»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg camelia perl6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by sorear on 25 June 2013.
00:05 hoverboard left 00:10 hoverboard joined 00:11 pippo left 00:19 rurban joined 00:23 rurban left 00:31 hoverboard left
lue wonders how hard it would be for postcircumfix:<{ }> not defined for type Str to instead be postcircumfix:«< >» not defined for type Str when <> was used instead of {} ... 00:33
timotimo why is IO::Spec::splitfile's nofile parameter not a named? 00:43
oh, it is 00:48
but not in the docs.
er, specs.
it seems like you cannot watch a non-existing file? 00:50
but nothing seems to complain about non-existing files 00:51
BenGoldberg Is file watching done by libuv? 00:53
timotimo yes
00:56 araujo left
timotimo modification watching is just ... ugh 00:56
bbl.
00:57 araujo joined 01:03 thou joined 01:05 Su-Shee_ joined 01:08 Su-Shee left 01:16 FROGGS left 01:18 klapperl joined 01:20 rurban joined 01:22 klapperl_ left 01:23 lue left 01:24 rurban left 01:28 FROGGS joined 01:36 lue joined 01:44 kurahaupo joined 01:57 plobsing left 01:58 plobsing joined 02:01 bjz left, telex left 02:02 telex joined
dalek osystem: f314638 | (Andrew Egeler)++ | / (2 files):
November has its own META; remove from SHELTER
 02:07
02:10 colomon left 02:11 colomon joined 02:13 kurahaupo_mobile joined 02:15 kurahaupo left 02:20 rurban joined, kurahaupo_mobile is now known as kurahaupo 02:22 hoverboard joined 02:25 rurban left 02:26 kst` joined 02:27 kst left 02:42 rurban joined 02:46 MikeFair_ left 02:47 bjz joined 02:51 virtualsue left 02:54 cooper- left
dalek ast: 52e0238 | (David Warring [email@hidden.address] | integration/advent2010-day (4 files):
adding 2010 advent days 6, 7, 8 10
 03:20
03:23 hoverboard left 03:42 hoverboard joined 04:03 dwarring left, logie_ joined 04:08 logie_ left 04:19 rurban left 04:24 cooper- joined 04:35 rurban joined 04:41 sftp left, sftp joined 04:44 Crazylemon64 left 04:55 rurban left 05:16 kaare_ joined 05:26 rurban joined 05:28 rurban1 joined, atta left 05:30 rurban left 05:32 rurban1 left 05:37 dotslashb left 05:55 hoverboard left 05:57 BenGoldberg left 05:58 SamuraiJack_ joined 06:36 atta joined 07:03 kurahaupo left 07:14 nnunley left
moritz \o 07:18
07:30 rurban joined
FROGGS morning moritz 07:34
07:34 rurban left
FROGGS if (uwex.result instanceof P6OpaqueBaseInstance) { 07:39
System.out.printf("is P6OpaqueBaseInstance%n");
vmex = (VMExceptionInstance)decont(uwex.result, tc);
jnthn: this explodes with: java.lang.ClassCastException: __P6opaque__72 cannot be cast to org.perl6.nqp.sixmodel.reprs.VMExceptionInstance
what am I doing wrong?
07:42 darutoko joined 08:01 Rotwang joined 08:21 Su-Shee_ is now known as Su-Shee 08:31 rurban joined 08:32 denis_boyun_ joined, rurban1 joined 08:35 rurban left 08:37 rurban1 left 08:40 anaeem1 joined 08:42 igorsutton joined 08:45 igorsutton left 08:48 igorsutton_ joined 08:50 dmol joined 08:51 igorsutton_ left 09:03 kivutar joined 09:05 denis_boyun___ joined, denis_boyun_ left 09:08 berekuk joined 09:12 kurahaupo joined 09:33 rurban joined, spider-mario joined 09:34 plobsing left 09:37 rurban left 09:38 SevenWolf left 09:46 kivutar left
masak good noon, #perl6 09:53
FROGGS morning masak 09:54
woolfy1 Is this in any way related to the use of the jvm by Perl 6? www.theverge.com/2014/5/9/5699958/f...e-v-oracle 09:55
www.businessinsider.com/google-orac...ing-2014-5 09:57
09:58 vendethiel left
masak woolfy1: don't think so. 10:01
FROGGS gah' another karma point for Oracle® on my companies-that-I-hate list...
masak woolfy1: we're not implementing any Java APIs. 10:02
10:02 vendethiel joined
FROGGS yeah 10:02
masak by the way, is there anything we can do to implement JSR 223 for r-j? www.drdobbs.com/jvm/jsr-223-scripti.../215801163
tadzik I think google is being punished by reimplementing Java SDK, sort of
we're not doing that, we're just using what java provides
FROGGS exactly
masak I know some people who would find it very convenient to plug Rakudo into the Eclipse platform through JSR 223.
just to be clear, what Oracle is doing (suing Google for implementing Java for Android), regardless of its legal implications, is what's commonly known as a "dick move". 10:03
(as in "a Richard (who is not a very nice guy) move") 10:04
10:05 kivutar joined
vendethiel
.oO( I never knew richard danced )
 10:06
masak they seem to want to send the signal "aaaargh Google how dare you use *our* open-sourced language for *your* purposes!?" 10:07
which will in the end hurt Java, Oracle, Google, and Open Source.
everybody loses.
vendethiel isn't google's "java" outdated ?
the one on android 10:08
masak no idea. does it make a difference?
vendethiel I think it does, I don't like seeing 3 different javas running around with differences for each one, restrictions, syntax which will never be implemented, etc
.oO( one is already bad enough )
10:14 berekuk left
Timbus huh.. android java compiles to normal java bytecode, and then the dex compiler converts the bytecode to android dex. you might be mistaken 10:17
10:17 berekuk joined
Timbus i wrote my android app in scala, so 10:18
vendethiel (same here) 10:19
10:31 berekuk left 10:34 rurban joined, Rotwang left 10:38 rurban left 10:40 anaeem1 left, igorsutton_ joined 10:41 igorsutton_ left 10:42 [Sno] left, [Sno] joined 10:44 _daniel-s__ joined 10:45 daniel-s__ left, denis_boyun___ left 10:46 thou left 10:50 kurahaupo left 11:02 molaf_ left 11:03 anaeem1 joined 11:08 brrt joined 11:11 xinming_ joined
masak ooh, writing an Android app in Scala. 11:14
11:14 xinming left
masak that sounds like it could actually be potentially pleasant! 11:14
11:15 virtualsue joined, virtualsue left, Rotwang joined 11:16 Rotwang left, denis_boyun_ joined 11:29 virtualsue joined
dalek p: eafcab0 | jnthn++ | src/vm/moar/QAST/QASTOperationsMAST.nqp:
Start using new getlexstatic_o op.
 11:33
p: 43fbe2b | jnthn++ | src/NQP/Optimizer.nqp:
Fix NQP analysis of when callstatic is valid.

Makes it match Rakudo's more closely. We can cheat in NQP, but not
  /that/ hard.
p: c9e9e90 | jnthn++ | src/vm/moar/stage0/ (11 files):
Update stage0 for MoarVM.

The previous bootstrap contained mis-generated code that upset a more aggressive spesh. Furthermore, the generated code is far better these days. Between spesh and the better stage0 (meaning we build stage 1 faster), the NQP build is now down to 41s (from around 50s) for me.
p: 7d6bcb2 | jnthn++ | / (6 files):
Merge remote-tracking branch 'origin/master' into lexopts
p: 25d9faa | jnthn++ | tools/build/MOAR_REVISION:
Bump to latest Moar for improved spesh, new lex op
kudo/nom: a8942e4 | jnthn++ | tools/build/NQP_REVISION:
Get latest NQP/MoarVM, with various optimizations.
 11:34
11:35 rurban joined 11:41 rurban left 11:55 thou joined
FROGGS jnthn: do you have a clue about the thing I pasted earlier? (I highlighted you) 11:59
12:00 thou left 12:07 anaeem1 left
jnthn FROGGS: Well, just 'cus the thing you have is a P6opaque doesn't mean it's a container 12:12
It looks like you're trying to cast something to a VMException that you aren't allowed to assume is one.
tadzik timotimo: when is p6weekly due? 12:18
12:21 brrt left
moritz tadzik: usually Monday, no? 12:27
FROGGS jnthn: yeah, makes sense 12:29
12:29 dalek left, dalek joined, ChanServ sets mode: +v dalek
timotimo yup, monday 12:32
tadzik ah, okay 12:33
plenty of time to release Space Invaders then :)
colomon any recent changes? (says the man who was not able to get it to work....) 12:34
FROGGS jnthn: can I dump the SixModelObject in uwex.result to get some information what is in there? 12:38
jnthn Not easily...I mean, you can look at its fields in the debugger I suess 12:40
*guess
You might be able to find something that tries to obtain the typename somewhere else in Ops.java
Like the method dispatch error
FROGGS k, will have a look for that... thank you :o)
typeName() it is 12:42
tadzik colomon: well, I only added xbox buttons to the menu 12:43
FROGGS it is an Int
tadzik and I think it's release-ready :)
12:45 pecastro left
FROGGS hmmm, and the Int's value in .result is 2 12:56
ahh *facepalm* 12:57
it has a block handler, not an unwind handler... 12:58
jnthn: am I allowed to give UnwindException a payload attribute? 12:59
12:59 dalek left
jnthn FROGGS: To carry the label? 12:59
13:00 dalek joined, ChanServ sets mode: +v dalek
FROGGS yes 13:00
because .result is already used for block handlers
jnthn Yeah...
I don't immediately see a problem with doing that.
FROGGS cool!
13:12 rurban joined 13:15 guru joined 13:16 guru is now known as Guest66359, Guest66359 is now known as ajr_
colomon I won! \o/ 13:21
tadzik: Had to do DYLD_LIBRARY_PATH=. ./SpaceInvaders to make it work. 13:22
tadzik hmm 13:26
13:29 grep0r left
FROGGS I think stuff like rpath can prevent that 13:29
13:32 grep0r joined 13:35 sqirrel joined 13:44 molaf joined 13:45 grep0r left 13:48 Pleiades` left 13:49 grep0r joined
colomon the problem seems to be the in SDL's PATH 13:51
take out the ./ and it just works
SDL.pm line 2, I mean 13:53
13:57 Pleiades` joined
tadzik oh! 13:58
13:59 rindolf joined
colomon btw, I can confirm the gamepad interface works under OS X. ;) 14:03
lizmat good *, #perl6 14:04
14:05 zakharyas joined
colomon \o 14:07
lizmat hmmm. I just pulled, completely rebuild, and the spectest is showing quite a few failures *and* is much slower :-( 14:08
Files=857, Tests=32597, 216 wallclock secs ( 8.29 usr 3.90 sys + 1349.56 cusr 160.93 csys = 1522.68 CPU)
t/spec/S04-declarations/will.t only shows 7 tests when run with make 14:09
but if I run the fudged .rakudo.moar, it completes
timotimo much slower? =o 14:12
well, we do specialize huge amounts of invocations now but only put in logging instructions rather than doing actual improvements 14:13
and spesh now does a fixed-point-iteration instead of one pass for eliminating dead code in specialization
which results in both better specialized code and longer specializer run time
tadzik colomon: nice :) 14:14
dalek p/loop_labels: 51b5ab0 | (Tobias Leich)++ | src/vm/jvm/ (4 files):
implement label handling for rakudo's 'for' loops (nqp::handle)
 14:15
woolfy1 Wow ARM support done by ncw10++ ++ ++ 14:22
14:23 gtodd joined
jnthn lizmat: The spectest timing is likely worse 'cus we optimize harder in various places now. In MoarVM and Rakudo's optimizer is doing a little extra work too. 14:24
lizmat: The flip side is we can do better in benchmarks for code that involves hot paths. 14:25
lizmat I can live with the worsening if it was to be expected 14:26
strange errors that weren't there before however :-) 14:27
jnthn Well, the other thing we can do is tweak various thresholds.
Oh?
Nobody else reproted those :(
What fails, exactly?
lizmat t/spec/S04-declarations/will.t 14:28
FROGGS I don't see it failing
lizmat somewhere after test 7 it dies when run under test harness
FROGGS lizmat: btw, there are also more advent calender related test files... so maybe it got slower due to that too 14:29
lizmat perhaps 14:30
14:30 Rotwang joined
jnthn Well, the S17 tests seem the big slowdown for me... 14:31
lizmat well, I can try to reduce the length of the sleeps, but then the chance of it failing under load increase 14:32
BTW, it's also the CPU that has increased significantly, not just wallclock 14:33
one from yesterday: Files=857, Tests=32624, 211 wallclock secs ( 8.05 usr 3.89 sys + 1311.61 cusr 152.84 csys = 1476.39 CPU) 14:34
1476 vs 1522 CPU 14:35
dalek kudo-star-daily: 45cf8a9 | coke++ | log/ (5 files):
today (automated commit)
jnthn lizmat: Well, as I mentioned before, I think we should probably write a TestScheduler where we write the tests in terms of virtual time. 14:36
lizmat: That is, the scheduler has an internal "clock" that we control
FROGGS damn, one label test fails when I run m-spectest, but passes when I run it directly... valgrind shows nothing :/ 14:37
jnthn lizmat: That way the tests can be fast *and* robust under load
lizmat jnthn: I know we discussed this, but to be honest, I don't see how we could get that to work
jnthn lizmat: Because all the time-based things we have in Supply should be taking a :$scheduler argument. 14:39
my $ts = TestScheduler.new; my $seen = 0; my $s = Supply.new; $s.delay(1, :scheduler($ts)).tap({ $seen = $_ }); $s.more(1); nok $seen; $ts.advance_by(1); ok $seen; 14:40
Rx has prior art 14:41
lizmat ok, so first point of order is to have all Supply methods have a :scheduler parameter 14:42
14:42 spider-mario left
jnthn All those that schedule, yes 14:43
14:43 Alina-malina left
jnthn So for example, map/grep don't need one 14:43
As they're pure and filtery
But uniq should 'cus of :expires, for example. 14:44
lizmat well, not really, as :uniq doesn't use any timing
jnthn Oh?
lizmat *:expires
timotimo oh btw
jnthn Ah, right...
lizmat it simply checks the time last seen
jnthn Yeah, forget how it was implemented. :)
timotimo may i suggest we make grep and other filtery ops on supplies take Mu instead of Callable?
14:45 sqirrel left
timotimo because smart match semantics are way cooler than only callables as filters 14:45
jnthn timotimo: Should be doable
timotimo maybe keep a Callable candidate around for optimization purposes if that makes sense
lizmat timotimo: will look at that also
14:48 Alina-malina joined, SamuraiJack_ left
timotimo Tack. 14:48
jnthn lizmat++ 14:49
dalek kudo/loop_labels_test: 9515682 | (Tobias Leich)++ | src/core/PseudoStash.pm:
enable Label.next/.redo/.last for jvm
timotimo yay
dalek ecs: 2281be7 | (Elizabeth Mattijsen)++ | S17-concurrency.pod:
A better example of signal().tap usage
 14:54
14:55 SamuraiJack_ joined 15:01 Alina-malina left 15:02 Alina-malina joined
timotimo m: say (@()).perl 15:04
camelia rakudo-moar a8942e: OUTPUT«().list␤»
15:07 ivanshmakov left
FROGGS seems legit 15:08
15:08 ivanshma` joined
lizmat $ perl6 src/core/SupplyOperations.pm 15:09
Segmentation fault: 11
wow
15:12 hoverboard joined 15:15 SamuraiJack_ left 15:16 hoverboard left, ajr_ left, guru joined 15:17 ivanshma` is now known as ivanshmakov, ivanshmakov left, ivanshmakov joined, guru is now known as ajr_
lizmat jnthn: fwiw, fixed the segfault by fixing wrongly named variables 15:21
jnthn Those probably shouldn't segfault...
lizmat indeed, not, they shouldn't
timotimo may want to rebuild a very latest moar with a fix for a dumb mistake i made? 15:22
lizmat: can you time the spectests again with MVM_SPESH_DISABLE=1 in your environment and see if it makes a big difference? 15:27
lizmat will do in a mo 15:33
now spectesting Supply.grep(Mu $test) :-) 15:34
timotimo yay :)
15:43 chenryn joined
FROGGS p: 1 ?? Q:PIR { say 42 } !! Q:PIR { say 21 } # :/ 15:48
camelia rakudo-parrot a8942e: OUTPUT«===SORRY!===␤Error while compiling op if (source text: "?? Q:PIR { say 42 } !!"): Cannot infer type from ''␤»
dalek kudo/nom: 32f0e40 | (Elizabeth Mattijsen)++ | src/core/Supply (2 files):
Supply.grep now allows simple smartmatch semantics
 15:49
ast: a92ff7c | (Elizabeth Mattijsen)++ | S17-supply/grep.t:
Add tests for Supply.grep(smartmatch)
jnthn FROGGS: Yeah, say is a void pirop and if you put it in a context that wants a result...boom like that. 15:51
15:52 darutoko- joined
jnthn Time for a walk, then I'll take a look at dealing with the n**2 behavior in CORE.setting compilation that's costing us a bunch... 15:52
timotimo oooh nice
15:53 darutoko left 15:58 berekuk joined
lizmat timotimo: Files=857, Tests=32634, 207 wallclock secs ( 8.16 usr 3.88 sys + 1290.51 cusr 148.27 csys = 1450.82 CPU) 15:59
quite a significant difference
and fewer failures
timotimo oh my :| 16:01
FROGGS jnthn: yeah, I don't bother much about Q:PIR in a ternary, since if/else does work 16:02
16:05 Rotwang left
JimmyZ lizmat: Do you have a number before the merging ? 16:05
lizmat Files=857, Tests=32624, 211 wallclock secs ( 8.05 usr 3.89 sys + 1311.61 cusr 152.84 csys = 1476.39 CPU) e.g. 16:06
dalek ecs: 63266bc | (Elizabeth Mattijsen)++ | S17-concurrency.pod:
Some more about Supply.grep
 16:08
JimmyZ sounds nice :P 16:17
lizmat github.com/rakudo/star/downloads is now empty 16:18
timotimo good
16:19 crazedpsyc is now known as crazedpsyc`, crazedpsyc` is now known as crazedpsyc
FROGGS lizmat++ 16:19
16:19 nnunley joined 16:25 chenryn left 16:26 Psyche^_ joined
lizmat jnthn: a method such as Supply.elems *does* take a $seconds, but has not need for a Scheduler 16:27
the only way to make that one faster, is to mess with time()
same with Supply.batch 16:28
(in its current implementation)
16:30 Psyche^ left 16:37 denis_boyun_ left
jnthn lizmat: What things fail for you in spectest with spesh turned on? 16:43
lizmat jnthn: not running S17-tests saves me about 30 seconds wallclock 16:48
lizmat runs spectest again with spesh 16:49
perhaps we should add a flag to not run #conc tests ?
16:50 Alula left
jnthn lizmat: Well, or maybe mark some as stress tets 16:50
16:50 Alula joined
lizmat well, I didn't add any stress tests really 16:50
the ones that come close, are actually skipped atm 16:51
16:51 virtualsue left
jnthn S17-scheduler/basic.t alone takes 47s to run o.O 16:52
lizmat checks 16:53
jnthn 20s for S17-promise/basic.t
lizmat yup, that's all the :at, :in, :every tests 16:54
jnthn lizmat: Maybe could have a in.t, at.t, every.t 16:56
lizmat sure 16:57
that's on my list of things to do
S17-supply just took a little longer to split
I'll split the scheduler and promise tests up 16:58
wrt to failing test, jnthn:
t/spec/S17-supply/uniq.t
Non-zero wait status: 11 16:59
Parse errors: Bad plan. You planned 83 tests but ran 66.
only fails when run under make, not when called with perl6 :-( 17:00
it's using the CurrentThreadScheduler at that time, oddly enough 17:01
17:09 btyler joined
dalek ast: 8d53b52 | (Elizabeth Mattijsen)++ | S17-promise/ (6 files):
Split Promise tests into separate files

For better parallel testing		 17:14
jnthn lizmat++
timotimo nice. 17:17
17:19 guru joined, guru is now known as Guest21672 17:20 ajr_ left, thou joined
dalek ast: e8ed156 | (Elizabeth Mattijsen)++ | S17-scheduler/ (3 files):
Create new test files for parallel testing
 17:22
17:24 thou left 17:25 _daniel-s__ left 17:26 _daniel-s__ joined 17:29 virtualsue joined 17:39 Guest21672 is now known as ajr_ 17:43 ajr_ left 17:48 zakharyas left
dalek ast: 4d4e22d | (Elizabeth Mattijsen)++ | S17-promise/start.t:
Make sure we don't get "out of order" errors
 17:55
kudo/nom: a3be5eb | (Elizabeth Mattijsen)++ | t/spectest.data:
Make sure we do all Promise tests
 17:56
17:56 kurahaupo joined 18:02 plobsing joined 18:10 thou joined
tadzik I blag: Polish Perl Workshop 18:26
bleh
mispaste :) ttjjss.wordpress.com/2014/05/11/new...-invaders/ 18:27
dogwalk &
18:30 kurahaupo left
FROGGS tadzik++ 18:30
18:31 kurahaupo joined 18:34 anaeem1 joined
dalek ast: b2d8100 | (Elizabeth Mattijsen)++ | S17-scheduler/ (5 files):
Spread out Scheduler tests

For better parallel testing		 18:37
18:38 igorsutton_ joined 18:39 Sqirrel joined
dalek kudo/nom: 9ca7c73 | (Elizabeth Mattijsen)++ | t/spectest.data:
Make sure we do all Scheduler tests
 18:42
lizmat hope this helps people spectesting on Win 18:43
18:53 darutoko- left
dalek ast: 56b0c83 | (Elizabeth Mattijsen)++ | S17-channel/winner.t:
Comment readability fix
 18:53
18:58 thou left
dalek kudo/nom: 5c160e7 | (Elizabeth Mattijsen)++ | t/spectest.data:
Add advent2010/day(6,7,8,10) tests
 19:04
lizmat dwarring++
19:09 dwarring joined 19:10 thou joined 19:16 dwarring left
nwc10 dear git experts, why does something keep deciding to delete src/gen/IGNOREME ? 19:18
git clean -dXf doesn't do this on any other platform, just the power box
src/gen is in .gitignore
which seems to be something to do with the problem 19:19
19:21 anaeem1 left
masak nwc10: the -X flag removes files ignored by Git. 19:29
nwc10: if you don't want that, don't provide -X or -x
nwc10 masak: but I'm doing exactly the same on other machines, and it doesn't get deleted 19:30
masak nwc10: if you want to track rather than ignore IGNOREME, don't have a .gitignore rule that ignores it :)
nwc10 masak: this is MoarVM, and the problem is src/gen/IGNOREME is a file in the repository
masak nwc10: if you want src/gen/ to be a tracked empty directory, may I suggest putting an empty .gitignore inside it and tracking that file?
nwc10 git clean -dXf is merrily deleting a tracked file
masak: well, this is actually jnthn's doing :-) 19:31
masak nwc10: I suspect you'll find that that particular box has an old Git.
nwc10 but the bit I dont't get is why I only get this happening on one machine
masak nwc10: this suddenly reminds me of a bug that I found once, but that since has been fix'd.
jnthn A .gitignore is probably better than an IGNOREME
FROGGS \o/ I got my tables in the pause/cpan database! Now I can write The Perl 6 Indexer™! \o/ 19:34
19:34 anaeem1 joined
jnthn FROGGS++ \o/ 19:34
masak woot
timotimo \o/
FROGGS attaching panda to it shouldn't be too hard then :D 19:35
tadzik niiice! 19:36
FROGGS++
FROGGS ANDK++, really :o)
19:38 anaeem1 left 19:39 sorear left 19:41 molaf_ joined 19:44 molaf left 19:48 dwarring joined
dwarring r: my @m <== sort <== ("c", "b", "a") 19:50
retupmoca github.com/rakudo/rakudo/pull/268 please :)
camelia rakudo-jvm 5c160e: OUTPUT«(timeout)»
..rakudo-{parrot,moar} 5c160e: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Unable to parse expression in quote words; couldn't find final '>'␤ ␤at /tmp/tmpfile:1␤------> my @m <== sort <== ("c", "b", "a")⏏<EOL>␤ exp…»
dwarring that's golfed from from 2010 advent day 10 19:51
looks to me like that's getting confused with with quoted-word syntax..
moritz m: my @m <== sort() <== ("c", "b", "a")
camelia ( no output )
dwarring r: say <== sort>
camelia rakudo-{parrot,jvm,moar} 5c160e: OUTPUT«== sort␤»
dalek kudo/nom: dd4a0e6 | (Andrew Egeler)++ | src/core/IO.pm:
Check for filename length instead of truthiness.

Otherwise, if a directory contains a file named '0', we assume no more files exist (while there could still be any number left)		 19:52
kudo/nom: c1845e2 | (Tobias Leich)++ | src/core/IO.pm:
Merge pull request #268 from retupmoca/nom

Check for filename length instead of truthiness.
timotimo oooh, good catch
moritz dwarring: yes; it expects a term after a listop
retupmoca FROGGS++ 19:53
FROGGS retupmoca++ # :P
moritz std: my @m <== sort <== ("c", "b", "a")
camelia std ec3d853: OUTPUT«ok 00:01 127m␤»
moritz dunno how STD manages it
probably "LTM" is the answer :-) 19:54
FROGGS std: <== sort>
camelia std ec3d853: OUTPUT«ok 00:01 121m␤»
FROGGS yeah
LTM FTW
masak huh. STD has the cake and eats it somehow. 19:57
dwarring starts looking at std 19:58
timotimo m: class Foo is rw { has @.test }; my Foo $f .= new; my @oldtest := $f.test; $f.test := @(); $f.test.push: "hi"; say $f.test; say @oldtest; 19:59
camelia rakudo-moar 5c160e: OUTPUT«===SORRY!=== Error while compiling /tmp/A2lAUPImPR␤Cannot use bind operator with this left-hand side␤at /tmp/A2lAUPImPR:1␤------> ; my @oldtest := $f.test; $f.test := @()⏏; $f.test.push: "hi"; say $f.test; say @…»
timotimo m: class Foo is rw { has @.test }; my Foo $f .= new; my @oldtest := $f.test; $f.test = @(); $f.test.push: "hi"; say $f.test; say @oldtest;
camelia rakudo-moar 5c160e: OUTPUT«hi␤hi␤»
timotimo how do i ...
20:00 grep0r left 20:05 grep0r joined, xinming_ left 20:15 btyler left
timotimo i should probably figure out how to create a new object at that point or something. 20:15
20:23 grep0r left 20:28 lichtkind joined
lichtkind soory for repeating me but was there another reason why now the jvm port was released 20:29
vendethiel I don't understand that question :/
FROGGS lichtkind: there are still a few module test failures 20:30
lichtkind: see gist.github.com/FROGGS/0ea5537eb675588baaa2
vendethiel oh, why was not ?
jnthn lichtkind: It's been in compiler releases for months and months.
lichtkind: Just not in the Star release.
FROGGS lichtkind: it looks worse than it is though, because when a dep fails, others fail too
jnthn Mostly due to nativecall and other bits taking longer than the initial compiler work. 20:31
20:31 kaare_ left
FROGGS lichtkind: you can still build it from latest star or latest compiler release, but not every module works 20:31
20:31 kaare_ joined, denis_boyun_ joined
jnthn Partly because it just turned out more efficient on development resources (which we're rather short on) to do the R* multi-backend work for JVM and Moar at the same time. 20:31
FROGGS ahh, that reminds me that my .msi's needs testing/uploading 20:32
lizmat r: say $*OUT.VAR.dynamic
camelia rakudo-{parrot,jvm,moar} c1845e: OUTPUT«False␤»
lizmat jnthn: shouldn't that be True ? 20:33
jnthn lizmat: Probably not
lizmat: There is not $*OUT declared anywhere
lizmat: It fails over to looking in PROCESS
(Which contains perfectly ordinary variables.)
lizmat aha, ok
FROGGS gnight all o/
lizmat so even if we define our own $*foo (which would have .dynamic True) 20:34
this would not be true for stuff in PROCESS::
FROGGS: gnight!
20:35 grep0r joined
jnthn Yeah, .VAR is about the variable we actually reach, not the way we looked it up 20:35
lichtkind lizmat: hope to show you praha 20:36
lizmat r: PROCESS::<$a> = "foo"; say $*a # indeed
camelia rakudo-{parrot,jvm,moar} c1845e: OUTPUT«foo␤»
lichtkind will arrive sa
lizmat lichtkind: will arrive mo evening, driving down from Poznan 20:37
lichtkind lizmat: you have my mobil? 20:38
20:42 igorsutton__ joined, igorsutton_ left
retupmoca does anyone in here have a commitbit to the November repo? (github.com/viklund/november/pull/28) 20:46
looks like viklund hasn't been active in...a whie
jnthn masak, surely...
retupmoca while*
looking at the commit log, moritz may as well 20:48
anyway, that PR makes november build/test on latest rakudo-moar
20:48 kaare_ left 20:51 rindolf left 20:56 virtualsue left
masak retupmoca: looking. 20:59
retupmoca: merged. 21:01
retupmoca: do you know how github.com/viklund/november/pull/27 relates to your effort? it seems it can't be merged automatically after your fixes?
moritz remembers that novmber existed, bu not much more. 21:05
retupmoca oh, I didn't even bother to look at other pending PRs 21:09
anyway, masak++ 21:15
I can make November use URI instead of it's own thing if you want/need it, but mostly I just wanted it installable
(I'm annoyed by all the red in colomon++'s ecosystem tests, so I'm trying to make them green) 21:16
masak: also, I have github.com/masak/csv/pull/10 for you if you're in a PR-merging mood :)
colomon retupmoca++
21:17 igorsutton__ left, igorsutton__ joined 21:19 kivutar left
retupmoca I have a few other PR's out as well. If nobody picks them up within a reasonable time, I assume I should point panda to my fixed version instead? 21:19
also, what guidelines do we have (if any) for removing things from the ecosystem? (I want to remove Enc::MIME::Base64, for example) 21:21
(it doesn't work, and the functionality it provides is in MIME::Base64)
21:21 igorsutton__ left
retupmoca (or instead of pointing to my fixed version of a module, should we re-point to a perl6-community-modules/foo repo of some kind?) 21:23
21:24 kurahaupo left, raiph joined, denis_boyun_ left 21:42 rurban left, rurban joined 21:47 rurban left 21:50 lichtkind left
dalek kudo/nom: b8567ca | (Elizabeth Mattijsen)++ | src/core/Deprecations.pm:
Add system var deprecation capability

And deprecate $*OS.		 21:58
ast: 5e70780 | (Elizabeth Mattijsen)++ | S02-types/deprecations.t:
Add test for $*OS deprecation
 21:59
lizmat this can probably be done better, but it is a start
colomon retupmoca: +1 for pointing things to perl6-community-modules/foo repo and getting deadish modules out of the ecosystem 22:02
22:05 xinming joined
flussence perl6: sub a(&b($, $)) { say b(1, 2) }; sub add($x, $y) { $x + $y }; a(&add) 22:07
camelia niecza v24-109-g48a8de3: OUTPUT«Unhandled exception: Unable to resolve method Capture in type Sub␤ at <unknown> line 0 (ExitRunloop @ 0) ␤ at /tmp/tmpfile line 0 (a @ 1) ␤ at /tmp/tmpfile line 1 (mainline @ 5) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4595 (ANON @ 3…»
..rakudo-{parrot,moar} c1845e: OUTPUT«Not enough positional parameters passed; got 0 but expected 2 in sub-signature of parameter &b␤ in sub a at /tmp/tmpfile:1␤ in block at /tmp/tmpfile:1␤␤»
..rakudo-jvm c1845e: OUTPUT«Flattening named argument must have VMHash REPR␤ in sub a at /tmp/tmpfile:1␤ in block at /tmp/tmpfile:1␤␤»
flussence is that just a NYI, or am I writing the bit between a(...) completely wrong? 22:08
22:08 rurban joined 22:12 labster left, xinming left 22:17 _daniel-s__ left, daniel-s joined 22:20 labster joined
timotimo t.h8.lv/p6bench/2014-05-12-spesh_lo...stuff.html - feel free to be the judges of whether or not spesh_log was a good merge to do or not :) 22:22
since this also includes improvements like the assign_key and assign_pos changes for hashes and arrays ... hmm 22:26
not sure how much of that is spesh_log and how much is the rest
t.h8.lv/p6bench/2014-05-12-spesh_lo...artup.html - for some reason i like this one better 22:28
22:29 daniel-s left, daniel-s_ joined
timotimo kinda sad ... and our benchmarks are still not very good >_> 22:30
as in: not very representational
huh, forest-fire didn't run? 22:31
dalek osystem: e170031 | (Andrew Egeler)++ | / (2 files):
MessagePack has its own META.info
 22:34
timotimo whoever wants to can take a look at why forest fire doesn't seem to have any spreading fire any more in the perl6-bench repo
please.
22:35 daniel-s_ left
skids Speaking of PRs, what exact repos/branches should I send PRs to for the nqp/parrot and moarvm bigint bitops fixes? 22:36
retupmoca skids: I would guess perl6/nqp branch master and moarvm/moarvm branch master 22:37
timotimo bigint ops are in nqp src/vm/parrot/ops/bigint_ops.ops or something
and the bigint ops of moarvm are inside moarvm
22:37 daniel-s_ joined
skids Right I have to patch nqp and moarvm. 22:38
Hopefully I'll not have to build an entire ecosystem to test, and can get away with pulling them into star's build. 22:39
timotimo aye :)
m: say (1, 2, 3, 4) ~~ 3..10; say (1, 2, 3, 4, 5) ~~ 1..4; 22:42
camelia rakudo-moar b8567c: OUTPUT«False␤True␤»
timotimo m: my @foo where 3..4; @foo = 1, 2, 3; @foo = <a b c>; @foo = <x y z a s d f g>;
camelia ( no output )
timotimo m: say <1 2 3 4 5>.^methods 22:43
camelia rakudo-moar b8567c: OUTPUT«BUILD Parcel Capture elems item flat list lol reverse rotate at_pos STORE FLATTENABLE_LIST FLATTENABLE_HASH fmt of Bool Numeric Str ACCEPTS gist perl DUMP␤»
22:45 rurban left
timotimo m: say "foobar baz".flip 22:45
camelia rakudo-moar b8567c: OUTPUT«zab raboof␤»
timotimo m: say <foobar baz>.flip
camelia rakudo-moar b8567c: OUTPUT«zab raboof␤»
timotimo m: say <foobar baz>.reverse
camelia rakudo-moar b8567c: OUTPUT«baz foobar␤»
22:46 rurban joined
lizmat gnight #perl6! 22:47
timotimo gnite lizmat 22:48
22:51 rurban left
dalek osystem: 9e383fb | (Andrew Egeler)++ | META.list:
Remove Enc::MIME::Base64

This looks like an old, unmaintained, not-working version of MIME::Base64		 23:04
23:15 rurban joined 23:17 rurban1 joined 23:20 rurban left 23:22 rurban1 left 23:33 xinming joined 23:38 dmol left 23:39 berekuk left 23:59 xragnar_ joined, xragnar left, xragnar_ is now known as xragnar

Powered by Cro and the Raku® Programming Language.

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

Thank you!