sorear good * #perl6 00:01
itz: hi! 00:03
dalek p/bs: 9b2a92b | jnthn++ | src/6model/reprs/P6bigint.c:
Seems P6bigint.copy_to passed the arguments to mp_init_copy backwards (not helped by the fact that it takes them the opposite way around to mp_copy). Sadly, this alone doesn't seem to fix the heap corruption.
00:09
00:25 att joined
dalek p/bs: 3553834 | jnthn++ | src/HLL/World.pm:
Code-gen building of list of static code refs in a different Parrot sub, to reduce load on IMCC register allocator a little.
00:34
00:42 mattp_ left 00:45 mattp_ joined 00:49 thou left 00:52 BooK left 01:00 BooK joined
dalek kudo/bs: fb6bc7d | jnthn++ | src/ (5 files):
Re-work the way we handle the fresh magicals flag, moving it into an int in the StaticLexPad. Saves several thousand lines of generated PIR in the fixups in CORE.setting, and probably around a thousand PCC method calls (yeah, the slow ones...) on startup.
01:05
01:08 woosley joined, scott_ joined 01:09 scott_ is now known as Guest61033
jnthn sleep & 01:15
TimToady masak: a slightly interesting monadic-ish logic toy in P6: gist.github.com/1946473 01:36
dinner &
01:38 benabik joined 01:44 wolfman2000 left 02:20 wolfman2000 joined 02:35 NamelessTee left 02:51 jaldhar joined
colomon ack, just realized I missed #phasers again 02:54
(and I will probably be more or less out of contact for the next three days, for good reasons. :) 03:00
dalek ast: d26054b | coke++ | S02-types/array.t:
pugs fudge
03:01
gs.hs: 0eb2863 | coke++ | t/spectest.data:
run fudged test
sorear o/ colomon 03:03
03:17 orafu left, orafu joined 03:23 odoacre left 03:41 sftp left 03:46 odoacre joined 04:10 noam__ left 04:21 Minthe joined 04:33 arnsholt_ joined, arnsholt left 04:47 alester joined 04:56 birdwindupbird joined 05:12 birdwind1pbird joined 05:13 birdwind1pbird left, birdwind1pbird joined, birdwind1pbird left, birdwind1pbird joined 05:37 am0c joined 05:44 birdwindupbird left 05:47 packetknife joined 05:51 thou joined 05:55 robinsmidsrod left 05:57 robins joined 06:02 alester left 06:07 havenn joined 06:11 am0c left
moritz \o 06:25
sorear o/ moritz
dalek ast: 844bcb2 | moritz++ | S03-operators/buf.t:
tests for Buf.subbuf
06:26
06:28 havenn left 06:29 havenn joined 06:34 kaleem joined 06:37 marmalade left
fsergot o/ 06:45
sorear hello fsergot! 06:46
fsergot: until the other day I always parsed your nick as something like "This is your /home on acid"
fsergot sorear:I thought about remove the first character. :-D 06:50
06:51 takesako left 06:52 kaleem left 06:53 takesako joined
fsergot sorear: Where are You from? 06:54
06:54 am0c joined
sorear fsergot: USA, southwest 06:56
06:57 kaleem joined, am0c left
Teratogen sorear, you in Arizona? 06:57
06:57 havenn left
TimToady sorear is elsewhere 06:57
Teratogen hi TimToady 06:58
TimToady I don't think sorear would want to live with that many rednecks
sorear "he prefers bluenecks" 06:59
Teratogen Actually Arizona has a lot of wetbacks
sorear I am not familiar with that term
Teratogen Urban Dictionary to the rescue!
TimToady yes, well, it's okay for white folks to call themselves rednecks, but it's not really okay to call a different ethnic group names 07:00
07:01 kaleem left
TimToady in general we should try to avoid denigrating terms, and in fact I apologize for using "rednecks" that way 07:02
Teratogen I wouldn't consider Arizona to be a redneck state, though. 07:03
Indiana is a redneck state, definitely. 07:04
07:12 wtw joined
fsergot Have to go to the school. Will be back later. Have a nice day|night #perl6! :) 07:14
07:27 Kharec joined
Kharec hi there 07:27
07:28 autark left
Kharec Is this time of exec is normal? paste.debian.net/158146/ 07:29
07:38 Kharec left 07:41 packetknife left
TimToady my system runs in in under a second, but that's still slow; jnthn's new serializing branch should help with that 07:44
07:45 rsimoes left
Teratogen nom: say 1+1 07:45
TimToady (I'm running the most recent release, so that's probably some improvement over November)
p6eval nom 4726db: OUTPUT«2␤»
TimToady in general, we'd recommend p6 more for longer running processes until we get the startup time lower 07:47
oh, wait, Kharec left already, duh... 07:48
08:01 rsimoes joined 08:09 brxue joined 08:10 TEST_ joined
TEST_ say 20 + 5 08:11
rakudo: say 29*3
p6eval rakudo 4726db: OUTPUT«87␤»
08:12 TEST_ left 08:14 brxue left 08:23 noam joined 08:24 pat_js joined
pat_js hey a quick question for building rakudo: is there already a way to pass the make '-j num' option to the parrot-build? 08:32
08:35 mj41 joined
sorear I don't know offhand (not really a rakudo person), but you can always build Parrot separately, which is what I did back when I could build rakudo 08:41
08:41 thou left 08:42 bkolera joined
pat_js ah ok thanks, was just examining that Configure-script, and it doesn't seem like. 08:44
08:46 stepnem joined
pat_js what does .pir stand for? is this the parrot language? 08:47
09:01 Minthe left 09:09 woosley left
sorear pat_js: Parrot Intermediate Representation 09:10
pat_js: don't think of it as a "language", except in a very technical sense
pat_js: it's more like the .s files that gcc creates in /tmp during compilations 09:11
pat_js ah ok 09:14
masak hola, #perl6. 09:16
Teratogen: why do you keep doing 'say 1 + 1' on p6eval?
tadzik hello #perl6
masak Teratogen: why not something cool, like 'class A { method foo { say "HAI" } }; class B is A { method foo { print "OH "; nextsame } }; B.new.foo'? 09:17
tadzik! \o/
tadzik \o/ 09:18
sorear \o/ 09:19
pat_js: I often see people talking about "Parrot" as if it were a language. 09:20
09:21 pat_js left
Teratogen because if say 1 + 1 works, it is ready to ship! =) 09:22
tadzik wow, rakudo builds pretty fast now
masak Teratogen: 'say 1 + 1' has worked since 2005.
Teratogen: and we *are* shipping both Rakudo and Niecza, since 2007 and 2011, respectively. 09:23
sorear strike 2. 09:25
masak sorear: um, don't do that, it's uncanny :/ 09:27
09:27 arnsholt_ is now known as arnsholt
sorear "suggesting the operation of supernatural influences"? 09:29
I am trying to make it clear to Teratogen that eir unfunny inanity is very, very tiring
and direct words don't seem to have helped
masak sorear: I'm just expressing my uneasiness at the indirect method. 09:30
sorear maybe I'll try super-direct.
hey, Teratogen, you're making it less fun to develop Perl 6
masak it feels more like you're justifying to yourself the right to boot him after three strikes.
yes, that was better.
sorear please stop it
masak Teratogen: I agree with sorear. your views on "ship it" lately are one-sided and not constructive. 09:31
Teratogen ok =( 09:32
masak write something cute in Perl 6 instead. there are lots of little applications that haven't been written yet. 09:37
it was the "uncomfortably strange" meaning of "uncanny" I was after. 09:42
jnthn morning, #perl6
tadzik morning jnthn 09:43
Files=616, Tests=22111, 900 wallclock secs ( 9.59 usr 1.84 sys + 1512.89 cusr 127.42 csys = 1651.74 CPU)
nom: 900 / 1050
p6eval nom 4726db: ( no output )
tadzik nom: say 900 / 1050 09:44
p6eval nom 4726db: OUTPUT«0.857142857142857␤»
tadzik 86% of what it was yesterday
09:44 noam left
jnthn tadzik: bs got faster since your last run? 09:46
tadzik jnthn: si
jnthn Oh, I guess it may be thanks to fb6bc7d49e. Nice.
tadzik I was checking because of github.com/rakudo/rakudo/commit/fb6bc7d49e
aye
jnthn nom: say 191 / 199 09:47
p6eval nom 4726db: OUTPUT«0.959798994974874␤»
jnthn Curious. It was a bigger difference for you than for me.
I mean, proportionally
Do you do a serial or parallel spectest run?
tadzik pararell, 3 jobs 09:49
masak jnthn! \o/ 09:50
jnthn masak! \o/
tadzik afks 09:53
09:55 dakkar joined 09:59 tty234 left, etneg left
masak TimToady: gist.github.com/1946473 -- nice 10:02
still no amb, though ;)
10:09 Guest61033 left
dalek p/bs: 47ed8e9 | jnthn++ | src/HLL/World.pm:
Don't need to natter about how much data we serialized in the build output now. :-)
10:10
10:10 noam joined
moritz wakes up a second time today, and feels much better 10:15
jnthn o/ moritz 10:16
masak good re-morning. 10:17
moritz \o/ all tests pass on rakudo/bs
spectests, that is
felher Nice :D 10:18
masak what? that's awesome!
jnthn++!
felher yeah, jnthn++ indeed 10:19
jnthn Yeah, for once the second 80% wasn't so bad. :)
moritz :-)
1083MB memory used for compiling the setting 10:23
and 'make podtest' also passes 10:24
jnthn yay
moritz t/02-embed/01-load.t aborts silently after printing the paln 10:25
*plan
jnthn Does that work in master?
*nom
Also, is not everything in t\pmc full of it... 10:26
moritz not ok 1 # TODO embedding
# Failed (TODO) test 'can load_bytecode perl6.pbc'
# unimplemented key type 1
is the output on nom
dalek kudo/bs: ff75374 | jnthn++ | t/pmc/ (7 files):
Toss a bunch of tests that test stuff that doesn't exist any more.
10:27
felher "A try block by default has a CATCH block that handles all fatal exceptions by ignoring them." -- is that why there is a try in TimToadys gist? And if it is (hope that is not a stupid question), what happens to non-fatal exceptions? 10:28
moritz which gist? 10:29
and about non-fatal exceptions 10:30
masak gist.github.com/1946473
moritz we have fail(), which ignores try blocks, because the exception isn't thrown yet
jnthn That gist is cute
10:30 cosimo joined
moritz and we have warnings, which don't interact with CATCH at all, because they are control exceptions 10:30
felher So fatal exceptions are all non-controll exceptions that were thrown? 10:32
moritz yes 10:33
felher moritz: okay, thanks a lot :)
Now that bs passes all spectests, will it be merged into nom? 10:40
jnthn felher: There's a few non-spectest issues to take care of first. One (which I'm looking into now) is that CORE.SETTING doesn't build with --optimize=3 10:44
*.setting
felher jnthn: okay, great :)
felher times his rakduo-build-process to be able to compare it to bs later on. 10:46
jnthn Got it narrowed down to a particular optimization.
moritz a2.sphotos.ak.fbcdn.net/hphotos-ak-...4111_n.jpg a nice little riddle 10:57
11:01 noam left 11:09 noam joined 11:10 ab5tract left
felher moritz: is there a solution i can check as soon as i think i've solved it? 11:11
s/solution/link to a solution/
moritz felher: no. Just tell me what you think the answer is, and I'll confirm it :-) 11:12
felher moritz: okay, great.
felher prints it and runs to catch the bus
bye folks
11:12 ab5tract joined
moritz ciao 11:13
11:17 daxim joined
dalek kudo/bs: 4f557f8 | jnthn++ | src/Perl6/Optimizer.pm:
Correct method call done in optimizer when optimizer private method calls.
11:19
kudo/bs: 13d4b0a | jnthn++ | src/Perl6/Optimizer.pm:
Correct an op signature. Sadly, doens't unbust immediate block inlining optimization, though.
kudo/bs: f173558 | jnthn++ | src/Perl6/ (2 files):
Take care not to attempt to inline blocks inside scopes that were dynamically compiled during BEGIN time; this reified their lexpads, and thus we should not go fiddling with them. We got away with this before, though it was probably never really right.
kudo/bs: 8af0430 | jnthn++ | tools/build/Makefile.in:
Now we can compile the setting with --optimize=3 again.
11:40 awoodland joined 12:06 NamelessTee joined
[Coke] # 02/29/2012 - niecza++ ; pugs (29.59%); rakudo (99.27%) 12:11
"niecza", 20302, 8, 751, 1539, 22600, 23752
"pugs" , 6008, 7, 2766, 729, 9509, 23561
"rakudo", 20154, 28, 605, 1864, 22651, 23762
dalek p/bs: bd82150 | jnthn++ | src/6model/serialization.c:
Add a diagnostic message rather than segfaulting on a closure having a missing outer context pointer.
12:25 Bzek joined 12:31 Trashlord joined 12:36 colomon left 12:38 bluescreen10 joined 12:43 pernatiy left
jlaire moritz: heh, funny puzzle 12:57
dalek kudo/bs: 7b5784b | jnthn++ | src/Perl6/World.pm:
Method to add a folded constant no longer needs to cheat epicly.
13:00
p/bs: d11dc37 | jnthn++ | t/serialization/01-basic.t:
Remove a now-bogus test - it relied on an implementation detail, so was probably a bad idea anyway.
p/bs: 52bf76b | jnthn++ | src/6model/serialization.c:
Seems that ignoring lack of outer_ctx serves us better. It semi-legitimately-ish comes up in some cases, and ignoring those is safe enough for the time being. This fixes pre-compilation of enums (and perhaps some other bits too), and is the last known blocker for merging bs.
moritz \o/
jlaire: did you solve it?
jnthn moritz: Any objections to going for the merge? 13:01
jlaire moritz: yep, and apparently I'm a pre-school child and not a programmer :|
jnthn There may be undiscovered issues, but I suspect we'll find them faster with people using it rather than it being sat in a branch. :)
moritz jnthn: -1 objections, +1 to merging :-) 13:02
jnthn Can I hear your negative objection? ;-)
OK, let's do this!
dalek p: 9ff82b3 | jnthn++ | / (36 files):
Merge branch 'master' into bs
13:04
p: 00c5509 | jnthn++ | src/6model/serialization.c:
Patch from tadzik++ to get bs building on bleed Parrot.
p: 9b2a92b | jnthn++ | src/6model/reprs/P6bigint.c:
Seems P6bigint.copy_to passed the arguments to mp_init_copy backwards (not helped by the fact that it takes them the opposite way around to mp_copy). Sadly, this alone doesn't seem to fix the heap corruption.
p: 3553834 | jnthn++ | src/HLL/World.pm:
Code-gen building of list of static code refs in a different Parrot sub, to reduce load on IMCC register allocator a little.
p: 47ed8e9 | jnthn++ | src/HLL/World.pm:
Don't need to natter about how much data we serialized in the build output now. :-)
p: bd82150 | jnthn++ | src/6model/serialization.c:
Add a diagnostic message rather than segfaulting on a closure having a missing outer context pointer.
p: d11dc37 | jnthn++ | t/serialization/01-basic.t:
Remove a now-bogus test - it relied on an implementation detail, so was probably a bad idea anyway.
p: 52bf76b | jnthn++ | src/6model/serialization.c:
Seems that ignoring lack of outer_ctx serves us better. It semi-legitimately-ish comes up in some cases, and ignoring those is safe enough for the time being. This fixes pre-compilation of enums (and perhaps some other bits too), and is the last known blocker for merging bs.
jnthn hm, little noisy... :)
dalek kudo/nom: c077570 | jnthn++ | / (20 files):
Merge branch 'nom' into bs
13:06
kudo/nom: 136ef41 | jnthn++ | tools/build/NQP_REVISION:
Update NQP_REVISION to one in the NQP bs branch, to ease trying out the bs branch a bit.

Remove a completed ROADMAP item.
13:06 dalek left, Trashlord left
jnthn huh, it got it the wrong way around... :) 13:06
masak moritz: I solved the riddle in 10 seconds. ;)
13:06 dalek joined, ChanServ sets mode: +v dalek
jnthn masak: You're SO childish :) 13:06
masak but I feel I was lucky, because the header made me switch to lateral mode first thing I did.
jlaire it took me several minutes 13:07
moritz jnthn: I think your merge was correct-ish
jnthn: you merged nom into bs, and then bs into nom. The latter was a fast-forward merge, so no merge commit was made for it
jnthn moritz: It was, it looks fine.
kudo/nom: 4536120 | jnthn++ | docs/ChangeLog:
A couple of ChangeLog notes about the effects of bs.
13:10
au hmm it took me a full six minutes :) 13:13
(solving the puzzle algebraically in a decidedly non-childish way)
13:13 Trashlord joined
moritz fwiw I stared at it for a minute or two, then did something else, and after a minutes not thinking about I knew the answer :-) 13:17
au finally "gets" the lateral reference right after solving it
heh. nice puzzle indeed :)
moritz *a few minutes 13:18
[Coke] moritz: perlgeek.de/en/imprint is stuck in historic pumpking times. 13:20
moritz jnthn: so, constant declarators should now be easy to do, right?
[Coke]: "historic pumpking times"? 13:21
[Coke] rants about irc again. I should be able to ignore dalek sends to #foo that i've already seen in #bar
daxim who is the current pumpking?
[Coke] moritz: Rafael Garcia-Suarez ?
moritz [Coke]: oh, you mean the copyright date in the footer? 13:22
daxim grep for pumpking on the page.
moritz oh, that too
thanks
13:23 Nalla joined, Nalla left
moritz fixing now 13:24
13:24 conntrack left 13:25 Bzek left 13:32 grondilu joined
grondilu perl6: package Foo { our multi sub talk { say "hi" } }; Foo::talk; 13:32
p6eval niecza v15-2-gd19c478: OUTPUT«Potential difficulties:␤ &talk is declared but not used at /tmp/StirSsDcvR line 1:␤------> [32mpackage Foo { our multi sub talk [33m⏏[31m{ say "hi" } }; Foo::talk;[0m␤␤hi␤»
..rakudo 453612: OUTPUT«Missing or wrong version of dependency 'src/gen/CORE.setting'␤current instr.: '' pc 1312 (src/gen/SAFE.setting.pir:499) (src/SAFE.setting:1)␤»
..pugs b927740: OUTPUT«hi␤»
13:32 sisar joined
sisar o/ 13:33
grondilu niecza--, rakudo--, pugs++ :)
jnthn heh, that looks like a Rakudo build fail.
Or something...
grondilu perl6: package Foo { our sub talk { say "hi" } }; Foo::talk; 13:34
p6eval rakudo 453612: OUTPUT«Missing or wrong version of dependency 'src/gen/CORE.setting'␤current instr.: '' pc 1312 (src/gen/SAFE.setting.pir:499) (src/SAFE.setting:1)␤»
..pugs b927740, niecza v15-2-gd19c478: OUTPUT«hi␤»
sisar is installing rakudo from scratch, coz he screwed up something and was unable to merge conflict
jnthn: build fails? huh, i'm cloning into rakudo now and will build soon, so will be able to confirm if build fails 13:35
grondilu nom: say "hello" 13:36
p6eval nom 453612: OUTPUT«Missing or wrong version of dependency 'src/gen/CORE.setting'␤current instr.: '' pc 1312 (src/gen/SAFE.setting.pir:499) (src/SAFE.setting:1)␤»
grondilu fail builed confirmed :)
sisar grondilu++
so maybe i should wait before it is resolved 13:37
grondilu indeed
sisar has been without rakudo for 24hrs now :( 13:38
jnthn I think it may only be SAFE.setting with issues.
sisar .o.O( time to use niecza more)
jnthn sisar: The build issue only affects SAFE.setting, which is used by p6eval. 13:39
sisar: It's highly unlikely you'll be using that locally 13:40
sisar: So go ahead and build.
sisar moritz: +1 #wish for UI designers...
jnthn Plus the bs branch just landed so the build should be faster than before :)
sisar jnthn: oh. great ! :)
13:41 conntrack joined
dalek kudo/nom: 2aa69ed | jnthn++ | src/Perl6/ (2 files):
Enable use of BEGIN phasers as r-values.
13:42
kudo/nom: 455035f | jnthn++ | src/Perl6/Actions.pm:
Quick crack at non-literal constants.
sisar have been hearing about bs for a long time. It used to take almost 20min on my *very* slow machine to build, so lets see now... jnthn++
dalek ast: f25bc31 | jnthn++ | S04-phasers/begin.t:
Unfudge a couple of spectests for BEGIN in r-value context.
ast: a84a160 | jnthn++ | S04-declarations/constant.t:
Unskip some tests for non-literal constants.
jnthn moritz: Seems so... :)
moritz \o/ 13:43
dalek atures: 7dd72b9 | jnthn++ | features.json:
Rakudo now more fully supports the constant declarator.
13:44
grondilu nom: constant C = <13 42>
p6eval nom 453612: OUTPUT«Missing or wrong version of dependency 'src/gen/CORE.setting'␤current instr.: '' pc 1312 (src/gen/SAFE.setting.pir:499) (src/SAFE.setting:1)␤»
grondilu niecza: constant C = <13 42> 13:45
p6eval niecza v15-2-gd19c478: ( no output )
grondilu niecza: say constant C = <13 42>
p6eval niecza v15-2-gd19c478: OUTPUT«13 42␤»
13:45 conntrack left
grondilu It really seems to me that niecza is much more advanced than rakudo, isn't it? 13:45
13:46 noam left
dalek atures: 03a13f2 | jnthn++ | features.json:
A couple more features tweaks for Rakudo.
13:46
jnthn grondilu: No.
grondilu: It is in some areas, and not in others.
13:47 noam joined
jnthn > say constant C = <13 42> 13:47
13 42
Works locally.
grondilu what's « > » ?
moritz grondilu: for example rakudo's object and meta-object model is more advanced, whereas niecza's grammar/rules engine is better
« » is a quoting construct
much like qw//, but more magical, and interpolates stuff etc. 13:48
grondilu no I mean: "> perlcode" here on this chan
oh hang on. I understand
moritz it's from the REPL
jnthn Just copy-paste from repl
13:49 pernatiy joined
grondilu Ok. Silly of me. 13:49
moritz jnthn: should I update the feature matrix?
grondilu It's good that these type of constants work now. I've been waiting for that for a while.
jnthn moritz: Already did. 13:50
moritz oh right
jnthn moritz: Oh, I updated features.json...the rest is automated, no
?
13:50 bkolera left
moritz jnthn: yes, it's updated hourly or so 13:50
jnthn ok
Think I see what's up with SAFE.setting. 13:53
13:53 mtk left
jnthn oh noes, I forgot to lunch... :) 13:53
dalek atures: 588fab1 | moritz++ | template.html:
note that the feature matrix is about development versions, not releases
13:53 mtk joined
tadzik sniff sniff, do I smell a branch merge? :) 13:54
13:54 Trashlord left
jnthn tadzik: Yes, it happened. :) 13:54
tadzik lol!
grondilu Is this feature matrix somewhere in the git directory (I don't have regular internet access so I'd like to look at it offline) ? 13:55
moritz grondilu: well, the HTML page at perl6.org/compilers/features is generated from features.json and template.html inside the perl6/features repo
dalek kudo/nom: 28f3d64 | jnthn++ | src/Perl6/World.pm:
Settings are load dependencies, not fixups. Seems to fix --setting=SAFE usage (and thus p6eval).
13:56
jnthn evalbot rebuild nom
p6eval NOT OK (maybe a rebuild is already in progress?)
jnthn evalbot rebuild rakudo
p6eval OK (started asynchronously)
dalek Rebuild of rakudo complete.
jnthn Uh. :)
I know I sped up the build but... P
:P
moritz grondilu: you just need Perl 5 with JSON and HTML::TEmplate::Compiled to generate it locally
jnthn evalbot rebuild rakudo
p6eval OK (started asynchronously)
dalek Rebuild of rakudo complete.
masak o.O
moritz jnthn: 'rakudo' is the 'b' branch
masak oh!
jnthn oh?
but 13:57
rakudo: say 42
moritz jnthn: and it checks for updates, and doesn't do a recompile unless necessary
p6eval rakudo 453612: OUTPUT«Missing or wrong version of dependency 'src/gen/CORE.setting'␤current instr.: '' pc 1312 (src/gen/SAFE.setting.pir:499) (src/SAFE.setting:1)␤»
jnthn moritz: ah, ok
[Coke] make: *** [src/stage2/nqpmo.pbc] Segmentation fault
jnthn So is it nom?
moritz jnthn: yes, the mappings in p6eval and of the build system have diverged a bit
jnthn ok
[Coke] tries a realclean and rebuild.
jnthn [Coke]: Suggest realclean et al. 13:58
13:58 grondilu left
jnthn [Coke]: It was a 200+ commit merge...quite a bit changed. :) 13:58
evalbot: rebuild nom
evalbot rebuild nom
p6eval NOT OK (maybe a rebuild is already in progress?)
jnthn Hmpfh. 13:59
Seems it musta started just before I did the commit to fix things...
moritz jnthn: patience :-) 14:00
[Coke] jnthn: same failure building nqp after removing npq/ and parrot/
(on feather) 14:01
jnthn Is p6eval not on feather?
It seems to have got rather further than that...
moritz no 14:02
[Coke] (so, rm -rf nqp/ parrot/; make realclean; perl Configure.pl --gen--parrot .... segfault.)
moritz it's hosted on a machine provided by diakopter++
jnthn ah, ok
moritz [Coke]: I'll try to reproduce 14:03
jnthn Curious. Backtrace most welcome.
I thought I had a feather account somewhere too...
moritz [Coke]: fwiw 'git clean -xdf' is usually more effective than 'make relclean' (deletes all files that are not under version control)
[Coke] gist.github.com/1950017 - backtrace 14:05
moritz: yes, but I have test scripts and other things about that I don't want to remove. (also, this failure is happening in nqp, which I cleaned pretty effectively)
moritz [Coke]: ok 14:06
sisar jnthn: build took 10 min. Almost a redution of 50% in build time for me :) 14:07
[Coke] tries a straight nqp build outside of rakudo.
tadzik now that the build time isn't annoying I could maybe hack something up :)
jnthn tadzik: Set! KeyBag :) 14:08
moritz steal them from niecza :-)
jnthn aye :)
14:09 agentzh joined
jnthn [Coke]: Hm...curious backtrace. 32-bit or 64-bit? 14:09
14:09 benabik left
[Coke] jnthn: feather is 32, I believe. 14:09
jnthn OK
bbiab, need food 14:10
moritz too gets a segfault on feather while building nqp 14:12
Linux feather 2.6.32-5-amd64 #1 SMP Mon Mar 7 21:35:22 UTC 2011 x86_64 GNU/Linux
looks like 64bit
14:15 havenn joined
flussence you can have a 32-bit userspace even with a x86-64 kernel, though it's pretty rare 14:16
14:17 conntrack joined 14:18 Ulti joined
Ulti irclog.perlgeek.de/perl6/2010-09-24 <rjbs> I've been waiting for someone to write a PILE OF POO operator for p6... 14:19
[Coke] (32/64) whoops.
Ulti I hate that I did this independetly blog.mattoates.co.uk/2012/02/perl6-...ators.html 14:20
took 2 years for operating systems to catch up I guess
14:21 tokuhirom joined
[Coke] Ulti++ 14:21
masak Util: Y U shuffle manually and not just .pick(*) ? 14:23
tadzik jnthn: exactly :) 14:25
Ulti masak: because I don't know any perl6 ;) 14:27
14:29 att left
Ulti there is also an out by one error on that shuffle iirc 14:29
but its not like it was a serious program :) 14:30
masak just saw that "Ulti" ne "Util"...
Ulti erk that's quite confusing, sorry 14:31
[Coke] should run the daily tests now, as pugs would beat rakudo!
Ulti I've had this nick a long while though
flussence nom: 'iltu'.comb.pick(*).join.ucfirst.say
p6eval nom 28f3d6: OUTPUT«Ulti␤»
14:31 Ulti is now known as Ultima 14:32 Trashlord joined, Ultima is now known as Ulti
flussence wait, that was... abnormally quick 14:32
nom: 'iltu'.comb.pick(*).join.ucfirst.say
[Coke] nom: 'anagram'.comb.pick(*).join.say
p6eval nom 28f3d6: OUTPUT«Util␤»
nom 28f3d6: OUTPUT«arngama␤»
masak Ulti: that's fine, keep "Ulti" :)
[Coke] perl6: 'anagram'.comb.pick(*).join.say
p6eval pugs b927740: OUTPUT«anagram␤»