Geth rakudo: d93f805389 | (Zoffix Znet)++ | tools/lib/NQP/Configure.pm
Fetch tags too to avoid potental missing tags

As here: irclog.perlgeek.de/perl6/2018-01-11#i_15676712		 00:46
nqp: fb5848819b | (Zoffix Znet)++ | tools/lib/NQP/Configure.pm
Fetch tags too to avoid potental missing tags

As here: irclog.perlgeek.de/perl6/2018-01-11#i_15676712
travis-ci NQP build failed. Zoffix Znet 'Fetch tags too to avoid potental missing tags 01:02
travis-ci.org/perl6/nqp/builds/327486698 github.com/perl6/nqp/compare/dfa8a...5848819b86
Geth rakudo: c6b7012a9c | (Zoffix Znet)++ | 2 files
Fix spurious warnings with colonpaired longnames

Fixes RT#131305: rt.perl.org/Ticket/Display.html?id=131305
With a simple colonpair, like `infix:<+>` we have the value right away and don't get warnings, but with things like `postfix:<[ ]>` the value has a complex value that we compile-time compile when ... (10 more lines)		 01:29
synopsebot RT#131305 [new]: rt.perl.org/Ticket/Display.html?id=131305 Useless Useless Use Warning when invoking colonpaired routines
travis-ci Rakudo build canceled. Zoffix Znet 'Comment out leftover debugging statement' 02:06
travis-ci.org/rakudo/rakudo/builds/327402532 github.com/rakudo/rakudo/compare/d...e9d55205d4
buggable [travis build above] ☠ Did not recognize some failures. Check results manually. 02:07
Geth rakudo: 0dc4f7318f | (Zoffix Znet)++ | src/Perl6/Actions.nqp
Use faster nqp::elems instead of + on QAST nodes

Bench shows it being 20% faster. There are more opportunities for this change in Rakudo, but converting all I found caused an explosion I had trouble locating the cause of, so doing it partially.		 03:48
nqp: 6b9d226f0c | (Zoffix Znet)++ | 18 files
Use faster nqp::elems instead of + on QAST nodes

Bench shows it being about 20% faster
rakudo: b6004362bf | (Zoffix Znet)++ | tools/build/NQP_REVISION
[NQP Bump] Brings 3 commits

NQP bump brought: github.com/perl6/nqp/compare/2017....g6b9d226f0 6b9d226f0 Use faster nqp::elems instead of + on QAST nodes f760f56f1 Remove trailing whitespace fb5848819 Fetch tags too to avoid potental missing tags		 03:49
rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2017....g6b9d226f0
631875fcdf | (Zoffix Znet)++ | src/core/allomorphs.pm

Phixes github.com/rakudo/rakudo/issues/1387
The methods use Rational's candidates, which create a new object using `self.new`, expecting it to take a numerator and denominator, but `self` is a RatStr, not a Rat. ... (7 more lines)
travis-ci Rakudo build canceled. Zoffix Znet 'Fetch tags too to avoid potental missing tags 03:59
travis-ci.org/rakudo/rakudo/builds/327486625 github.com/rakudo/rakudo/compare/0...3f80538934
buggable [travis build above] ☠ Did not recognize some failures. Check results manually.
travis-ci Rakudo build passed. Zoffix Znet 'Remove trailing whitespace' 06:06
travis-ci.org/rakudo/rakudo/builds/327514299 github.com/rakudo/rakudo/compare/c...65efe8b8f5
Rakudo build failed. Zoffix Znet 'Use faster nqp::elems instead of + on QAST nodes 07:15
travis-ci.org/rakudo/rakudo/builds/327524245 github.com/rakudo/rakudo/compare/8...c4f7318f6a
buggable [travis build above] ☠ Did not recognize some failures. Check results manually.
[Tux] I plan to do a distribution update of my workstation this weekend from openSUSE Leap 42.2 to 42.3. I do not think it will have any influence on the timings, but it might 08:17
Rakudo version 2017.12-169-gb6004362b - MoarVM version 2017.12.1-26-g65098f05b
csv-ip5xs1.096 - 1.172
csv-ip5xs-2012.795 - 12.842
csv-parser12.617 - 12.719
csv-test-xs-200.450 - 0.476
test11.372 - 11.556
test-t2.865 - 2.870
test-t --race1.146 - 1.173
test-t-2050.993 - 52.320
test-t-20 --race18.009 - 18.023
 08:29
travis-ci Rakudo build failed. Zoffix Znet '[NQP Bump] Brings 3 commits 08:48
travis-ci.org/rakudo/rakudo/builds/327524372 github.com/rakudo/rakudo/compare/0...004362bf82
buggable [travis build above] ☠ Did not recognize some failures. Check results manually.
lizmat good *, #perl6-dev 08:55
on modules.perl6.org/search/ , is there a way to show only core modules?
if not, maybe core modules should have "core" tag ? 08:56
DrForr That almost seems important enough to be its own (optional) flag in META6.json or friends. 09:01
lizmat but *is* there a META6.json file for those ? 09:02
if so, where does that live ?
DrForr Aah, so I fail at knowing... right. 09:03
That does sound similar to dual-life modules in p5, but the circumstances are clearly different. Dual-life modules exist in core *and* CPAN, perl6 core modules aren't yet broken out onto CPAN. 09:05
Which leads me to the inevitable reminder to get what I have into CPAN over the weekend or similar. 09:06
lizmat using App::Mi6 it's trivial :-) 09:10
DrForr Duly noted, and installing when I get home. 09:24
lizmat Files=1231, Tests=76004, 330 wallclock secs (14.77 usr 5.40 sys + 2272.39 cusr 232.07 csys = 2524.63 CPU) 09:32
afk& 09:40
Zoffix lizmat: use from:core modules.perl6.org/search/?q=from%3Acore 12:51
tbrowder .tell AlexDaniel success!! tidying up and adding some new tests...will ping when ready again for a review.
yoleaux tbrowder: I'll pass your message to AlexDaniel.
Zoffix hm, looks like modules.perl6.org/help doesn't list that one
lizmat Zoffix: also, Telemetry appears to be missing, how do I add that ? 12:52
and snapper for that matter
Zoffix lizmat: they need to be listed in this hash: github.com/perl6/modules.perl6.org....pm#L8-L22 12:53
lizmat ok, will add them 12:54
AlexDaniel` . 13:21
... Ah...
lizmat m: my $a = <2.2>; dd $a; $a++ # this should work, shouldn't it ? 13:23
camelia RatStr $a = RatStr.new(2.2, "2.2")
Type check failed in binding to parameter '$r'; expected Rat but got Int (16)
in block <unit> at <tmp> line 1
lizmat m: my $a = 2.2; dd $a; $a++; dd $a # this does 13:24
camelia Rat $a = 2.2
Rat $a = 3.2
lizmat m: my $a = <2.2>; dd $a; $a += 1; dd $a # and this does 13:25
camelia RatStr $a = RatStr.new(2.2, "2.2")
Rat $a = 3.2
lizmat RatStr seems to be the only one with this problem, NumStr, IntStr and ComplexStr are all ok 13:27
Zoffix s: &postfix:<++>, \(<2.2>) 13:30
SourceBaby Zoffix, Something's wrong: ␤ERR: Could not find candidate that can do \(RatStr.new(2.2, "2.2"))␤ in sub sourcery at /home/zoffix/services/lib/CoreHackers-Sourcery/lib/CoreHackers/Sourcery.pm6 (CoreHackers::Sourcery) line 37␤ in block <unit> at -e line 6␤␤
Zoffix s: &postfix:<++>, \(my $ = <2.2>)
SourceBaby Zoffix, Sauce is at github.com/rakudo/rakudo/blob/b600...ny.pm#L512
Zoffix s: <2.2>, 'succ', \() 13:31
SourceBaby Zoffix, Sauce is at github.com/rakudo/rakudo/blob/b600...al.pm#L202
lizmat GH #1387
Zoffix m: my $x = <2e0>; dd $x; $x++; dd $x 13:34
camelia NumStr $x = NumStr.new(2e0, "2e0")
Num $x = 3e0
Zoffix All allomorphs are affected looks like; it's just RatStr is the only one to explode.
[Coke] .ask zoffix do you have an LC_COLLATE variable set? your comment on doc's 287bd033 indicates you're using sort -u (which is what I use), but when I do it, I end up not collating, just sorting by codepoint. 14:16
yoleaux [Coke]: I'll pass your message to zoffix.
[Coke] .tell zoffix (like your sort order better, want to keep it when I resort in the future)
yoleaux [Coke]: I'll pass your message to zoffix.
AlexDaniel . 14:19
yoleaux 12:51Z <tbrowder> AlexDaniel: success!! tidying up and adding some new tests...will ping when ready again for a review.
Zoffix . 14:45
yoleaux 14:16Z <[Coke]> Zoffix: do you have an LC_COLLATE variable set? your comment on doc's 287bd033 indicates you're using sort -u (which is what I use), but when I do it, I end up not collating, just sorting by codepoint.
14:16Z <[Coke]> Zoffix: (like your sort order better, want to keep it when I resort in the future)
Zoffix [Coke]: I don't have LC_COLLATE set
[Coke]: here's my entire env: gist.github.com/zoffixznet/2b52b3b...36c020b4bd 14:47
$ sort --version 14:48
sort (GNU coreutils) 8.21
lizmat afk& 14:54
dogbert2_ hmm, is github having problems, I get a page with a unicorn 15:20
MasterDuke dogbert2_: same here 15:21
dogbert2_ what a meltdown :)
[Coke] .tell zoffix thanks, it's probably the sort binary itself then. 15:55
yoleaux [Coke]: I'll pass your message to zoffix.
geekosaur $LANG is set, this will set $LC_COLLATE implicitly 16:03
and most unicode locales have a default collation that, among other things, case-folds 16:04
[Coke] geekosaur: my LANG is set to the same thing. 16:18
but my sort version is "2.3-Apple (99)" 16:19
timotimo ooh, fascinating 16:22
Geth roast: 4e1bcfccc6 | (Zoffix Znet)++ | S02-literals/allomorphic.t
Test .succ/.pred on allomorphs

Closes github.com/rakudo/rakudo/issues/1387 Rakudo fix: github.com/rakudo/rakudo/commit/631875fcdf
nqp: a620559a78 | (Zoffix Znet)++ | tools/build/MOAR_REVISION
[MoarVM Bump] Brings 5 commits

MoarVM bump brought: github.com/MoarVM/MoarVM/compare/2...-ged7c1234 ed7c1234 Optimize not_i when its input is a known value 671bf1db Add some missing fact dependencies 38a7f577 Remove bogus comment and known value fact tweak c1da6f96 Make instruction deletion within dead BBs a no-op 65098f05 Make log facts tolerate the odd outlier		 16:57
¦ nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2...-ged7c1234
rakudo: 31b3b77318 | (Zoffix Znet)++ | tools/build/NQP_REVISION
[NQP Bump] a620559a7 [MoarVM Bump] Brings 5 co […]

NQP bump brought: github.com/perl6/nqp/compare/2017....ga620559a7
MoarVM bump brought: github.com/MoarVM/MoarVM/compare/2...-ged7c1234 ed7c1234 Optimize not_i when its input is a known value 671bf1db Add some missing fact dependencies 38a7f577 Remove bogus comment and known value fact tweak c1da6f96 Make instruction deletion within dead BBs a no-op 65098f05 Make log facts tolerate the odd outlier
¦ rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2017....ga620559a7
rakudo/supply-syntax-opts: 55e56393d7 | (Jonathan Worthington)++ | 2 files
Compile time error for misplaced whenever

I thought we already were doing this, but: make it a compile time error to have a `whenever` block that is not lexically within a
  `supply` or `react` block. Hopefully nobody was relying on this not
being an error; if they are, we can ban it in 6.d, and conditionally do the opts that rely on it only for 6.d also. However, it seems highly unlikely.		 17:05
rakudo/supply-syntax-opts: d41acccd5e | (Jonathan Worthington)++ | 2 files
Optimize `supply emit <arg>`

This is a convenient way of getting a Supply that will simply emit a single argument and then be done (or quit if the process of producing the argument throws an exception). Given the program `my $s = supply emit 42; for ^100_000 { react whenever $s { } }`, this results in it running in 78% of the time it would without this optimization.
Kaiepi sorry for the newby question and compile moarvm, nqp, and rakudo? 18:12
i want to script this but working out how to build them in the right order 18:14
timotimo there's a script for you! 18:15
github.com/zoffixznet/z - i think this is the one
Kaiepi perfect you're a lifesaver 18:16
timotimo other than that it's not very difficult. just got to run Configure.pl with the same --prefix in moarvm, then nqp, then rakudo. and after Configure.pl in each of those folders you'll have to make install, so the next configure script can find what you've built 18:17
Kaiepi yeah, just the order was confusing a bit 18:21
remembering which prefis is which
timotimo the prefixes should be the same or it won't work :)
Kaiepi hm 18:23
what was the order?
nvm read
i was doing nqp first, not moar
timotimo you can build nqp without moar, but then you'll have to use either jvm or js as backend (or both) 18:34
Kaiepi i need to check if the jdk plays nicely with it 18:35
timotimo rakudo on jvm is known to have problems :( 18:38
and rakudo on js doesn't pass the whole perl6 spec test suite yet
Kaiepi idr what happened on freebsd but the install failed
timotimo oh, freebsd
Kaiepi i'm on openbsd now 18:39
timotimo i wonder if some security feature is being too restrictive, like, we ask for executable stack because of our nativecall stuff
Kaiepi the most annoying thing i dealt with was tweaking /etc/login.conf to let me compile stuff without giving up 18:40
timotimo huh? 18:41
Kaiepi its default max-prox, open file limit , proc count, file 18:42
they were preeetty modest with the limits
thank god i wasn't building the userland discovering 18:43
*discovering that
timotimo open files, yeah, that's fun 18:48
we have multiple open files for every thread the user code starts, and we start out with two threads already
Geth rakudo: 70c57c3a92 | usev6++ | src/Perl6/Metamodel/BOOTSTRAP.nqp
[jvm] Hide unimplemented op captureinnerlex

The code refactoring in 773d2b9e7b exposed nqp::captureinnerlex for all backends -- even though the new method !clone_phasers is called on MoarVM only.
Fixes the JVM build.		 20:57
lizmat bartolin++ 21:06
Geth rakudo: ac2e50c855 | (Zoffix Znet)++ | src/Perl6/Actions.nqp
NO MONKEY BUSINESS!! 🙈🙉🙊

Fixes github.com/rakudo/rakudo/issues/1386
Not adding any tests.		 21:34
lizmat .tell Zoffix modules.perl6.org/search/?q=List%3A%3AUtil shows List::Util 2x, did I do something wrong ? 22:31
yoleaux lizmat: I'll pass your message to Zoffix.
Zoffix lizmat: nope, and the duplicate is gone now. It's a bug on the site. See #2 on this issue: github.com/perl6/modules.perl6.org/issues/95 22:37
lizmat ok *phew* 22:38
travis-ci Rakudo build passed. Zoffix Znet '[NQP Bump] a620559a7 [MoarVM Bump] Brings 5 co […] 23:37
travis-ci.org/rakudo/rakudo/builds/327767945 github.com/rakudo/rakudo/compare/6...b3b7731804

Powered by Cro and the Raku® Programming Language.

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

Thank you!