00:36 Kaiepi left 00:56 benjif left 00:59 ZzZombo_ joined 01:02 ZzZombo left, ZzZombo_ is now known as ZzZombo 01:50 softmoth left 02:14 softmoth joined 02:33 Kaiepi joined 03:27 ZzZombo_ joined 03:30 ZzZombo left, ZzZombo_ is now known as ZzZombo 03:32 Kaiepi left 05:31 Kaiepi joined 05:46 Kaiepi left 05:51 Kaiepi joined 06:13 softmoth left 06:21 [Tux] left
nine vrurg: that sounds promising! The only thing making me nervous is that the comment above that method talks about precomp and our minimal test case doesn't actually use precompilation. But the comment may just be misleading 06:49
07:39 ZzZombo_ joined 07:42 ZzZombo left 07:43 ZzZombo_ is now known as ZzZombo 08:11 ZzZombo left
Geth ¦ problem-solving: JJ self-assigned Generic name for FALLBACK and CALL-ME github.com/perl6/problem-solving/issues/119 08:30
08:32 ufobat_ is now known as ufobat
lizmat Files=1276, Tests=109517, 202 wallclock secs (28.33 usr 7.31 sys + 2893.61 cusr 213.88 csys = 3143.13 CPU) 10:06
tellable6 2019-10-18T08:38:24Z #moarvm <brrt> lizmat I know, I pushed those accidentally, then couldn't push back, or didn't want to
2019-10-18T09:08:20Z #moarvm <brrt> lizmat I'm decently sure that I fixed *those* bugs now
Geth nqp: 0d7250f00d | (Elizabeth Mattijsen)++ | tools/templates/MOAR_REVISION
Bump NQP to get latest fixes
10:13
lizmat .tell tbrowder looks like your changes broke code at $project 10:19
tellable6 lizmat, I'll pass your message to tbrowder
lizmat .tell tbrowder looks like an empty #| at the start of a line in a sequence of #| comments, breaks compilation 10:21
tellable6 lizmat, I'll pass your message to tbrowder
lizmat m: #| foo 10:25
#|
#| bar
camelia ( no output )
lizmat hmmm
Geth rakudo: 33192b5e93 | (Elizabeth Mattijsen)++ | tools/templates/NQP_REVISION
Bump NQP to get latest MoarVM fixes
10:29
lizmat bisectable6: "#| foo\n#|\n#| bar".EVAL 10:31
tellable6 lizmat, I'll pass your message to bisectable6
lizmat hmmm... looks like we lost bisectable ?
.tell tbrowder github.com/rakudo/rakudo/issues/3242 10:32
tellable6 lizmat, I'll pass your message to tbrowder
Geth roast: f39e92ac2f | (Elizabeth Mattijsen)++ | S26-documentation/why-leading.t
Add test for R#3242
10:44
lizmat m: "#| foo\n#|\n#| bar".EVAL 10:58
camelia 5===SORRY!5=== Error while compiling /home/camelia/EVAL_0
Opening bracket required for #| declarator block
at /home/camelia/EVAL_0:2
------> 3#|7⏏5<EOL>
11:19 MasterDuke left 11:44 pmurias joined 11:47 ufobat left
tbrowder lizmat: oops! i'll see if i can fix that. good catch. 11:49
tellable6 2019-10-18T10:19:51Z #raku-dev <lizmat> tbrowder looks like your changes broke code at $project 11:50
2019-10-18T10:21:29Z #raku-dev <lizmat> tbrowder looks like an empty #| at the start of a line in a sequence of #| comments, breaks compilation
2019-10-18T10:32:22Z #raku-dev <lizmat> tbrowder github.com/rakudo/rakudo/issues/3242
12:02 MasterDuke joined 12:08 pmurias left 12:09 pmurias joined 12:16 pmurias left 12:17 pmurias joined 12:22 pmurias left
tbrowder .ask lizmat can you revert my merge? 12:25
tellable6 tbrowder, I'll pass your message to lizmat
12:27 [TuxCM] joined
[TuxCM] Rakudo version 2019.07.1-414-g33192b5e9 - MoarVM version 2019.07.1-283-ge594fbcbd
csv-ip5xs0.746 - 0.756
csv-ip5xs-206.357 - 6.371
csv-parser20.835 - 20.851
csv-test-xs-200.417 - 0.428
test7.195 - 7.228
test-t1.730 - 1.748
test-t --race0.780 - 0.781
test-t-2029.152 - 29.358
test-t-20 --race9.157 - 9.236
12:41
12:44 Kaiepi left, pmurias joined 12:53 Kaiepi joined 13:11 bisectable6 joined
AlexDaniel tbrowder: how do you write your commit messages? 13:13
tbrowder: do you use git from the command line or some tool?
13:18 ufobat joined
Geth rakudo: Kaiepi self-assigned signal has odd signature github.com/rakudo/rakudo/issues/3035
9f1794cf36 | (Aleks-Daniel Jakimenko-Aleksejev)++ | 2 files

This reverts commits:
  * 24bd36bca3869d1b5529c36c34a3523e6b1fd81c.
  * 03fb9e6ed1179acd665d0f6f2ac4382e7b5e50d2.
  * 749831b2d20855b6d07ee8d288ec9bbf77c7d0bb.
13:20
vrurg nine: it wasn't even close. Just spotted a potential issue.
tbrowder git -m at cli 13:28
vrurg nine: just for the record: what I currently surmise is that core.setting is not serialized with a unit sc. But when unit itself is deserialized it fixes this with load deps task. But if an object from that unit is repossessed – there is nobody to attach correct core to its outers. 13:34
tbrowder AlexDaniel: nice work, thnx! 13:44
AlexDaniel tbrowder: don't use -m
tbrowder: github.com/rakudo/rakudo/blob/mast...t-messages 13:45
tbrowder: first line is a title, then an empty line, then a useful description
realistically there's no easy way to do that if you use -m
if you don't use -m, git will open a text editor where you can write the commit message properly 13:46
which text editor it opens is configurable, in my case for example that's emacs because that's what I use
tbrowder what i've done in the past is use "git commit -F file.log" 13:49
with this pr i had to commit in pieces because of various mistakes. but it was to my branch and i've seen git allows an overall merge msg so i thought that was satisfactory. 13:53
AlexDaniel -F – Yeah, well, that's a hard way to do it
as for fixing mistakes in your branch `git commit --amend` and `git rebase -i` 13:54
tbrowder in the future, on my end, should i incrementally merge master into my branch during work for the pr?
AlexDaniel no
tbrowder i'm really sorry about the snafu. i usually do a full spec test before the pr but got sloppy. 13:55
AlexDaniel what you can do is rebase your commits on top of master 13:56
but generally you don't need that
well, you do if your commits are very old… 13:57
AlexDaniel` to visualize, these are your commits: 13:59
AlexDaniel` uploaded an image: image.png (2KB) < matrix.org/_matrix/media/r0/downlo...rIhZffKoJw >
AlexDaniel that blue lines goes back several *months*
and everything else looks confusing too
functionally it doesn't matter, but if everyone starts doing that our git history will look like something from nightmares :) 14:00
tbrowder aarg!
AlexDaniel tbrowder: I think this was a nice resource, probably still is: learngitbranching.js.org/ 14:01
there are a few great excercises on how to rebase and juggle with commits
tbrowder ok, but the -F log seems easier than the interactive--keep notes as one goes along. 14:02
AlexDaniel tbrowder: perhaps, but I never needed that because I can always just look at the diff of what I've done 14:04
`git diff` to see what was changed and `git diff --staged` to see what I'm about to commit
and after commiting `git show` to see the last commit before I push
tbrowder yeah, i use the git diff regularly, but haven't used --staged 14:06
nine vrurg: also my message was bogus anywhere. Your bug is all about precompilation. I just confused it with the other issue I'm looking into
vrurg I should not forget to commit the fix later anyway. 14:08
14:15 pmurias left
vrurg If I'm correct about core setting not been re-attached to resposessed closures, then it means same bug must happen if closure is referring to a symbol from another module. Need to check it when have time... 14:16
14:21 pmurias joined
AlexDaniel kawaii: any news? :) 14:23
14:24 lucasb joined
Geth nqp/truffle: 4 commits pushed by (Paweł Murias)++ 14:29
kawaii For you, Alex? Always.
usercontent.irccloud-cdn.com/file/...-18.tar.gz
AlexDaniel: ^ 🥳
AlexDaniel kawaii: holy moly 14:30
Geth roast: Kaiepi++ created pull request #584:
Add tests for R#3035
14:38 pmurias left 14:50 pmurias joined 14:51 [TuxCM] left 14:58 pmurias left 15:06 Guest13443 left 15:29 ufobat left
Geth ¦ problem-solving: JJ assigned to AlexDaniel Issue Wikipedia and Wikidata curation github.com/perl6/problem-solving/issues/120 15:46
16:24 softmoth joined 16:33 [Tux] joined
AlexDaniel kawaii: btw I looked through these and filed github.com/rakudo/rakudo/issues/3243 and github.com/rakudo/rakudo/issues/3244 17:00
everything else is probably fine
17:40 Kaiepi left 17:45 benjif joined 18:04 Kaiepi joined
nine Finally! 18:21
The new magic incantation that fixes things is: $*W.context().sub_id_to_code_object(){nqp::getcodecuid(nqp::getattr(self, Code, '$!do'))} := -> {} if $*W; 18:25
AlexDaniel: this is about the NativeCall regression ^^^ 18:41
So what's happening is that while we replace the routine's $!do (the actual code body) when we apply the native trait, $*W.compile_in_context will replace it again with the original routine. compile_in_context is used for evaluating BEGIN blocks and also role bodies during compilation. 18:42
The line I just posted deflects that into the empty code block, so our replacement $!do stays intact. 18:43
19:14 ggoebel left
Geth rakudo: e45bb3418d | (Stefan Seifert)++ | 2 files
Fix native subs declared in BEGIN blocks and role bodies

When a native sub was declared within a BEGIN block or a role body, the replacement code body did not stick but was replaced with the original stub when the containing block was compiled dynamically.
Fix by removing the code object from the map of objects needing fixup. Fixes GH #3235 and #3243
19:24
nine AlexDaniel: ^^^
19:25 Kaiepi left 19:26 Kaiepi joined
tbrowder see rakudo issue #2332 for my original proposal 19:44
nine Anyone else seeing t/08-performance/05-processkeys.t failing during make test? Can't make it fail when run individually but it reproducibly fails in make test 19:51
lizmat nine: nope, can't reproduce, even under load 19:53
tellable6 2019-10-18T12:25:02Z #raku-dev <tbrowder> lizmat can you revert my merge?
lizmat tbrowder: looks like AlexDaniel beat me to it
tbrowder yep, sorry for the snafu 19:54
if you get a chance i've proposed a change to handling leading declarator blocks over on #raku 19:56
MasterDuke nine: not happening for me 20:00
but i think i've seen some random fails in t/02-rakudo/repl.t
20:05 pmurias joined
Kaiepi nine, can't reproduce either 20:16
Geth ¦ problem-solving: JJ assigned to jnthn Issue Check MoarVM out using OSS-fuzz github.com/perl6/problem-solving/issues/121 20:17
20:29 pmurias left
Geth roast: Altai-man++ created pull request #585:
Test Parameter subclassing
20:30
20:31 pmurias joined
Geth roast: Altai-man++ created pull request #586:
Test runtime-created classes being attributes
20:40
rakudo: Altai-man++ created pull request #3245:
Test rakudo-specific warning for shape specifier
20:54
21:01 Kaiepi left
Geth nqp: xkr47++ created pull request #585:
Update README.pod: irc #perl6 -> #raku
21:01
21:02 Kaiepi joined
Geth roast: Altai-man++ created pull request #587:
Test hyperop results in a slip
21:05
21:29 softmoth left
Geth nqp/truffle: c1706c6733 | (Jonas Berlin)++ (committed using GitHub Web editor) | README.pod
Update README.pod

  #perl6 -> #raku
21:52
nqp/truffle: 12a1676ba8 | (Jonathan Worthington)++ (committed using GitHub Web editor) | README.pod
Merge pull request #585 from xkr47/patch-1

Update README.pod: irc #perl6 -> #raku
AlexDaniel nine: niiice! 22:07
.seen cygx
tellable6 AlexDaniel, I saw cygx 2019-08-24T08:52:30Z in #perl6-dev: <cygx> Well, there are still some moving pieces, but in contrast to my last attempt a couple of years ago, I have a better idea what needs to happen
AlexDaniel vrurg: any news on github.com/rakudo/rakudo/issues/3190 ? 22:09
Geth ¦ rakudo: AlexDaniel assigned to jnthn Issue Regression in Cro (fetching a json file eats RAM) github.com/rakudo/rakudo/issues/3223 22:11
22:12 ggoebel joined 22:17 ggoebel left
Geth rakudo/new-extensions: daa16b9c92 | (Aleks-Daniel Jakimenko-Aleksejev)++ | src/core.c/CompUnit/Repository/FileSystem.pm6
Add support for .rakumod extension
22:18
rakudo: AlexDaniel++ created pull request #3246:
Add support for .rakumod extension
22:26 [TuxCM] joined 22:34 [TuxCM] left 22:54 pmurias left
vrurg AlexDaniel: no news. Frankly, I forgot about it. Installing openbsd now... Again... :( 23:39