ugexe use local CURIs as a faux metacpan microservice: gist.github.com/ugexe/10a1bf67f884...08e3577b2a 00:54
idea is to eventually act as both a web front-end to zef/cur as well as easily exposing it as a microservice to act as a darkpan 00:57
AlexDaniel- “35768 snippets tested, 10524 are known to be unstable” 02:18
wait what
no way I'm going through 10524 snippets
:S 02:20
so can we finally open rakudo/issues? I don't think I can take it anymore 02:36
.tell Zoffix Have you seen RT #132280 ? It looks like something you might enjoy working on (judging by the recent-ish optimizer ticket you resolved) 02:40
yoleaux AlexDaniel-: I'll pass your message to Zoffix.
synopsebot RT#132280 [new]: rt.perl.org/Ticket/Display.html?id=132280 [REGRESSION][PERF] Cool methods that take and pass a Capture are uber slower (.contains: 'meow' for ^10000_00)
AlexDaniel- .tell lizmat fwiw please feel freer to reject my latest regression tickets 02:44
yoleaux AlexDaniel-: I'll pass your message to lizmat.
AlexDaniel- this is probably one of the funniest snippets I've seen 02:53
e: %*ENV<RAKUDO_NO_DEPRECATIONS>:delete; DEPRECATED("the garbage bin")
evalable6 Saw 1 occurrence of deprecated code.
=====================================================…
AlexDaniel-, Full output: gist.github.com/98bf0ca4675e32ef98...239a1f4544
AlexDaniel- “Please use the garbage bin instead.”
.tell MasterDuke FWIW RT #132291 03:14
yoleaux AlexDaniel-: I'll pass your message to MasterDuke.
synopsebot RT#132291 [new]: rt.perl.org/Ticket/Display.html?id=132291 [LTA] Too few positionals passed to '' (for [:a] X [:b] -> ($i, $j) { })
AlexDaniel- “No regressions in stresstest from this, which is a promising sign.” github.com/rakudo/rakudo/commit/4a...eb98002608 03:26
how is it called when bisectable points at some commit? “unpromising sign”? :) 03:27
.tell fwiw RT #132292. IMO it's blocker-worthy, but not marking it as such just yet. What do you think? 03:44
yoleaux AlexDaniel-: I'll pass your message to fwiw.
synopsebot RT#132292 [new]: rt.perl.org/Ticket/Display.html?id=132292 [REGRESSION] Recursively .emit-ing from the tap of the same supply bails out
AlexDaniel- /o\
.tell jnthn fwiw RT #132292. IMO it's blocker-worthy, but not marking it as such just yet. What do you think?
yoleaux AlexDaniel-: I'll pass your message to jnthn.
AlexDaniel- .tell lizmat does this look alright? Let me know. c: 21b9a720c7^,21b9a720c7 dd Set.Mix 04:36
yoleaux AlexDaniel-: I'll pass your message to lizmat.
AlexDaniel- ouch 04:45
ah, nvm 04:46
.tell Zoffix fwiw RT #132293 06:08
yoleaux AlexDaniel-: I'll pass your message to Zoffix.
synopsebot RT#132293 [new]: rt.perl.org/Ticket/Display.html?id=132293 [REGRESSION] It is no longer possible to create native ints with .new (int32.new(42))
AlexDaniel- m: 0xD4DB.uniname.say 06:15
camelia Hangul Syllable PWILH
AlexDaniel- .tell samcv hey, is it correct that uninames of Hangul syllables have no <…> around them? c: 2017.08,HEAD 0xD4DB.uniname.say 06:16
yoleaux AlexDaniel-: I'll pass your message to samcv.
samcv AlexDaniel-, yes
yoleaux 06:16Z <AlexDaniel-> samcv: hey, is it correct that uninames of Hangul syllables have no <…> around them? c: 2017.08,HEAD 0xD4DB.uniname.say
AlexDaniel- OK, thanks
samcv the previous name was a placeholder name, and they're generated at database generation time
since it's algorithmic the name data 06:17
technically we could do that, but there aren't THAT many hangul characters. so i just added it easily. though if we wanted to save a tiny bit of space we could do the algorithmic aproach. but then i was concerned about then having to add extra logic to distinguish that codepoint range and would have been much more work 06:18
AlexDaniel- samcv: but I was thiking, don't we use <> for all autogenerated stuff that's in a way unofficial? 06:19
samcv no it's official 06:20
AlexDaniel- ah, okay
samcv the < > are for ones that have shared names
AlexDaniel- hmmmmm
samcv as in more than one cp with the same designation
AlexDaniel- m: say 0x10FFFF.uniname
camelia <noncharacter-10FFFF>
AlexDaniel- but that's unique now, isn't it?
samcv well yes
that's how unicode designates it with those numbers 06:21
but i meant before the hex digits-
the < > designates it doesn't have an assigned name
or at least we are still using the placeholder names in UnicodeData.txt. so like for hangul, even though it shows <Hangul Syllabl> there. they have always had names. just not listed in that file
AlexDaniel- hmm 06:22
samcv but the names are determined by essentially the same algorithm that's used to decompose hangul's
so we follow that spec for hangul
does that make some sense? 06:23
AlexDaniel- yes, thank you very much
samcv ok :)
AlexDaniel- samcv: any plans for other grant applications? :)
I'm just so happy with your work :) 06:24
samcv i need to decide what it will be though!
i am very happy with my work too :)
AlexDaniel- like, I'm going through tons of snippets that have different output, and not all of them regressions of course, most are actually bug fixes
and unicode related stuff comes up *very* often
something about looking at what we had previously and what we have now makes me think that another grant is a very good idea :) 06:25
uhhh 06:31
c: 79b8ab9d3f^,79b8ab9d3f say $*KERNEL.signal: SIGUSR1
committable6 AlexDaniel-, ¦79b8ab9d3f^: «10» ¦79b8ab9: «30»
AlexDaniel- c: 79b8ab9d3f^,79b8ab9d3f say +$*KERNEL.signal(SIGUSR1)
committable6 AlexDaniel-, ¦79b8ab9d3f^: «10» ¦79b8ab9: «30»
AlexDaniel- .tell Zoffix maybe you can also take a look at rt.perl.org/Ticket/Display.html?id...xn-1499238 06:35
yoleaux AlexDaniel-: I'll pass your message to Zoffix.
AlexDaniel- w… what 06:38
c: 2015.12 my int @x[-2**63]; say @x.shape 06:39
committable6 AlexDaniel-, ¦2015.12: «(-9223372036854775808)»
AlexDaniel- c: 2016.06 my int @x[-2**63]; say @x.shape
committable6 AlexDaniel-, ¦2016.06: «Illegal dimension in shape: -9223372036854775808. All dimensions must be integers bigger than 0␤ in block <unit> at /tmp/I1tG8GtUgl line 1␤ «exit code = 1»»
AlexDaniel- so it was fixed? but
c: HEAD my int @x[-2**63]; say @x.shape
committable6 AlexDaniel-, ¦HEAD(012c80f): «(-9223372036854775808)»
AlexDaniel- (屮ಠ益ಠ)屮
arrrrrrrrggggh 06:41
.tell lizmat fwiw RT #131966 06:43
yoleaux AlexDaniel-: I'll pass your message to lizmat.
synopsebot RT#131966 [new]: rt.perl.org/Ticket/Display.html?id=131966 [REGRESSION][9999] Shaped native arrays with negative sizes (my int @a[-9999999999999999])
AlexDaniel- c: c3b044df169^,c3b044df169 my $a = rx/aaaaaab/; 'a' x 999999 ~ 'b' ~~ /$a/; say now - INIT now 06:52
errr, what do I do with this
committable6 AlexDaniel-, ¦c3b044df169^: «3.29355215» ¦c3b044d: «4.84434404»
AlexDaniel- MasterDuke is maybe interested in INTERPOLATE stuff 06:53
hmm
.tell MasterDuke hey, maybe you can take a look at RT #132294 06:56
yoleaux AlexDaniel-: I'll pass your message to MasterDuke.
synopsebot RT#132294 [new]: rt.perl.org/Ticket/Display.html?id=132294 [REGRESSION][PERF] INTERPOLATE-related change made matching of large strings slower
AlexDaniel- mch: say (await (^10000).map({ start { $*THREAD.id } })).unique 06:59
committable6 AlexDaniel-, ¦2015.12: «(3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18)» ¦HEAD(012c80f): «(4)»
AlexDaniel- is it supposed to be this way? 07:01
it is! 07:02
mch: say (await (^10).map({ start { sleep 0.3; $*THREAD.id } })).unique
committable6 AlexDaniel-, ¦2015.12: «(3 4 5 6 7 8 9 10 11 12)» ¦HEAD(012c80f): «(4 6 7 8 9 10 11 12 13 14)»
AlexDaniel- ok nvm
.tell Zoffix maybe also RT #132295 although it is low priority stuff 07:54
yoleaux AlexDaniel-: I'll pass your message to Zoffix.
synopsebot RT#132295 [new]: rt.perl.org/Ticket/Display.html?id=132295 [REGRESSION][LTA] :2(…) radix syntax no longer suggests using .base (:2(42))
AlexDaniel- c: 2017.06,HEAD my $v = "\c[ZWJ]" x 100; say $v.chars 08:07
committable6 AlexDaniel-, ¦2017.06: «100» ¦HEAD(012c80f): «1»
AlexDaniel- samcv: what about this? ↑
is it correct that you can chain ZWJ with itself?
Geth_ rakudo/nom: 7 commits pushed by (Daniel Green)++, lizmat++ 09:03
timotimo ugexe: that microservice looks cool! 09:31
Zoffix . 10:00
yoleaux 02:40Z <AlexDaniel-> Zoffix: Have you seen RT #132280 ? It looks like something you might enjoy working on (judging by the recent-ish optimizer ticket you resolved)
06:08Z <AlexDaniel-> Zoffix: fwiw RT #132293
06:35Z <AlexDaniel-> Zoffix: maybe you can also take a look at rt.perl.org/Ticket/Display.html?id...xn-1499238
07:54Z <AlexDaniel-> Zoffix: maybe also RT #132295 although it is low priority stuff
synopsebot RT#132280 [new]: rt.perl.org/Ticket/Display.html?id=132280 [REGRESSION][PERF] Cool methods that take and pass a Capture are uber slower (.contains: 'meow' for ^10000_00)
RT#132293 [new]: rt.perl.org/Ticket/Display.html?id=132293 [REGRESSION] It is no longer possible to create native ints with .new (int32.new(42))
RT#132295 [new]: rt.perl.org/Ticket/Display.html?id=132295 [REGRESSION][LTA] :2(…) radix syntax no longer suggests using .base (:2(42))
Zoffix c: 2017.06 multi x (int) { say "native" }; multi x (Int) { say "here" }; x int32.new(42) 10:04
committable6 Zoffix, ¦2017.06: «here»
Zoffix AlexDaniel: looks like it wasn't possible before either
c: 2017.04 multi x (int) { say "native" }; multi x (Int) { say "here" }; x int32.new(42)
committable6 Zoffix, ¦2017.04: «here»
jnthn m: say int32.new(42).WHAT 10:14
yoleaux 03:44Z <AlexDaniel-> jnthn: fwiw RT #132292. IMO it's blocker-worthy, but not marking it as such just yet. What do you think?
camelia No such method 'BUILDALLPLAN' for invocant of type 'Perl6::Metamodel::NativeHOW'
in block <unit> at <tmp> line 1
synopsebot RT#132292 [new]: rt.perl.org/Ticket/Display.html?id=132292 [REGRESSION] Recursively .emit-ing from the tap of the same supply bails out
jnthn AlexDaniel: Actually I'd count that as a fix. 10:15
You're not meant to be allowed to be in the closure passed to tap more than once at a time 10:16
This could accidentally happen 10:17
Due to the implmentation detail of us using a recursive mutex
Zoffix m: use nqp; say nqp::box_i(42, int32.WHAT) 10:20
camelia This representation (P6int) cannot unbox to other types (for type int32)
in block <unit> at <tmp> line 1
jnthn int32 is not meant to be something you can box 10:21
It's a native type
Zoffix Yeah. Was just trying to figure out where the error's coming from
jnthn Calling .new on it is very dubious also
Though since type objects are objects, you can do it
Zoffix jnthn: so should int32.new(42) return a native int32 42? 10:22
jnthn No 10:23
IMO it should blow up
Though likely with a better error than today
"Cannot make an instance of a native type" would do it
Zoffix ok
jnthn As to how the error happens: the type objects of native types claim to be subtypes of, say, Int 10:25
This is only true in a type checking sense, not in a representation sense
It's why int32 comes out narrower than Int when doing the multi sort though
[TuxCM] This is Rakudo version 2017.09-404-gd56151743 built on MoarVM version 2017.09.1-594-gb9d3f6da
csv-ip5xs 1.175 - 1.184
test 13.997 - 14.336
test-t 3.071 - 3.071
csv-parser 0.854 - 0.863
Zoffix :o so close to sub-3 10:26
[TuxCM] how's that for a small range :)
timotimo might want to deprecate method new in native ints
so that existing code will just take a performance hit at first
rather than straight-up dying
then kill it with fire in a later release
Zoffix timotimo: but it never worked 10:27
timotimo it gave you an Int
Zoffix ok
timotimo it probably didn't truncate to the right size though?
Zoffix timotimo: it was the same as Int.new 10:28
timotimo right
Zoffix m: use MONKEY; augment class Int { proto method new2(|) {*}; multi method new2(int \SELF: $) { say "here" }; multi method new2(Int \SELF: $) { say "there" } }; int32.new2(42)
camelia there
Zoffix Don't even know how to make the error better anyway :}
m: use MONKEY; augment class Int { proto method new2(|) {*}; multi method new2(int32 \SELF: $) { say "here" }; multi method new2(Int \SELF: $) { SELF.^name.say; say "there" } }; int32.new2(42) 10:30
camelia int32
there
Zoffix k, tht works
"works".... trying to figure out if it's a native by name will make it so slow, it'd be no point in using Int.new: native int (if people are even using it) 10:33
releasable6: status 10:34
releasable6 Zoffix, Next release in 7 days and ≈8 hours. 1 blocker. Changelog for this release was not started yet
Zoffix, Details: gist.github.com/8a5dc4a8820d6d5b16...7c56ce871f
Zoffix Gonna work on this later.
timotimo you can figure out if self.HOW is NativeHOW
Zoffix Ah, thanks. 10:35
jdv79 Zoffix: do you sleep? 10:39
Zoffix What do I need to look up to learn how to make stuff like we have for dalek, where it shows up as a service, like `sudo service dalek restart` ? 10:40
jdv79: robots don't sleep
I wanna make a geth service
on hack
timotimo yeah, that's "system file" stuff from systemd 10:42
jnthn Zoffix: fwiw, probably the best way to do it is to make NativeHOW do MethodContainer and accounting for that in its find_method/publish_method_cache. Then we can just write "cannot make an instance of this" methods in the type object 10:43
Zoffix ok 10:44
jnthn A moderately easy way to earn the MOP Hacker badge :-) 10:45
Zoffix \o/ 10:46
jnthn (If you didn't already :))
Geth_ geth: 5bb36864d3 | (Zoffix Znet)++ | lib/Geth/Plugin/GitHub.pm6
Handle missing chans gracefuler

Any.split still got a proto for split to the "safe" call fails there and still tries to call Any.split triggering no-multi		 11:18
geth: c362e66b4b | (Zoffix Znet)++ | lib/Geth/GitHub/Hooks.pm6
Update to handle latest HTTP::Server::Tiny
 11:19
geth: 5536d31b03 | (Zoffix Znet)++ | util-dev/emulate-github-event.p6
Add a dev util

For simulating a GitHub push event hook arriving to our server
geth: bb70c2cde0 | (Zoffix Znet)++ | META6.json
Remove bogus `test-depends`
 11:21
geth: 5615eae1f1 | (Zoffix Znet)++ | META6.json
Ensure prereqs match our use
 11:25
geth: 8ca11367aa | (Zoffix Znet)++ | META6.json
Use semver
Geth__ geth: 8ca11367aa | (Zoffix Znet)++ | META6.json
Use semver
 11:49
specs: tbrowder++ created pull request #123:
changes for Texas => ASCII		 11:55
specs: 1c979ab5a5 | (Tom Browder)++ | 2 files
changes for Texas => ASCII
specs: f52230c29b | (Zoffix Znet)++ (committed using GitHub Web editor) | 2 files
Merge pull request #123 from tbrowder/texas

changes for Texas => ASCII
roast: 0bdfb1b157 | usev6++ | packages/Test/Util.pm
Use longer timeout for doesn't-hang on JVM
 11:56
geth: 13becdb8c0 | (Zoffix Znet)++ | config.json
Hang in #perl6-toolchain too
 11:58
MasterDuke c: c3b044df169^,c3b044df169,HEAD my $a = rx/aaaaaab/; 'a' x 999999 ~ 'b' ~~ /$a/; say now - INIT now 12:02
yoleaux 03:14Z <AlexDaniel-> MasterDuke: FWIW RT #132291
06:56Z <AlexDaniel-> MasterDuke: hey, maybe you can take a look at RT #132294
synopsebot RT#132291 [new]: rt.perl.org/Ticket/Display.html?id=132291 [LTA] Too few positionals passed to '' (for [:a] X [:b] -> ($i, $j) { })
RT#132294 [new]: rt.perl.org/Ticket/Display.html?id=132294 [REGRESSION][PERF] INTERPOLATE-related change made matching of large strings slower
committable6 MasterDuke, ¦c3b044df169^: «3.42546156» ¦c3b044d: «5.072197» ¦HEAD(d561517): ««timed out after 10 seconds» «exit signal = SIGHUP (1)»» 12:03
MasterDuke hm, that's unfortunate, not what my PR was supposed to do 12:04
Geth geth: 13becdb8c0 | (Zoffix Znet)++ | config.json
Hang in #perl6-toolchain too
 12:06
Zoffix .tell nine I moved geth to hack. Something was wrong with the geth box. GitHub couldn't connect and I couldn't telnet to geth port's either (said connection refused). I tried restarting geth several times and connecting using localhost to geth's listener worked, but not from outside network. So I stopped the geth service on geth box (but did not remove it entirely) and set up a geth service on hack and 12:09
yoleaux Zoffix: I'll pass your message to nine.
Zoffix moved webhooks of all of our repo's to hack's geth.
.tell nine moved webhooks of all of our repo's to hack's geth.
yoleaux Zoffix: I'll pass your message to nine.
Zoffix Can someone with keys to rakudo/star change the geth web hook to geth.niner change the URL to hack.p6c.org:8888/?chan=#perl6-dev (or give me access and I'll set it up myself) 12:10
PSA: Geth now lives on hack, in /home/geth/geth There's a service setup, so if a restart of it is needed, you'd just do `sudo service geth restart` 12:12
jnthn Zoffix: Given you access 12:13
Zoffix \o/ 12:14
MasterDuke c: c3b044df169^,c3b044df169,HEAD my $a = "aaaaaab"; 'a' x 999999 ~ 'b' ~~ /$a/; say now - INIT now 12:33
committable6 MasterDuke, ¦c3b044df169^: «5.4635297» ¦c3b044d: «7.4579997» ¦HEAD(d561517): «4.9643206» 12:34
MasterDuke right, that's the case my recent PR made faster
goodness, the slow rx version spends 3s in the proto for Match's Bool and then another 1.5s in the multi it chose 12:41
2999980 calls to both 12:42
Zoffix I'm having a hard time understanding where Perl6::Metamodel::MethodContainer fits into the picture of disabling int32.new... If I make NativeHOW do it, seems all the methods need to overridden. The method_table and submethod_table already exist in NetiveHOW without that role (they return an empty nqp::hash). 12:44
timotimo Zoffix: then those can probably be thrown out 12:46
Zoffix They fix a bug, I think: github.com/rakudo/rakudo/commit/9a...bc8dbe31e0 12:47
Geth rakudo: MasterDuke17++ created pull request #1193:
Whitespace fix, convert tabs to spaces		 12:48
rakudo/nom: a0279e561a | MasterDuke17++ (committed using GitHub Web editor) | src/core/Match.pm
Whitespace fix, convert tabs to spaces
 12:49
rakudo/nom: 5092d6e60e | (Zoffix Znet)++ (committed using GitHub Web editor) | src/core/Match.pm
Merge pull request #1193 from MasterDuke17/patch-1

Whitespace fix, convert tabs to spaces
rakudo: MasterDuke17++ created pull request #1194:
Rest of tabs -> spaces in Match.pm		 12:54
rakudo/nom: 0262a997aa | MasterDuke17++ (committed using GitHub Web editor) | src/core/Match.pm
Rest of tabs -> spaces in Match.pm
 12:55
rakudo/nom: 0d217357aa | (Zoffix Znet)++ (committed using GitHub Web editor) | src/core/Match.pm
Merge pull request #1194 from MasterDuke17/patch-1

Rest of tabs -> spaces in Match.pm
Zoffix greppable6, \b(int32|int|int64|int16|int8)\.new 13:01
greppable6 Zoffix, Found nothing!
Zoffix timotimo: do we really need a deprecation for it?
It's be a deprecation for a feature so rarely used that it's not used in the ecosystem and one that never worked right
c: 2017.06 say int8.new: 999999999999999999999999999999999999999999999999999999999999999999999999999999999999 13:02
committable6 Zoffix, ¦2017.06: «999999999999999999999999999999999999999999999999999999999999999999999999999999999999»
Zoffix Oh damn: github.com/nkh/P6-Data-Dump-Tree/b...Call.t#L45 13:03
nqp: nqp::getstaticcode(sub ( | ) { nqp::die('Cannot instantiate native type') }) 13:05
camelia Unable to parse expression in routine_def; couldn't find final ')' at line 2, near "| ) { nqp:"
at gen/moar/stage2/NQPHLL.nqp:707 (/home/camelia/rakudo-m-inst-1/share/nqp/lib/NQPHLL.moarvm:panic)
from gen/moar/stage2/NQPHLL.nqp:714 (/home/camelia/raku…
Zoffix any way to spec "take any args" in nqp?
More of them: github.com/wentam/P6-Graffiks/blob...ra.pm6#L86 13:06
So no use of ints but nums quite a few :(
moritz nqp: sub f(*@a) { nqp::say(@a) }; f(1, 2) 13:07
camelia cannot stringify this
at <tmp>:1 (<ephemeral file>:f)
from <tmp>:1 (<ephemeral file>:<mainline>)
from gen/moar/stage2/NQPHLL.nqp:1542 (/home/camelia/rakudo-m-inst-1/share/nqp/lib/NQPHLL.moarvm:eval)
from gen/moar/stage2/NQPHLL.nqp:1779 (/home/c…
moritz nqp: sub f(*@a) { nqp::say(nqp::join(', ', @a)) }; f(1, 2)
camelia This representation (P6int) cannot unbox to a native string (for type BOOTInt)
at <tmp>:1 (<ephemeral file>:f)
from <tmp>:1 (<ephemeral file>:<mainline>)
from gen/moar/stage2/NQPHLL.nqp:1542 (/home/camelia/rakudo-m-inst-1/share/nqp/lib/NQPHLL.moarv…
moritz nqp: sub f(*@a) { nqp::say(nqp::join(', ', @a)) }; f("a", "b")
camelia a, b
Zoffix nqp: sub f(*@a) { nqp::say(nqp::join(', ', @a)) }; f("a", "b", :foo(42)) 13:08
camelia Unexpected named argument 'foo' passed
at <tmp>:1 (<ephemeral file>:f)
from <tmp>:1 (<ephemeral file>:<mainline>)
from gen/moar/stage2/NQPHLL.nqp:1542 (/home/camelia/rakudo-m-inst-1/share/nqp/lib/NQPHLL.moarvm:eval)
from gen/moar/stage2/NQPHLL.n…
Zoffix nqp: sub f(*@a, *%_) { nqp::say(nqp::join(', ', @a)) }; f("a", "b", :foo(42))
camelia a, b
Zoffix ok thanks
Well, I done this and it works: gist.github.com/zoffixznet/19825d7...b33f0c3b01 13:13
make method_table return an nqp::hash that got a blowy method new in it (in the past it was just an empty nqp::hash) 13:14
.in 1d irclog.perlgeek.de/perl6-dev/2017-...i_15301581 13:19
yoleaux Zoffix: I'll remind you on 15 Oct 2017 13:19Z
Zoffix I've no idea how to do deprecation from there 13:21
from NativeHow
What does getstaticcode do? Like what's a "static code"? 13:22
moritz when you have a closure, you have a pointer to a thing that holds the captured lexpad and the "actual" code 13:24
that's what "static code" - code reference that's not a closure
Zoffix Ah 13:25
m: use nqp; say nqp::getlexdyn('&DEPRECATED') 13:31
camelia (signal SEGV)
Zoffix :)
Filed as rt.perl.org/Ticket/Display.html?id=132300 13:33
MasterDuke what causes lots of time spent in a proto? in my experiments with INTERPOLATE, it was because the find_best_dispatchee wasn't able to cache the right method. but in the profile i'm looking at now, find_best_dispatchee doesn't even appear 13:39
github.com/rakudo/rakudo/blob/nom/...#L772-L774 the proto on line 772, and it always calls the multi on line 774 13:44
buggable New CPAN upload: Config-Parser-toml-1.0.1.tar.gz by TYIL cpan.metacpan.org/authors/id/T/TY/...0.1.tar.gz 13:50
New CPAN upload: Config-Parser-yaml-1.0.1.tar.gz by TYIL cpan.metacpan.org/authors/id/T/TY/...0.1.tar.gz
Zoffix Weird. `DEPRECATED "foo"; DEPRECATED "bar"` just shows 1 deprecation :/ 14:14
Is it possible to add some sort of annotations to a method? Roughly thinking about how Cool methods that pass captures could be fixed up to be faster and I thought maybe we could stick some annotation to the relevant methods in Cool and then the optimizer could just rewrite the calls to Cool.meow(<some args>) to Cool.Stringy.meow(<some args>) 14:29
Geth rakudo/nom: 9d9c7f9c3b | (Zoffix Znet)++ | src/Perl6/Metamodel/NativeHOW.nqp
Deprecate .new on native types

  - For now, just maintain old behaviour of returning a boxed version
  - After 2017.12, toss the deprecation and make the methods die
  - Fixes RT#132293: rt.perl.org/Ticket/Display.html?id=132293		 14:43
synopsebot RT#132293 [open]: rt.perl.org/Ticket/Display.html?id=132293 [REGRESSION] It is no longer possible to create native ints with .new (int32.new(42))
Geth rakudo/nom: cc6c05586b | (Zoffix Znet)++ | t/05-messages/01-errors.t
Test behaviour of .new on native types

RT#132293: rt.perl.org/Ticket/Display.html?id=132293 Rakudo fix: github.com/rakudo/rakudo/commit/9d9c7f9c3b		 14:44
Zoffix .in 70d take care of tossing deprecation in github.com/rakudo/rakudo/commit/9d9c7f9c3b and github.com/rakudo/rakudo/commit/cc6c05586b , bruh 14:49
yoleaux Zoffix: I'll remind you on 23 Dec 2017 14:49Z
Geth roast: 4043a63b60 | (Zoffix Znet)++ | packages/Test/Util.pm
Revert "Use longer timeout for doesn't-hang on JVM"

This reverts commit 0bdfb1b157b5286480d8355cf62ce1a76fd445ea.
Will rewrite a better version in next commit		 15:18
roast: bdf8904da6 | (Zoffix Znet)++ | 2 files
Rescale ROAST_TIMING_SCALE for JVM

  'cause it takes longer to fire up procs
Zoffix (this version will scale doesn't-hangs with custom wait times too; not just the default times) 15:19
I think the most annoying thing about "ANNOYING" ticket tags is the tag, not the bug the ticket describes 15:22
Like wtf is RT#132012 an "annoying" ticket? 15:23
synopsebot RT#132012 [new]: rt.perl.org/Ticket/Display.html?id=132012 [ANNOYING] Numeric values of signals are wrong (say +SIGUSR1)
Zoffix .tell AlexDaniel [ANNOYING] ticket tag is annoying
yoleaux Zoffix: I'll pass your message to AlexDaniel.
Zoffix "Show message only if value has method .base" "if $.num.perl.^can('base')" why yes, .perl it first :P 15:49
ZofBot: good job!
ZofBot Zoffix, weigh my friend's affection with mine own
dogbert17 ZOFFLOP t/spec/S02-types/baghash.t 16:07
'Cannot unbox 64 bit wide bigint into native integer' 16:08
Zoffix dogbert17: is it actually a flop? 16:12
As in, are you sure it fails only some times and not all the time?
Also, do you have latest and greatest rakudo? 16:13
And if yes, what's the full output
Geth rakudo/nom: 441fc9b32b | (Zoffix Znet)++ | t/05-messages/01-errors.t
Do not add more stuff into 01-errors.t

It takes forever to run and there's no reason we can't split it up into multiple files.		 16:14
rakudo/nom: de2b9ff720 | (Zoffix Znet)++ | 2 files
Remove unwanted .perl call

Fixes RT#132295: rt.perl.org/Ticket/Display.html?id=132295		 16:28
synopsebot RT#132295 [new]: rt.perl.org/Ticket/Display.html?id=132295 [REGRESSION][LTA] :2(…) radix syntax no longer suggests using .base (:2(42))
ugexe m: package Foo { our &foo := -> $a { say $a }; }; Foo::foo(1); # should this work? if not, does it differ from (the working) `our sub ...`? 16:38
camelia Cannot invoke this object (REPR: Uninstantiable; Callable)
in block <unit> at <tmp> line 1
Zoffix Seems to work if you don't use binding: 16:40
m: package Foo { our &foo = -> $a { say $a }; }; Foo::foo(1);
camelia 1
Zoffix No idea if it should work with it as well
Geth rakudo: MasterDuke17++ created pull request #1195:
Optimize Regex multi for INTERPOLATE		 16:41
ugexe cool
timotimo i thought "zofflop" is just "flops on zoffix' system", did dogbert root zoffix' box? :) 16:47
Zoffix no, zofflop is "flop" with a prefix that makes it easier to search for in the log 16:48
This looks bogus: github.com/rakudo/rakudo/commit/ec...bce05122f9
m: my $x = "abc"; $x ~~ tr|abc|{$/.uc}|; say $x 16:51
camelia {$/
Zoffix Just gonna toss all the $/ fumbling in Str.trans and see if any tests explode
(it's also not done for Cool.trans, which is how I noticed it)
travis-ci Rakudo build failed. Zoffix Znet 'Do not add more stuff into 01-errors.t 17:01
travis-ci.org/rakudo/rakudo/builds/287962876 github.com/rakudo/rakudo/compare/c...1fc9b32b70
buggable [travis build above] ☠ Did not recognize some failures. Check results manually.
Zoffix why do travis reports now have these slanted lines across them :? 17:03
One Linux job segfaulted during make Segmentation fault (core dumped)
Tossing everything makes this test fail: 17:06
m: say '"foo & bar"'.trans(/(' '+)/ => {' ' ~ ('&nbsp;' x ($0.chars - 1))}, /(\W)/ => {"&#{ord($0)};"})
camelia &#34;foo &nbsp;&#38; &nbsp;&nbsp;bar&#34;
Zoffix m: say '"foo & bar"'.trans(/(' '+)/ => {' ' ~ ('&nbsp;' x ($0.chars - 1))}, /(\W)/ => {"&#{ord($0)};"}); say $/ 17:07
camelia &#34;foo &nbsp;&#38; &nbsp;&nbsp;bar&#34;
Nil
Zoffix But then, I still don't get why it's grabbing caller's $/
Gonna just file it for now
Filed as rt.perl.org/Ticket/Display.html?id=132302 17:15
Geth roast: c4f9ac0bbd | usev6++ | 8 files
[jvm] Skip hanging tests
 17:20
rakudo/nom: c1e5cfcfa5 | usev6++ | t/spectest.data
[jvm] Don't run files when all tests are hanging
 17:22
Zoffix I don't get how this works at all. It's missing a decont: github.com/rakudo/rakudo/blob/nom/...tr.pm#L150 17:37
m: use nqp; class Foo { method x ($suffix) { say nqp::getattr($suffix,Str,'$!value') } }.x: "x"
camelia P6opaque: no such attribute '$!value' in type Str when trying to get a value
in method x at <tmp> line 1
in block <unit> at <tmp> line 1
Zoffix m: my $x = "x"; say "x".ends-with: $x
camelia True
Zoffix yet it doesn't explode. Huh?
MasterDuke and nothing changes if you add the decont? 17:39
Zoffix Changes how?
It's now behaving as if the decont was already there 17:40
MasterDuke .ends-with still produces the same results?
Zoffix Haven't tried. I'd expect it to. Don't really wanna pay the compilation price 17:41
dogbert17 Zoffix: it's not a flapper, it fails every time, I suspect it's 32 bit related, it's the very last test which causes the problem
t/spec/S02-types/baghash.rakudo.moar (Wstat: 65280 Tests: 292 Failed: 0)
Non-zero exit status: 255
Parse errors: Bad plan. You planned 293 tests but ran 292.
Zoffix m: use nqp; class Foo { method x (Str:D $suffix) { say nqp::p6bool(nqp::getattr($suffix,Str,'$!value')) } }.x: "x"
camelia True
Zoffix m: use nqp; class Foo { method x (Str $suffix) { say nqp::p6bool(nqp::getattr($suffix,Str,'$!value')) } }.x: "x"
camelia True
Zoffix m: use nqp; class Foo { method x (Str:D $suffix) { say nqp::p6bool(nqp::getattr($suffix,Str,'$!value')) } }.x: "x"
camelia True
Zoffix k, looks like Str type constraint does some sort of decont voodoo 17:42
dogbert17 m: use Test; is-deeply ("foo" => 10000000000000000000).BagHash.grab(1), ("foo",),"can .grab() a BagHash key with weight larger than 64 bits"
camelia ok 1 - can .grab() a BagHash key with weight larger than 64 bits
Zoffix oh wait
Yeah
m: use nqp; class Foo { method x ($suffix) { say nqp::p6bool(nqp::getattr($suffix,Str,'$!value')) } }.x: "x"
camelia P6opaque: no such attribute '$!value' in type Str when trying to get a value
in method x at <tmp> line 1
in block <unit> at <tmp> line 1
dogbert17 this fails for me with
Cannot unbox 64 bit wide bigint into native integer
in sub _is_deeply at /home/dogbert/repos/rakudo/install/share/perl6/sources/C712FE6969F786C9380D643DF17E85D06868219E (Test) line 652
in sub is-deeply at /home/dogbert/repos/rakudo/install/share/perl6/sources/C712FE6969F786C9380D643DF17E85D06868219E (Test) line 598
Zoffix dogbert17: well, is your rakudo a HEAD? It was fixed recently
dogbert17 2017.09-410-gcc6c05586 17:43
Zoffix m: say $PERL.compiler.version
camelia 5===SORRY!5=== Error while compiling <tmp>
Variable '$PERL' is not declared. Did you mean 'Perl'?
at <tmp>:1
------> 3say 7⏏5$PERL.compiler.version
Zoffix m: say $*PERL.compiler.version
camelia v2017.09.413.gc.1.e.5.cfcfa
Zoffix dogbert17: and that's on 32-bit box?
dogbert17 yes
dogbert17 pull latest Rakudo
Zoffix 410 would have had the fix already
dogbert17: you could reopen the ticket and say it still fails on 32-bit: rt.perl.org/Ticket/Display.html?id=132279 17:44
dogbert17 parsing ...
Zoffix and include --ll-exception of perl6 -e '("foo" => 10000000000000000000).BagHash.grab(1)' 17:45
dogbert17 false alarm, it works with Rakudo head. Sorry for the noise 17:49
Zoffix oh, *phew* 17:50
What do the cool kids use to measure something too close to noise? Like which one's faster: method x($x) { $x.uc } or method x(\x) { x.uc } 17:54
MasterDuke callgrind will give you the actual instruction count
Zoffix I get from 4.6x faster to 2x slower with Benchy -_-
sudo aptitude search callgrind... nothing found 17:55
MasterDuke it's part of valgrind
valgrind --tool=callgrind <...>
Zoffix ok
MasterDuke and then kcachgrind is a nice viewer for the results if you want more than just the instruction count 17:56
Zoffix doesn't seem to print anything for me. Creates an empty file :/
dogbert17 how did you run it? 17:57
Zoffix dogbert17: valgrind --tool=callgrind ./perl6 foo.p6 17:58
MasterDuke oh, you might need to call the moar executable directly, not the perl6 shell script
Zoffix Oh
sounds hard
Zoffix puts it on the back burner for now
MasterDuke eh, i just cat perl6 and copy it onto the command line 17:59
dogbert17 make a copy of perl6-valgrind-m and stick in the '--tool=callgrind' there
MasterDuke and i also have some bash aliases that do that for me
Zoffix ok
yoleaux Zoffix: see if nqp::eqaddr deconting stuff needs to be done anywhere in core 18:00
Zoffix MVM_VALGRIND_OPTS=--tool=callgrind ./perl6-valgrind-m foo.p6 works too
yoleaux: meh
MasterDuke heh, good timing past Zoffix
dogbert17 Zoffix: I guess you know that the generated files are, in essence, unreadable. KCachegrind, like MasterDuke wrote or callgrind_annotate if you want to stay in the shell 18:04
Zoffix I opened cachegrind, but I don't know what to look for. The numbers fluctuate on each run of the script 18:05
There are 19 fewer `optimize_decont` calls and that's all I can interpret. I guess you need to know what to look for 18:07
dogbert17 I belive that the recommended cmdline should be something like 'MVM_SPESH_BLOCKING=1 valgrind --tool=callgrind ./perl6 foo.p6' the fluctuations might go down
yoleaux Zoffix: fix (my @a is default(42) = <a b c>)[1]:delete; @a.Slip.List.say, bruh
Zoffix yoleaux: dude, leave me be :( 18:08
dogbert17 correction 'MVM_SPESH_BLOCKING=1 MVM_VALGRIND_OPTS=--tool=callgrind ./perl6-valgrind-m foo.p6'
Zoffix I need a better reminder bot 18:09
I was gonna write all the tests today... Still haven't started :/ 18:11
Geth rakudo/nom: 04b171bd12 | (Daniel Green)++ | src/core/Match.pm
Optimize Regex multi for INTERPOLATE

Move some code into a conditional and don't allocate some variables if not needed.
rakudo/nom: 0bd2b27df5 | (Aleks-Daniel Jakimenko-Aleksejev)++ (committed using GitHub Web editor) | src/core/Match.pm
Merge pull request #1195 from MasterDuke17/optimize_Regex_multi_for_INTERPOLATE

Optimize Regex multi for INTERPOLATE
Zoffix dogbert17: still fluctuating enough to wash the difference 18:13
dogbert17 :(
Zoffix Maybe there ain't any. Just gonna leave it./
AlexDaniel- . 18:20
AlexDaniel .
yoleaux 15:23Z <Zoffix> AlexDaniel: [ANNOYING] ticket tag is annoying
AlexDaniel- Zoffix: RT #132012 is annoying for the same reason described in rt.perl.org/Ticket/Display.html?id...xn-1496218 18:21
synopsebot RT#132012 [open]: rt.perl.org/Ticket/Display.html?id=132012 [ANNOYING] Numeric values of signals are wrong (say +SIGUSR1)
AlexDaniel- some signals have correct values and you don't notice it right away, and there's practically no workaround
ok you can hardcode the values for a particular system, but uhhh…
that's an annoying workaround :) 18:22
Zoffix Seems a pointless tag and it certainly makes me want to fix those less.
" 18:23
"» is no longer shuffled (<a b c d>».say)" why is this annoying?
This tag is nonsense
AlexDaniel- because we keep explaining to people “hey actually it doesn't work and it just happens that these are returned in the same order right now” 18:24
timotimo Zoffix: if you want less noisy results from callgrind, set the spesh blocking env var 18:25
AlexDaniel- the tag differentiates painful issues with other bugs that are mostly unnoticeable (or, when noticed, are easily workaroundable)
MasterDuke i think you could replace annoying with lta 18:26
AlexDaniel- it's other way round
LTA = it's ok but could be better
Zoffix timotimo: I did. Still too noisy 18:27
AlexDaniel- ANNOYING = something that keeps getting in your way and makes rakudo look like garbage
timotimo damn
Zoffix AlexDaniel-: the tag differentiates issues I'll personally gonna be avoiding. It's like a bug reporter who's shouting at how pissed off they are about a bug. You just phase them out and go fix something nicer. 18:28
AlexDaniel- Zoffix: I'm sorry that it makes you feel that way
Zoffix: maybe if you suggest some other name for that purpose, I can retag those 18:29
Zoffix AlexDaniel-: anything that doesn't require you to explain why we're using such a tag. So far you had to explain "ANNOYING" to at least two people. 18:32
AlexDaniel- it's ok for a new tag
I had to explain what a REGRESSION is a couple of times also…
and it's ok for people to ask why some particular issue is annoying (maybe they don't see the actual issue, it's ok to ask for clarification) 18:33
MasterDuke m: class A { method x($x) { $x.uc } }; my $a; $a = A.x("aBc") for ^10_000_000; say $a; say now - INIT now 18:36
camelia ABC
5.4048479
MasterDuke m: class A { method x(\x) { x.uc } }; my $a; $a = A.x("aBc") for ^10_000_000; say $a; say now - INIT now
camelia ABC
3.85693285
MasterDuke Zoffix: ^^^ doesn't look like noise? or was that not what you were actually testing? 18:37
Zoffix MasterDuke: you just ran it twice though. 18:39
MasterDuke timotimo: do you know anything about protos?
Zoffix m: class A { method x(\x) { x.uc } }; my $a; $a = A.x("aBc") for ^10_000_000; say $a; say now - INIT now
camelia ABC
3.40128488
Zoffix m: class A { method x(\x) { x.uc } }; my $a; $a = A.x("aBc") for ^10_000_000; say $a; say now - INIT now
camelia ABC
3.7919970
Zoffix m: class A { method x($x) { $x.uc } }; my $a; $a = A.x("aBc") for ^10_000_000; say $a; say now - INIT now
MasterDuke i ran it a bunch of times locally
Zoffix m: class A { method x($x) { $x.uc } }; my $a; $a = A.x("aBc") for ^10_000_000; say $a; say now - INIT now
camelia ABC
5.3459838
ABC
5.359951
Zoffix MasterDuke: ah. I ran them both in one program, so I guess that's what wiped the difference in Benchy
ok 18:40
MasterDuke huh, wouldn't have expected that to matter
Zoffix well, GC is involved no? 18:41
MasterDuke right, that could effect things 18:42
Zoffix Weird, this script consistently shows 2017.09-173-g4d0ead2 to be 1-15ms seconds faster than HEAD gist.github.com/zoffixznet/988071e...ef0bc88db7
Running as ./perl6 bar.p6 -e '42.starts-with: "x" for ^100_00; say now - INIT now' 18:43
bar.p6 being the script and ./perl6 being head
MasterDuke benchable6: 4d0ead2,HEAD 42.starts-with: "x" for ^100_000 18:44
benchable6 MasterDuke, starting to benchmark the 2 given commits
MasterDuke, benchmarked the given commits, now zooming in on performance differences 18:45
MasterDuke, ¦4d0ead2: «1.0593» ¦HEAD: «1.1259»
MasterDuke benchable6: 4d0ead2,HEAD 42.starts-with: "x" for ^500_000 18:46
benchable6 MasterDuke, starting to benchmark the 2 given commits
MasterDuke, benchmarked the given commits, now zooming in on performance differences 18:47
MasterDuke, ¦4d0ead2: «4.7527» ¦HEAD: «4.8870»
Zoffix ZOFFLOP: t/spec/S17-procasync/kill.t 18:58
Geth rakudo/nom: 1275fd4ab7 | (Zoffix Znet)++ | 2 files
Make Cool.[starts|ends]-with 12x faster

Part of fixing RT#132280:
  rt.perl.org/Ticket/Display.html?id=132280		 18:59
synopsebot RT#132280 [new]: rt.perl.org/Ticket/Display.html?id=132280 [REGRESSION][PERF] Cool methods that take and pass a Capture are uber slower (.contains: 'meow' for ^10000_00)
Zoffix Gonna do rest of 'em later today.
MasterDuke nice 19:00
Zoffix Not really, 'cause many of these is just gaining back speed lost earlier in 2017. But we might break the 3s test-t barrier, 'cause I think I recall it getting slower when these methods were fixed to be captures, but I didn't know why it'd get slow 19:01
travis-ci Rakudo build passed. usev6 '[jvm] Don't run files when all tests are hanging' 19:14
travis-ci.org/rakudo/rakudo/builds/287979127 github.com/rakudo/rakudo/compare/d...e5cfcfa595
Rakudo build errored. Zoffix Znet 'Merge pull request #1194 from MasterDuke17/patch-1 19:41
travis-ci.org/rakudo/rakudo/builds/287916515 github.com/rakudo/rakudo/compare/5...217357aa80
buggable [travis build above] ✓ All failures are due to: missing build log (1 failure).
timotimo MasterDuke: not much; what's up? 19:43
haha. "not much" is also the common answer to "what's up?" 19:54
i was going to ask how to best turn "while $data-which-is-a-Channel.receive -> $blop { }" into something that uses await to free up the thread in question ... then i remembered we have "react whenever", which ought to be better 19:55
oh wow, that's quite the performance hit 20:02
1_500_000 messages with a while try receive gives me 8.95s user, 6.15s elapsed; using react whenever gives me 10s user, 5.88s user, but for only 50_000 messages 20:05
30x difference ;(
Zoffix :o 20:06
AlexDaniel-: whatdyumean reject RT#131515? If that's the right commit, then it should be optimized. The assumtion in the commit is we can be expensive because it won't impact anything, but if it is the right commit, then the assumption is wrong 20:09
synopsebot RT#131515 [open]: rt.perl.org/Ticket/Display.html?id=131515 [BISECTME][PERF] Performance loss in some commits
AlexDaniel- Zoffix: I was thinking that it only impacts the spectest and normal users will never notice it otherwise… now that you said it, I realize it may impact module tests and anything else that tests exceptions 20:11
Zoffix: so you're right
timotimo *sigh*, i'm glad someone is going to make profiler work with multithreade programs ...
Zoffix AlexDaniel-: ah, never mind, it's in .message which has since been optimized away to being called only when actually needed, so I'll close the ticket 20:12
AlexDaniel-: well, bisect it and if it is the commit, then reject it
AlexDaniel- Zoffix: oh! OK
I'll do it right now if it's that simple 20:13
timotimo adding --profile to this program gives me "inconsistent bind result" m(
Zoffix c: 2500e50,2500e50^,HEAD for ^1000 { try "foo".meow } 20:14
committable6 Zoffix, ¦2500e50: ««timed out after 10 seconds» «exit signal = SIGHUP (1)»» ¦2500e50^,HEAD(1275fd4): «»
AlexDaniel- right
Zoffix c: 2500e50,2500e50^,HEAD for ^100 { try "foo".meow }; say "done"
committable6 Zoffix, ¦2500e50: ««timed out after 10 seconds» «exit signal = SIGHUP (1)»» ¦2500e50^,HEAD(1275fd4): «done»
Zoffix c: 2500e50,2500e50^,HEAD for ^10 { try "foo".meow }; say "done" 20:15
committable6 Zoffix, ¦2500e50,2500e50^,HEAD(1275fd4): «done»
Zoffix c: 2500e50,2500e50^,HEAD for ^10 { try "foo".meow }; say now - INIT now
committable6 Zoffix, ¦2500e50: «3.46179999» ¦2500e50^: «0.00295601» ¦HEAD(1275fd4): «0.0024989»
Zoffix rejects
AlexDaniel- OK
or maybe resolve instead?
it *is* resolved after all 20:16
Zoffix right
Done.
AlexDaniel- yay
Geth rakudo/nom: a7b0449375 | usev6++ | src/vm/jvm/runtime/org/perl6/rakudo/RakudoContainerSpec.java
[jvm] Move handling of Nil back to 'store'
 20:29
Zoffix ZOFVM: Files=1278, Tests=152667, 151 wallclock secs (21.59 usr 3.47 sys + 3259.22 cusr 180.73 csys = 3465.01 CPU) 20:34
Geth rakudo/nom: 530e129461 | (Zoffix Znet)++ | 2 files
Make Cool.contains 11.5x faster

Part of fixing RT#132280:
  rt.perl.org/Ticket/Display.html?id=132280		 20:36
synopsebot RT#132280 [new]: rt.perl.org/Ticket/Display.html?id=132280 [REGRESSION][PERF] Cool methods that take and pass a Capture are uber slower (.contains: 'meow' for ^10000_00)
Geth roast: 611988f01f | usev6++ | S02-types/mix.t
[jvm] Skip test that blows up whole test file
 20:44
Zoffix weird.. Patched Cool.split and `42.split: "x" for ^1000_00` now got 2.7x faster, but `42.split: /\d/ for ^1000_0` saw no change whatsoever :S 20:54
Oh, prolly 'cause it's so damn slow, the dispatch is drowned 20:55
timotimo i'd assume so 20:56
Geth rakudo/nom: 785d2f251d | (Zoffix Znet)++ | 2 files
Make Cool.split 0x-2.7x faster

Split with Str matcher is 2.7x faster, but looks like Regex matcher is just so slow that dispatch improvement is drowned so there's no noticalbe improvement there.
Part of fixing RT#132280:
  rt.perl.org/Ticket/Display.html?id=132280
synopsebot RT#132280 [new]: rt.perl.org/Ticket/Display.html?id=132280 [REGRESSION][PERF] Cool methods that take and pass a Capture are uber slower (.contains: 'meow' for ^10000_00)
rakudo/nom: 43e7b893b1 | (Elizabeth Mattijsen)++ | src/core/Rakudo/Iterator.pm
Fix for RT #132288

Make sure we check in HLL before committing to the natives.		 21:14
synopsebot RT#132288 [new]: rt.perl.org/Ticket/Display.html?id=132288 [REGRESSION][9999] .tail with large negative arguments (say (4,5,6).tail(-9999999999999999999))
Geth rakudo/nom: d79ac9704d | (Zoffix Znet)++ | 2 files
Revert fixes to Cool methods with captures

Do not sweet the problem under the rug and instead fix slippy method caching
Zoffix *sweep 21:18
Geth rakudo/nom: b6982e6892 | (Elizabeth Mattijsen)++ | src/core/traits.pm
Streamline Routine traits

  - give name to role for better gisting
  - make sure we only mixin if a trueish value was specified
  - traits affected:
   - is hidden-from -backtrace
   - is hidden-from-USAGE
... (7 more lines)		 22:04
AlexDaniel- samcv: hello? :) 22:11
samcv: nevermind! 22:12
samcv hi!
AlexDaniel- samcv: actually I have two questions 22:15
samcv go ahead
AlexDaniel- I feel like I'm bothering you too much though
anyway
samcv i am at your disposal
not at all
i enjoy sharing my knowledge
AlexDaniel- c: releases "a".uniprop('Numeric').say
what is “Numeric” property? What does it stand for?
committable6 AlexDaniel-, ¦2015.12,2016.01.1,2016.02,2016.03,2016.04,2016.05,2016.06,2016.07.1,2016.08.1,2016.09,2017.09,HEAD(b6982e6): «Lower» ¦2016.10,2016.11,2016.12,2017.02,2017.03,2017.04.3,2017.05,2017.06,2017.07,2017.08: «ALetter» ¦2017.01: «None»
AlexDaniel- it's not the same as Numeric_Type 22:16
I guess?
samcv ah 22:17
AlexDaniel- and second question: is there some indexed documentation for unicode properties so that I can add a link to every row here: gist.github.com/Whateverable/4f40f...8cd62d663d
(and also so that I don't have to ask the first question ever again :) ) 22:18
samcv numeric just means it has a numeric type
so numeric_type != "None" 22:19
AlexDaniel- ok, but what is “Lower”?
Geth rakudo/nom: bb1df2cb0c | (Elizabeth Mattijsen)++ | 3 files
Streamline "is pure" trait handling

  - rename role's method to "is-pure", to be consistent with other traits
  - don't bother calling the method: the existence is the flag
samcv well Numeric is not really an official property. and i don't know when that happened but it should be Numeric_Type != None 22:20
AlexDaniel- want me to submit it as a ticket?
samcv hold on let me check something
well it seems to be doing MVM_UNICODE_PROPERTY_SENTENCE_BREAK when you ask for Numeric 22:21
so yeah make a ticket i guess
if numeric should be a property that is up to perl 6 i guess 22:23
ah i know what's going on
Sentence Break has a Numeric property value
i'm guessing
so anyway it's just an artifect of us looking up property values and not just property names 22:24
AlexDaniel- buggable: tags 22:26
buggable AlexDaniel-, Total: 1631; 6.D: 2; 9999: 11; @LARRY: 27; ANNOYING: 8; BOOTSTRAP: 4; BUG: 592; BUILD: 12; CONC: 45; DOCS: 1; EXOTICTEST: 3; FLAP: 1; GLR: 3; IO: 20; JVM: 48; LHF: 7; LTA: 176; MATH: 5; META: 2; MOAR: 1; MOLD: 233; NATIVECALL: 21; NYI: 57; OO: 13; OPTIMIZER: 8; OSX: 2; PARSER: 5; PERF: 26; POD: 19; PRECOMP: 15; REGEX: 46; REGRESSION:
AlexDaniel-, 39; REPL: 6; RFC: 61; RT: 2; SECURITY: 2; SEGV: 28; SINK: 1; SITE: 1; SPESH: 1; STAR: 7; TESTCOMMITTED: 12; TESTNEEDED: 38; TODO: 13; UNI: 26; UNTAGGED: 282; WEIRD: 2; WINDOWS: 4; See fail.rakudo.party/ for details
AlexDaniel- samcv: RT #132305 22:29
synopsebot RT#132305 [new]: rt.perl.org/Ticket/Display.html?id=132305 [UNI] In some cases ‘Numeric’ uniprop returns “Lower” but it should be “None” instead (say ‘a’.uniprop(‘Numeric’)) 22:30
AlexDaniel- c: 2016.06,HEAD my @a = ^50; my $a = now; for ^100_000 { my $b = Junction.new("all",@a) }; say now - $a; $a = now; for ^100_000 { Junction.new("all",@a) }; say now - $a 22:53
committable6 AlexDaniel-, ¦2016.06: «0.18258788␤0.160125» ¦HEAD(bb1df2c): «0.0761381␤0.4081705»
AlexDaniel- c: 2016.06,HEAD my @a = ^50; my $a = now; for ^100_000 { Junction.new("all",@a) }; say now - $a 22:54
committable6 AlexDaniel-, ¦2016.06: «0.2100153» ¦HEAD(bb1df2c): «0.37509849»
AlexDaniel- c: 2016.06,HEAD my @a = ^50; my $a = now; for ^1000_000 { Junction.new("all",@a) }; say now - $a
committable6 AlexDaniel-, ¦2016.06: «1.53712789» ¦HEAD(bb1df2c): «3.37815925»
AlexDaniel- bisect: old=2016.06 my @a = ^50; my $a = now; for ^1000_000 { Junction.new("all",@a) }; say (now - $a) > 2 22:55
bisectable6 AlexDaniel-, Bisecting by output (old=2016.06 new=bb1df2c) because on both starting points the exit code is 0
AlexDaniel-, bisect log: gist.github.com/eb832cc5450405ae96...094a0535cb 22:56
AlexDaniel-, (2016-08-11) github.com/rakudo/rakudo/commit/97...e6ff5f620a
AlexDaniel- c: 977797fa40^,977797fa40 my @a = ^50; my $a = now; for ^1000_000 { Junction.new("all",@a) }; say now - $a 22:57
committable6 AlexDaniel-, ¦977797fa40^: «1.67519507» ¦977797f: ««timed out after 10 seconds» «exit signal = SIGHUP (1)»»
AlexDaniel- moves to #whateverable
I'm really not sure what to make out of it 23:08
c: 2016.06,HEAD my @a = ^50; my $a = now; for ^1000_000 { @a.all }; say now - $a 23:09
committable6 AlexDaniel-, ¦2016.06: «1.7508351» ¦HEAD(bb1df2c): «4.41190691»
AlexDaniel- c: 2016.06,HEAD my @a = ^50; my $a = now; for ^1000_000 { $ = @a.all }; say now - $a
committable6 AlexDaniel-, ¦2016.06: «1.7970691» ¦HEAD(bb1df2c): «0.8887317»
AlexDaniel- so sinking a junction is now slower, but not sinking it is faster
why would somebody sink a junction, right?
AlexDaniel- shrugs 23:10
timotimo use the coverage functionality on all regressable snippets :3 23:11
buggable New CPAN upload: App-Bob-0.2.2.tar.gz by TYIL cpan.metacpan.org/authors/id/T/TY/...2.2.tar.gz 23:20
New CPAN upload: Hash-Merge-0.1.1.tar.gz by TYIL cpan.metacpan.org/authors/id/T/TY/...1.1.tar.gz
New CPAN upload: IRC-Client-Plugin-Github-0.1.4.tar.gz by TYIL cpan.metacpan.org/authors/id/T/TY/...1.4.tar.gz
New CPAN upload: MPD-Client-0.1.3.tar.gz by TYIL cpan.metacpan.org/authors/id/T/TY/...1.3.tar.gz
MasterDuke timotimo: was just hoping you know something about why this `my $a = rx/aaaaaab/; 'a' x 999999 ~ 'b' ~~ /$a/` spends a couple seconds in Match.pm's Bool proto 23:26
after my recent commit that created a Regex multi for INTERPOLATE 23:28

Powered by Cro and the Raku® Programming Language.

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

Thank you!