00:00 lizmat joined 00:01 lizmat_ joined 00:02 woolfy joined 01:01 lizmat joined 01:52 colomon joined 02:35 JimmyZ joined
JimmyZ jnthn: looks like this comment is out of date? github.com/MoarVM/MoarVM/blob/mast...rap.c#L312 02:36
dalek arVM: 5a347d4 | jimmy++ | src/strings/unicode_ops.c:
return str_consts.empty directly
03:45
arVM: bd9d5af | jimmy++ | src/strings/unicode_ops.c:
fix alignment
03:50
04:45 ggoebel1111115 joined
timotimo good catch o nthe s = "" thing; but i'm not sure we'd been running into that case a lot so far? 06:23
06:30 Ven joined 06:52 lizmat joined 07:11 zakharyas joined
dalek arVM: 4a91df9 | jimmy++ | src/core/alloc.h:
Added MVM_calloc function
07:16
07:18 travis-ci joined
travis-ci MoarVM build failed. Jimmy Zhuo 'Added MVM_calloc function' 07:18
travis-ci.org/MoarVM/MoarVM/builds/37573456 github.com/MoarVM/MoarVM/compare/b...91df93e9a4
07:18 travis-ci left
dalek arVM: a1ab118 | jimmy++ | / (2 files):
Add src/core/alloc.h to Makefile and fix build
07:25
07:34 travis-ci joined
travis-ci MoarVM build passed. Jimmy Zhuo 'Add src/core/alloc.h to Makefile and fix build' 07:34
travis-ci.org/MoarVM/MoarVM/builds/37573884 github.com/MoarVM/MoarVM/compare/4...ab118178d7
07:34 travis-ci left
dalek arVM: 7c13f31 | jimmy++ | src/ (36 files):
s/calloc/MVM_calloc/g
07:36
07:37 brrt joined 07:40 lizmat joined 07:53 Ven joined 08:23 cognome joined 09:02 woolfy joined 09:32 JimmyZ joined 10:44 daxim_ joined 10:47 kjs_ joined 11:09 brrt joined
brrt timotimo: no 11:10
it's a long story
but getobjsc is ultimately a pretty simple op if you trace it down to its roots
what i intend to do is ignore the assert statements 11:12
which simplifies things considerably
11:31 Ven joined
timotimo hm. have i asked about this already? how difficult will it be to jit nativecallinvoke properly? 13:09
are we likely to have a KNOWN_VALUE for the thing that we want to invoke? 13:10
13:13 brrt joined
jnthn timotimo: It's somewhat tricky to do really well 14:06
timotimo that's not surprising :) 14:13
jnthn We may also want to revisit the API if we find it's too JIT hostile. :) 14:15
14:30 JimmyZ joined
JimmyZ jnthn: I had a question for you .... : irclog.perlgeek.de/moarvm/2014-10-10#i_9484107 14:30
jnthn JimmyZ: It seems that ended up being done elsewhere 14:33
(further donw in the ifle)
So that probably can go
JimmyZ OK, thanks
dalek arVM: c27f1aa | jimmy++ | src/6model/bootstrap.c:
remove a obsolete comment
14:42
14:56 Ven joined 14:57 cognome joined 15:22 cognome joined 15:53 woolfy left 16:19 Ven joined 16:20 Ven_ joined 17:54 cognome joined 18:05 brrt joined 19:01 brrt joined 19:30 colomon joined
brrt ok, MVM_USE_OVERFLOW_SERIALIZATION_INDEX is never defined for x64 because MVM_PTR_SIZE must be 8 20:10
20:22 FROGGS joined 20:35 lizmat joined
dalek arVM/jit-getobjsc: b172709 | (Bart Wiegmans)++ | src/jit/ (2 files):
Add getobjsc.

This causes two spectest failures
  * S06-operator-overloading/sub.rakudo.moar
  * S17-procasync/print.rakudo.moar
So this is pushed into a separate branch until further notice.
20:38
brrt a little present for APW :-) 20:39
lizmat brrt: fwiw, S17-procasync is a flapper
brrt oh
it worked this morning so i assumed it to be fixed
anyway, rather safe than sorry 20:40
and branches are cheap
lizmat yup :-)
brrt i'll be weekending now. have fun :-) 20:43
20:43 brrt left 20:44 kjs_ joined 20:51 Ven joined