timotimo gnite jnthn 00:03
i'm also going to bed soon
AlexDaniel Geth: ver github.com/rakudo/rakudo/commit/23...b95339af34 00:09
Geth AlexDaniel, version bump brought in these changes: github.com/perl6/nqp/compare/2015....5-g02910ff
AlexDaniel Geth: ver github.com/perl6/nqp/commit/02910f...75056265ca 00:10
Geth AlexDaniel, version bump brought in these changes: github.com/MoarVM/MoarVM/compare/2...8-g3240047
AlexDaniel buggable: tags 00:32
buggable AlexDaniel, Total: 1647; 6.D: 2; 9999: 10; @LARRY: 24; ANNOYING: 7; BOOTSTRAP: 5; BUG: 601; BUILD: 12; CONC: 48; DOCS: 1; EXOTICTEST: 1; GLR: 3; IO: 23; JVM: 53; LHF: 2; LTA: 177; MATH: 4; META: 1; MOAR: 1; NATIVECALL: 22; NYI: 55; OO: 13; OPTIMIZER: 8; OSX: 2; PARSER: 5; PERF: 28; POD: 19; PRECOMP: 15; REGEX: 48; REGRESSION: 37; REPL: 6; RFC:
AlexDaniel, 63; RT: 1; SECURITY: 1; SEGV: 25; SINK: 1; SITE: 1; SPESH: 1; STAR: 7; TESTCOMMITTED: 10; TESTNEEDED: 50; TODO: 13; UNI: 26; UNTAGGED: 416; WEIRD: 2; WINDOWS: 3; See fail.rakudo.party/ for details
AlexDaniel buggable: tags DOCS 00:33
buggable AlexDaniel, There is 1 ticket tagged with DOCS; See fail.rakudo.party/t/DOCS for details
AlexDaniel maybe we should start assigning runnable snippets to tickets 00:43
j: my int $namcu; $namcu = 2**100 02:05
camelia Error while reading '/home/camelia/p6eval-token': No such file or directory at /home/camelia/rakudo-j-inst/bin/eval-client.pl line 10.
AlexDaniel aww…
ppppppppffffttt I'd much rather close a ticket next time than cause this horrible mess by merging it 02:13
RT should not attempt to keep everything in chronological order 02:14
or maybe I should simply use “Refers to” “Referred to by” 02:16
yea, that works better
Ha. This is one of the funniest tickets: RT #130092 04:20
synopsebot RT#130092 [new]: rt.perl.org/Ticket/Display.html?id=130092 [LTA] error message when misspelling “method” (meathod bar($test))
AlexDaniel buggable: tag TESTNEEDED 04:30
buggable AlexDaniel, There are 54 tickets tagged with TESTNEEDED; See fail.rakudo.party/t/TESTNEEDED for details
AlexDaniel 1.6k don't look so scary when you realize that a lot of them are already resolved and just need tests 04:40
buggable: tag MACRO 04:42
buggable AlexDaniel, There are no tickets tagged with MACRO
AlexDaniel m: say “\0” ~~ /<[\0]>/ 04:52
camelia ｢␀｣
AlexDaniel … did it just print a null byte?
gfldex yes
AlexDaniel RT #132205 04:56
synopsebot RT#132205 [new]: rt.perl.org/Ticket/Display.html?id=132205 .gist of a Match should not print null bytes, and maybe other control chars too (say “\0” ~~ /<[\0]>/)
[Tux] This is Rakudo version 2017.09-193-gfcbd8adbe built on MoarVM version 2017.09.1-62-g89ca8eb0 07:06
csv-ip5xs 1.201 - 1.217
test 10.654 - 10.777
test-t 3.278 - 3.579
csv-parser 12.948 - 13.217
lizmat Files=1226, Tests=75553, 331 wallclock secs (15.07 usr 5.35 sys + 2279.67 cusr 226.19 csys = 2526.28 CPU) 07:19
Geth roast: e5747f0484 | (Elizabeth Mattijsen)++ | S12-introspection/methods.t
Make tests less sensitive to autogenerated BUILDALL
 08:21
roast/6.c-errata: f50d6e43ef | (Elizabeth Mattijsen)++ | S12-introspection/methods.t
Make tests less sensitive to autogenerated BUILDALL
 08:26
roast: 603026d44a | (Elizabeth Mattijsen)++ | S12-introspection/methods.t
Make tests less sensitive to autogenerated BUILDALL

Oops, missed a few.		 08:28
rakudo/nom: 21788c8909 | (Elizabeth Mattijsen)++ | src/Perl6/World.nqp
Make all classes with empty BUILDALLPLAN share BUILDALL

  - so they don't need to go to Mu.BUILDALL and run the plan		 08:31
lizmat afk for a few hours& 08:48
jnthn morning, #perl6-dev o/ 09:30
Zoffix \o
jnthn decides not to get distracted by the SEGV this time 09:36
dogbert17 wonders what jnthn is up to 09:43
jnthn Debugging the cro Travis failure I meant to yesterday, before I was blocked by the Rakudo bug I ended up fixing :) 09:47
Geth rakudo/nom: e2f8a57dcc | (Elizabeth Mattijsen)++ | src/Perl6/Metamodel/BUILDPLAN.nqp
Make sure all empty BUILDPLANs are shared

  - this should save a bit of memory for each class created		 10:31
lizmat m: dd Date.^methods(:local).grep(*.name eq q/BUILDALL/)[0].signature # jnthn: the invocant type is not what I expected 11:05
camelia :(Rakudo::Internals::SupplySequencer:D $: ;; %init, *%_)
lizmat yet all works fine
fwiw, I think Rakudo::Internals::SupplySequencer is the *first* class that has a BUILDALL generated 11:06
Geth rakudo/nom: b58bd8fb1e | (Elizabeth Mattijsen)++ | src/Perl6/Metamodel/BUILDPLAN.nqp
Additional BUILDPLAN sharing

  - if a class has an empty BUILDPLAN, it will not add to the BUILDALLPLAN
  - therefore its BUILDALLPLAN is the same as its first parent in MRO
This should save memory for all mixins that don't add any attributes.		 11:09
timotimo bisectable6: my &s = 1 < *.elems < 3 11:11
jnthn lizmat: An empty one, or?
bisectable6 timotimo, Bisecting by output (old=2015.12 new=e2f8a57) because on both starting points the exit code is 1
timotimo, bisect log: gist.github.com/6b4cbeb27129c2825d...dd43c23494
timotimo, (2015-12-25) github.com/rakudo/rakudo/commit/07...dc61f84053
timotimo, The result looks a bit unrealistic, doesn't it? Most probably the output is different on every commit (e.g. ｢bisect: say rand｣)
lizmat jnthn: no, *not* empty
jnthn oh...so why does it get the type wrong in the signature?! 11:12
lizmat hmmm
timotimo ah good old times where our cuids were long
lizmat jnthn: re-investigating
jnthn: seems to only go wrong for builtins from the setting 11:15
if I precomp a module with an empty class, the BUILDALL sig is Any:D:, and if it's not empty, it's class:D: 11:16
timotimo jnthn: how does limiting to 32k sound? 11:17
jnthn timotimo: context?
tadzik . o O ( a 32kbps mp3 can't sound too good... )
yoleaux 27 Sep 2017 20:45Z <[Coke]> tadzik: if p6prove is busted with "replace this Array is copy logic"
timotimo oh, this was for moarvm
Geth rakudo/nom: 0dd6af716d | (Elizabeth Mattijsen)++ | src/Perl6/World.nqp
Set up auto-generated signature correctly

  - the first array parameter was set up correctly in declarations
  - but not in the signature, so introspection showed you the wrong thing		 11:44
AlexDaniel squashable6: status 11:45
squashable6 AlexDaniel, ⚠🍕 Next SQUASHathon in 2 days and ≈22 hours (2017-10-07 UTC-12⌁UTC+14). See github.com/rakudo/rakudo/wiki/Mont...Squash-Day
travis-ci Rakudo build passed. Elizabeth Mattijsen 'Make sure all empty BUILDPLANs are shared 12:06
travis-ci.org/rakudo/rakudo/builds/282668356 github.com/rakudo/rakudo/compare/2...f8a57dcc04
AlexDaniel buggable: tag TESTNEEDED 12:49
yoleaux 12:09Z <Zoffix> AlexDaniel: looks like squashathon PRs won't count towards Hacktoberfest for many people: irclog.perlgeek.de/perl6/2017-10-03#i_15250088
buggable AlexDaniel, There are 54 tickets tagged with TESTNEEDED; See fail.rakudo.party/t/TESTNEEDED for details
AlexDaniel awwwwww…
Geth rakudo/nom: f946bd35dc | (Elizabeth Mattijsen)++ | src/Perl6/Metamodel/ClassHOW.nqp
Don't bother with BUILDALL if BUILDPLAN is empty

  - because the next class with a BUILDALL in ^mro will do the right thing		 12:53
travis-ci Rakudo build passed. Elizabeth Mattijsen 'Additional BUILDPLAN sharing 12:56
travis-ci.org/rakudo/rakudo/builds/282680557 github.com/rakudo/rakudo/compare/e...8bd8fb1e2e
jnthn Apparently I'm better at building stuff that keeps running than I am at updating websites. The Cro web app behind its website ran for 24 days without a restart. :) 13:41
travis-ci Rakudo build passed. Elizabeth Mattijsen 'Set up auto-generated signature correctly 13:42
travis-ci.org/rakudo/rakudo/builds/282692144 github.com/rakudo/rakudo/compare/b...d6af716d7e
jnthn Now it's running 2017.09, so we'll see how that fares :)
travis-ci Rakudo build canceled. Elizabeth Mattijsen 'Should check both (sub)method tables for BUILDALL 13:44
travis-ci.org/rakudo/rakudo/builds/282223826 github.com/rakudo/rakudo/compare/d...13b857c015
buggable [travis build above] ✓ All failures are due to: missing build log (1 failure).
AlexDaniel Zoffix: review and ideas plz github.com/rakudo/rakudo/wiki/Raku...thon-Guide 15:12
Zoffix AlexDaniel: can regular people even comment on RT tickets they didn't create? 15:16
AlexDaniel aaaaaaaaaaahhhhhhhhhhhhh that has to be covered
yes, IIRC you just write an email and include [perl #132210] in the subject
synopsebot RT#132210 [new]: rt.perl.org/Ticket/Display.html?id=132210 Can't use `!` or `is required` with named Callable param constrainted by signature 15:17
AlexDaniel the comment will appear in a few minutes
or maybe it will be lost, this happens sometimes too
AlexDaniel facepalms on the harsh reality that we have imposed on us by RT
Zoffix Also, maybe 1st paragraph should say something about that people can just come to #perl6 and talk to other humans. ATM that page looks pretty intimidating 15:21
AlexDaniel Zoffix: I agree
Zoffix buggable: tag lhf
buggable Zoffix, There are 3 tickets tagged with LHF; See fail.rakudo.party/t/LHF for details
AlexDaniel wonder why he added ｢LHF｣ tag to RT #126669 15:22
synopsebot RT#126669 [new]: rt.perl.org/Ticket/Display.html?id=126669 [LHF][LTA] error with "need 6"/"use 6" (no "v")
Zoffix well, it is pretty easy to fix :) 15:23
Though the fix would block people from declaring `need`/`use` routines that take numbers, I guess 15:24
m: sub need (\v) { dd v }; need Int
camelia ===SORRY!===
Int is a builtin type, not an external module
Zoffix I guess it's already blocked :D
AlexDaniel /o\
m: sub need (\v) { dd v }; &need Int
camelia 5===SORRY!5=== Error while compiling <tmp>
Two terms in a row
at <tmp>:1
------> 3sub need (\v) { dd v }; &need7⏏5 Int
expecting any of:
infix
infix stopper
statement end
statement modifier
…
AlexDaniel m: sub need (\v) { dd v }; need(Int)
camelia Int
Zoffix m: sub need (\v) { dd v }; need 42 15:25
camelia 42
Zoffix this would get blocked
AlexDaniel right
Zoffix Basically, I imagine the fix would just be shoving a case that matches <[\d.]>+ here: github.com/rakudo/rakudo/blob/nom/....nqp#L1605 and complain about it 15:26
And here github.com/rakudo/rakudo/blob/nom/....nqp#L1547
m: need v6.c
camelia ( no output )
Zoffix m: need v6.e 15:27
camelia ( no output )
AlexDaniel Maybe you can fix by adding some logic to “undeclared routine” processing?
Zoffix Well, there's also this other bug
AlexDaniel it's a compile-time error, so it's ok to take some time to do an extra check or whatever
Zoffix That feels a bit spaghettified 15:28
AlexDaniel ah, but again I'm thinking that the error message can be better, not that it should work…
Zoffix m: use 6.000 15:29
camelia 5===SORRY!5=== Error while compiling <tmp>
Undeclared routine:
use used at line 1
AlexDaniel I was thinking about “did you mean use v6.000” or whatever, but now I realize that's not what the ticket is asking for
Zoffix What is it asking? I thought that's what it was asking 15:30
AlexDaniel oh… /o\ yes! 15:31
\o/
Zoffix m: need v6.d; say $*PERL.version 15:32
camelia v6.c
jnthn Not quite sure it makes sense with need... :)
AlexDaniel m: y 15:33
camelia 5===SORRY!5=== Error while compiling <tmp>
Undeclared routine:
y used at line 1 (in Perl 6 please use tr instead)
AlexDaniel m: qr
camelia 5===SORRY!5=== Error while compiling <tmp>
Unsupported use of qr for regex quoting; in Perl 6 please use rx//
at <tmp>:1
------> 3qr7⏏5<EOL>
Zoffix jnthn: yeah, I think maybe `<version>` should just be removed here: github.com/rakudo/rakudo/blob/f946....nqp#L1548 15:34
AlexDaniel so you just add use and need here: github.com/rakudo/rakudo/blob/f946...1025-L1031
or… maybe not
Zoffix Filed as LHF ticket: rt.perl.org/Ticket/Display.html?id=132214 15:35
AlexDaniel: looks like it yeah
AlexDaniel “maybe not” because you can't really get the argument there, but maybe we don't need that 15:38
buggable: tag META 15:41
buggable AlexDaniel, There is 1 ticket tagged with META; See fail.rakudo.party/t/META for details
AlexDaniel buggable: tag RT
buggable AlexDaniel, There is 1 ticket tagged with RT; See fail.rakudo.party/t/RT for details
Geth rakudo/nom: c91c40115e | (Jonathan Worthington)++ | src/core/ThreadPoolScheduler.pm
Eliminate dupe and slightly optimize cue

  * Only do a closure clone if we really need it
  * Use implicit return		 16:15
travis-ci Rakudo build failed. Jonathan Worthington 'Eliminate dupe and slightly optimize cue 17:41
travis-ci.org/rakudo/rakudo/builds/282806177 github.com/rakudo/rakudo/compare/f...1c40115e59
buggable [travis build above] ☠ Did not recognize some failures. Check results manually.
Zoffix 1 job on OSX: "Stage parse : This type (Submethod) does not support elems" 17:46
lizmat that is weird 17:53
Zoffix ZOFFLOP: t/spec/S11-modules/nested.t 17:58
ZOFVM: Files=1276, Tests=152519, 155 wallclock secs (21.68 usr 3.43 sys + 3297.98 cusr 246.05 csys = 3569.14 CPU)
Geth nqp: a6a1aa0b48 | (Zoffix Znet)++ | tools/build/MOAR_REVISION
Bump MoarVM
 17:59
¦ nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2...3-ga4fef0b
rakudo/nom: 2724a85144 | (Zoffix Znet)++ | tools/build/NQP_REVISION
Bump NQP
rakudo/nom: version bump brought these changes: github.com/perl6/nqp/compare/2017....0-ga6a1aa0
fa8fe84b2f | (Elizabeth Mattijsen)++ | src/core/Rakudo/Internals.pm
Zoffix m: say 553-62
camelia 491
Zoffix that's a lot of commits :) brrt++
travis-ci NQP build failed. Zoffix Znet 'Bump MoarVM' 18:12
travis-ci.org/perl6/nqp/builds/282847986 github.com/perl6/nqp/compare/289f7...a1aa0b4869
Zoffix One Linux job: t/nqp/059-nqpop.t ...................... 1/333 java.lang.ClassCastException: org.perl6.nqp.sixmodel.reprs.P6strInstance cannot be cast to java.lang.String 18:13
in ok (gen/jvm/stage2/NQPCORE.setting:1069)
lizmat pulls, builds and spectests 18:21
[TuxCM] This is Rakudo version 2017.09-200-g2724a8514 built on MoarVM version 2017.09.1-553-ga4fef0bd 18:22
csv-ip5xs 1.169 - 1.182
test 9.672 - 9.722
test-t 3.191 - 3.212
csv-parser 0.822 - 0.878
AlexDaniel evalable6: say “are you alive?” 18:28
evalable6 are you alive?
AlexDaniel hm. Bisectable also tracks moarvm repo so these 491 commits should be a lot of fun for the server :D 18:29
bloatable6: HEAD 18:30
bloatable6 AlexDaniel, No! It wasn't me! It was the one-armed man! Backtrace: gist.github.com/713193dc91917afb0d...4521a1bea0
AlexDaniel :S
Zoffix buggable: speed 20 :4 18:33
buggable Zoffix, ↑ ↑ dates: 2017-09-29–2017-10-03
Zoffix, ▆█ ▄ █ range: 3.154s–3.579s
Zoffix, ▃ ▃ ▁██▅█▃▇█▅ speed: 1% slower
Zoffix, ▄▆█▅█▅▆█████████▃▁▄▆
Zoffix buggable: speed 100 :2
buggable Zoffix, ▂▂▃▄▂ █▅ ▁ █▂ ▂↑▅↑▃ ▂▁ ▄ ▁ ▂▅▃▂↑↑▂▂ ▁ ▃ ▁▂ ▂ ▂↑▇▃ ▂ ▃ dates: 2017-09-01–2017-10-03
Zoffix, ██████▇▆██▇█▇█▇▇█▆▆▇██████████████▇█▆▆██████████▇▆▇█▆█▆▆█▇▆█▇▇▆▆▆▇▆▇██▇█▇▄████▄▃▂▂▃▂▃▂▂▃▅█▄▅▃▄█▄▁▁▂▂ range: 3.154s–4.237s; speed: 10% faster
Zoffix lookin' good \o/
AlexDaniel MasterDuke: a critical flaw was found in *able error messages: github.com/perl6/whateverable/issues/242 19:04
travis-ci Rakudo build passed. Zoffix Znet 'Bump NQP' 19:31
travis-ci.org/rakudo/rakudo/builds/282848093 github.com/rakudo/rakudo/compare/c...24a8514474
lizmat decommute& 20:02
gfldex what are the chances that Grammars benefit from future optimisations? 20:35
Geth nqp: 39666cb5b0 | usev6++ | src/vm/jvm/runtime/org/perl6/nqp/runtime/Ops.java
[jvm] Fix 1 ** NaN

Makes recently added test (commit 4cca702363) in t/nqp/059-nqpop.t pass.		 20:43
timotimo gfldex: good chances 20:45
do you know a bit about how the NFA stuff works?
AlexDaniel squashable6: next 20:47
squashable6 AlexDaniel, ⚠🍕 Next SQUASHathon in 2 days and ≈13 hours (2017-10-07 UTC-12⌁UTC+14). See github.com/rakudo/rakudo/wiki/Mont...Squash-Day
Geth roast: d95020aac0 | usev6++ | S32-num/power.t
Add test for 1 ** NaN giving 1

This was discussed in RT #129894 (Zoffix++) but not tested in roast.		 20:53
synopsebot RT#129894 [rejected]: rt.perl.org/Ticket/Display.html?id=129894 1**NaN is 1, but any numeric operation with NaN should be NaN (say 1**NaN)
travis-ci NQP build failed. usev6 '[jvm] Fix 1 ** NaN 20:56
travis-ci.org/perl6/nqp/builds/282919231 github.com/perl6/nqp/compare/a6a1a...666cb5b093
gfldex timotimo: i have no idea. Just asking because I'm preparing my next blog post. 20:57
timotimo ah, OK
one detail of a future optimization opportunity is that when you have multis or proto tokens a bit deeply nested, it currently will make a decision concerning many nested ones on the first layer, then call into the right next layer and do the same calculation again 20:58
that's probably the worst way to phrase it
imagine you have a "term" rule that can call into, among other things, a "number" rule that has a "decimal number" rule it calls etc
it'll figure out - in the very first step - that it'll have to call term -> number -> decimal number, but it'll immediately forget that after calling into the next step 20:59
and re-do the rest of the calculation
MasterDuke timotimo: where is that done? 21:01
timotimo search for nfa_run or something?
MasterDuke k 21:02
gfldex i'm toying with specialising a JSON grammar by subclassing from JSON::Tinys grammar. Works nicely but it comes with a speed hit of ~50%. 21:03
timotimo or alt_run
travis-ci Rakudo build failed. Elizabeth Mattijsen 'Make dynamic variable debugging simpler'
travis-ci.org/rakudo/rakudo/builds/282879726 github.com/rakudo/rakudo/compare/2...8fe84b2ff5
buggable [travis build above] ✓ All failures are due to: GitHub connectivity (1 failure).
Geth rakudo/nom: df5899d7df | (Elizabeth Mattijsen)++ | src/core/HyperSeq.pm
Promise is already stubbed at this time
 21:21
rakudo/nom: 98fae3d84e | (Elizabeth Mattijsen)++ | 3 files
Move up initialization of %*ENV and $*SCHEDULER

  - these were being requested at each startup before being initialized
  - saves creation of 2 X::Dynamic::NotFounds and 2x Failure
nqp: ec53b8b934 | usev6++ | 2 files
[jvm] Die with 'Division by zero' in div_i

Without this change a division by zero led to a Java exception: java.lang.ArithmeticException: / by zero
For some reasons that didn't work well with 'dies-ok'. As a result t/nqp/059-nqpop.t was blowing up on the jvm backend. With this patch 'make test' is clean, again. Also, the error message is now identical to the other two backends.		 22:07
travis-ci NQP build failed. usev6 '[jvm] Die with 'Division by zero' in div_i 22:11
travis-ci.org/perl6/nqp/builds/282953403 github.com/perl6/nqp/compare/39666...53b8b9348b
Geth nqp: 3b1356871e | usev6++ | 2 files
Revert "[jvm] Die with 'Division by zero' in div_i"

This reverts commit ec53b8b9348b12febd545c13b806e9dfc2822928.		 22:14
MasterDuke that was fast 22:15
bartolin sorry for the noise, looks like i've messed something up. need to take a closer look (but not today).
yeah, travis++
travis-ci NQP build failed. usev6 'Revert "[jvm] Die with 'Division by zero' in div_i" 22:30
travis-ci.org/perl6/nqp/builds/282955586 github.com/perl6/nqp/compare/ec53b...1356871ea7
MasterDuke ran a spectest about an hour ago at head, got flops in t/spec/S10-packages/basic.rakudo.moar and t/spec/S28-named-variables/init-instant.t 22:58
Zoffix init-instant is false alarm 23:05
Zoffix fixes
MasterDuke yeah guessed that was just a too tight tolerance (haven't looked at the test though) 23:06
Geth roast: a0afd2f582 | (Zoffix Znet)++ (committed using GitHub Web editor) | S28-named-variables/init-instant.t
Allow wider difference between $*INIT-INSTANT and INIT now
travis-ci Rakudo build passed. Elizabeth Mattijsen 'Move up initialization of %*ENV and $*SCHEDULER 23:11
travis-ci.org/rakudo/rakudo/builds/282935765 github.com/rakudo/rakudo/compare/f...fae3d84efe

Powered by Cro and the Raku® Programming Language.

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

Thank you!