Zoffix nqp-m: say(nqp::isnull(NQPMu.new)) 03:39
camelia nqp-moarvm: OUTPUT«0␤»
Zoffix Getting closer with that bug! :)
Gonna eat a nice bug stew today :) 03:40
MasterDuke m: my $e = 28; my str $a = "a" x 2**28; say now - INIT now 04:04
camelia rakudo-moar ff8362: OUTPUT«Memory allocation failed; could not allocate 1073741824 bytes␤»
MasterDuke m: my $e = 28; my str $a = "a" x 2**27; say now - INIT now
camelia rakudo-moar ff8362: OUTPUT«Memory allocation failed; could not allocate 536870912 bytes␤»
MasterDuke m: my $e = 28; my str $a = "a" x 2**25; say now - INIT now 04:05
camelia rakudo-moar ff8362: OUTPUT«0.04169808␤»
MasterDuke m: my $e = 25; my str $a = "a" x 2**$e; say now - INIT now
camelia rakudo-moar ff8362: OUTPUT«0.00155113␤»
MasterDuke a profile when using the variable has a whole lot of info in it 04:06
but a profile when using the constant has three lines, with all the exclusive time spent in <unit> (-e:1) 04:07
maybe i'm just tired, but it's not making any sense to me 04:08
geekosaur hm, is any kind of constant folding implemented as yet? 04:15
(although if that's the problem, would seem LTA to not have at least the SORRY scaffold...) 04:16
...and if it is constant folding, why would it make things slower 04:17
[Tux] This is Rakudo version 2016.06-104-gff83625 built on MoarVM version 2016.06-9-g8fc21d5 08:01
test 16.091
test-t 9.404
csv-parser 16.657
FROGGS o/ 10:32
timotimo MasterDuke: i expect the int is built at compile time there 10:39
dalek kudo/nom: 9180464 | lizmat++ | src/core/List.pm:
Speed up foo xx Int about 10%
10:52
kudo/nom: f878186 | lizmat++ | src/core/Map.pm:
Streamline Map.Capture & Map.clone
11:16
kudo/nom: 855f4d4 | (Brad Gilbert)++ | src/core/IO/ArgFiles.pm:
Fix .slurp on an IO::ArgFiles that has already been read from

It was calling $!io.readall but that method was renamed to .slurp-rest
kudo/nom: a2ff54c | (Brad Gilbert)++ | src/core/IO/ArgFiles.pm:
c318601 | (Daniel Green)++ | src/ (3 files):

Added $SIG_ELEM_CODE_SIGIL
timotimo oh, nice 11:18
masak TimToady_: ooc, do you agree that tracing a grammar breaks LTA matching? github.com/jnthn/grammar-debugger/...-229940842 11:27
jnthn LTM? :P 11:28
timotimo less than matching?
masak er
jnthn It's LTA though.
:P
masak LTM, naturally
masak 's brain is on vacation
jnthn The solution, for whoever fancies implementing it, is probably to mix a role into the tracing closure that we hand back.
Where said role has the methods like !NFA or whatever on it and forwards them to the underlying code-ref
(That of the original rule)
Since the reason for the problem is, afaict, that the closure we return instead of the method lacks the NFA 11:29
OTOH, we could switch the NFA resolution to use lookup instead of find_method since it's after the original declarator 11:30
Then Grammar::Tracer will just work and everyone won't have to re-solve the same problem if doing similar things.
In fact that's probably the better way, and then Grammar::Tracer will Just Work again.
So let's do it that way :)
Then it'll freely fix other modules like Grammar::Debugger and Grammar::Profiler too I suspect. 11:31
masak adds this insight to the RT ticket
timotimo but that's a github ticket
jnthn plz can haz thunderstorm already? 11:32
Sticky weather is sticky
masak jnthn: um, where would the solution reside? in Rakudo, or in Grammar::Tracer?
timotimo we have nice cold weather in germany today
i think it'll reside in nqp
jnthn masak: NQP
kudo/nom: 23daf3b | lizmat++ | src/ (3 files):
Merge pull request #791 from MasterDuke17/RT128392

Fix RT #128392
ast: 3fe0101 | (Daniel Green)++ | S06-signature/types.t:
Test for RT #128392

Requires a Rakudo built with PR #791 (github.com/rakudo/rakudo/pull/791)
11:37
ast: 04f2229 | lizmat++ | S06-signature/types.t:
Merge pull request #130 from MasterDuke17/RT128392

Test for RT #128392
jnthn lunch &
masak ok, then it belongs in the (closed) RT ticket 11:51
timotimo: it's a github ticket and an RT ticket, each linking to the other 11:59
timotimo OK
masak urh, stale backlog, sorry :)
MasterDuke is anybody else here having problems with RT? 13:09
i've emailed perlbug-admin three times over the past five weeks about my account, but it still isn't fixed (i can't see tickets when logged in) 13:10
timotimo :[
MasterDuke and now my emails to perl6-bugs-followup aren't showing up in tickets either
maybe i got stuck in some spam filter recently? 13:11
there are a couple RT tickets that could be closed, but i can't comment on them 13:12
perhaps i should comment on the merged PRs asking for someone else to close/comment on the RT tickets? 13:13
timotimo i think it'll likely be fine to just post the ticket numbers here for someone to pick up? 13:17
oh, synopsebot6 is down again
MasterDuke RT #128392 13:20
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=128392
MasterDuke RT #125489
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=125489
MasterDuke RT #128097 13:22
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=128097
Zoffix Well, I give up. It's too hard :| 13:55
jdv79 what is it? 14:10
Zoffix RT #128457 14:12
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=128457
Zoffix I think I spent about 4 days on it :)
And this is all I found: rt.perl.org/Ticket/Display.html?id...xn-1406988 14:13
It seems like the fix is around the corner. But ... ¯\_(ツ)_/¯
I don't know about lexpads and context, so it's a waste of my life trying to debug this further :) 14:14
jdv79 probably just another 4 days to fix it:) 14:18
nice troubleshooting in any case. similar to a utf8 bug i was looking at. i got blocked by lack of domain knowledge. 14:20
AlexDaniel MasterDuke: have you tried commenting by sending emails? 14:31
MasterDuke AlexDaniel: yeah, i sent some emails to 125489 on June 17 14:33
AlexDaniel … :-/ 14:34
MasterDuke it's worked before, but not recently
tbrowder re RT: maybe it's time to revisit bugzilla... 14:47
AlexDaniel #128520 14:48
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=128520
tbrowder i find it hard to believe companies pay to use RT
Zoffix Heh. And who's going to change all the references to RT in roast, commits, etc? :) 14:50
Or is that a proposal to basically use two bug trackers?
AlexDaniel also, I really hate the fact that I cannot edit my tickets. I can never get it right from the first time…
Zoffix: the proposal is to stop using RT, because this hell must end some day
mst could be worse. could be github. 14:51
Zoffix There was a conversation about this already and I doubt this RFC will get any further.
AlexDaniel Zoffix: whether it means that we will be using two bug trackers at the same or not is a different question, I think. But feel free to leave a comment
Zoffix I think if the logs are grepped for me saying "why the hell" you'd find my comments :P 14:52
AlexDaniel if it wont get any further then it is pretty sad.
tbrowder i'm sure most know the definition of idiocy attributed to Einstein... 14:53
Zoffix doesn't
mst tbrowder: companies pay for RT because it's extremely good at what it does; JIRA can come close but takes if anything even longer to configure 14:54
tbrowder something like keep doing the same thing over and over again expecting different results
AlexDaniel mst: well, at least people can login on github…
Zoffix Ah. AlexDaniel found the previous conversation on the topic. It's around this line: irclog.perlgeek.de/perl6/2016-02-04#i_11989733
mst AlexDaniel: you have a login problem with rt? did you talk to the admins? 14:55
AlexDaniel mst: did you read the ticket?
tbrowder if RT is so good at what it does why so many complaints?
AlexDaniel mst: see this: well probably not, it's Mast
brr
irclog.perlgeek.de/perl6-dev/2016-0...i_12773781
pasted the wrong thing
mst tbrowder: because everybody hats ticketing systems 14:56
tbrowder: they're either completely feature free like github and we hate them because we can't keep the tickets organised
tbrowder: or they're full of stuff like bugzilla but it often isn't the right stuff 14:57
AlexDaniel mst: what features of RT help to keep the tickets organized?
mst tbrowder: or they're configurable as hell like RT but then everybody's upset about the current configuration
Zoffix AlexDaniel, tags? 14:58
nine The only thing I miss in RT is some git integration
mst the fixed in/broken in stuff is nice, having a status other than 'closed' is nice
AlexDaniel Zoffix: have you seen anybody actually using tags? :)
mst tags are what people don't ever use on github 14:59
Zoffix AlexDaniel, yes, we have tests needed/bug tags for Rakudo
mst I've seen all the features of rt.cpan.org used for organising things on sane CPAN dists
AlexDaniel right. And other stuff is just part of the title
mst nine: I could get you git integration easily ifg you used a real git server rather than github
but I can't reach into the proprietary goo :( 15:00
Zoffix If you use [BUG] in the title, the system adds bug tag for for you for free
AlexDaniel anyway, you can use tags on Github so that's not an argument I guess. Not that I'm saying that we should use github, but if you say that RT helps to keep tickets organized because of tags… well… I don't know…
mst that isn't what we said
you're strawmanning everything people say because you've decided RT sucks and are uninterested in making things better
AlexDaniel /o\
mst only in the "this project sucks, we should rewrite it in a different language" type solution
if, instead of concentrating on bitching about RT, you were to try and make a list of things you wish you had, I can figure out what RT can already do with you 15:01
and it's entirely possible to customise things
nine mst: at work we host our git repos on an internal server accessible over ssh. In fact it's the same server RT is running on. But I haven't found any git plugin or extension for RT?
mst last time I had a missing feature on rt.cpan.org, I wrote the feature and the lovely volunteers who maintain it for us were happy to review and deploy it 15:02
but I suppose bashing their work is more -Ofun if you're AlexDaniel
AlexDaniel mst: what's the solution to CSRF thingy?
not that it causes many problems, but it is a bit annoying 15:03
tbrowder i'll start a list of things that bug me or i would like to see on RT...
AlexDaniel mst: or how can I edit a ticket?
MasterDuke i think bashing is an overstatement. he's raised some concerns, which i share (i'll admit mostly because they directly effect me) 15:04
nine AlexDaniel: you cannot edit a ticket and that's a good thing. When you add a comment or reply to a ticket, RT sends emails with your content. What should it do when you edit?
AlexDaniel nine: should not do anything, as these are typically just changes to grammar or formatting fixes 15:05
or if I pasted some wrong output or something
mst AlexDaniel: oh fucking hell github's ticket editing 15:06
you realise there's no history there, so people can retroactively wipe out bug reports and comments
Zoffix mst, what was the name of the dude who refused to deal with RT tickets?
mst if only they'd heard of version control
Zoffix And was ranting about it
mst Zoffix: no idea, they all blur into one eventually ;)
Zoffix heh
mst along with the people who won't use github because github's ToS aren't properly open
Zoffix MLEHMANN maybe 15:07
mst and the people who give up on bugzilla because after half an hour they can't figure out how to do a standard search
AlexDaniel yes, github is less than awesome
mst and then JIRA, well, java, so you need four times as many servers and a full time workflow admin
though admittedly RT kinda needs a half-time workflow admin
nine: throwing something together with RT::CLient::REST would be pretty quick I thinkm, but, yeah, surprised there's not already 15:16
nine mst: that reminds me of the idea to have an intern work on this :) 15:17
Shouldn't $failure.defined defuse the Failure instead of making it explode? 15:22
psch m: sub f { fail "foo" }; my $f = f; $f.defined; $f # it does? 15:27
camelia rakudo-moar 23daf3: OUTPUT«WARNINGS for <tmp>:␤Useless use of $f in sink context (line 1)␤»
psch m: sub f { fail "foo" }; my $f = f; $f # it does?
camelia rakudo-moar 23daf3: OUTPUT«WARNINGS for <tmp>:␤Useless use of $f in sink context (line 1)␤»
psch ah, sinking doesn't show that :| 15:28
m: sub f { fail "foo" }; my $f = f; say $f
camelia rakudo-moar 23daf3: OUTPUT«foo␤ in sub f at <tmp> line 1␤ in block <unit> at <tmp> line 1␤␤Actually thrown at:␤ in block <unit> at <tmp> line 1␤␤»
psch m: sub f { fail "foo" }; my $f = f; $f.defined; say $f
camelia rakudo-moar 23daf3: OUTPUT«(HANDLED) foo␤ in sub f at <tmp> line 1␤ in block <unit> at <tmp> line 1␤␤»
nine Not in the setting it seems. Even assigning the Failure to a variable makes it explode
psch oh
mst AlexDaniel: also, if perlbug-admin didn't reply, please check with #rt or the perl NOC
AlexDaniel MasterDuke: ↑ 15:29
MasterDuke is there a standard procedure for contributing when you have a Perl 6 commit bit?
mst AlexDaniel: ... oh fuck.
nine Or maybe it's because the failure is thrown by a Proxy's FETCH
MasterDuke until now i have been submitting PRs which have been merged by someone else, since i couldn't commit anything directly
mst MasterDuke: er, looking at irclog.perlgeek.de/perl6-dev/2016-0...i_12773781
psch MasterDuke: if it's potentially controversial a branch (or, well, PR) plus RFC is a good idea
mst you had a problem getting email through to RT
MasterDuke AlexDaniel, mst: thanks, i'll give them a try
mst and tried to get it solved by sending email to the same server 15:30
I think I'm willing to take a guess why perlbug-admin didn't get back to you :(
psch MasterDuke: if it's something that's pretty clearly just a bugfix or NYI you can probably just commit to nom, depending on scope i guess
mst has anybody tried emailing them the details from a *different* email address?
MasterDuke emailing used to work
mst and? 15:31
MasterDuke doesn't now
mst yes, exactly my point.
let me try this again
if your problem is: my emails are no longer getting to perlbug-followup
this means that the underlying cause may be: my emails are no longer getting through the mail server
which means that if you try to report it by: sending an email to that mail server 15:32
your report probably got eaten by whatever's eating your followup comments
and that would explain why they're 'ignoring' you
MasterDuke well, i believe my email to perlbug-admin was before my emails to perlbug-followup stopped working, but yes, i suspect a larger email problem (on whose end i have no idea) 15:33
so i'll follow up at #rt 15:34
mst that *may* work if the rt.cpan.org people are there
but I think the mail servers are handled by the NOC
err, rt.perl.org
fuck
I always typo that, too used to dealing with cpan stuff
MasterDuke: if you don't get anywhere, ping me, I can probably chase the relevant people down out of band if I have to 15:35
AlexDaniel: please do remember this isn't a black box and that there are people around who can escalate - especially those of us who straddle the perl5 community and therefore are in a bit closer touch with the people involved 15:36
AlexDaniel mst: ok, I'll keep that in mind. 15:38
tbrowder re editors: is anyone using neovim? if so, how do you like it for p6 and utf8 support? 15:53
Woodi so, 2015.xmass we had 6.c spec and implementation. what we could expect on next xmass ? :) 15:54
jnthn Woodi: Sleep. Relaxation. Food. Drink. :P 15:55
masak I can sign up on that :) 15:56
Woodi is it officially sheduled ? ;) probably is now...
masak Woodi: what do you mean "official scheduled"?
masak .oO( I hereby officially schedule Christmas to fall somewhere near the end of December ) 15:57
Woodi masak: confirmation that other #perl6ixer don't get gifts this year ;)
Woodi actually didn't have expectations, just question without answet :) 15:59
*ver
jnthn Woodi: Will we time *another* language version release to coincide with a time when most people want to rest and do family stuff? No, or at least, now when I'm on the hook for some of it. :-) 16:00
It was worth doing once because of the Christmas meme. :) 16:01
masak .oO( the unintended consequences of the Christmas meme intersecting with the realities of a release ) 16:02
masak .oO( yeah, so, Perl 7 will be released on, um, Talk Like a Pirate Day ) 16:04
TimToady_ and v6.28 will be releast on Tau Day 16:28
*released
masak .oO( to say the leased! )
dalek kudo/nom: ca67be5 | (Zoffix Znet)++ | docs/release_guide.pod:
List Zoffix Znet to do #101 release

Assigned by [Coke] irclog.perlgeek.de/perl6/2016-07-01#i_12771625
19:22
kudo/nom: 02c7fed | lizmat++ | docs/release_guide.pod:
Merge pull request #807 from zoffixznet/patch-1

List Zoffix Znet to do #101 release
lizmat Files=1109, Tests=52368, 246 wallclock secs (14.47 usr 4.21 sys + 1522.26 cusr 126.43 csys = 1667.37 CPU) 19:31
^^^ long time ago since I've seen CPU time below 1700 seconds
:-) 19:32
dalek kudo/nom: d358ee6 | lizmat++ | src/core/Map.pm:
Make Map eqv Map about 3x faster

For the worst case, with all keys/values being the same.
19:43
timotimo cool! 20:00
i don't eqv maps often, but damn. that's a fine improvement
lizmat well, there's something weird there, though 20:01
because Map are supposed to be immutable
I'm not sure what Map.clone should do other than return self ?
my benchmark being: my $m = Map.new(^1000); my $n = $m.clone; for ^1000 { my $d = $m eqv $n } 20:02
will think about this some more 20:03
vendethiel- lizmat: if it's immutable, then definitely return self 20:06
[Coke] m: say CompUnitRepo; 20:16
camelia rakudo-moar d358ee: OUTPUT«===SORRY!=== Error while compiling <tmp>␤Undeclared name:␤ CompUnitRepo used at line 1␤␤»
psch m: say CompUnit::Repository 20:17
camelia rakudo-moar d358ee: OUTPUT«(Repository)␤»
[Coke] docs has some references to some old stuff.
sub die_on_fail(int $fail=1) { - that shouldn't take an int. 20:22
(from Test)
we also apear to be missing "sub fail" 20:23
psch m: say &fail
camelia rakudo-moar d358ee: OUTPUT«sub fail (| is raw) { #`(Sub|55506600) ... }␤»
psch in the docs i assume?
[Coke] that fail is for Failures. Wondering about how to deliberate fail a test. 20:26
*deliberately 20:27
psch ohh, right. there was a contextual hint there in the backlog :|
[Coke] ok False is close enough for now. 20:31
timotimo that's not True 20:33
lizmat m: use Test; flunk("dog at my homework") 20:35
camelia rakudo-moar d358ee: OUTPUT«not ok 1 - dog at my homework␤␤# Failed test 'dog at my homework'␤# at <tmp> line 1␤»
lizmat in other words: it's flunk in tests 20:36
dalek kudo/nom: 8f440ea | lizmat++ | src/core/ (2 files):
Cloning a Map is a noop, a Hash *should* clone

Because a Map is immutable, and a Hash *is* mutable
20:37
[Coke] lizmat++
lizmat
.oO( I even flunked writing ate )
20:38
geekosaur ...clearly you need to eat 20:46
:)
dalek kudo/nom: 4622c50 | lizmat++ | src/core/ (2 files):
Simplify Hash|Map.STORE_AT_KEY

  - no need to check for definedness of $!storage
   STORE_AT_KEY is only called from Hash.STORE, and that initializes $!storage
  - remove unneeded native str for keeping .WHICH
22:33
lizmat more work on Hash.STORE tomorrow
first some sleep
timotimo \o/
sleep well!
lizmat so good night, #perl6-dev
dalek kudo/nom: 3fda0b0 | coke++ | lib/Test.pm6:
Add a note to help devs find "fail"
23:10
travis-ci Rakudo build errored. Elizabeth Mattijsen 'Simplify Hash|Map.STORE_AT_KEY 23:31
travis-ci.org/rakudo/rakudo/builds/141916851 github.com/rakudo/rakudo/compare/8...22c5042118
timotimo seems spurious ^ 23:44