dynamic # so far so good

13:58

camelia

lizmat

m: my $a = 42; with $a { dd $_.VAR.dynamic } # huh ?

13:59

camelia

Zoffix

lizmat: it gets made into a lexical

If you look at --target=ast

14:01

lizmat

m: my $a is dynamic; dd $a.VAR.dynamic

camelia

lizmat

lexical doesn't prevent it from being dynamic

psch

m: my $a is dynamic; with $a { dd $_.VAR.dynamic }

14:02

camelia

( no output )

psch

:l

m: my $a is dynamic; $a = 42; with $a { dd $_.VAR.dynamic }

camelia

lizmat

m: my $a is dynamic = 42; with $a { dd $_.VAR.dynamic }

camelia

Zoffix

I notice tho `with 42 { .say }` just has lexical $_ param, but `.say with 42` has lexical $_ var, but then also lowered param that's then bound to lexical $_ is that really a win? Wouldn't just binding to $_ straight away be better?

lizmat

the reason I'm asking is this:

m: sub a { dd CALLER::<$_> }; my $a = 42; with $a { a }

14:03

camelia

Cannot access '$_' through CALLER, because it is not declared as dynamic
in sub a at <tmp> line 1
in block <unit> at <tmp> line 1

Zoffix

m: sub a { dd CALLERS::<$_> }; my $a = 42; with $a { a }

14:04

camelia

Int $a = 42

Zoffix

(why? no idea...)

lizmat

hmmm...indeed

m: sub a { dd CALLER::<$_> }; with 42 { a } # and this works ?

14:05

camelia

42

lizmat

CALLERS it is then

14:08

psch

hm, looks like AT-KEY on PseudoStash is actually buggy there?

14:09

as in, we only check for REQUIRE_DYNAMIC when we have PRECISE_SCOPE

...if i can still read nqp that is :X

Zoffix

Ah, then it might be R#1372

14:10

synopsebot

R#1372 [open]: github.com/rakudo/rakudo/issues/1372 Map instead of Hash from sub EXPORT results in missing error about non-dynamics via CALLER::

psch

around src/core/PseudoStash.pm:203

Zoffix

"I think the problem here is that dynamic is presently a property on a Scalar (or Array or Hash)."

So `with 42` doesn't have a scalar, but `with $a` does

psch

Zoffix: i'm about CALLER vs CALLERS

Zoffix

ah

m: sub a { dd CALLER::<$_> }; my $a := 42; with $a { a }

14:11

camelia

42

psch

PseudoStash.AT-KEY only checks for REQUIRE_DYNAMIC when we have PRECISE_SCOPE, but CALLERS doesn't

Zoffix

m: sub a { use nqp; dd nqp::getlexdyn('$_') }; my $a = 42; with $a { a }

camelia

Int $a = 42

lizmat

AlexDaniel: it's test *605* that fails consistently

14:17

AlexDaniel

lizmat: ok that makes more sense

lizmat

sorry for the confusion :-)

afk for Brussels.PM meeting

14:42

[Coke]

have fun!

14:44

dogbert2

m: say (* + $_)(32) given 10

14:49

camelia

Use of uninitialized value of type Any in numeric context
32
in whatevercode at <tmp> line 1

psch

m: given 10 { say (* + $_)(32) }

14:53

camelia

42

psch

m: say (* + CALLER:<$_>)(32) given 10

camelia

5===SORRY!5=== Error while compiling <tmp>
Undeclared name:
CALLER:<$_> used at line 1

psch

m: say (* + CALLER::<$_>)(32) given 10

camelia

42

psch

/o\

j: say (* + $_)(32) with 10

14:55

camelia

Use of uninitialized value of type Any in numeric context
in whatevercode at <tmp> line 1
32

Zoffix

dogbert17: that's all part of R#1212 I already got a rough fix; just need to make it more robust

15:08

synopsebot

R#1212 [open]: github.com/rakudo/rakudo/issues/1212 Uninitialized $_ and weird blowage (<a b c>[$_ xx 2] with 1)

dogbert2

Zoffix: cool, I stole the example from RT #126413

15:10

synopsebot

RT#126413 [new]: rt.perl.org/Ticket/Display.html?id=126413 [BUG] WhateverCode loses topic inside given modifier

dogbert2

I got the impression that you were working on a similar/same problem

Zoffix

Ah, nice. Gonna add it to my monster-kill list :)

dogbert2

:-)

m: say ("{$_}") for <aa bb> # is this something similar ?

15:11

camelia

Use of uninitialized value $_ of type Any in string context.
Methods .^name, .perl, .gist, or .say can be used to stringify it to something meaningful.

in block at <tmp> line 1
Use of uninitialized value $_ of type Any in string context…

dogbert2

Zoffix: ^^ RT #128054

15:12

synopsebot

RT#128054 [open]: rt.perl.org/Ticket/Display.html?id=128054 [PARSER] In a string enclosed in parens, a {}-interpolation can't access the topic $_ of a statement modifier

Zoffix

dogbert2: yeah, that too. These 7 tickets is what I currently have: github.com/rakudo/rakudo/issues/12...-359869111

(8 if you count the OP Issue)

dogbert2

very cool, do you know what's wrong ?

15:13

Zoffix

Blocks in wrong scope.

dogbert2

Is it a Rakudo problem or nqp?

15:14

Zoffix

Rakudo.

dogbert2

I guess it will be part of your post-release branch then :-)

15:15

Zoffix

Yeah.

dogbert2

m: sub foo($x) { say (* ~ $x)($_) given $x }; foo(1); foo(2)

15:16

camelia

11
21

dogbert2

RT #126984

15:17

synopsebot

RT#126984 [open]: rt.perl.org/Ticket/Display.html?id=126984 [BUG] WhateverCode in given doesn't get refreshed when entering surrounding block in Rakudo

dogbert2

perhaps that's something different though

m: sub x { my $ver = .lines.uc with "blazr".IO.open orelse note "meow {.exception.message}" and return 42; "meow$ver" }; dd x

15:22

camelia

No such method 'exception' for invocant of type 'Any'
in sub x at <tmp> line 1
in block <unit> at <tmp> line 1

dogbert2

RT #132172

synopsebot

RT#132172 [new]: rt.perl.org/Ticket/Display.html?id=132172 [BUG] `orelse` + block in a string gets the wrong $_

dogbert2

Zoffix: is the last two RT's suffers from the same problem you'll be able to close 10 bugs

15:24

s/is/if/

15:26

Zoffix

Yeah look to be the same

15:27

Thanks.

pmurias

when is 6.d planned? (what's the deadline for throwing away not yet specced bad tests :) )

15:34

Zoffix

pmurias: oh, I thought I responded, but forgot.... For overflow tests, yeah kinda implementation specific. I think these should go into optional Appendix. Kinda like out error message testing currently in Rakudo's test suite. As in: implementations can adhere to these to be more similar to other implementations, but they can still call themselves Perl 6 even if they don't pass optional Apendixes

15:36

pmurias

Zoffix: github.com/perl6/roast/commit/7143a330e9 seems a super bad test

15:37

Zoffix

pmurias: as for 6.d. It's currently blocked by R#1289 . If that's resolved, I imagine 6.d can happen rather fast, as pretty much everything else I have a good idea how to implement

15:38

synopsebot

R#1289 [open]: github.com/rakudo/rakudo/issues/1289 [6.d][severe] [6.d BLOCKER] Implement a Way to Know Caller's Language

pmurias

for some large powers we expect hanging, for other bigger ones we expect and exceptions

Zoffix

yeah

15:39

psch

m: say $?LANG.WHAT

15:40

camelia

No such method 'gist' for invocant of type 'Perl6::Grammar'. Did you mean any of these?
ast
isa
list

in block <unit> at <tmp> line 1

psch

...yeah, .gist on .WHAT >_>

hm, but that doesn't really help with 1289 anyway, unless we expand Grammar to have some kind of version info

Zoffix

Like here: github.com/perl6/roast/blob/master...#L131-L136

R#1289

15:41

Ah the blocker.

psch

right

Zoffix

psch: the grammar has version info. It's during runtime that it needs to be available

15:42

m: use nqp; dd nqp::getcomp('perl6').language_version

camelia

"6.c"

psch

plus i imagine "callers language" might eventually not necessarily be Perl6::Grammar but e.g. Perl5::Grammar

e.g. when 'use v5' gets finished

Zoffix

^ that can't be used, because after we compile with whatever language a method may be called by either 6.d or 6.c caller

15:43

pmurias

Most of the time I can port MoarVM specific stuff to the JS (or JVM) but I think that tying our language to what early 2018 MoarVM does is a bad idea

15:44

Zoffix

So it could be shoved into a variable of block/comp_unit, but then still a problem exists: (6.c user code calls open "-") -> (6.d-compiled core's &open calls IO::Handle.new(:path<->).open) -> (.open thinks caller is 6.d and uses 6.d semantics, rather than the 6.c semantics user intended)

15:46

pmurias: I'm +1 on ripping out tests like you pointed out + string content of error messages out of roast and shoving them somewhere separate

like any tests that expect some limits. maximum RAM/maximum power op size etc

15:47

psch

Zoffix: the grammar in $?LANG has version info?

15:48

m: use nqp; say nqp::getcomp('perl6') eqv $?LANG

15:49

camelia

Type check failed in binding to parameter '<anon>'; expected Any but got Perl6::Compiler (?)
in block <unit> at <tmp> line 1

Zoffix

Hm. We could shove version into special dyn var stuffed into blocks/comp_unit, except when we're compiling core stuff. Then dynlookup would always reach to place where user would have ability to set version

getcomp gives Perl6::Compiler

psch

hm, maybe i'm wrong tying "language" to "grammar" there anyway

Zoffix

At least currently, we're not making `use v6.blah` lexical. It's comp-unit scoped and from what I can see `nqp::getcomp('perl6').language_version` gives right version then.

15:50

m: use v6.d.PREVIEW; dd nqp::getcomp('perl6').language_version

camelia

5===SORRY!5=== Error while compiling <tmp>
Could not find nqp::getcomp, did you forget 'use nqp;' ?
at <tmp>:1
------> 3e v6.d.PREVIEW; dd nqp::getcomp('perl6')7⏏5.language_version

Zoffix

m: use v6.d.PREVIEW; use nqp; dd nqp::getcomp('perl6').language_version

camelia

"6.d"

pmurias

Zoffix: re error messages a lot of them seem like places where we might want to use a structured exception

15:51

psch

m: use v6.d.PREVIEW; use nqp; dd nqp::getcomp('perl6').language_version; { use v6.c; use nqp; dd nqp::getcomp('perl6').language_version }; # /o\

15:52

camelia

5===SORRY!5=== Error while compiling <tmp>
No compiler available for Perl v6.c
at <tmp>:1
------> 3mp('perl6').language_version; { use v6.c7⏏5; use nqp; dd nqp::getcomp('perl6').lang

Zoffix

pmurias: yeah, but there are a lot of tests that test exact text content of what `.message` returns. IMO they should just check the type of the exception and any attributes that are meant to have some sort of objects, but not the exact text, because it can be reworded/clarified/additional info attached to it, etc.

psch

+1 on that

15:53

that was kinda the first thing i contributed too, typed exceptions from some parts of Perl6::Grammar heh

Zoffix

:)

psch

though i might've been overzealous with .message with some tests

15:54

$ grep -RIn '.message' t/spec | grep -v '.git' | wc -l

15:55

513

hummm

though some of them probably are throws-like with e.g. message => /.../

...or however the &throws-like interface exactly looks like

and also false positives vOv

15:56

Zoffix &

psch stops blabbing

tbrowder

nqp err msg: no registered operation handler for spawn

16:20

it’s documented in the opcodes listing. looking for a test...

16:21

no test found except for spawnprocasync

16:23

doc error or nqp error?

psch

Proc::Async seems to use it just fine

16:26

pmurias

tbrowder: hasn't it been removed?

16:27

psch

m: use nqp; nqp::spawnprocasync

camelia

===SORRY!===
Arg count 0 doesn't equal required operand count 6 for op 'spawnprocasync'

tbrowder

pmurias: i didn’t see anything in GH nqp issues or PRs, but it may have been. i’m not real familiar with the code.

16:29

pmurias

psch: nqp::spawnprocasync is the thing nqp::spawn and nqp::shell where replaced with

16:30

psch

ohh

pmurias: right, apparently i misread tbrowders messages :)

tbrowder

the docs still show spawn and shell

psch

right, so missing docs and tests for spawnprocasync is the bugs

16:31

tbrowder

a

psch

and maybe still having tests and docs for shell and spawn is more bugs, i suppose

pmurias

psch: we don't have the tests anymore

16:32

and I wrote a spawnprocasync test

tbrowder

no, tests are there for spawnprocasynch, just missing docs...i’ll enter bug but i would appreciate the document update...

psch

pmurias++

also my patches to the GlobalContext and EvalServer don't really seem to help with memory usage of the latter...

16:36

plus jhat is kinda difficult to handle

and then there's the whole "*well* business apps on the jvm should have between 12 and 16gb of heap anyway"

not instantiating IgnoreNameClassLoader seemed to have done *something* at least, unfortunately this something includes "looping in S29-os-system.t" :/

16:41

but yeah, it looks to me like we keep runtime generated classes around indefinitely in memory, probably via class loader references

tbrowder

hm, spawn would sure be handier to use for a simple test of a short prog,

16:42

can you show an example of spawnprocasynch for a simple command of “nqp myfile.nqp”? blocking until completion is desired if that is a factor.

16:43

my use case is to add an nqp test for a known pod parse failure.

16:45

anyone?

16:46

jnthn

tbrowder: Do you really need to write it in NQP?

17:02

Anyway, you'll need a concurrent queue, to set up a completion callback that will be pushed onto that queue, then I guess shift from the queue which will block

17:03

It's possible, though not convenient :)

tbrowder

well i guess i could add another nqp test file that uses rakudo’s run routine to drive nqp but i don’t think that would make sense. i think the existing nqp pod test are weak, even for the limited pod allowed.

17:14

*tests

17:15

psch

oh hm, the loop in S29-os/system.t is expected :o

i suppose i have to rerun spectest then /o\

tbrowder

they don’t test the existing nqp grammar imho.

17:16

psch

plus run it again without my changes to see if those do anything :l

DrForr

tbrowder: The tests certainly aren't complete. Feel ree to pick over what I've got in drforr/perl6-Pod-To-HTMLBody if you need better coverage.

17:29

tbrowder

DrForr: thanks for the offer!

17:32

but the nqp tests can’t use p6.

17:34

however, i believe i can use p5!

17:35

DrForr

Oh, wrong layer, right.

17:36

tbrowder

correct.

17:41

pmurias

tbrowder: for an example of nqp::spawnasync look at the tests (or how the js backend call beautify)

17:45

jnthn

Oh, there's a best for it? Neat; I shoulda thought of that. pmurias++

17:47

tbrowder

thanks, but pretty ugly for me. any objections to using p5 “system”for execing nqp file?

Geth

nqp/post-release: 0264b23793 | (Zoffix Znet)++ | 2 files
Easyfy matching Block WVals with their QAST::Blocks…

  …when looking through QAST dumps.
  - Teach QAST::Node.dump to dump value for statement_id
  - Teach QAST::WVal.dump_extra_node_info to rake for CUID
   from past_block annotation, if one's present

18:31

[Coke]

tbrowder: do you have a sentence or two on why you're doing this in nqp & not rakudo?

18:46

(I imagine it would be easier in rakudo?)

18:48

tbrowder

[Coke] yes: nqp and other low-level code allows simple pod but the allowable pod formats are not fully tested. it seems to me that that should be tested along with the other tests, thus using nqp (or perl 5). unless there is another way, i would like to test known failure conditions by executing an nqp program with known bad code and test the exec return for failure. we already have pod tests, they just need to

19:03

be more robust. my plan is to use nqp to learn how to add a pod slang. i think that is easier than working first with rakudo when i'm trying to learn from scratch--MUCH easier build cycle for one thing.

that is, much easier to build nqp than rakudo.

19:04

of course if someone wants to add the p6 pod-slang and let me then putter with it and extend it that would be wonderful!

19:06

japhb

AlexDaniel: I'm still catching up in the backlog, so apologies if this is resolved now, but in a local build a few minutes ago, I also got the '===> Failed to find dependencies: DateTime::Parse' error

19:13

AlexDaniel

japhb: dammit :)

19:14

yea it still happens here too

japhb

Unfortunately it's part of a larger "rebuild everything" script, so I'm not sure what got the spurious dependency. :-(

19:15

AlexDaniel

actually, it installed just fine on my computer

but it failed in toaster again when I reran everything

no idea really

japhb: while we're at it, can you try `zef install CSS::Grammar`? :)

19:18

japhb

Ah, forgot about `zef rdepends`. Looks like HTTP::UserAgent and Task::Popular

AlexDaniel: install of CSS::Grammar succeeded, but that's using a 1-2 day old Rakudo build (though with fully updated zef indexes)

19:20

AlexDaniel

can you retry with a new one? It's not a zef issue with CSS::Grammar

19:21

but what I really don't understand is that it succeeded on both 2017.12 and the very HEAD in toaster

but I can't install it…

19:22

sure, looking at the code now, but what's going on…

both CSS::Grammar and JSON::Fast (its dep) were not touched in the last few hours

19:23

so can't be that

japhb is doing a full rebuild with HTTP::UserAgent commented out of the module install list

19:24

dogbert17

AlexDaniel: ===> Testing: CSS::Grammar:ver<0.3.1>:auth<github:p6-css>

19:26

===> Testing [OK] for CSS::Grammar:ver<0.3.1>:auth<github:p6-css>

===> Installing: CSS::Grammar:ver<0.3.1>:auth<github:p6-css>

AlexDaniel

argh

dogbert17

can try it on my 64 vit vm as well

*bit

AlexDaniel

maybe that looks familiar to somebody?

19:27

gist.githubusercontent.com/AlexDan...tfile1.txt

I've already tried rm -rf-ing the whole installation

19:28

aaaaaaaaaaaaaaaah

19:29

japhb

Was that a scream or a sigh of relief?

AlexDaniel

relief

19:30

I nuked ~/.zef

japhb

oh?

AlexDaniel

now it works

OK although I still don't understand why that was needed

japhb

I was about to ask if you'd deleted ~/.zef and ~/.perl6/precomp too

Gah, zef dependency search is S-L-O-W

19:32

timotimo

yeah, rather not quite very fast

19:42

dogbert17

m: use Test; is "\t \t" ~~ /<:space>+/, " ", "Space chars"

19:45

camelia

not ok 1 - Space chars
# Failed test 'Space chars'
# at <tmp> line 1
# expected: " "
# got: Match.new(list => (), made => Any, pos => 3, hash => Map.new(()), orig => "\t \t", from => 0)

japhb

jnthn: Commented (in support, of course) on your grant extension request -- and I just checked, my donation through work to the fund was paid too.

dogbert17

failing test, #605, in S05-mass/properties-general.t

19:46

AlexDaniel

dogbert17: yes, known

dogbert17: thanks

dogbert17: I pinged samcv so I guess she'll take care of that

dogbert17

there's a comment in the file claiming that the test seems wrong

ah, ok :-)

japhb

BTW, anyone here who has company match on donations, and your company uses Benevity for delivering those -- you should now be able to donate directly to the Perl 6 Core Development Fund. :-)

psch

" " being the desired result seems wrong to me

19:47

AlexDaniel

so apart from DateTime::Parse failing under some miraculous conditions, and one test failing obviously, I see no issues

19:56

AlexDaniel afk for ≈2 hours

samcv

ok i am awake

20:02

yoleaux

13:43Z <AlexDaniel> samcv: S05-mass/properties-general.t file has a failing test

Geth

roast: 144670d13b | (Samantha McVey)++ | 2 files
Correct <:space> test and add tests for White_Space regression

<:space> should match \t, so the test was incorrect as previously
written. Also test '1''s White_Space property to make sure it is false.

20:06

samcv

AlexDaniel: ok test should be correct now

japhb

AlexDaniel: Complete rebuild

20:09

Erg

AlexDaniel: Complete rebuild without HTTP::UserAgent (and thus the DateTime::Parse dependency) worked fine, as did installing CSS::Grammar on the freshly build Rakudo

20:10

AlexDaniel

samcv: then I think we're good

samcv

nice :)

[Coke]

tbrowder: sounds like a reasonable plan.

20:12

jnthn

japhb: Thanks for both of those. :-)

20:24

MasterDuke

jnthn: what's the link for the grant request?

20:27

jnthn

MasterDuke: news.perlfoundation.org/2018/01/gra...l-6-1.html

20:28

MasterDuke

thanks

20:30

samcv

jnthn: nice :)

20:31

AlexDaniel

anyway, afk for a bit

20:35

or a kilobit, but hoping to get the release going when I'm back

*wink-wink*

20:36

[Coke]

jnthn: I tried to give you a positive comment, but am unable to login at the moment. Uhoh.

20:41

AlexDaniel

.oO( give a negative then xD )

20:43

.oO( anonymously )

20:44

tbrowder

[Coke]: thnx

21:52

pmurias

jnthn:

23:03

sorry

jnthn

What for? :P

23:04

Oh, the highlight :)

psch

well, without my patches spectest got stuck S17-supply/throttle.t for a significant amount of time

23:27

+in

23:28

i'm guessing those go into post-release or i sit on them until after the release? or did i miss it happening?

23:29

...curiously i'm also seeing a continuous increase in mem usage with 'say "foo"' as the to-be-eval'd program, which bartolin says in RT#132104 doesn't happen :/

23:42

synopsebot

RT#132104 [open]: rt.perl.org/Ticket/Display.html?id=132104 [JVM] EvalServer leaks threads and memory when using Proc::Async

psch

my patches do impact the rate of mem usage growth at least vOv

23:43

Powered by Cro and the Raku® Programming Language.

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

Thank you!