releasable6 Next release in ≈6 days and ≈15 hours. There are no known blockers. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft 03:00
Geth ¦ problem-solving: JJ assigned to jnthn Issue The raku-community-modules organization and its problems. github.com/Raku/problem-solving/issues/210 07:27
nine Why is jnthn assigned to this issue? 07:37
ShimmerFairy I think the thing just auto-assign jnthn bugs in that repo. 07:47
Geth ¦ problem-solving: lizmat unassigned from jnthn Issue The raku-community-modules organization and its problems. github.com/Raku/problem-solving/issues/210 09:15
lizmat Files=1307, Tests=113021, 219 wallclock secs (29.25 usr 8.57 sys + 3061.00 cusr 290.06 csys = 3388.88 CPU) 09:39
[Tux] Rakudo version 2020.06-65-g90f86d4c9 - MoarVM version 2020.06-22-gaf5b89b64
csv-ip5xs0.820 - 0.849
csv-ip5xs-207.835 - 7.852
csv-parser22.905 - 25.569
csv-test-xs-200.389 - 0.411
test7.538 - 7.897
test-t1.826 - 1.882
test-t --race0.825 - 0.936
test-t-2030.110 - 30.974
test-t-20 --race9.041 - 9.475
 10:36
I'll be away most of the upcoming week, so I cannot guarantee timings
Geth_ rakudo: 9e9851a808 | (Elizabeth Mattijsen)++ | 8 files
CompUnit::Repository.need to return Nil on failure

  - moving the onus of throwing an exception to World.load_module
  - CU:Repository implementations need not know what exception to throw
  - should allow CU::Repository implementations to be smarter handling
   missing dependency specifications.
  - added ":first" flag to initial call to $*REPO.head.need, which
... (7 more lines)		 12:47
lucs Are these logs still available somewhere?: irclog.perlgeek.de/perl6-dev/2017-...i_14372945 13:10
lucs (link found on stackoverflow) 13:12
Altai-man_ lucs, colabti.org/irclogger/irclogger_lo...2017-04-04 <- 13:13
lucs Ah, thanks. I'd tried a similar, but failing, URL with "raku-dev" instead of "perl6-dev". 13:15
Geth_ nqp-configure/usev6-patch-1: 98c295ee5d | (Christian Bartolomäus)++ (committed using GitHub Web editor) | lib/NQP/Config.pm
Quote command line parameters (except for Windows)

This compensates for github.com/Raku/nqp/commit/1517cf77db and unbreaks nqp-j on non-Windows systems. Fixes github.com/Raku/nqp/issues/644		 13:45
nqp-configure: usev6++ created pull request #23:
Quote command line parameters (except for Windows)
Geth_ rakudo: 3674f812bb | (Elizabeth Mattijsen)++ | src/core.c/Str.pm6
Make sure naive-word-wrapper handles empty strings
 13:54
rakudo: c389cb5279 | (Elizabeth Mattijsen)++ | src/core.c/Exception.pm6
Add support for suggestions to X::CompUnit::UnsatisfiedDependency
rakudo: 2f6677b2d1 | (Elizabeth Mattijsen)++ | src/core.c/CompUnit/Repository/FileSystem.pm6
Make CU:R:FileSystem support the "first" flag

  - add extensive suggestion if file not found, and it is the first repo,
   and there was a META6.json file available.
  - fixes R#3787
linkable6 R#3787 [open]: github.com/rakudo/rakudo/issues/3787 [docs] Cannot load modules from a directory which contains a META6.json without a provides key indicating the location of files.
ugexe so we are changing the CUR interfaces it seems without review, can we not do that? 14:04
tellable6 2019-09-29T19:12:30Z #perl6-dev <tbrowder> ugexe when zef installs a module, is the on-disk module compressed? i.e., does a module author need to worry about a large resource set?
2019-09-30T01:09:28Z #perl6-dev <tbrowder> ugexe ^^^
2019-10-02T11:24:54Z #perl6-dev <tbrowder> ugexe does Build.pm6 need to be listed somehow in the META6.json file?
2019-10-02T17:50:36Z #perl6-dev <tbrowder> ugexe the zef build problem seems to be an incompletely unquoted cmd string in the legacy build module. see new issue #319
2019-10-06T23:40:04Z #perl6 <AlexDaniel> ugexe colabti.org/irclogger/irclogger_lo...10-06#l298
2019-10-15T19:20:06Z #perl6 <jdv79> ugexe github.com/perl6/ecosystem/issues/468 ?
hey ugexe, you have a message: gist.github.com/3fa09f519769ea6397...b7fff0604d
hey ugexe, you have a message: gist.github.com/7fd74ebce90a0de956...c66a4f4147
hey ugexe, you have a message: gist.github.com/1eecc243ce6125f370...f4136f6777
hey ugexe, you have a message: gist.github.com/79661347251b69d02d...e0fc846e35
ugexe :P
i dont want to review 5 separate commits to master 14:05
lizmat ugexe: you only need to review github.com/rakudo/rakudo/commit/9e9851a808 and github.com/rakudo/rakudo/commit/2f6677b2d1 really 14:07
afk for a few hours&
ugexe the other commits are related
there are 3 related commits 14:08
one as a reviewer i do not like the feeling of the clock counting down 14:10
and that is certainly the case when commiting directly to master
i have to not only make sure the code isn't going to cause a regression in a case that isn't covered obviously somewhere, but to make sure the design out of that subsystem itself is ideal (which is completely separate) 14:11
to get a better error message for a pretty non-common case? 14:12
anywho i have a problem-solving ticket open that asks why we don't require review process for interface changes, so i've tried to address this more generally already 14:13
Geth_ rakudo: ugexe++ created pull request #3789:
Revert broken master		 15:21
rakudo: f6acd50c38 | (Nick Logan)++ (committed using GitHub Web editor) | 9 files
Revert broken master (#3789)

  * Revert "Make CU:R:FileSystem support the "first" flag"
This reverts commit 2f6677b2d161e5dd0743eb6fce6b0719a0628fdf.
  * Revert "Add support for suggestions to X::CompUnit::UnsatisfiedDependency"
... (6 more lines)
linkable6 RAKUDO#3789 [closed]: github.com/rakudo/rakudo/pull/3789 Revert broken master
Geth_ Blin: 9ad9b7d2a8 | (Aleks-Daniel Jakimenko-Aleksejev)++ | META6.json
Fix source-url
 16:02
Geth_ roast: 69ebf6db69 | (Elizabeth Mattijsen)++ | S22-package-format/local.t
Make check for non-existing candidate less specific

The fact that a repo throws a specific error, is not part of any API. So instead just try the code, and see if we get Nil.		 16:55
Geth_ rakudo: 8e2bc22793 | (Elizabeth Mattijsen)++ | src/core.c/Exception.pm6
Add support for suggestions to X::CompUnit::UnsatisfiedDependency

This does not change any API, it adds an optional parameter.		 17:53
ugexe so, for the second time today master is failing its own tests 18:04
disable commits to master please
because the problem is apparently not going to get solved any other way 18:05
master is not the integration branch 18:06
i need someone to explain to me why we feel commiting directly to master is in our best interest? 18:10
Geth_ rakudo: 85320f4663 | (Elizabeth Mattijsen)++ | src/core.c/Exception.pm6
Revert "Add support for suggestions to X::CompUnit::UnsatisfiedDependency"

This reverts commit 8e2bc2279372a94c30f11d702210f8b7a4066ce6.		 18:11
lizmat because it evidently ensures immediate feedback ? 18:13
ugexe yeah i love having my priorities forced to me. i do have other stuff i planned to review today that has been patiently waiting in a PR itself 18:14
lizmat which one? 18:15
ugexe alternative dependencies in zef from nine
sena_kun I can see benefit of using branches. For example, with the match refactoring branch we could try out some ideas safely and seeing it isn't really simple to fix, can just put it away until a better time. No need to revert-re-revert in master. 18:30
Xliff m: class A does ::('B') { } 18:36
camelia 5===SORRY!5=== Error while compiling <tmp>
('B') is not composable, so A cannot compose it
at <tmp>:1
tellable6 2020-07-11T12:27:50Z #raku <holyghost> Xliff, thanks for the quick response, there's more developer kits on the server now
hey Xliff, you have a message: gist.github.com/e6f1fcdca3f996361a...35eae0db34
Xliff m: class A does ::('B') { } role B { }
camelia 5===SORRY!5=== Error while compiling <tmp>
('B') is not composable, so A cannot compose it
at <tmp>:1
Xliff m: role B { }; class A does ::('B') { }
camelia 5===SORRY!5=== Error while compiling <tmp>
('B') is not composable, so A cannot compose it
at <tmp>:1
Xliff m: class A does ::('B') { }; role B { method b { }; }
camelia 5===SORRY!5=== Error while compiling <tmp>
('B') is not composable, so A cannot compose it
at <tmp>:1
Xliff m: role B { method b { }; }; class A does ::('B') { };
camelia 5===SORRY!5=== Error while compiling <tmp>
('B') is not composable, so A cannot compose it
at <tmp>:1
Xliff Hrm. 18:37
m: role B { method b { }; }; class A { also does ::('B'); };
camelia 5===SORRY!5=== Error while compiling <tmp>
('B') is not composable, so A cannot compose it
at <tmp>:1
Xliff jnthn timotimo vrurg: So roles cannot be late-bound.
lizmat sena_kun: my mistakes were a. not pushing a roast change making the check less specific, b. not testing the cherry-pick I did afterwards 18:41
this broke CI twice, and made people angry 18:42
sena_kun lizmat, I wonder what is the policy about exception types.
ugexe it broke CI twice for what value? the second time it was added it is essentially dead code
lizmat [20:04:35] <ugexe>so, for the second time today master is failing its own tests 18:43
ugexe sure it might make sense in the context of the solution to the greater problem, but then it belong grouped with that solution
sena_kun Because it is clearly the gray zone and can be argued either way "Yes, it bends the canonical interpretation of roast" and "No, that should not be specced". 18:44
lizmat sena_kun: wrt exception types, I also wonder
but I guess the booboo was that something that *was* an exception before, wasn't anymore 18:45
in any case, I will give myself some time off of doing any commits until at least the release
sena_kun The release is pretty clear, by the way. 18:46
Blin is happy, so needs some packing.
Xliff m: say 1 +< 1 +| 1 +< 3 18:48
camelia 10
vrurg_ Xliff: 'does' is a compile-time thing. 18:55
Xliff vrurg: Yeah. Finally worked that out. Thanks. :) 19:41
vrurg Xliff: BEGIN? 19:42
Xliff vrurg: Huh. That might work.
vrurg: Also, reworked the PR to remove the .precomp nolock. 19:43
R#3785
linkable6 R#3785 [open]: github.com/rakudo/rakudo/pull/3785 Revised version of the Precomp Store with file, rather than directory locking
releasable6 Next release in ≈5 days and ≈19 hours. There are no known blockers. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft 23:00

Powered by Cro and the Raku® Programming Language.

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

Thank you!