Perl 6 language and compiler development | Logs at colabti.org/irclogger/irclogger_log/perl6-dev | For toolchain/installation stuff see #perl6-toolchain | For MoarVM see #moarvm
Set by Zoffix on 27 July 2018.
00:26 SporkWitch4 joined 00:27 p6bannerbot sets mode: +v SporkWitch4 00:32 SporkWitch4 left 00:40 Ricardus2 joined 00:41 p6bannerbot sets mode: +v Ricardus2 00:45 Ricardus2 left 01:02 Kaiepi left, Kaiepi joined 01:03 p6bannerbot sets mode: +v Kaiepi 01:25 GTAXL10 joined, GTAXL10 left 01:44 DanPin16 joined 01:45 p6bannerbot sets mode: +v DanPin16 01:50 DanPin16 left 02:03 meti24 joined 02:04 meti24 left
Geth rakudo/fix-windows-repositories: c3e1ec0bb8 | (Rob Hoelz)++ | src/core/CompUnit/RepositoryRegistry.pm6
Use $*SPEC.dir-sep for separators in CompUnit::RepositoryRegistry

On Windows, keys to the $repos hash are constructed exclusively with backslash separators, so keys look something like this:
   C:\rakudo\share\perl6\site
... (12 more lines)		 02:37
03:20 j3nnn1 left 03:37 lizmat left
Geth rakudo: c3e1ec0bb8 | (Rob Hoelz)++ | src/core/CompUnit/RepositoryRegistry.pm6
Use $*SPEC.dir-sep for separators in CompUnit::RepositoryRegistry

On Windows, keys to the $repos hash are constructed exclusively with backslash separators, so keys look something like this:
   C:\rakudo\share\perl6\site
... (12 more lines)		 03:43
04:23 lizmat joined, p6bannerbot sets mode: +v lizmat 05:51 brackets7 joined, brackets7 left 06:26 robertle joined 06:27 p6bannerbot sets mode: +v robertle 07:07 pmurias joined, p6bannerbot sets mode: +v pmurias 07:10 Adbray16 joined 07:11 p6bannerbot sets mode: +v Adbray16, Adbray16 left 07:49 dct joined, p6bannerbot sets mode: +v dct 08:03 pmurias left 08:04 pmurias joined 08:05 p6bannerbot sets mode: +v pmurias, pmurias left 08:11 pmurias joined, p6bannerbot sets mode: +v pmurias, pmurias left 08:18 dct left 08:23 pmurias joined, p6bannerbot sets mode: +v pmurias
lizmat good *, #perl6-dev! 08:24
Files=1248, Tests=75965, 365 wallclock secs (15.87 usr 5.73 sys + 2602.09 cusr 236.21 csys = 2859.90 CPU)
08:25 dct joined, p6bannerbot sets mode: +v dct
Geth rakudo: a9c44f4987 | (Elizabeth Mattijsen)++ | src/core/Map.pm6
Fix for R#2195

Not sure why the fix is needed, but at least we can continue with this now.		 08:29
synopsebot R#2195 [open]: github.com/rakudo/rakudo/issues/2195 [⚠ blocker ⚠] False positive throwage of "flattening of other hash reprs NYI"
08:40 infina8 joined, infina8 left 08:47 pmurias left 08:51 Ven` joined, p6bannerbot sets mode: +v Ven` 08:58 pmurias joined, p6bannerbot sets mode: +v pmurias 09:48 planetmaker5 joined 09:49 planetmaker5 left
Geth rakudo: 681ab71224 | (Jonathan Worthington)++ | src/core/Map.pm6
Revert "Fix for R#2195"

This reverts commit a9c44f4987fa7b255013a5505563bc9a547c9aa0, which worked around the problem rather than fixed it.		 09:50
rakudo: ec7de8417b | (Jonathan Worthington)++ | src/Perl6/Metamodel/BOOTSTRAP.nqp
Never bind a Mu into Hash's $!storage

Just leave it unset, and it'll vivify on demand if needed.
synopsebot R#2195 [open]: github.com/rakudo/rakudo/issues/2195 [testneeded] False positive throwage of "flattening of other hash reprs NYI"
09:52 patrickz joined, p6bannerbot sets mode: +v patrickz 09:55 Ven` left
timotimo # that. Otherwise, putting Mu in there is fine; Hash is smart 09:58
??? # enough to know what to do.
evalable6 Stub code executed
in block <unit> at /tmp/WnnJaU5mMW line 1
timotimo ???
evalable6 Stub code executed
in block <unit> at /tmp/kE3V1cpxTR line 1
timotimo i'll update that :)
thanks evalable6 09:59
Geth rakudo: 8d8c755b9b | (Timo Paulssen)++ | src/Perl6/Metamodel/BOOTSTRAP.nqp
update comment on Mu in Map's storage
 10:01
jnthn timotimo++ 10:02
timotimo proud
dogbert2 good morning 10:06
10:07 perl6nopath joined, p6bannerbot sets mode: +v perl6nopath
jnthn o/ dogbert2 10:07
dogbert2 hi jnthn, had any coffee yet
jnthn Yup :) 10:08
dogbert2 :)
timotimo man, i wish coffee tasted as good as it smells
perl6nopath reboot did solve the problem with PATH, but not the set up problem with Linennoise / editor
timotimo yes, that's a bug in rakudo that just got fixed about 8 hours ago 10:09
dogbert2 so what are you guys up to today, more cool opts perhaps
jnthn Trying to decide which difficult problem to solve next :)
dogbert2 want an appetizer ? if so try running ./perl6-valgrind-m t/spec/S06-signature/introspection.rakudo.moar 10:10
timotimo say, jnthn, do we want to have an annotation for ops that are fine to re-execute after deopt? like, a bindattr with the same values would be fine to re-do if that lets us put a deopt target earlier in the original code, and maybe that would let us get rid of, for example, the getspeshslot/wval that we sometimes keep around for bindattr's sake 10:11
like an "idempotent" annotation
jnthn I was just going to use :pure for that, tbh 10:12
But actually we keep a lot less getspeshslot/wval around nowadays
timotimo hm, OK
jnthn Partly thanks to the better deopt use tracking
timotimo i thought there's many ops that aren't :pure, but could still benefit from being considered pure for this purpose
jnthn But yesterday afternoon I added a PHI elimination pass *before* deopt tracking
Which let us kick out quite a lot more 10:13
timotimo cool, i didn't pull those changes yet
jnthn Currently staring at an bonkers guard sequence where it even duplicately guards the same thing 10:15
dogbert2 is it github.com/MoarVM/MoarVM/blob/mast...ne.c#L1161 10:16
Geth rakudo/master: 8 commits pushed by (Elizabeth Mattijsen)++
timotimo what is "it"?
lizmat that shaved off 1056 bytes 10:17
jnthn CORE.setting is shrinking :)
dogbert2 "it" refers to ==7408== Conditional jump or move depends on uninitialised value(s)
==7408== at 0x41AF2FD: MVM_spesh_inline (inline.c:1161)
from running the instrospection test file 10:18
jnthn How the hell can that be uninitialized...
dogbert2 is innocent :)
jnthn It's set at github.com/MoarVM/MoarVM/blob/mast...ine.c#L653 10:19
timotimo i'd kind of expect that to have been allocated from a zeroed block of memory
or perhaps the proxy_deopt_idx was the uninitialized thing?
jnthn Yeah, but it's unconditionally set
That ain't involved in the condition though
timotimo right 10:20
but who knows what the optimized code looks like :)
10:22 travis-ci joined, p6bannerbot sets mode: +v travis-ci
timotimo ah dang it. the profiler can still go off into outer space with the call graph, even after throwpayloadlexcaller has been made to prevent inlines while profiling 10:22
travis-ci Rakudo build errored. Jonathan Worthington 'Never bind a Mu into Hash's $!storage
travis-ci.org/rakudo/rakudo/builds/414441687 github.com/rakudo/rakudo/compare/a...7de8417b9c
10:22 travis-ci left
buggable [travis build above] ☠ Did not recognize some failures. Check results manually. 10:22
timotimo so there must be another op that does the same thing 10:23
10:25 perl6nopath left
lizmat commute to Glasgow& 10:38
10:38 lizmat left 10:40 planigan18 joined 10:41 p6bannerbot sets mode: +v planigan18, planigan18 left
Geth ¦ rakudo: zoffixznet self-assigned False positive throwage of "flattening of other hash reprs NYI" github.com/rakudo/rakudo/issues/2195 11:26
11:30 brrt joined 11:31 p6bannerbot sets mode: +v brrt, pmurias left, pmurias joined 11:32 p6bannerbot sets mode: +v pmurias
Geth roast: 27d34e73c3 | (Zoffix Znet)++ | S06-signature/slurpy-params.t
Cover crash in a slurpy slippage

Closes https//github.com/rakudo/rakudo/issues/2195 R#2195 Rakudo fix: github.com/rakudo/rakudo/commit/ec...cb16f07436		 11:33
synopsebot R#2195 [open]: github.com/rakudo/rakudo/issues/2195 [testneeded] False positive throwage of "flattening of other hash reprs NYI"
11:44 pmurias left 11:45 pmurias joined 11:46 p6bannerbot sets mode: +v pmurias 12:01 pmurias left 12:11 pmurias joined, p6bannerbot sets mode: +v pmurias 12:16 dct left, pmurias left 12:17 pmurias joined 12:18 p6bannerbot sets mode: +v pmurias, pmurias left 12:25 Zoffix joined, p6bannerbot sets mode: +v Zoffix
Zoffix is about to beat some records 12:26
Google has a 160CPU/3.8TB RAM VM
*heavy breathing*
|Tux| Rakudo version 2018.06-403-g79d7c4bec - MoarVM version 2018.06-399-g4fc267b60
csv-ip5xs0.925 - 0.939
csv-ip5xs-207.351 - 7.658
csv-parser22.622 - 23.055
csv-test-xs-200.407 - 0.452
test9.117 - 9.282
test-t2.112 - 2.231
test-t --race0.911 - 0.965
test-t-2036.216 - 36.247
test-t-20 --race11.794 - 12.014
 12:27
2018-08-09 19:12:48 test-t 2.162
2018-08-07 16:28:38 test-t 2.159
2018-08-09 16:34:28 test-t 2.133
2018-08-09 17:17:53 test-t 2.129
2018-08-10 14:15:17 test-t 2.112
12:27 pmurias joined
Zoffix ZOFFLOP: t/spec/S17-procasync/stress.t 12:28
12:28 p6bannerbot sets mode: +v pmurias
Zoffix ZOFFLOP: t/spec/S17-supply/syntax.t 12:28
Geth nqp: 295a580866 | (Zoffix Znet)++ | tools/build/MOAR_REVISION
[MoarVM Bump] Brings 4 commits

MoarVM bump brought: github.com/MoarVM/MoarVM/compare/2...9-g4fc267b 4fc267b Account for inlines that may cause deopts efbc6c6 Fix an uninitialized memory access bfbbb0f Do dead instruction elim pass after usage setup 210a402 Make inline logging available through an env var
¦ nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2...9-g4fc267b
rakudo: 0b3fc077e5 | (Zoffix Znet)++ | tools/build/NQP_REVISION
[NQP Bump] Brings 3 commits

NQP bump brought: github.com/perl6/nqp/compare/2018....7-g295a580 295a580 [MoarVM Bump] Brings 4 commits b23941d Often do a decont less in getattr/bindattr codegen 2bb945d Improve getattr/bindattr code gen quality
... (5 more lines)
rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2018....7-g295a580
AlexDaniel self-assigned 2018-08-10 Toast results / 29 burns github.com/rakudo/rakudo/issues/2196

Reduction spree in Baggy
Zoffix boooo 12:30
"Quota CPUs exceeded: your limit 64"
(you have to put a manual request to have your quota bumped)
12:31 brrt left 12:32 brrt joined, travis-ci joined, p6bannerbot sets mode: +v travis-ci
travis-ci NQP build failed. Zoffix Znet '[MoarVM Bump] Brings 4 commits 12:33
travis-ci.org/perl6/nqp/builds/414493737 github.com/perl6/nqp/compare/b2394...5a5808664b
12:33 p6bannerbot sets mode: +v brrt, travis-ci left 12:34 frmus6 joined 12:35 p6bannerbot sets mode: +v frmus6 12:36 frmus6 left
Zoffix Got a non-pre-emptive VM, so toast is comming up in ~1.5hr 12:37
12:41 Chex19 joined, pmurias left 12:42 brrt left, p6bannerbot sets mode: +v Chex19 12:43 Chex19 left 12:45 pmurias joined 12:46 p6bannerbot sets mode: +v pmurias
Zoffix releasable6 status 12:47
releasable6: status
releasable6 Zoffix, Next release in ≈8 days and ≈6 hours. 3 blockers. 81 out of 404 commits logged (⚠ 9 warnings)
Zoffix, Details: gist.github.com/7b12cb75563a4fcc99...24a79b7a17
Zoffix Damn. I can't find any of the commits! *snort* *snort* 12:48
12:49 pmurias left
Zoffix The R#2126 looks hard 12:49
synopsebot R#2126 [open]: github.com/rakudo/rakudo/issues/2126 [regression][⚠ blocker ⚠] Weird failure in Path::Router module
Zoffix and the other two blockers are OS/VM/compiler specific 12:50
12:51 Ven` joined 12:52 p6bannerbot sets mode: +v Ven`
jnthn Zoffix: I've a good idea how to fix that one but yes it's tricky and it's been too hot to think straight all week :/ 12:52
Will try and deal with it over the weekend or early next weke.
*week
Zoffix \o/ 12:53
12:54 pmurias joined, p6bannerbot sets mode: +v pmurias 12:56 Ven` left 12:57 sjohnson4 joined 12:58 sjohnson4 left 13:17 Guest16309 joined 13:18 Guest16309 left 13:19 pmurias left 13:26 Ven` joined 13:27 p6bannerbot sets mode: +v Ven`, robink joined, robink is now known as Guest28594, p6bannerbot sets mode: +v Guest28594, Guest28594 left 13:32 pmurias joined 13:33 p6bannerbot sets mode: +v pmurias, pmurias left
Zoffix buggable: toast 13:37
buggable Zoffix, Between 2018.06-404-g0b3fc07 and 2018.06: 29 (2.58%) modules got burnt; 45 (4.01%) got unsucced; 271 (24.13%) out of 1123 modules appear unusable. See toast.perl6.party/ for details.
13:37 pmurias joined
Zoffix "Type check failed for return value; expected CompUnit::Handle:D but got Nil (Nil)" 13:37
13:38 p6bannerbot sets mode: +v pmurias
Zoffix .tell AlexDaniel fresh toast: toast.perl6.party/ 13:38
yoleaux Zoffix: I'll pass your message to AlexDaniel.
jnthn *sigh* 13:39
It's stupid that we allowed that
Now we'll be stuck with it
Zoffix That :D on type constraint allows Nil? 13:40
13:40 k6ka22 joined, k6ka22 left
jnthn Yes. 13:41
13:41 x49F joined
jnthn If you'd asked me before this week if it did, I'd have said "of course not" 13:41
Zoffix Kinda makes sense IMO. Otherwise you can't `fail` out of :D constraints
13:41 x49F left
jnthn Yes, that's what the :D would mean 13:41
timotimo an argument could be made that :D means "shouldn't fail"
jnthn Otherwise it's meaningless
timotimo does having nil fail the type check blow up any spec tests at all though? 13:42
jnthn Apparently not.
timotimo so are we *really* stuck with it? :)
jnthn Otherwise I'd have known about the regression.
Zoffix No, it's not meaningless. -> --> Int:D { Int } fails type constraint 13:43
jnthn Well, it's like my Proxy returning bug fix.
Zoffix I don't undertand why it's fine for Failure to bypass all type constraints, EXCEPT the :D.
jnthn Ecosystem forces us to carry on with it that way
Zoffix: We'd kinda counted Nil and Failure as kinds of undefined-ish
Anyway, I can't be bothered to argue it 13:44
Zoffix I see
jnthn A golf would be welcome, as would a spectest
Zoffix Does the ecosystem really force anything onto us? If even the rakudo's behaviour doesn't define the Perl 6 language, the ecosystem certainly shouldn't. 13:45
jnthn Well, if we just break it then we'll get a reputation we don't want of breaking stuff.
So my strategy is to be grumpy here and then patch stuff up to keep working. :P 13:46
13:48 jimbeamm joined, skids joined, p6bannerbot sets mode: +v jimbeamm 13:49 p6bannerbot sets mode: +v skids, jimbeamm left 13:53 Zoffix left 13:54 brrt joined, robertle left 13:55 p6bannerbot sets mode: +v brrt 14:00 dogbert17 left 14:25 ExtraCrispy left 14:26 pmurias left, pwillard25 joined 14:27 pwillard25 left 14:38 enchi joined, enchi left
Ulti jnthn: for what its worth renaming a method for no reason feels like breaking stuff, suring up semantics of something I might have used without really thinking feels like fixing stuff 14:41
I have :D in the ecosystem where I sort of chucked out for the sake of it, and I wasnt aware it worked how just described
so its not like I was assuming that behaviour
AlexDaniel . 14:44
yoleaux 13:38Z <Zoffix> AlexDaniel: fresh toast: toast.perl6.party/
AlexDaniel Zoffix: I guess R#2196 is for me?
synopsebot R#2196 [open]: github.com/rakudo/rakudo/issues/2196 [⚠ blocker ⚠] 2018-08-10 Toast results / 29 burns
15:13 j3nnn1 joined, p6bannerbot sets mode: +v j3nnn1 15:35 lizmat joined 15:36 p6bannerbot sets mode: +v lizmat
AlexDaniel Bytecode stream corrupt (missing magic string) 15:50
oh wow, wow
jnthn o.O
diakopter make nukefromorbit 15:51
AlexDaniel that's from toast.perl6.party/module?module=Sp...4-g0b3fc07
anyway, gonna take a look at these failures later this evening
jnthn Do we know if they were done with MoarVM HEAD or MOAR_REVISION? 15:52
I fixed a deopt bug last night that was revealed by more PHI pruning (which in turn means more aggressive instruction deletion) 15:53
And I think the current bump was inbetween the two 15:54
AlexDaniel MOAR_REVISION
jnthn OK, that may account for some of 'em. 15:55
16:01 Zoffix joined, p6bannerbot sets mode: +v Zoffix
Zoffix AlexDaniel: well, for whomever. I just wanted to notify devs there were some burnt toasts 16:01
Won't have the time to look through the failures and file individual tickets myself. 16:02
jnthn: I bumped before toasting, so MOAR_REVISION is Moar HEAD ATM 16:03
16:03 Sebastien23 joined, Sebastien23 left 16:05 Zoffix left
jnthn Ah, OK 16:05
16:06 ZofBot left, ZofBot joined, ChanServ sets mode: +v ZofBot 16:07 p6bannerbot sets mode: +v ZofBot, Ven` left 16:20 diakopter left 16:27 brrt left
AlexDaniel Zoffix: ok, thanks! I'll go through the failures like usually 16:27
16:41 lizmat left 16:42 lizmat joined, p6bannerbot sets mode: +v lizmat, diakopter joined, p6bannerbot sets mode: +v diakopter 17:03 Ven` joined 17:04 p6bannerbot sets mode: +v Ven` 17:25 oldschool^11 joined 17:26 p6bannerbot sets mode: +v oldschool^11 17:29 oldschool^11 left 17:45 astj left
rakudo: ce5a011379 | (Elizabeth Mattijsen)++ | src/core/Mixy.pm6
Reduction spree in Mixy
lizmat which shaved off 520 from the core setting 17:49
dinner&
17:49 lizmat left 17:55 montag4516 joined, montag4516 left 17:57 astj joined 17:58 p6bannerbot sets mode: +v astj 18:01 astj left 18:04 travis-ci joined, p6bannerbot sets mode: +v travis-ci
travis-ci Rakudo build errored. Elizabeth Mattijsen 'Reduction spree in Mixy' 18:04
travis-ci.org/rakudo/rakudo/builds/414626074 github.com/rakudo/rakudo/compare/0...5a0113794c
18:04 travis-ci left
buggable [travis build above] ☠ Did not recognize some failures. Check results manually. 18:04
18:11 JoseyWales left 18:26 Ven` left
Geth ¦ rakudo: AlexDaniel assigned to samcv Issue The Regional Indicator Symbols seem to get mangled when flipped sometimes github.com/rakudo/rakudo/issues/2198 18:53
19:37 pmurias joined 19:38 p6bannerbot sets mode: +v pmurias
Geth nqp/truffle: 7504bf7d2e | (Paweł Murias)++ | 2 files
[truffle] Implement nqp::null_s
 19:48
nqp/truffle: a81fec369b | (Paweł Murias)++ | 2 files
[truffle] Implement nqp::list_s
nqp/truffle: 1fd7f51fd4 | (Paweł Murias)++ | 2 files
[truffle] A primitive version of nqp::die/nqp::die_s
19:52 lizmat joined 19:53 p6bannerbot sets mode: +v lizmat 20:26 wget13 joined 20:27 wget13 left 20:35 dogbert17 joined, p6bannerbot sets mode: +v dogbert17
Geth rakudo: 3301ddb1c6 | (Elizabeth Mattijsen)++ | 2 files
Reduction spree in natiev_array
 20:48
21:07 skids left 21:08 lizmat_ joined 21:09 p6bannerbot sets mode: +v lizmat_ 21:11 lizmat left 21:15 travis-ci joined, p6bannerbot sets mode: +v travis-ci
travis-ci Rakudo build passed. Elizabeth Mattijsen 'Reduction spree in natiev_array' 21:15
travis-ci.org/rakudo/rakudo/builds/414690258 github.com/rakudo/rakudo/compare/c...01ddb1c650
21:15 travis-ci left 21:19 patrickz left 21:22 Ven` joined 21:23 p6bannerbot sets mode: +v Ven` 21:27 Ven` left 21:56 annieslmaos joined, annieslmaos left 22:24 lizmat_ left 22:25 Kaiepi left 22:28 PufferBot joined, Kaiepi joined 22:29 p6bannerbot sets mode: +v PufferBot, p6bannerbot sets mode: +v Kaiepi 22:41 lizmat joined, CrunchyChewie15 joined, p6bannerbot sets mode: +v lizmat 22:42 p6bannerbot sets mode: +v CrunchyChewie15, CrunchyChewie15 left 22:59 pmurias left 23:00 pmurias joined, p6bannerbot sets mode: +v pmurias 23:06 __idiot__ joined 23:07 __idiot__ left 23:12 pmurias left 23:23 Ven` joined, p6bannerbot sets mode: +v Ven` 23:27 Ven` left
AlexDaniel .tell jnthn GH#2201 23:46
yoleaux AlexDaniel: I'll pass your message to jnthn.
synopsebot GH#2201 [open]: github.com/rakudo/rakudo/issues/2201 [regression][⚠ blocker ⚠] CompUnit::Util module failing after "use a spesh plugin for return value type checks" commit
AlexDaniel fwiw pretty much everyone can help with golfing the blockers 23:47
jnthn AlexDaniel: Thanks for the ticket; that one should be an easy fix 23:53
yoleaux 23:46Z <AlexDaniel> jnthn: GH#2201
synopsebot GH#2201 [open]: github.com/rakudo/rakudo/issues/2201 [regression][⚠ blocker ⚠] CompUnit::Util module failing after "use a spesh plugin for return value type checks" commit
jnthn Fixed 3 issues in MoarVM this evening, 2 of them mighta had toaster impact, one was harmless 23:55
AlexDaniel not sure if I'll finish this then: github.com/rakudo/rakudo/issues/2196 23:57
results from a rerun after GH#2201 is fixed will be easier to work with 23:58
synopsebot GH#2201 [open]: github.com/rakudo/rakudo/issues/2201 [regression][⚠ blocker ⚠] CompUnit::Util module failing after "use a spesh plugin for return value type checks" commit
AlexDaniel synopsebot: omg no need to repeat that 3 times
jnthn Yeah, probably easier for me to fix that tomorrow 23:59

Powered by Cro and the Raku® Programming Language.

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

Thank you!