El_Che New 2017.11 Linux packages: github.com/nxadm/rakudo-pkg/releas...g/v2017.11 00:17
samcv interesting testing related problem semi humorous unix.stackexchange.com/questions/4...me-at-0030 01:06
looks like a `man` easter egg caused someones tests to fail but only at 00:30
ugexe i dont think it occured under valid flag/arg combinations 01:07
heh maybe it did 01:08
ilmari[m] ugexe: it wasn't meant to, but accidentally did under -w. fixed in git master now 01:44
See the reply/comment from Colin Watson
samcv i personally think time based easter eggs that are not identical each run are a bad idea 02:20
Geth rakudo: 0a24efc32d | usev6++ | src/vm/jvm/runtime/org/perl6/rakudo/Binder.java
[jvm] Throw X::Parameter::InvalidConcreteness

and give better error messages. This follows the code for MoarVM in src/Perl6/Metamodel/BOOTSTRAP.nqp		 06:53
lizmat Files=1228, Tests=75867, 313 wallclock secs (14.50 usr 5.18 sys + 2164.05 cusr 212.16 csys = 2395.89 CPU) 07:20
afk for most of the rest of the day&
[Tux] I stopped my browser(s) before running tests this time. I see no diff: 07:27
Rakudo version 2017.11-3-g0a24efc32 - MoarVM version 2017.11-2-g26ad7cc55
csv-ip5xs1.105 - 1.114
csv-ip5xs-2012.804 - 13.195
csv-parser12.299 - 12.583
csv-test-xs-200.440 - 0.457
test12.093 - 12.182
test-t3.220 - 3.269
test-t --race1.329 - 1.340
test-t-2059.915 - 60.010
test-t-20 --race20.152 - 20.849
and something broke HP-UX: smoke.procura.nl/ 07:34
|Tux| that last line is obviously in the wrong channel 10:42
pmurias in docs/op.markdown should we also mention the return type of the op? 15:57
yoleaux 20 Nov 2017 17:57Z <samcv> pmurias: you want "ducet" for co
[Coke] pmurias: yes, probably 16:45
pmurias samcv: thank you 17:13
samcv: what is that shifter vs non-ignorable stuff we have a test for? 17:15
Geth nqp/master: 5 commits pushed by pmurias++
El_Che 20:06
Geth rakudo: f40c381811 | usev6++ | tools/build/Makefile-JVM.in
[jvm] Don't pass --nqp-lib=blib in 'make install'

This unbreaks the installed versions of perl6 and perl6-j.		 21:10
nqp: 3927afb302 | pmurias++ | docs/ops.markdown
Add more return types to op docs, fix some signatures
 21:43
lizmat . 22:39
Geth rakudo: 2c1a2860b8 | (Elizabeth Mattijsen)++ | src/core/Rakudo/Iterator.pm
Add While/Until iterators

  - While: pass on values until the first false value returned by block
  - Until: don't pass values on until the first true value returned by block		 22:43
rakudo: 76fcea1e18 | (Elizabeth Mattijsen)++ | src/core/Any-iterable-methods.pm
Expose new R:I:While/Until in .head/.tail

  - .head taking a Callable:D: block should return trueish while values
   should be passed on. The first falsish value will end the iterator.
  - .tail taking a Callable:D: block should return falsish while values
   are being skipped. The first trueish value will start passing on
   values until iterator is exhausted.
... (7 more lines)		 22:49
jnthn star: dd (1..10).head(*-2) # curious 22:51
camelia (1, 2, 3, 4, 5, 6, 7, 8).Seq
jnthn star: dd (1..10).head({ $_ -2 }) # curious
camelia (1, 2, 3, 4, 5, 6, 7, 8).Seq
jnthn Ah, that was in the bit of the commit message I had to follow the link to see. 22:52
lizmat yup
jnthn Yeah, I'm a bit uneasy about that
head-while and head-until are a little long but workable 22:53
.while and .until don't let us have a subroutine form
Which in turn don't let us use them with ==>
Took me a while to realize that conflict was why I was so uneasy about those names
Naming aside, I'm in agreement we want methods that do this, though 22:54
Plus they're desirable on Supply
MasterDuke skip-while and skip-until (with reversed meaning)? 22:57
lizmat well, actually, I see .skip as the reverse of .grep 22:58
a bit like .split and .comb
with grep you specify what you want, with .skip you specify what you don't want
but that's for later... 22:59
jnthn .skip is culturally a bit awkward in that it's typically been used for .skip($number-of-elements) 23:01
And .skip(2) would, like grep(2), "work" by smart-matching against 2...which will probably give people some headaches 23:02
Actually skip-while and skip-until, with skip($n) being the numeric thing, could work I guess 23:10
lizmat I see 4 types of semantics: 23:13
1. pass on values while condition is true, end iterator on first false
2. pass on values while condition is false, end iterator on first true 23:14
3. skip values while condition is true, pass on rest of iterator on first false 23:15
4. skip values while condition is false, pass on rest of iterator on first true
or do you think 2 and 4 should not be possible and should be handled by the developer negating the condition? 23:16
jnthn Well, if we decide that the developer should negate the conditions, then we get .grep, .initial (which probably needs a better name), and .skip 23:19
Oh no, .skip can't really if it should do the numeric thing
Hmm
I guess an interesting question would be: do people find the lack of negative grep and having to negate a pain point? 23:20
lizmat well, apparently not :-)
or not a lot anyway
jnthn Yeah, I haven't personally felt it alot 23:21
m: say (1, 1.5, 2, 2.5).grep(* !~~ Int)
camelia (1.5 2.5)
lizmat is tired though after getting up waaay to early, then being on the road for ~7 hours, and then doing some shopping to top it off
jnthn Since that worked, it's been less annoying
timotimo opposite-of-grep could be called ĘƒÉ¹Ēd
jnthn /o\ 23:22
jnthn revokes timotimo's language design license :P
timotimo it's not hard to use grep(none(blah)) to get the negative of grep(blah)
it's not very performant, of course :)
jnthn ah, the junction approach...I'd forgotten that one
lizmat yeah, that's not the opposite, that's the upsidedownite
jnthn Well, true
Lots to think about, anyway :) 23:23
lizmat yup
meanwhile, I'll revert that last commit as we appear to agree it's too magic, right ?
jnthn I think it's too surprising, yeah
MasterDuke i seem to recall a couple people asking how to negate a grep in #perl6, but i could be imaging it (don't really remember any details) 23:24
Geth rakudo: c967cb6cc8 | (Elizabeth Mattijsen)++ | src/core/Any-iterable-methods.pm
Revert "Expose new R:I:While/Until in .head/.tail"

This reverts commit 76fcea1e18455fac9e71ab4ecc2ef42789d7035e.
   irclog.perlgeek.de/perl6-dev/2017-...i_15484290		 23:25
jnthn If we do decide we want an inverse grep, omit would be a nice name 23:28
Though I'm not convinced we do 23:29
Discovered that while trying to figure out something better than .initial(&cond)
Best I found so far is .leading(&cond)
timotimo omit is fantastic, especially because it's just "timo" backwards
MasterDuke clojure has some good names 23:31
drop-while, drop-last
jnthn Does it have drop taking an integer number to drop too? 23:32
timotimo if you decide upon it based on "is it integer?" that can lead to surprises when you change your calculation to accidentally give a Rat and suddenly you're smartmatching against Rats rather than taking the first $n 23:33
MasterDuke that's nthnext
jnthn That's almost got jnthn in it :P 23:34
MasterDuke oO(have Rick Hickey and jnthn ever been seen in the same room?) 23:35
lizmat goes to bed& 23:39

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!