jnthn Other than that, seems sane at first glance 00:00
FROGGS cool :o) 00:04
jnthn 'night o/ 00:20
FROGGS gnight 00:21
01:07 odc joined
JimmyZ Good news, compiling core.setting works on linux now. 02:21
nwc10++ 02:22
jnthn++ 02:23
TimToady the gcc docs say: Please note the warning under -fgcse about invoking -O2 on programs that use computed gotos. 03:22
do we use computed gotos?
JimmyZ yes. but not by default 03:23
used by `make cgoto` 03:24
TimToady tries -O2 -fno-gcse to see what happens 03:26
JimmyZ I see -fgcse doces there 03:32
-fgcse Perform a global common subexpression elimination pass. This pass also performs global constant and copy propagation.
Note: When compiling a program using computed gotos, a GCC extension, you may get better run-time performance if you disable the global common subexpression elimination pass by adding -fno-gcse to the command line.
TimToady doesn't make any difference though :) 03:34
JimmyZ just may get better run-time -fgcse Perform a global common subexpression elimination pass. This pass also performs global constant and copy propagation.
oh
bad keyboard
04:44 cognominal joined 05:22 ingy joined 05:55 eternaleye joined 05:56 eternaleye joined 05:58 benabik joined 06:05 jnap joined 06:23 ssutch joined 07:06 jnap joined
nwc10 Elizabeth, 4.5, has a feature request. We want more colours. 07:38
[the build output is too monochrome :-)]
moritz it's dichrome here :-) 07:54
08:07 jnap joined
FROGGS the absolute_path function in rakudo's ML does not work, it prepends the cwd even to absolute paths 08:53
perl6-m -e 'say so "/home/froggs/dev/v5/lib" ~~ m/^\//' 08:54
False
perl6-p -e 'say "/home/froggs/dev/v5/lib" ~~ m/^\//' 08:55
ļ½¢/ļ½£
ummm, how can this be false?
Match.new(orig => "/home/froggs/dev/v5/lib", from => 0, to => 1, ast => Any, list => ().list, hash => EnumMap.new()) 08:56
$/.Bool says True 08:57
hmpf, most what I said was using perl6-p >.< 08:58
perl6-m -e 'say so "/home/froggs/dev/v5/lib" ~~ m/^\//; say so $/' 09:00
False
True
ohh 09:29
--ll-exception reveals it...
perl6-m --ll-exception -e 'say "/home/froggs/dev/v5/lib".match(/./)'
use of uninitialized value of type Any in numeric context
p6sort NYI
it is the sort in Match.caps btw 09:31
so it is say() only :/ (aka false positive) 09:32
btw, I am collecting absurdities now here: gist.github.com/FROGGS/cbdcd4054a450d0faa22 09:35
timotimo hm. is our regex codegen wrong? 09:37
FROGGS maybe for the dot
somehow a smartmatch against a regex does not update $/ early enough 09:51
yes, changing a $path ~~ m/^\// to $path.match(/^\//) helps 09:54
the method form should be cheaper anyway 10:01
timotimo wouldn't it compile-time dispatch anyway?
10:08 jnap joined
FROGGS the smartmatch version build an ast around the method call plus additional stuff 10:24
timotimo ah, yes. at some point we really should build clever code to only change the topic if it's going to be used and things like that 10:25
(also: compile-time dispatch won't happen for methods, duh) 10:51
though since it's a WVal in the first argument, we actually *could* compile-time dispatch, maybe
on the other hand, it should be caught by jnthn's invokedynamic work? i'm not sure if methods got the invokedynamic treatment yet 10:52
OTOH, /.../.match(...) won't compile-time dispatch either
12:10 jnap joined
dalek arVM/moar_libpaths: 80a60fc | jonathan++ | / (3 files):
Fix op signature of SC WB ops.

They were declared as handing back obj, but actually gave int64. This could leave invalid pointers around in registers.		 13:49
arVM/moar_libpaths: becd41e | (Tobias Leich)++ | / (3 files):
Merge branch 'master' of github.com:MoarVM/MoarVM into moar_libpaths
arVM/moar_libpaths: 592f914 | (Tobias Leich)++ | src/ (5 files):
git figure_filename a proper name

and also move it to io/fileops.
arVM: e884128 | (Tobias Leich)++ | src/ (6 files):
allow several --libpath options

And also search for libperl6_ops_moar.so in these paths.		 13:58
FROGGS yay, my first rebase -i 14:01
timotimo :)
14:11 jnap joined 14:19 jnap joined
FROGGS ===SORRY!=== 14:57
Object conflict detected during deserialization.
(Probable attempt to load two modules that cannot be loaded together).
:o(
timotimo :(
FROGGS that happens when compiling Perl5::Config 14:58
timotimo some merging resolution feature missing from mokudo still?
FROGGS no idea
ahh, it is about object repossession 15:03
15:50 pryamogon joined 16:28 jnap joined, rurban joined
FROGGS btw, I believe we pass about 18k tests 17:23
perl6-m -e 'use v5; say $^O' 17:37
linux
\o/
18:20 jnap joined 18:23 dalek joined 18:31 timotimo joined 19:19 masak joined, colomon joined
masak I want to run just the GC of MoarVM in isolation. 19:20
19:20 jnap joined
masak maybe extract the C files related to the GC and run them alone, instrumenting them in a harness. 19:21
can that be done? anyone know, before I go ahead and try? :)
I'm experiencing errors building MoarVM master :/ gist.github.com/masak/f76917b23911061f3d29 19:49
FROGGS masak: make clean or realclean 19:50
looks like your 3rdparty libs were built ages ago
masak realcleans and tries again 19:55
is it unreasonable to expect the Configure.pl process to stop-and-emit something like "looks like your 3rdparty libs were built ages ago" ? 19:56
(worked now.) FROGGS++ 19:57
FROGGS I guess we make it nicer when moarvm has more than ten users :o) 19:58
masak there is that. 19:59
masak files a moarvmissue
diakopter masak: hi 20:06
20:09 FROGGS[mobile] joined 20:21 jnap joined
masak diakopter: ahoj :) 20:23
diakopter: many happy &returns
diakopter haopy 2015 20:25
2013
2014
happy 20:26
FROGGS[mobile] diakopter: happy happy to you too 20:30
21:22 jnap joined 22:23 jnap joined
lue Unhandled exception: Bytecode validation error at offset 24, instruction 4: 22:32
operand type 32 does not match register type 64
While building nqp
22:52 colomon joined
lue Is it the result of old files hanging around somewhere perhaps? 23:45

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!