01:10 AlexDaniel left 01:11 AlexDaniel joined, AlexDaniel left, AlexDaniel joined 01:14 Altai-man_ joined 01:17 sena_kun left 02:00 softmoth left 02:01 softmoth joined
Geth_ nqp: 5db59e7973 | Coke++ | docs/ops.markdown
add unbox_u

moar variant opcode
[Coke] Down to 306 nqp doc failures - pick an easy one and doc it, get us below 300! (gist.github.com/coke/325a98074f317...266df7da4) 02:51
03:15 sena_kun joined 03:16 Altai-man_ left 05:14 Altai-man_ joined 05:16 sena_kun left 06:11 softmoth left 07:15 sena_kun joined 07:17 Altai-man_ left 07:27 lichtkind joined 08:05 [TuxCM] joined 09:08 pmurias joined 09:14 Altai-man_ joined 09:16 sena_kun left 09:51 Kaiepi left, Kaeipi joined 09:53 Kaeipi left, Kaiepi joined
Altai-man_ AlexDaniel, I truly appreciate your blin fixes! It turns gist.github.com/Altai-man/50db11ae...e91209ae80 into gist.github.com/Altai-man/5b2e7a12...e6ffeaf017 10:05
Which is a very nice improvement to have.
10:06 pmurias left 10:07 pmurias joined
Altai-man_ Slippy::Semilist has a PR, Carp was discussed and fixed, same with Grammar::BNF, CSS::Grammar. November is, erm, very odd one and two ones seem like a regression. I wonder how reproducible would `MoarVM panic: Internal error: invalid thread ID -1469900288 in GC work pass` be, though... 10:07
Xliff++ AlexDaniel++ 10:08
By the way, gist.github.com/Altai-man/b67996b0...772c88d744 <- I am seeing git-reference warning 17 (!) times when building rakudo, what can be at fault? 10:13
10:20 lichtkind left
lizmat Files=1306, Tests=111236, 214 wallclock secs (28.91 usr 8.19 sys + 3033.63 cusr 264.01 csys = 3334.74 CPU) 10:42
AlexDaniel Altai-man_: “I wonder how reproducible would `MoarVM panic: Internal error: invalid thread ID -1469900288 in GC work pass` be, though...” 11:06
Altai-man_: you can check the bisect log
or, hmm… if it bisected to fdfa6ac85 then maybe it's not reproducible at all 11:07
linkable6 (2020-04-17) github.com/rakudo/rakudo/commit/fdfa6ac85c Bump NQP
Altai-man_ I tried it locally a couple of times and no luck.
Which is, of course, not your best test case, but still it is not _so_ reproducible to be easy to catch. 11:08
AlexDaniel right right 11:09
I see, ok
[TuxCM] Rakudo version 2020.02.1-332-gfdfa6ac85 - MoarVM version 2020.02.1-96-g4ddc33f3d
csv-ip5xs0.813 - 0.824
csv-ip5xs-208.122 - 8.402
csv-parser23.733 - 25.013
csv-test-xs-200.389 - 0.390
test7.691 - 7.742
test-t1.946 - 1.959
test-t --race0.856 - 1.133
test-t-2031.595 - 32.723
test-t-20 --race9.887 - 10.253
AlexDaniel Altai-man_: OK, that's pretty good, so November is the only one that still has a problem
Altai-man_: we're not testing November anyway: github.com/Raku/Blin/blob/8dda1b4c...p6#L43-L61 11:13
so perhaps we can just move it into another list and skip it altogether
11:15 sena_kun joined 11:16 Altai-man_ left 11:41 dogbert17 joined
nine AlexDaniel: I think --git-reference got renamed 12:13
AlexDaniel sena_kun: ↑
nine Superseeded by --git-cache-dir 12:14
sena_kun Seems so, but the question is why don't we have it in rakudo? Or is it not really backward compatible for old git versions? I am even more surprised we have github.com/rakudo/rakudo/pull/3474 which updates this option and it's merged.
13:14 Altai-man_ joined 13:17 sena_kun left 13:22 Geth_ left, Geth joined 13:23 finsternis left 13:24 finsternis joined 13:33 Geth left 13:34 Geth joined 13:49 lichtkind joined
Geth nqp: 13de5f5d9b | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/templates/MOAR_REVISION
[MoarVM Bump] Brings 21 commits

MoarVM bump brought: github.com/MoarVM/MoarVM/compare/2...gab089e947 ab089e947 Merge branch 'wictory-align_st_size' 02486a319 fixup! Adding to assertions of alignment invariant to st->size in P6opaque.c 1eba4e8b3 I missed this align_to before. e932be9ac This align_to is unnecessary. ... (17 more lines)
rakudo: e5ecdc4382 | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/templates/NQP_REVISION
[NQP Bump] Brings 5 commits

NQP bump brought: github.com/perl6/nqp/compare/2020....g13de5f5d9 13de5f5d9 [MoarVM Bump] Brings 21 commits 5db59e797 add unbox_u 1f8664432 Document isttyfh op b6c4b56f1 document totalmem op 25e3ccd00 document jvmgetproperties
14:11 donaldh left
AlexDaniel goddamn GitHub PR user interface I hate it 14:12
you can resolve conflicts (nice!) but then can't choose what's going to happen exactly with the merge (no way to squash or anything like that) 14:13
if you do it manually then github has no idea that the PR was merged, of course
I don't want a line going two years back in git pretty log
Geth nqp/master: 4 commits pushed by (Martin Ryan)++, (Aleks-Daniel Jakimenko-Aleksejev)++ 14:19
AlexDaniel so one page of PRs in nqp, that's sorta OK 14:22
two pages in MoarVM, still somewhat fine 14:23
106 pull requests in rakudo? Whaaaaat
greppable6: PrivateAccessor 14:27
greppable6 AlexDaniel, 29 lines, 2 modules: gist.github.com/0e28684ce688667177...84aa58ddf6
AlexDaniel m: my @​a = 1,2,3,4,5; @​a[1]​:delete; .say for @​a[0..*] 14:44
camelia 5===SORRY!5=== Error while compiling <tmp>
Name must begin with alphabetic character
at <tmp>:1
------> 3my @7⏏5​a = 1,2,3,4,5; @​a[1]​:delete; .say for
expecting any of:
AlexDaniel what
m: my @a = 1,2,3,4,5; @a[1]​:delete; .say for @a[0..*]
camelia 5===SORRY!5=== Error while compiling <tmp>
Bogus postfix
at <tmp>:1
------> 3my @a = 1,2,3,4,5; @a[1]7⏏5​:delete; .say for @a[0..*]
expecting any of:
infix stopper
statement end
AlexDaniel m: my @a = 1,2,3,4,5; @a[1]:delete; .say for @a[0..*]
camelia 1
AlexDaniel right
m: my @a = 1,2,3,4,5; .say for @a[0..*]
camelia 1
AlexDaniel m: my @a = 1,2,3,4,5; .say for @a[-5..*]
camelia ( no output )
AlexDaniel m: my @a = 1,2,3,4,5; .say for @a[-Inf..*] 14:45
camelia Cannot convert -Inf to Int:
in block <unit> at <tmp> line 1
AlexDaniel o_o
lizmat: can you take a look at this change, please? github.com/rakudo/rakudo/pull/1172
lizmat: it looks like something you'd normally fix :)
lizmat: the `pos.min ~~ -Inf` check is just wrong, I think, so we can't merge it as is 14:46
sourceable6: 5.5.round() 15:06
sourceable6 AlexDaniel, github.com/rakudo/rakudo/blob/e5ec...l.pm6#L387
AlexDaniel sourceable6: (-i).round() 15:07
sourceable6 AlexDaniel, github.com/rakudo/rakudo/blob/e5ec...x.pm6#L228
AlexDaniel sourceable6: (5e5).round()
sourceable6 AlexDaniel, github.com/rakudo/rakudo/blob/e5ec...al.pm6#L41
AlexDaniel jnthn: I don't have the courage to shut down R#2826 altogether and tell them that it should be done in a module instead, but you might have better luck. If so then it's 1 ticket less :) 15:10
linkable6 R#2826 [open]: github.com/rakudo/rakudo/issues/2826 [math] Rounding behaviors
15:15 sena_kun joined 15:16 Altai-man_ left 15:25 ufobat__ joined
AlexDaniel looks at whateverable 15:26
250 open issues
AlexDaniel closes the tab
15:30 ufobat_ left
AlexDaniel 6c: Nil for ^10_000_000 + 1 15:56
15:56 softmoth joined
committable6 AlexDaniel, gist.github.com/c12b3d30db026a7779...585f2e2a79 15:58
AlexDaniel 6c: Nil for ^10_000_000 + 1; say now - INIT now 16:03
6c: Nil for ^10_000_000; say now - INIT now
committable6 AlexDaniel, gist.github.com/8f3b64d59221b24931...b5b32672b7 16:05
AlexDaniel, gist.github.com/f359cfbc5bbabb4ee9...21963453ce
AlexDaniel 6c: my $x; $x++ for ^10_000_000 + 1; say $x; say now - INIT now 16:07
6c: my $x; $x++ for ^10_000_000; say $x; say now - INIT now
committable6 AlexDaniel, gist.github.com/61fd9fb1060883ab14...ba5f8fdc07 16:11
AlexDaniel, gist.github.com/beb90de373a1b07fca...3e83f928f0
16:14 softmoth left 17:14 Altai-man_ joined 17:17 sena_kun left
Geth roast: 7033b07bbb | (Christian Bartolomäus)++ | S04-statements/gather.t
[JVM] Update reference for skipped test
¦ problem-solving: alabamenhu assigned to jnthn Issue Rounding issues github.com/Raku/problem-solving/issues/176 18:13
18:23 softmoth joined
Kaiepi uh 18:54
so this line in Perl6::Actions marks variables for lexical usage by default github.com/rakudo/rakudo/blob/mast....nqp#L3897
but none of Variable's traits set it to False, they set it to True instead 18:55
it's supposed to be set to False here right?
timotimo what kind of stuff is in late traits? 19:00
github.com/rakudo/rakudo/commit/b4...2c28f5c1aa - did you look at this? 19:01
that's where you got the "none set it to false"?
Kaiepi yes 19:04
it's still like that now
there isn't any code elsewhere that touches that attribute either, according to git grep 19:07
19:15 sena_kun joined 19:16 Altai-man_ left 19:31 wildtrees joined 19:47 guifa2 joined
Xliff Kaiepi: If the variable isn't initialized, it will be Falsy. Those are the only situations in which it is set to be true. 20:09
Are you saying there's no where in the code where that attribute is even referenced/checked? 20:10
20:11 softmoth left
timotimo i think only written to 20:16
the optimizer for sure reads it
Xliff Then you'd only need to write a true. There would be no need to ever set it false. 20:28
It's not like it's a flag or a sentinel.
timotimo well, that isn't the only place where traits can be called from 20:29
this could be about meta-programming situations? 20:30
21:02 Xliff left 21:14 Altai-man_ joined 21:16 sena_kun left 22:03 pmurias left 22:11 guifa2 left 22:37 Altai-man_ left
Kaiepi xliff, i was thinking that what's happening is Perl6::Actions sets this to True -> some traits reassign this to true -> lexical_used_implicitly gets annotated for *all* variables, not just those that get used implicitly by blocks in the `will` traits 22:39
so the optimizer might never be lowering any variables
i'll check if that's actually the case or not though,gimme a sec 22:41
yeah, when it's set to False using `will check { ... }` will still lower the variable, but not when it's set to True 22:46
22:54 lichtkind left 23:45 guifa2 joined