MasterDuke IOninja: re Signature.$!count, pretty sure i tried to make that an Int, but the possibility of being an Inf prevented it 00:46
i think there was even some discussion here about it (back when i was making a bunch of changes to BOOTSTRAP.nqp), but it was decided that keeping the possibility of Inf was the best option 00:48
so that rebind looks suspect to me, unless maybe it's an optimization? 00:49
IOninja But Num is not an Int... 00:51
So I'm not surprised your making it an Int failed.
MasterDuke well, no, i wasn't either after finding out Inf was a possible value 00:53
IOninja and I forget what that rebind looked like, but likely just sets the type contraint to Any 00:55
MasterDuke nqp::bindattr($signature, $sig_type, '$!count', self.add_constant('Int', 'int', $count).value);
IOninja m: use nqp; class Foo { has Num $.x is rw = 0e0; }; my \x = Foo.new; nqp::bindattr(x, Foo, '$!x', $ = 42); x.x = 'x'; dd x.x 00:56
camelia Str $ = "x"
IOninja m: use nqp; class Foo { has Num $.x is rw = 0e0; }; my \x = Foo.new; nqp::bindattr(x, Foo, '$!x', $ = 42); x.x = 'x'; dd x.x.VAR.of
camelia Mu
IOninja or Mu even
m: use nqp; class Foo { has Num $.x is rw = 0e0; }; my \x = Foo.new; nqp::bindattr(x, Foo, '$!x', my Int $ = 42); x.x = 'x'; dd x.x.VAR.of
camelia Type check failed in assignment; expected Int but got Str ("x")
in block <unit> at <tmp> line 1
IOninja m: use nqp; class Foo { has Num $.x is rw = 0e0; }; my \x = Foo.new; nqp::bindattr(x, Foo, '$!x', 42); x.x = 'x'; dd x.x.VAR.of
camelia Cannot modify an immutable Int
in block <unit> at <tmp> line 1
MasterDuke m: say &log.signature.count.WHAT 00:57
camelia (Int)
MasterDuke m: say &say.signature.count.WHAT
camelia (Num)
MasterDuke perl6 -e '' creates 3,286 Signatures, maybe making the $!count's Int is better for performance 00:58
IOninja Why would make it an int be better for performance? 01:04
*Int
MasterDuke i don't know about this specific case, but integer operations are almost always faster than floating point operations 01:06
IOninja :/ 01:07
IOninja still stands behind all the reasons stated last time $!count was proposed to be made an Int 01:08
MasterDuke i'm not re-proposing to make it an Int 01:09
i said rebinding it to an Int was suspect
and if there's no noticeable performance reason for doing so, i would say it should stay a Num 01:10
IOninja But my point was there's no point in it being a Num, because that type constraint appears to be removed. 01:11
MasterDuke `nqp::bindattr($signature, $sig_type, '$!count', self.add_constant('Int', 'int', $count == -1 ?? nqp::inf() !! $count).value);`
IOninja &
MasterDuke sure, or leave it a Num and don't remove the constraint 01:12
oops
`nqp::bindattr($signature, $sig_type, '$!count', self.add_constant('Num', 'num', $count == -1 ?? nqp::inf() !! $count).value);`
that's what i meant ^^^
samcv NeuralAnomaly, status 01:57
NeuralAnomaly samcv, [✘] Next release will be in 1 week and 1 day. Since last release, there are 48 new still-open tickets (9 unreviewed and 1 blockers) and 60 unreviewed commits. See perl6.fail/release/stats for details
samcv luckily have almost finished work on fixing that blocker with the case insensitive regex
MasterDuke nice 01:58
travis-ci Rakudo build passed. TimToady 'propagate purity up through meta-meta asts 02:21
travis-ci.org/rakudo/rakudo/builds/209506945 github.com/rakudo/rakudo/compare/f...a40f7d95eb
Geth nqp: 50233099d5 | TimToady++ | 4 files
move %*HOW into the braid

We'll just reuse the same slang hash to simplifiy braid cloning, but prefix with a 'H:' to prevent symbol clashes.		 06:07
rakudo/nom: 49b1b32345 | TimToady++ | 3 files
move %*HOW (and %*HOWUSE) into the braid

  %*HOWUSE just uses the same braid mechanism with a "U:" on the front.
rakudo/nom: version bump brought these changes: github.com/perl6/nqp/compare/2017....7-g5023309
9d7c0ab8ff | TimToady++ | tools/build/NQP_REVISION
rakudo/nom: version bump brought these changes: github.com/perl6/nqp/compare/2017....7-g5023309
f190f241e0 | (Elizabeth Mattijsen)++ | src/core/List.pm

For very large integer values, like: 42 xx 99999999999 . The thunk version is not yet lazy, still looking at two spectest failures :-(
[Tux] This is Rakudo version 2017.02-226-g9d7c0ab8f built on MoarVM version 2017.02-35-gdc40845b 07:09
csv-ip5xs 2.863
test 12.565
test-t 4.964 - 5.006
csv-parser 12.707
samcv o/ [Tux] 08:09
[Tux], does yo,ur CVS script use case insensitive regex anywhere?
cause that's about to get a lot faster
IOninja Leaving a Num and not removing the constraint won't work, because Ints aren't Nums :/ 12:04
The closest constraint would be Real, Numeric, or Cool, but I doubt that has any real benefit.
naxieAlDle does anybody have any ideas about RT #130910 ? 15:52
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=130910
IOninja shakes head 15:53
timotimo naxieAlDle: did you get any interesting results with the heapanalyzer btw? 15:55
naxieAlDle timotimo: I don't think I ever tried it
timotimo oh, OK 15:56
hey folks, y'all may want to make sure to run perl Configure.pl inside MoarVM the next time you "git pull" 16:32
i just turned 3rdparty/libtommath into an actual submodule 16:33
lizmat timotimo: this doesn't matter for people just doing dev work in rakudo, right ? 17:20
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=130956
timotimo it'll be interesting the moment moar gets bumped i assume
lizmat but then a configure.PL at the rakudo level should be enough then, no? 17:21
or should we nuke install to be sure ?
timotimo hm. to be fair, it'll probably not be a problem at all
IOninja perl6.party/post/Tag-Your-Dists 17:29
New post: "Camelia wants YOU to tag your dists": perl6.party/post/Tag-Your-Dists
timotimo ugh, the tag "cloud" on that page is ... massive 17:31
IOninja timotimo: <IOninja> Well, I'm gonna leave it like that for now. And will sort out something more usable in ~7-8 hours. 17:32
timotimo oh 17:33
sorry :)
a bit distracted by impending doom 17:34
IOninja doom?
timotimo oh, you know, just someone i don't know coming into my home and judging stuff in it
IOninja huh
[Tux] samcv, yes 17:39
timotimo well, i live in an apartment for rent. the current owner is selling all the small apartments they own and instead buying houses
so naturally, someone who wants to buy my apartment has to be found
and they're not going to buy it without having a look at it first. for good reasons.
perlpilot IOninja: "camelion" makes me think of a camel with a lion's head or something :)
IOninja: If you don't already have a suggestion for the tag cloud, I suggest making it a drop down. 17:42
IOninja gross
timotimo i'd prefer a click-to-expand. it's already starting with "all" selected anyway 17:43
perlpilot or make a link for the tag cloud, that when hovered over, give you the traditional cloud of words at different font sizes
timotimo why not make deselecting "all" cause the tabs display to open up
IOninja tabs display? 17:44
timotimo well, a box full of tabs 17:45
or whatever we end up with
perlpilot that also seems LTA
right now, the results of the search box are pushed off the bottom of my screen because of the tags. I think a box full of tabs would have a similar problem. 17:46
timotimo that box-full-of-tabs can of course already be half-visible, so that you get the clue of "you can expand this, there's more stuff to be seen!"
perlpilot IOninja++ for the tags though, that's awesome. 17:47
timotimo hm, will the search filter also thin out the list of tags?
IOninja No 17:49
Unless you visit the /q/ page (non-JS search)
timotimo i mean more like: do we want that to happen 17:50
perlpilot huh ... why do all of the modules show up in the list twice?
timotimo oh, huh, they do 17:51
but not immediately
perlpilot sorry, not all, but most
timotimo and only if no tags are selected
IOninja I'll shove the 1-2 item tags to the bottom of the list and make the tag box grow-on-hover/tap. If anyone else wants to make it different, they're welcome to it
timotimo thank you for your work! 17:52
perlpilot IOninja++ sounds good.
IOninja ... not sure what's up with duplication... 17:53
perlpilot I would have guessed some sort of join error, but I dunno
IOninja probably 'cause primary keys got changed :|
or maybe not 17:54
gonna check back in ~2 hours to see if it grows on each db rebuild and will debug it after work
IOninja &
and if I die before that... old db build is in old.db in modules.perl6.org dir :) just copy it over to modulesperl6.db and checkout a commit before my tags workk 17:57
japhb IOninja: Love that Camelion logo. :-) 18:26
IOninja :/ 19:33
buggable is nomming all the ram again. 19:34
Stupid OS. Kill buggable, not Geth!
nine IOninja: ulimit is your friend :) 19:42
IOninja too hard. I don't want to learn things. 19:44
japhb IOninja: Better stop working on Perl 6 then. Hazard of the hobby. :-) 20:10
IOninja my google-foo tells me ulimit can't restrict resident memory or something and that's what buggable is noming. I need cgroups or something. 20:22
samcv morning all 20:45
IOninja \o 20:47
samcv gotta bump nqp and perl6 now to get the m:i fix 21:47
IOninja cool 21:52
Geth nqp: b083e3471a | (Samantha McVey)++ | tools/build/MOAR_REVISION
Bump Moar to get case insensitive string compare fix

58457845 Fix string_equal_at_ignore_case when string `a` changes length (#548) 1288e7ed update libtommath fork to have expmod hang fix fbb4c272 turn libtommath into a submodule, use MoarVM's fork dc40845b Merge pull request #546 from MasterDuke17/add_checks_on_resulting_size_to_string_concatenating_and_repeating d4d9eb8c Do gen2 sweep before letting stolen thread go. ffd32f53 MVMROOT decoder when building string result. beefd2b1 Check result of repeat/concat fit in an MVMString
nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2...-g58457845
4ceee06de6 | (Samantha McVey)++ | t/moar/07-eqatic.t

Add some tests that test when expanding characters are present and when they are not.
jnthn wonders if we have clean spectest6 now :) 21:54
Geth rakudo/nom: 22f00cd72d | (Samantha McVey)++ | tools/build/NQP_REVISION
Bump Moar/NQP to get case insensitive string compare fix

4ceee06d Add tests for nqp::eqatic case insensitive string compare b083e347 Bump Moar to get case insensitive string compare fix		 21:58
¦ rakudo/nom: version bump brought these changes: github.com/perl6/nqp/compare/2017....-g4ceee06d
IOninja jnthn: were there fixes for stability of spectest6? As in... should I fire up my 32-core VM and try to make it crash?
samcv tests should now pass
IOninja samcv++
IOninja fires up the VM to do some stresstesting :)
OHOHOH 22:00
Google got an upgrade, apparently.
I can fire up a 64-core VM :o 22:01
IOninja checks the pricing on that one before doing so... :}
samcv more cores == able to crash faster right? 22:02
hah
lizmat hmmm... git pull / perl config breaks ? 22:04
nuking install doesn't help 22:05
IOninja m: say "${3.97/4.345} per hour for 64-core 416GBRAM VM"
camelia 5===SORRY!5=== Error while compiling <tmp>
Unsupported use of ${3.97/4.345}; in Perl 6 please use $(3.97/4.345) for hard ref or $::(3.97/4.345) for symbolic ref
at <tmp>:1
------> 3say "${3.97/4.345}7⏏5 per hour for 64-core 416GBRAM VM"
lizmat Generating tools/check.mk .............................. OK
Generating build/mk-moar-pc.pl ......................... OK
Configuration FAIL. You can try to salvage the generated Makefile.
IOninja m: say "\${3.97/4.345} per hour for 64-core 416GBRAM VM"
camelia $0.913694 per hour for 64-core 416GBRAM VM
lizmat is this a known issue ?
git error: fatal: destination path '3rdparty/libtommath' already exists and is not an empty directory. 22:06
Clone of 'git://github.com/MoarVM/libtommath' into submodule path '3rdparty/libtommath' failed
IOninja m: say "\${2.57/4.345} per hour for 64-core 57GB RAM VM"
camelia $0.591484 per hour for 64-core 57GB RAM VM
IOninja lizmat: that looks like what timotimo++ was doing earlier today
timotimo: ping... any idea how to fix ^ breakage? 22:07
samcv well it got upgraded lizmat and uh looks like it become a submodule instead of actually being in the Moarvm/moarvm repo
jnthn IOninja: Yeah, I found a really nasty parallel GC bug
lizmat yeah, building rakudo is bored atm
jnthn (And fixed it)
lizmat *borked
jnthn Heh, I mentally corrected to boned...
lizmat: Toss your MoarVM checkout
samcv lizmat, hmm it worked fine for me pulling mvm and getting the change. but yeah do what jnthn said maybe 22:08
or uh
maybe git reset --hard
jnthn Or that
samcv and then git pull
jnthn Oh, may have to clean
lizmat nuked nqp
getting past the submodule update now 22:09
jnthn phew :) 22:10
IOninja dammit "Starting VM instance "perlbuild2" failed. Error: Quota 'CPUS' exceeded. Limit: 32.0" 22:14
wtf
jnthn 32 CPUs should be enough for anybody :P 22:16
IOninja heh 22:17
eh, can't see a region with 64-cpu quota :/... gonna try to figure out it some other day. 32-cores it is 22:22
timotimo lizmat: just have to rm -rf 3rdparty/libtommath 22:27
lizmat I nuked nqp
:-)
IOninja nuked everything 22:28
lizmat seems to also fixed the problem, albeit a bit more rigorously :-)
but not quite as rigorous as IOninja
timotimo if MoarVM lives under nqp/ on your end, that's fine 22:32
lizmat timotimo: it does, as I use the default build instructions of rakudo all the time 22:36
timotimo understood
samcv yay blocker fix. now what else to do today. i'm hungry 22:39
IOninja samcv++ awesome
NeuralAnomaly: blockers
NeuralAnomaly IOninja, There is 1 release blocker. See perl6.fail/release/blockers
IOninja, perl6.fail/130953 : chars that look different in .fc break :i regex matches
IOninja m: say "ststststT" ~~ m:i/T/ 22:40
camelia False
samcv NeuralAnomaly, blockers
NeuralAnomaly samcv, There are no release blockers
IOninja how... did you... :o
samcv it just took a sec to update
or maybe because i visited the webpage idk
camelia is not up to date yet. should hopefully start working again shortly 22:41
IOninja Oh. I thought you were able to unmark it :)
I think camelia will need to nuke its nqp too, won't it?
samcv well i marked it as resolved in rt
i didn't have to nuke anything
IOninja or nqp/MoarVM/3rdparty/libtommath or whatever
samcv i have moarvm rakudo and nqp in seperate directories under ~/git 22:42
IOninja m: $*PERL.compiler.version.say
camelia v2017.02.227.gf.190.f.24
samcv and had no problems just pulling
but camelia sheould have updated by now
maybe it only breaks if you build rakudo with moarvm and nqp automade 22:43
instea dof having seperate repos
IOninja hm, commitable updated itself just fine too
samcv dunno
IOninja c: HEAD say "ststststT" ~~ m:i/T/
committable6 IOninja, ¦HEAD(22f00cd): «｢T｣»
IOninja \o/
stresstest6: Files=1209, Tests=132708, 277 wallclock secs 22:44
IOninja sticks it to run in a look and goes to eat
Geth roast: d30ef47f07 | (Elizabeth Mattijsen)++ | S03-operators/repeat.t
Unfudge now passing tests for RT #127971, RT #130924

  samcv++ I believe		 23:00
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=127971
Link: rt.perl.org/rt3//Public/Bug/Displa...?id=130924
IOninja Hm. harness6 doesn't pickup passing TODOs then? 'cause I ran it on current HEAD without those roast changes :o 23:01
Geth roast: 5aa268236e | (Elizabeth Mattijsen)++ | S05-substitution/match.t
Unfudge now passing tests for RT #130953

Also samcv++ I believe		 23:02
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=130953
lizmat IOninja: no idea really
I was waiting for a stable harness6 before delving deeper into it
IOninja looks like not indicating a failed run with an exit code either; I had while HARNESS_TYPE=6 make stresstest; do true; done loop running and on one run t/spec/S10-packages/precompilation.t flooped, but it kept chugging along 23:04
Geth rakudo/nom: db1f689b60 | (Elizabeth Mattijsen)++ | src/core/Rakudo/Iterator.pm
Introducing R::It.Callable-xx-Times

Basically the 'rand xx 42' functionality.
Also remove specific sink-all in Callable-xx-Whatever, as it needs to execute the thunks for their possible side-effects		 23:05
IOninja Anyway. 6 runs on 32-cores, no weird failures with harness6 \o/
Geth rakudo/nom: 1754dc5f7e | (Elizabeth Mattijsen)++ | src/core/List.pm
Make thunk xx 42 use the new iterator

This breaks 4 tests:
   t/spec/S03-metaops/cross.t (56-57)
   t/spec/S03-metaops/zip.t (56-57)
I've spent several days trying to figure out why: ... (16 more lines)		 23:13
lizmat I hope someone with more inspiration can figure out why those tests fail ^^^
meanwhile, I'm going to get some shuteye 23:14
good night, #perl6-dev!
samcv night lizmat 23:17
IOninja night
MasterDuke samcv: do you think you've completely addressed RT #126793 ? 23:34
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=126793
samcv i have not tested ignoremark + ignorecase which has different codepaths in nqp and rakudo, but i will take an educated guess and say yes. but ticket will stay open until some tests get made 23:35
oh looks like there's still some bugs in it. but no more problem than there was before 23:38
before we used foldcase i mean.
m: say 'aaaasta' ~~ m:i/st/
camelia False
samcv think that's an old commit actually on camelia
MasterDuke yeah, that's the one before your bump 23:39
samcv evalable6, $*PERL.compiler.version.say 23:40
IOninja .ask nine camelia seems to be stuck on an old commit. Needs to nuke nqp/MoarVM/3rdparty/libtommath to get it going again
yoleaux2 IOninja: I'll pass your message to nine.
evalable6 v2017.02.230.g.1754.dc.5.f.7
samcv evalable6, say 'st' ~~ m:i/st/ 23:41
evalable6 ｢st｣
samcv evalable6, say 'aaast' ~~ m:i/st/
evalable6 False
samcv pretty sure that's not moar's fault on this one and the issue here is somewhere in rakudo
so that's fine since it's not a regression 23:42
m: use nqp; say nqp::eqatic('aaast', 'st', 3) 23:43
camelia 1
samcv yeah that works fine
but that ticket i guess is that we should do fc instead of lc. and i changed all of those to foldcase 23:44
bisectable6, say 'ABstEFG' ~~ m:i/st/ 23:45
bisectable6 samcv, On both starting points (old=2015.12 new=1754dc5) the exit code is 0 and the output is identical as well
samcv, Output on both points: «False»
samcv bisectable6, say 'stEFG' ~~ m:i/st/ 23:46
bisectable6 samcv, Bisecting by output (old=2015.12 new=1754dc5) because on both starting points the exit code is 0
samcv, bisect log: gist.github.com/3e30751ae31d2808d6...eee7c05ca6
samcv, (2017-02-28) github.com/rakudo/rakudo/commit/26...2c543a7d95
samcv if st is at the start of the string it works fine, but otherwise has issues
MasterDuke looks like uc/lc are being used here: github.com/perl6/nqp/blob/master/s...A.nqp#L284 23:47
samcv oh yeah i forgot that one
yeah that i didn't touch cause i didn't want to break anything 23:48
but maybe it makes character classes with lowercase and uppercase letters?
and then doesn't use nqp::eqatic
(which is only availible on moar)
MasterDuke well, both the `charrange` and `literal` methods use uc/lc, seems like `literal` might be a simpler fix 23:50
Geth nqp: daea00bd47 | (Samantha McVey)++ | src/QRegex/NFA.nqp
Fit line to 80 columns width in QRegex/NFA.nqp
 23:59
nqp: c8b630155e | (Samantha McVey)++ | src/QRegex/NFA.nqp
Strip extra whitespace

Powered by Cro and the Raku® Programming Language.

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

Thank you!