01:01 lucasb left 04:27 yuplushi left 05:35 frost-lab joined
Geth_ z: 6124ea79c7 | (Stoned Elipot)++ | 2 files
Fix 'q' command

While here add support for VISUAL environment variable.
z: 48852abc58 | (Juan Julián Merelo Guervós)++ (committed using GitHub Web editor) | 2 files
Merge pull request #8 from Raku/fix-q

Fix 'q' command
z: 07924c4446 | (Trey Harris)++ | README.md
Add install instructions for calling z by another name

The [/rupa/z](/rupa/z) script for maintaining a dynamic directory cache is quite popular and included in some Linux distributions.
So the README should probably explain how to install this `z` under a different name (since _that_ `z`, being a shell function, can't be easily renamed without forking the code).
z: 4a0959a676 | (Juan Julián Merelo Guervós)++ (committed using GitHub Web editor) | README.md
Merge pull request #6 from treyharris/docs-aliasing-script

Add install instructions for calling z by another name
moritz what's z? 06:32
Geth_ roast: dumarchie++ created pull request #698:
Cover push/unshift with multiple values
rakudo: dumarchie++ created pull request #4035:
Don't flatten values passed to push/unshift
08:01 Merfont left, Merfont joined
Geth_ rakudo: 07c951314e | (Peter du Marchie van Voorthuysen)++ | 2 files
Add Windows specific build instructions
rakudo: 5b282cfc15 | (Peter du Marchie van Voorthuysen)++ | CAVEATS.md
Assert that building with MinGW should be possible
rakudo: 9070d5cb77 | (Patrick Böker)++ (committed using GitHub Web editor) | 2 files
Merge pull request #4030 from dumarchie/Windows-Build

Add Windows specific build instructions
08:10 domidumont joined 08:29 Kaeipi joined, Merfont left 08:30 patrickb joined
patrickb moritz: z (github.com/Raku/z) is a script to support a raku core dev to deal with the rakudo/nqp/moar repos. I automates things like bumping, testing, pulling. 08:33
tellable6 2020-11-15T18:16:49Z #raku-dev <[Coke]> patrickb - got the latest invite... brings me to a 401 page still. ah well
moritz patrickb: ah, thanks 08:34
patrickb lizmat, Altai-man: AzureCI does indeed cover all test scenarios Travis (and CircleCI and AppVeyor) covers. The only reason we didn't shut the others down yet, is that I wanted to give us some time to make sure Azure is not breaking on us. The capability of doing a re-test seems to be a bit fragile (possibly only because people tend to fail to find 08:38
the button). But I'm a bit reluctant to look into this, as I'm working on a tool to orchestrate the entire CI stack anyways.
Btw. I'm thinking about applying for a grant for the CI manager program in the hopes that will allow me to work on it more often. (I do tend to finish stuff I set out to do. But experience has shown that I often need a long time, because I can't allocate the time to work things.) 08:43
08:47 frost-lab left 08:49 frost-lab joined 08:51 patrickb left 08:57 travis-ci joined
travis-ci Rakudo build errored. Patrick Böker 'Merge pull request #4030 from dumarchie/Windows-Build 08:57
travis-ci.org/rakudo/rakudo/builds/743887592 github.com/rakudo/rakudo/compare/1...70d5cb7778
08:57 travis-ci left 09:15 eater left 09:16 eater joined 09:59 sena_kun joined 10:07 maggotbrain left
AlexDaniel` 6c: my @a = 42; push @a, (42, 50), (50, 30); dd @a 11:10
committable6 AlexDaniel`, gist.github.com/97df52a4d67a587a07...11fe72670e
AlexDaniel` bisect: my @a = 42; push @a, (42, 50), (50, 30); dd @a 11:11
bisectable6 AlexDaniel`, Will bisect the whole range automagically because no endpoints were provided, hang tight
AlexDaniel`, Output on all releases: gist.github.com/27b6feb10e9d566fc5...01265d6ac1 11:12
AlexDaniel`, Bisecting by output (old=2020.02.1 new=2020.05.1) because on both starting points the exit code is 0
AlexDaniel`, bisect log: gist.github.com/a0cafa0b4f2f8a6ace...dfed98c740
AlexDaniel`, (2020-03-31) github.com/rakudo/rakudo/commit/38...49f2d96639
AlexDaniel`, Output on all releases and bisected commits: gist.github.com/95e3671da87c452aab...90c60699d4
11:28 MasterDuke left
Geth_ rakudo: lizmat self-assigned Inconsistent flattening by push and unshift github.com/rakudo/rakudo/issues/4034
e45bcd53da | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | 13 files

This doesn't change any functionality, it just makes the code better maintainable and changes some names:
   TypedArray -> Array::Typed
   ShapedArray -> Array::Shaped
... (15 more lines)
AlexDaniel` hm, is append even working 11:47
6c: my @a = 42; @a.append: (42, 50), (60, 60); dd @a 11:48
committable6 AlexDaniel`, ¦6c (48 commits): «Array @a = [42, (42, 50), (60, 60)]␤»
AlexDaniel` absolutely not what I expected but OK, maybe it's meant to be this way 11:49
lizmat append uses the single arg rule ? 11:52
my @a = 42; @a.append: (42, 50); dd @a
evalable6 Array @a = [42, 42, 50]
AlexDaniel` 6c: my @a = 42; push @a, (), (), (); dd @a 11:53
lizmat 6c: my @a = 42; push @a, (), (), (); dd @a
committable6 AlexDaniel`, gist.github.com/ef9af3d06ffd701322...b9f02b794e
AlexDaniel` whoops
committable6 lizmat, gist.github.com/d706445497e5394db4...9dc60b417c
AlexDaniel` I gtg and won't file a ticket for this, but maybe it's related to that last issue 11:54
lizmat guess it's the same issue
12:10 Altai-man joined 12:13 sena_kun left 12:29 domidumont left 12:30 domidumont joined
linkable6 RAKUDO#4016 [closed]: github.com/rakudo/rakudo/pull/4016 Reorganize array shaping
rakudo/expose-objecthash-role: 50 commits pushed by (Vadim Belman)++, (Elizabeth Mattijsen)++, (Peter du Marchie van Voorthuysen)++, (Stoned Elipot)++, (Patrick Böker)++
review: github.com/rakudo/rakudo/compare/f...bb71a039ae
rakudo: 1a6cf78770 | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | 13 files
Expose Hash::Object/Hash::Typed roles (#4013)

These roles used to be internal to the Hash class. By exposing them it is possible to use them for easy typechecking and dispatch.
This makes the rather esoteric check:
... (15 more lines)
linkable6 RAKUDO#4013 [closed]: github.com/rakudo/rakudo/pull/4013 Expose Hash::Object/Hash::Typed roles
14:23 frost-lab left
Geth_ ¦ rakudo: Altai-man assigned to lizmat Issue Blin 2020.11, round 2 github.com/rakudo/rakudo/issues/4037 14:29
¦ rakudo: Altai-man assigned to vrurg Issue Blin 2020.11, round 2 github.com/rakudo/rakudo/issues/4037
14:29 Geth_ left, Geth joined
lizmat bisectable6: use nqp; dd nqp::getattr(nqp::create(Parameter),Parameter,q/$!nominal_type/) 14:47
bisectable6 lizmat, Will bisect the whole range automagically because no endpoints were provided, hang tight
lizmat, Output on all releases: gist.github.com/0f171b5189afaea2bb...2b7690238f
lizmat, Bisecting by exit code (old=2020.10 new=1a6cf78). Old exit code: 0
lizmat, bisect log: gist.github.com/2f9c139b30e08531b5...1db7172c28 14:48
lizmat, (2020-11-15) github.com/rakudo/rakudo/commit/17...4fe4025233
lizmat, Bisecting by output (old=2016.10 new=2016.11) because on both starting points the exit code is 0
lizmat, bisect log: gist.github.com/ca1749435fa82ea021...11901f32d7
lizmat, (2016-10-20) github.com/rakudo/rakudo/commit/94...090e31e6eb
lizmat, Output on all releases and bisected commits: gist.github.com/68f36d72990eb11465...2f40d2f3ad
vrurg lizmat: Parameter now has just `$!type`. 14:58
lizmat yeah, so the module was doing naughty things and is now caught :-( 14:59
vrurg We knew it back then when once was discussing it. Don't remember the context, but it was talked about. :)
lizmat warning this is module is experimental and subject to change 15:00
warning this module uses unspec'd rakudo internals and could break without warning
vrurg Anyway, I don't know how you use the attribute, but if $!type is a cercion then you have.^nominal_type, .^target_type, and .^constraint_type. 15:01
lizmat I think it goes further than that: nqp::bindattr_i($param,Parameter,'$!flags',128); 15:02
didn't the values of flags also change ? 15:03
vrurg Only added one for coercive parameter.
lizmat github.com/LLFourn/p6-DispatchMap/issues/2 15:05
afk for a bit&
15:10 lucasb joined
[Coke] test 15:27
nwc10 [Coke]: colabti.org/irclogger/irclogger_lo...11-16#l148 15:29
nine Pretty bizarre: 15:53
m: module A::B { class A::B { }; class C::A::B { }; class X::A::B { }; }; my A::B $b; my X::A::B $x; my C::A::B $c;
camelia 5===SORRY!5===
Type 'C::A::B' is not declared
at <tmp>:1
------> 3}; my A::B $b; my X::A::B $x; my C::A::B7⏏5 $c;
Malformed my
at <tmp>:1
------> 3 }; my A::B $b; my X::A::B $x; my C::A::7⏏5B $c;
[Coke] nwc10: thank you, was a wrong window. :) 15:55
16:11 sena_kun joined 16:12 Altai-man left
nine vrurg: I seem to recall you working on these multi part name declaration issues at some point? 16:12
vrurg nine: The only related thing coming to mind are relaxed resolutions to simplify references to parameterized roles. 16:14
Ah, nested modules... It was all wrong was I was doing, to start with. 16:15
*what I was doing
m: module A::B { class A::B { }; class C::A::B is export { }; class X::A::B { }; }; my A::B $b; my X::A::B $x; my C::A::B $c; 16:16
camelia 5===SORRY!5===
Type 'C::A::B' is not declared
at <tmp>:1
------> 3}; my A::B $b; my X::A::B $x; my C::A::B7⏏5 $c;
Malformed my
at <tmp>:1
------> 3 }; my A::B $b; my X::A::B $x; my C::A::7⏏5B $c;
vrurg m: module A::B { class A::B { }; class C::A::B is export { }; class X::A::B { }; }; my A::B $b; my X::A::B $x; my B $c; 16:17
camelia 5===SORRY!5===
Type 'B' is not declared
at <tmp>:1
------> 3 { }; }; my A::B $b; my X::A::B $x; my B7⏏5 $c;
Malformed my
at <tmp>:1
------> 3:B { }; }; my A::B $b; my X::A::B $x; my7⏏5 B $c;
vrurg m: module A::B { class A::B { }; class C::A::B { }; class X::A::B { }; }; my A::B $b; my X::A::B $x; my A::B::C::A::B $c;
camelia ( no output )
nine But I guess you learned a bit about how this stuff is supposed to work? Because...I cannot imagine what we have now is really that
vrurg Actually, I don't see what's wrong about it. C::A::B is declared under A::B namespace. So, it's actually A::B::C::A::B 16:18
m: module A::B { class A::B { }; class C is export { }; class C::A::B is export { }; class X::A::B { }; }; my A::B $b; my X::A::B $x; my C::A::B $c;
camelia ( no output )
vrurg ^ this is how you get C::A::B available outside of A::B 16:19
nine That's actually the one that I understand :) By the same logic though X::A::B should become A::B::X::A::B and class A::B should become A::B::A::B
vrurg Ah, A::B... Sure, it's all wrong.
m: module A::B { class A::B { }; class C is export { }; class C::A::B is export { }; class X::A::B { }; }; my A::B::A::B $b; my X::A::B $x; my C::A::B $c;
camelia 5===SORRY!5===
Type 'A::B::A::B' is not declared
at <tmp>:1
------> 3{ }; class X::A::B { }; }; my A::B::A::B7⏏5 $b; my X::A::B $x; my C::A::B $c;
Malformed my
at <tmp>:1
------> 3 { }; class X::A::B { }; }; my A::B::A::7⏏5B…
vrurg It's no doubt a bug. 16:20
nine m: class A::B { method Str() { "right" }; class A::B { method Str() { "wrong!" } } }; say A::B.Str
camelia wrong!
vrurg I'd say they whole namespacing needs at least some refurbishing.
nine Trouble is that this bug is actually depended on in the wild by modules like OpenSSL: unit module OpenSSL::Stack; class OpenSSL::Stack { ... } 16:21
vrurg I think the problem is caused by the outer A::B being stubbed at the point where the inner one is declared.
Then we'd have to patch OpenSSL. Because it just must not be this way. 16:22
nine github.com/rakudo/rakudo/blob/mast....nqp#L1651
vrurg Unfortunately, I have to go. And yes, it's about the location I was stumbling upon long ago and thinking that there is a problem with it. 16:24
nine That ^^^ is the code responsible for this surprising behaviour. Looks pretty intentional. Also the X::A::B part is depended on in the wild by e.g. Inline::Perl5.
vrurg I guess it was supposed to handle stubs, not incomplete outers. 16:25
vrurg is afk
16:30 MasterDuke joined
vrurg nine: I got back for a minute. As I think about it more and more, the thing looks entirely wrong to me. Names must be scoped. Perhaps worth a problem-solving to think more about it, but so far I'd be fore removing that functionality. 16:43
vrurg is afk again
nine Good thing is that it's easy to do in a language version dependent way
Why oh why am I fighting with merge_globals when I could have the deserialization code that I already generate have add those package symbols to the target's globals directly? 18:14
Oh I know: because it's yet another intrusive change on top of all the other changes and I'd rather have things settle down 18:16
18:43 domidumont left
releasable6 Next release in ≈4 days and ≈23 hours. 3 blockers. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft 19:00
lizmat and another Rakudo Weekly News hits the Net: rakudoweekly.blog/2020/11/16/2020-...n-renewed/ 19:02
MasterDuke lizmat++ (if you care, `sub foo(Int() $f {` is missing a closing paren) 19:10
lizmat oops :-) fixed # MasterDuke++ 19:11
lizmat wonders how many people will recognize where that picture was taken :-) 19:13
MasterDuke not i 19:14
nwc10 oh, am I supposed to? :-) 19:15
if so, was it the venue for GPW 2018?
lizmat nope 19:17
the building was completed in 1966
nine Florida?
lizmat yup 19:18
nine It is kinda hard to think about anything else today - at least if you're watching the live stream all day :) 19:26
lizmat you mean the spacex launch ? 19:28
nwc10 That's so last week, er last night. The real hardcore fun is watching the crane, I mean this thing: www.youtube.com/watch?v=Ky5l9ZxsG9M 19:31
19:31 MasterDuke left
lizmat aaaahhhh 19:31
nine Ah, tank watchers! 19:32
nwc10 Yes. Sometimes the tanks fly.
lizmat but... that's Texas ?
nwc10 Other times they explode.
nine it is
lizmat so what was nine watching in FL ?
nwc10 Yes. Almost Mexico
nine Crew-1
lizmat ah, ok, watched the launch live last night 19:37
afk for a bit& 19:38
Geth roast: fc3f24592a | (Christian Bartolomäus)++ | 3 files
[JVM] Fudge for fallout of new coercion semantics

These tests are dying after (or have been added with) the merge of
bartolin vrurg: ^^ spectest showed some fallout on the JVM backend. nothing serious, I'd say. I'll try to understand what goes wrong 19:44
Geth roast: mienaikage++ created pull request #699:
Add test for trailing declarator on conditionally set attribute
20:10 Altai-man joined 20:12 sena_kun left
Geth rakudo: a3ed02042e | (Christian Bartolomäus)++ | src/vm/jvm/runtime/org/raku/rakudo/Binder.java
[JVM] Recompute HOW before calling coerce

In the case of a generic type that needs instantiation (::T, T) paramType gets a new value. Therefore it wasn't correct to only compute HOW once, at the top of bindOneParam (cmp. of 34bc0b9147).
This fixes test failures in S06-signature/type-capture.t.
roast: 3c9cf75f1e | (Christian Bartolomäus)++ | S06-signature/type-capture.t
[JVM] Unskip tests that find coercion method again

Fixed with github.com/rakudo/rakudo/commit/a3ed02042e
bartolin sorry for the noise (fudge/unfudge), but I didn't really expect to find a quick fix for any of the failing tests 21:07
vrurg bartolin: thanks. On my side I was never even able to complete `make test`. So, I don't even try spectesting on JVM. 21:09
nine To answer my own question: if I forgo merge_globals the code I generate has to become a lot smarter. Since there's no merge I cannot blindly add stub packages but have to use existing ones if they already exist - or actually still merge if the existing one is a stub. I.e. I'd have to generate code that does much of merge_globals. 21:13
Might still be worth it as it has the potential to fix existing bugs, but boy, that's a lot of work again... 21:14
vrurg nine: worse if the work would have to be re-done later. The documentation states that `use` is a lexical things. But, in fact, it is not. And worse, I suspect the effects of a use inside a lexical block could propagate to a 3rd party module code. 21:18
bartolin vrurg: you're welcome. Sadly 'make test' doesn't work (it uses the EvalServer which leaks threads and memory). But it is possible to run the test files from roast one after the other. (I have to admit that I don't do that for Rakudo's test ...) 21:19
vrurg bartolin: maybe it worth re-implementing harness to make it get around the EvalServer problem? 21:20
nwc10 and sometimes they melt the engines: www.youtube.com/watch?v=1yjAife58e...be#t=4m20s 21:25
bartolin vrurg: I took a quick look and I'm actually using the script t/spec/test_summary without further tweaks. It shouldn't be too hard to do something similar for 'make test'. (But I'm kind of afraid to see the results ;)) 21:27
nwc10 The astute quote being: Failure is an option here. If things are not failing, you are not innovating enough. 21:28
21:48 MasterDuke joined 22:08 wildtrees joined 22:18 Xliff left 22:36 eater left 22:37 eater joined 23:05 MasterDuke left 23:19 Altai-man left 23:26 raku-bridge left, raku-bridge joined, raku-bridge left, raku-bridge joined 23:36 travis-ci joined
travis-ci Rakudo build passed. Christian Bartolomäus '[JVM] Recompute HOW before calling coerce 23:36
travis-ci.org/rakudo/rakudo/builds/744043062 github.com/rakudo/rakudo/compare/1...ed02042e2d
23:36 travis-ci left