[Tux] This is Rakudo version 2016.10-28-gb1f77c8 built on MoarVM version 2016.10-15-g715e39a 06:19
csv-ip5xs 3.213
test 17.388
test-t 6.920
csv-parser 17.286
FROGGS o/ 08:13
tbrowder hi #perl6: o/ 11:23
I'm seeing warnings again on travis: warning: format ‘%llu’ expects argument of type ‘long long unsigned int’, but argument 8 has type ‘MVMuint64’ [-Wformat]
timotimo ... why doesn't that match? 11:26
is long long 128bit?
anyway, those kinds of things are fixed by putting "blah %" PRIu64 " blah" into the format string 11:27
or whatever that spelling is
tbrowder www.irccloud.com/pastebin/I48gQyCp/ 12:24
timotimo we never use functions without prototypes in our C code 12:26
tbrowder i didn't imagine so...
might have to do a cast to shut gcc up 12:28
lizmat Files=1150, Tests=53676, 215 wallclock secs (12.86 usr 3.50 sys + 1312.37 cusr 106.54 csys = 1435.27 CPU) 12:42
timotimo tbrowder: these errors all seem to come from the dynamic variable cache log, which isn't critical at all; i don't think it's used by anyone except core devs, and it's super rarely used even then 12:45
nine Pro tip: when you wonder why your code suddenly got a lot slower, check if you compiled moar with --optimize=0 --debug when investigating an issue... 13:59
tbrowder timotimo: ok 14:03
iBakeCake spots another trig bug o/ 15:12
m: say acosh -1e100
camelia rakudo-moar b1f77c: OUTPUT«-Inf␤»
iBakeCake m: say acosh -1e200
camelia rakudo-moar b1f77c: OUTPUT«Inf␤»
iBakeCake :D Both should be NaNs
m: say acosh(5+5i) ≅ log(5+2*sqrt(6))+5i 15:20
camelia rakudo-moar b1f77c: OUTPUT«False␤»
iBakeCake Seems complex acosh is broken entirely (if WolframAlpha's results are correct)
...and coverage results say it's covered :\ 15:22
ZOFVM: Files=1198, Tests=130073, 145 wallclock secs (20.07 usr 3.45 sys + 2326.37 cusr 238.93 csys = 2588.82 CPU) 15:52
dalek rakudo/nom: 5fe8cf7 | (Zoffix Znet)++ | src/core/Num.pm: 16:01
rakudo/nom: Fix acosh for large negative numbers
rakudo/nom:
rakudo/nom: The IEEE 754-2008[^1], Section 9.2, says acosh for values < 1 is a NaN.
rakudo/nom: We do that for numbers with small maginitude and -∞ itself, but for
rakudo/nom: large negative numbers (e.g. -1e100 and -1e200) we end up returning ±∞
rakudo/nom:
rakudo/nom: Fix by detecting out-of-domain input and returning NaN without performing
rakudo/nom: any calculations. This commit fixes Num and num versions, but not Complex.
rakudo/nom: I need to do further research on complex numbers to know how to fix them.
rakudo/nom:
rakudo/nom: [1] www.csee.umbc.edu/~tsimo1/CMSC455/I...4-2008.pdf
iBakeCake github.com/rakudo/rakudo/commit/5f...c0882ac08d
MasterDuke hope this isn't considered spamming, but trying to catch the earlier-in-the-day crowd 16:30
i've run this as a completely fresh checkout on a different machine. there were some failed tests when running j-spectest, but they all passed when i ran them individually
nine++ for already taking a look 16:31
github.com/MasterDuke17/rakudo/tre...nd_attempt
TimToady obviously there needs to be a unicode character for NaN
perlpilot
.oO( and it should be shaped like a cat )
TimToady or a bat 16:32
iBakeCake batman :D 16:33
TimToady NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN Batman!
iBakeCake EGITHUBDOWN
tadzik dnses are being funny 16:34
iBakeCake yeah
tadzik works for me on 8.8.8.8
TimToady whole interwebs is gettin' hammered lately
moritz I'd say it's getting trumped :-) 16:35
tadzik internet is a fragile lil thing
TimToady russian hackers throwing temper tantrums, I guess
perlpilot tadzik: less so than when we had to sling around hosts files though 16:36
tadzik perlpilot: heh, today it doesn't seem that different :)
dalek ast: 528d66c | (Zoffix Znet)++ | S02-types/num.t:
[coverage] Cover nummy trig: cosh and acosh

Also includes coverage for two acosh bugs[^1][^2] fixed in the past couple of days.
  [1] github.com/rakudo/rakudo/commit/a7...2087187834
  [2] github.com/rakudo/rakudo/commit/5f...c0882ac08d
16:43
TimToady .u ꀞ 16:46
yoleaux2 U+A01E YI SYLLABLE BAT [Lo] (ꀞ)
TimToady looks more like an Enterprise
nine MasterDuke: do you have such j-spectest failures also on an unmodified checkout? 16:51
iBakeCake MasterDuke: Files=1185, Tests=129989, 160 wallclock secs (22.11 usr 2.92 sys + 2453.95 cusr 247.08 csys = 2726.06 CPU) 16:53
MasterDuke: all passed, except what I'm assuming the new num tests. (that's on moar)
Still waiting for JVM
MasterDuke nine: should have thought to check that. i believe so, but not 100% sure 16:56
iBakeCake: cool, thanks
i probably won't have access to a computer to test anything/make changes for a day or two, but i'll check back if there are any comments 16:57
iBakeCake MasterDuke: it's not looking good :) 17:18
MasterDuke: native call tests fail: fpaste.scsys.co.uk/537398 and about 30 or so first stresstests all say No subtests run 17:19
.tell MasterDuke well, I've just killed it. Almost everything fails and I suspect it's just my box trying to do too many tests at once: fpaste.scsys.co.uk/537400 17:24
yoleaux2 iBakeCake: I'll pass your message to MasterDuke.
iBakeCake I think I have TEST_JOBS=30 set
dalek rakudo/nom: 66726e8 | (Zoffix Znet)++ | src/core/Num.pm: 17:55
rakudo/nom: Make atanh(num 1e0) give Inf instead of throwing
rakudo/nom:
rakudo/nom: The 2008 IEEE 754[^1] section 9.2, states atanh(x) with |x| = ∞ is a divideByZero
rakudo/nom: exception. In section 9.1.1, it states this is indicated by a function, which then is
rakudo/nom: to return ∞. And section 7.3 describes the default value of divideByZero as ∞.
rakudo/nom:
rakudo/nom: The speculation[^2] states "floating-point modes do not throw exceptions but rather
rakudo/nom: propagate ∞ and NaN," and in fact, atanh(-1) does return a -∞. Furthermore, the
rakudo/nom: following languages and tools almost universally give a ∞ for atanh(1):
rakudo/nom: R: ∞; Rust: ∞; C: ∞; Ruby: ∞; Perl 5: ∞; WolframAlpha: ∞; Python: math domain error
rakudo/nom:
rakudo/nom: Related fix for Num: github.com/rakudo/rakudo/commit/90...ec90ebe425
rakudo/nom:
iBakeCake github.com/rakudo/rakudo/commit/66...f48e294b9d 17:56
dalek ast: a8e616b | (Zoffix Znet)++ | S02-types/num.t:
[coverage] Cover nummy trig: tanh and atanh

Also covers atanh(num 1e0) bug fixed in:
  github.com/rakudo/rakudo/commit/66...f48e294b9d
18:06
FROGGS o/ 18:07
iBakeCake looks at S32-trig/generate-tests.pl 18:11
incomprehensible 18:12
dalek ast: 163d683 | (Zoffix Znet)++ | S02-types/num.t:
Cover 1e0.atanh

Fixed by github.com/rakudo/rakudo/commit/90...ec90ebe425
The S32-trig/* space is more appropriate for this test, but the tests there
  are generated by a 500-line script (S32-trig/generate-tests.pl) that I don't know.
18:17
travis-ci Rakudo build errored. Zoffix Znet 'Make atanh(num 1e0) give Inf instead of throwing 18:24
travis-ci.org/rakudo/rakudo/builds/169588766 github.com/rakudo/rakudo/compare/5...726e8d6f4a
iBakeCake buggable: yo 18:28
buggable: tags
buggable iBakeCake, Total: 1391; BUG: 943; UNTAGGED: 284; LTA: 121; NYI: 91; RFC: 58; JVM: 56; CONC: 55; SEGV: 37; REGEX: 36; UNI: 29; PERF: 28; @LARRY: 22; NATIVECALL: 19; POD: 19; IO: 17; TODO: 14; PRECOMP: 12; TESTCOMMITTED: 12; BUILD: 11; OO: 9; MATH: 8; STAR: 6; BOOTSTRAP: 5; GLR: 5; OSX: 4; TESTNEEDED: 4; DOCS: 3; OPTIMIZER: 3; REPL: 3; WEIRD: 3; REGRESSION: 2; SPESH: 2; CONFIGURE: 1; CUR: 1; LIBRARY: 1; RT: 1; SIT
iBakeCake stupid robot
hm :/ that travis URL gives me "The repository at rakudo/rakudo was not found."
geekosaur probably failed to reach github 18:29
iBakeCake prolly
geekosaur and that looks like the travis failures were the same thing (I did get it to load)
FROGGS m: sub foo($a:(Int)) { say $a(42) } # TimToady: that's meant to work, right? I mean, mit ampersand *and* dollar as sigil... 18:47
camelia rakudo-moar b1f77c: OUTPUT«===SORRY!=== Error while compiling <tmp>␤Variable '$a' is not declared␤at <tmp>:1␤------> sub foo($a:(Int)) { say ⏏$a(42) } # TimToady: that's meant to wor␤»
FROGGS s/mit/with/
FROGGS .oO( strangely inconsistent ) 18:48
jnthn fwiw, I thought it was conditional on the & sigil 19:15
FROGGS yes it was... I'm not quite sure how it parses it... 19:22
though I'm not seeing why we should not allow bot $ and & 19:23
dalek kudo/nom: f75da76 | lizmat++ | src/core/ (2 files):
{s/Seq/Iterate/}NextNFromIterator

A utility function working on iterators creating iterators is much more generally usable.
20:04
kudo/nom: 8d357af | lizmat++ | src/core/ (2 files):
Introduce Iterator.skip-at-least

Being able to skip a number of values on an iterator can be useful thing to have. Generalised the skip-at-least-pull-one functionality into skip-at-least and pull-one. Removed the now superfluous SeqSkipNFromIterator method from Rakudo::Internals.
20:16
lizmat good night, #perl6-dev! 20:21
timotimo gnite lizmat! 20:40
skip-at-least? why not skip-at-most?
or you'll end up skipping more than you wanted?
travis-ci Rakudo build errored. Elizabeth Mattijsen 'Introduce Iterator.skip-at-least 21:05
travis-ci.org/rakudo/rakudo/builds/169612371 github.com/rakudo/rakudo/compare/f...357affe41b
buggable [travis build above] ☠ Did not recognize some failures. Check results manually.