Geth rakudo: Kaiepi++ created pull request #1609:
Fix Metamodel shortname assignments		 00:07
MasterDuke jnthn: around? 00:38
jnthn MasterDuke: Going to bed right now, I'm afraid 00:45
Will be about tomorrow
MasterDuke k, think i figured things out anyway 00:46
later...
jnthn 'night o/ 00:47
Geth roast: Kaiepi++ created pull request #400:
Implement Metamodel shortname tests		 01:05
rakudo: ce08683fbf | (Ben Davies)++ | src/Perl6/Metamodel/Naming.nqp
Fix Metamodel shortname assignments

Fixes #1606		 01:21
rakudo: 470b38e186 | (Zoffix Znet)++ (committed using GitHub Web editor) | src/Perl6/Metamodel/Naming.nqp
Merge pull request #1609 from Kaiepi/metamodel

Fix Metamodel shortname assignments
synopsebot RAKUDO#1606 [closed]: github.com/rakudo/rakudo/issues/1606 [LHF] Wrong output of in CArray.^shortname
roast: 862ed15a49 | (Ben Davies)++ | S02-names-vars/names.t
Implement Metamodel shortname tests

  See github.com/rakudo/rakudo/issues/1606		 01:22
roast: 2018ca5be6 | (Zoffix Znet)++ (committed using GitHub Web editor) | S02-names-vars/names.t
Merge pull request #400 from Kaiepi/metamodel

Implement Metamodel shortname tests
Kaiepi PufferBot, help 01:55
PufferBot Kaiepi, address me with "build", "test", or "spectest" to test building Rakudo, running tests, and running Roast's suite respectively on openbsd. Address me with "all" to attempt to run all three sequentially.
Kaiepi PufferBot, spectest
PufferBot [openbsd] Running Roast test suite... (this will take a while)
Kaiepi on top of making it able to run roast, i think i'll make it so it will upload the entire output if builds or tests fail to hastebin, since i can't grep all the useful information to send when that happens 02:01
PufferBot [openbsd] | t/spec/S03-operators/bit.rakudo.moar (Wstat: 1280 Tests: 54 Failed: 5) 02:05
[openbsd] | t/spec/S32-io/IO-Socket-Async.t (Wstat: 256 Tests: 17 Failed: 0)
[openbsd] | t/spec/S32-io/open.rakudo.moar (Wstat: 512 Tests: 52 Failed: 2)
[openbsd] Roast tests failed.
Kaiepi the IO-Socket-Async tests fail because my network doesn't support ipv6 atm, but i'm not sure about the others 02:06
Zoffix Kaiepi: are you using IRC::Client for your bot? There's a copy-pasteable example of using :filters to pastebin large output here: rakudo.party/post/IRC-Client-Perl-...hyourmouth (last few lines in code examples) 02:08
eco: Pastebin::Gist
buggable Zoffix, Pastebin::Gist 'Use gist.github.com/ as a pastebin': github.com/zoffixznet/perl6-Pastebin-Gist
Zoffix And you can use that module instead of Pastebin::Shadowcat
Kaiepi yeah, i am 02:09
Zoffix AlexDaniel should also have this feature already implemented somewhere in github.com/perl6/whateverable
travis-ci Rakudo build failed. Zoffix Znet 'Merge pull request #1609 from Kaiepi/metamodel 02:39
travis-ci.org/rakudo/rakudo/builds/352139546 github.com/rakudo/rakudo/compare/2...0b38e18622
buggable [travis build above] ☠ All failures are due to: failed make test (1 failure). Across all jobs, only t/04-nativecall/09-nativecast.t test file failed.
Kaiepi ? 02:42
PufferBot, build 03:01
PufferBot [openbsd] Building Rakudo...
[openbsd] Build successful! 03:03
Kaiepi PufferBot, test 03:08
PufferBot [openbsd] The current branch has uncommitted changes. Please tell Kaiepi to commit or reset any changes made.
Kaiepi PufferBot, test
PufferBot [openbsd] Testing Rakudo...
[openbsd] Tests passed! 03:09
Kaiepi weird
PufferBot, spectest 04:01
PufferBot [openbsd] Running Roast test suite... (this will take a while)
Kaiepi shit, typo prevented the output from getting uploaded 04:14
alright this should upload the output now 04:52
PufferBot, spectest
PufferBot [openbsd] Running Roast test suite... (this will take a while)
[openbsd] | t/spec/S02-lexical-conventions/finish-pod.t (Wstat: 65280 Tests: 0 Failed: 0) 05:03
[openbsd] | t/spec/S03-operators/bit.rakudo.moar (Wstat: 1280 Tests: 54 Failed: 5)
[openbsd] | t/spec/S32-io/IO-Socket-Async.t (Wstat: 256 Tests: 17 Failed: 0)
[openbsd] | t/spec/S32-io/open.rakudo.moar (Wstat: 512 Tests: 52 Failed: 2)
[openbsd] Roast tests failed. See the output at fpaste.scsys.co.uk/575866
Kaiepi whoops, forgot to change t/spec/S02-lexical-conventions/finish-pod.t back from when i was testing 05:20
AlexDaniel weekly: dogbert++ worked on old rakudo tickets, closing 7 and marking 22 with `testneeded` 08:32
notable6 AlexDaniel, Noted!
AlexDaniel hmmm reportable is showing new→open instead of new→testneeded :S 08:33
weekly: reportable: gist.github.com/fc1121feadf9daa523...9412a048b6 08:54
notable6 AlexDaniel, Noted!
AlexDaniel (already fixed that reportable issue mentioned above) 09:05
dogbert2_ AlexDaniel: the "is showing new→open instead of new→testneeded" is probably my fault, I have a tendency to change 'new' to 'open'. Perhaps I should refrain from doing that. 10:25
Geth rakudo/master: 5 commits pushed by (Daniel Green)++, (Elizabeth Mattijsen)++ 11:47
AlexDaniel dogbert2_: it happens automatically when you write a message 11:58
so whether you change that or not shouldn't matter
… I think…
anyway it's already fixed so you can continue doing whatever you want! ♥
Geth nqp: 12d2482d45 | usev6++ | src/vm/jvm/QAST/Compiler.nqp
[JVM] Use sub needs_cond_passed for repeated tests

That's what we do for MoarVM, too.		 12:12
nqp: 868bd54ff8 | usev6++ | src/vm/jvm/QAST/Compiler.nqp
[JVM] Fix slurpies in conditional ops

This mirrors the fix for MoarVM, that offers some explanation, also:
  github.com/perl6/nqp/commit/8e7201eb64
Fixes RT #105872.
synopsebot RT#105872 [resolved]: rt.perl.org/Ticket/Display.html?id=105872 [BUG] 42, 44, 22 -> *@a { say @a.perl }
travis-ci Rakudo build passed. Elizabeth Mattijsen 'Merge pull request #1608 from MasterDuke17/make_2_elem_infix_comma_special_cases_more_generic 12:31
travis-ci.org/rakudo/rakudo/builds/352292045 github.com/rakudo/rakudo/compare/4...e66ad30c05
dogbert2_ lizmat: are you around? 12:38
lizmat sorta 12:55
dogbert2_ lizmat: time for a question? 12:56
lizmat sure
dogbert2_ cool, could you try the code example in RT #127030 in order to see if it works on your Mac?
synopsebot RT#127030 [open]: rt.perl.org/Ticket/Display.html?id=127030 [BUG] await start { qx/../ } doesn't wait 12:57
lizmat say (await start { qx/echo foo/ }).perl ? 12:58
dogbert2_ yes
lizmat $ 6 'say (await start { qx/echo foo/ }).perl'
"foo\n"
dogbert2_ which means that it works :-)
lizmat okidoki :-) 12:59
dogbert2_ many thanks
jnthn That probably got fixed when Proc (and so qx, run, etc.) were re-implemented in terms of Proc::Async 13:14
dogbert2_ jnthn: thx, we had some problems bisecting this with the bot 13:19
[Coke] AlexDaniel: we don't have one yet 13:36
dogbert2_ lizmat: another question, is RT #122137 still relevant? 13:43
synopsebot RT#122137 [open]: rt.perl.org/Ticket/Display.html?id=122137 Type object seem to bind in item or push
lizmat looks to me it can be closed 13:45
m: my @a; my $c = Int; @a.push($c); $c = Str; dd @a
camelia Array @a = [Int]
dogbert2_ lizmat: thx, I'll close it unless you change your mind :-) 13:46
lizmat m: my @a; my $c = Int; @a.push: $($c,1); $c = Str; dd @a # maybe not
camelia Array @a = [(Str, 1),]
dogbert2_ perhaps the issue just need an update ... 13:47
lizmat m: my $c = Int; my $b := ($c,); $c = Str; dd $b
camelia (Str,)
lizmat seems to be in the List creation
hmmm
but that is correct
in List creation, we don't decont 13:48
in Array creation, we recont
m: my $c = Int; my $b := [$c,]; $c = Str; dd $b
camelia [Int]
lizmat so yes, it can be closed 13:49
timotimo jnthn : could --profile interfere with JITting > 13:51
?
dogbert2_ lizmat: thx, closed 13:54
lizmat hmmm.,.. looking at a --profile of 'for ^1000000 { my $b = a => 42 }', it appears that: 14:11
1. a => 42 is not constant folded
2. that it calls Pair.new directly, rather than infix:«=>»
which I think is wrong in case someone would want to add an infix:«=>» candidate 14:12
m: sub infix:«=>»(Str:D,Int:D) { "foo" }; dd ("a" => 42) 14:14
camelia "foo"
lizmat m: sub infix:«=>»(Str:D,Int:D) { "foo" }; dd (a => 42)
camelia :a(42)
[Coke] wonders what a bare a is in that case. 14:15
lizmat Str:D
m: dd (a => 42).value 14:16
camelia 42
lizmat m: dd (a => 42).key # duh
camelia "a"
lizmat oddly enough, using a => 42 as a named parameter, *does* constant fold apparently 14:18
dogbert2_ m: multi infix:<+>(Real, Pair $ where {.value === True}) { 42 }; say 1 + :a 14:26
camelia ===SORRY!===
Circularity detected in multi sub types for &infix:<+>
Zoffix Looks like that's 'cause `a => 42` compiles to Pair.new, while when it's in an arglist it compiles to QAST::SpecialArg 14:42
m: sub infix:«=>» (|) {say "my pair"}; a => 42
camelia WARNINGS for <tmp>:
Useless use of "a => 42" in sink context (line 1)
lizmat I would argue that a => 42 compiles to Pair.new is wrong 14:43
1: it should probably constant fold
2: it should call infix:«=>» 14:44
Zoffix +1
lizmat looking into that now 14:45
Zoffix cool 14:46
timotimo a => 1 in a call shouldn't cause a call to &infix:«=>» because it's syntactically Something Else 14:50
but "a" => 1 definitely should
lizmat hmmm... 14:52
dogbert2_ Zoffix, lizmat: Rt #131574
synopsebot RT#131574 [new]: rt.perl.org/Ticket/Display.html?id=131574 [BUG] Incorrect circularity detection with infix:<+> 14:53
lizmat hmmm.. both 'foo( a => 42)' and '$b = a => 42' are handled by "fatarrow" and I don't seem to be able to find a way to distinguish the two cases :-( 15:04
dogbert2_ ZOFFLOP: t/spec/S10-packages/basic.rakudo.moar tests #47-48 # ok ? 15:15
MasterDuke oh, i had a flop the other day with t/spec/S17-supply/stable.t 15:19
lizmat hmmm.. given a QAST::Op, how can I determine its :name parameter ? 15:46
Zoffix lizmat: .name method
lizmat that's what I though, but apparently not or there is something else going on :-) 15:47
Zoffix m: use QAST:from<NQP>; my $o := QAST::Op.new: :op<meows>, :name<foos>; say $o.op; say $o.name
camelia meows
foos
lizmat yeah :-( 15:48
ok, maybe some fresh air will help& 15:49
japhb lizmat: Stand in the middle of a grove of trees for extra oxygen content? ;-) 15:59
Kaiepi m: my int $a := int.Range.max; say $a +> 16 18:26
camelia 5===SORRY!5=== Error while compiling <tmp>
Cannot bind to natively typed variable '$a'; use assignment instead
at <tmp>:1
------> 3my int $a := int.Range.max7⏏5; say $a +> 16
Kaiepi m: my int $a = int.Range.max; say $a +> 16
camelia 140737488355327
Kaiepi m: my Int $a = int.Range.max; say $a +> 16
camelia 140737488355327
Kaiepi m: my int $a = int.Range.min; say $a +> 64 18:28
camelia -9223372036854775808
Kaiepi m: my Int $a = int.Range.min; say $a +> 64
camelia -1
Kaiepi ok so it's not just me that the roast tests for that fail for
geekosaur I think there's an open bug for that one 18:29
Zoffix Kaiepi: aren't they fudged? 18:30
How are you running the tests?
Kaiepi they fail with both make spectest and running them manually with perl6
Zoffix Kaiepi: if you're running manually with perl6 by using `perl6 t/spec/S03-operators/bit.t` then yeah, it'll fail, because fudge doesn't get applied. You need to use `t/fudgeandrun t/spec/S03-operators/bit.t` don't know why they'd show up failing with `make spectest` 18:32
Ahh.
Kaiepi: are you using a fork? Perhaps it's not up to date with latest of stuff from origin? 18:33
~4-10 days ago bitshift changes were made and more tests in that group were fudged.
git pull --rebase github.com/perl6/roast/
Kaiepi ohh, that's it 18:34
Zoffix \o/
samcv Zoffix: is the result from nqp::getcomp('perl6').language_version in some special variable or do i need to query it myself 18:35
to see what version is being run
or does it only working during compilation?
Zoffix samcv: it works only during compilation and nqp::getcomp('perl6').language_version is all I know of, but I don't know much about this area. 18:36
samcv well as long as it works inside a method and returns the version that will work for me
Zoffix If it's for a core method, then it'd just be whatever version core was compiled in. 18:37
s: Perl, 'version', \() 18:39
SourceBaby Zoffix, Something's wrong: ␤ERR: Could not find candidate that can do \()␤ in sub sourcery at /home/zoffix/services/lib/CoreHackers-Sourcery/lib/CoreHackers/Sourcery.pm6 (CoreHackers::Sourcery) line 29␤ in block <unit> at -e line 6␤␤
Zoffix s: Perl, 'version'
SourceBaby Zoffix, Sauce is at github.com/rakudo/rakudo/blob/52e6...erl.pm6#L1
Zoffix just does `Version.new(nqp::p6box_s(nqp::getcomp('perl6').language_version()))`
[Coke] github.com/rakudo/rakudo/blob/52e6...m6#L16-L21 - that list used to be if def'd because the parrot list was different. Can we collapse it back down to a single path now? (does js change the picture?) 18:41
Zoffix Yeah, looks identical 18:42
Geth rakudo: 71012beb59 | (Zoffix Znet)++ (committed using GitHub Web editor) | src/core/Perl.pm6
Merge duplicate `if`/`endif`ed content
 18:43
Zoffix Kaiepi: FWIW, I see you're a member of perl6 github org and you should have a commit bit to roast repo. So you don't need to use a fork. Just use the master and commit what you can directly and what you can't commit in a branch and PR 18:50
Kaiepi alright 18:51
samcv there doesn't seem much documentation of how to use Encoding::Encoder object to use replacements in docs.perl6.org 18:53
lizmat notable6: weekly 18:56
notable6 lizmat, 3 notes: gist.github.com/0b1ae482bc7e612797...2c36657862
samcv or if anybody is using that feature or if it's specced hm.
Kaiepi m: my \PATH = '.tmp'; my $fh = open PATH, :truncate, :create; say $fh; $fh.close if defined $fh 18:57
camelia open is disallowed in restricted setting
in sub restricted at src/RESTRICTED.setting line 1
in sub open at src/RESTRICTED.setting line 9
in block <unit> at <tmp> line 1
samcv i would like to have a unified method of decoding/encoding with/without replacement and with/without strict/permissive decoding/encoding. but since there's no documentation on it, people may not be using it yet
plus the fact there is Encoding::Encoder for replacements and without replacements. which seems unneeded. and i was ENcoding::Decoder object which didn't need two different classes one for replacement and one without 18:59
greppable6: Encoding::Encoder
greppable6 samcv, 1 lines, 1 modules: gist.github.com/a1820fd73a164df388...e7a6d58547 19:00
samcv and i'm not seeing any spectest coverage 19:01
releasable6: status 19:07
releasable6 samcv, Next release in 4 days and ≈23 hours. Blockers: github.com/rakudo/rakudo/issues?q=...%9A%A0%22. Changelog for this release was not started yet
samcv, Details: gist.github.com/636141c434c991c83e...1b0ad1b654
mst blinks 19:10
for some reason I read that as Encoding::Foobar 19:11
and thought "yes, it always is when I try"
Geth rakudo: samcv++ created pull request #1611:
Add support to Encoding::Decoder and Buf.decode to use replacements		 19:13
samcv hmm something happened with that PR 19:14
[Coke] there are a LOT of commits on that PR.
ah, too slow.
samcv oops what have i done 19:15
i think i wrote over my branch
oh well that's fine
though not sure where those comments came from..... 19:16
oh it's going into nom... oops 19:19
not sure how that happned
`git remote set-head origin master` fixd it :) 19:23
so now shouldn't ever have that happen again
travis-ci Rakudo build failed. Zoffix Znet 'Merge duplicate `if`/`endif`ed content' 19:34
travis-ci.org/rakudo/rakudo/builds/352485546 github.com/rakudo/rakudo/compare/5...012beb59bb
buggable [travis build above] ☠ All failures are due to: failed make test (1 failure). Across all jobs, only t/04-nativecall/09-nativecast.t test file failed.
lizmat notable6: weekly 20:12
notable6 lizmat, 3 notes: gist.github.com/b5dc27c1c26f13ef24...8fdab58d7b
lizmat notable6: clear weekly
notable6 lizmat, Moved existing notes to “weekly_2018-03-12T20:12:08Z”
Kaiepi the tests in t/spec/S32-io/open.t that were failing on my system were failing because opening files with :truncate and :create only works if the file is opened as wo or rw 21:15
is this something that should be fixed from rakudo or from roast?
i have a fix worked out for rakudo, but i'm not sure if making it possible to write to files opened like this is a side effect that could break stuff or not 21:29
Zoffix Kaiepi: what kind of fix? 21:30
To make the tests pass I mean? 21:31
I think I'm just gonna throw them out. It doesn't make sense for something you opened in "read only" mode to be created and especially truncated
The fix would be to throw on such use 21:32
Kaiepi what i wrote was to make the file open with rw instead of ro if :truncate and :create are passed, but i didn't think it was quite right 21:33
geekosaur Zoffix is right, those should throw on ro 21:36
bartolin fwiw, I just added a workaround for the JVM backend to make the tests for opening a file with :truncate and :create pass: github.com/perl6/nqp/commit/a90c01152f. We can can remove that workaround, it the test is removed. 21:38
Zoffix Oh, i was wondering why I couldn't repro that on camelia anymore :) 21:39
bartolin Well, the nqp version hasn't been bumped, yet.
Zoffix I mean repro the throwage 21:40
r: say "/tmp/zxczxsdzdasdsadsdasdasdsads".IO.open: :truncate, :create;
camelia IO::Handle<"/tmp/zxczxsdzdasdsadsdasdasdsads".IO>(opened)
Zoffix R#1614 21:42
synopsebot R#1614 [open]: github.com/rakudo/rakudo/issues/1614 .open with :truncate/:create violates "read only" default open mode
bartolin hmm, a92950fb4 is quite old. It throws for me locally on 2a3947816f
Geth roast: 24822cf173 | (Zoffix Znet)++ (committed using GitHub Web editor) | S32-io/open.t
Throw out questionable proptests

  - They fail on JVM and apparently on MoarVM on some OSes too:
   irclog.perlgeek.de/perl6-dev/2018-...i_15913886
  - The implied mode is "read only", which doesn't jive much with creation
   and especially with trucation. Filed as github.com/rakudo/rakudo/issues/1614		 21:43
roast: 9a417ad619 | (Zoffix Znet)++ (committed using GitHub Web editor) | S32-io/open.t
Adjust plan
Zoffix .seen cygx 21:46
yoleaux I saw cygx 17 Feb 2018 13:22Z in #perl6: <cygx> Xliff: I left a comment on that gist
bartolin just for the records: on the JVM those options are translated to StandardOpenOption CREATE and TRUNCATE_EXISTING. Both are just ignored when a file is opened only for reading: docs.oracle.com/javase/8/docs/api/...ribute...-
Zoffix .tell cygx I think you designed .open() modes. Would you like to add any comments on github.com/rakudo/rakudo/issues/1614 ?
yoleaux Zoffix: I'll pass your message to cygx.
Zoffix & 21:47
Geth rakudo: 0d796fb0da | (Samantha McVey)++ | src/core/Buf.pm6
Add support for replacement and strict setting with Buf.decode

Allows the user to set a replacement that is used when encountering an undecodable character. Allows setting strict to throw (or replace if a replacement is provided) on an undecodable character.
Replacement support currently only working with windows-1252 and windows-1251 at the moment.		 22:15
rakudo: ea92f550af | (Samantha McVey)++ | 2 files
Add support to Encoding::Decoder to use replacements

This also makes sure we setup decodestream on MoarVM to explicitly use permissive decoding, so that is set properly when my decodestream patch lands in MoarVM.
rakudo: 015d5a191d | (Samantha McVey)++ (committed using GitHub Web editor) | 2 files
Merge pull request #1611 from samcv/decodestream

Add support to Encoding::Decoder and Buf.decode to use replacements
lizmat and another Perl 6 Weekly hits the Net: p6weekly.wordpress.com/2018/03/12/...less-gems/ 22:22
Geth rakudo: Kaiepi++ created pull request #1615:
Throw when opening read-only files with truncate and create flags		 22:24
travis-ci Rakudo build passed. Samantha McVey 'Merge pull request #1611 from samcv/decodestream 23:01
travis-ci.org/rakudo/rakudo/builds/352574108 github.com/rakudo/rakudo/compare/7...5d5a191d72
AlexDaniel notable6: weekly 23:24
notable6 AlexDaniel, No notes for “weekly”
AlexDaniel notable6: weekly clear
notable6 AlexDaniel, No notes for “weekly”
AlexDaniel notable6: clear weekly
notable6 AlexDaniel, No notes for “weekly”
AlexDaniel weekly: clear
notable6 AlexDaniel, No notes for “weekly”
AlexDaniel lizmat: ↑ note that all variants work now, no need to worry :) 23:25

Powered by Cro and the Raku® Programming Language.

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

Thank you!