01:56 TEttinger joined 02:48 vendethiel joined 02:50 colomon joined 03:29 colomon joined 03:40 colomon joined 05:22 vendethiel joined 06:04 rurban joined 06:09 dalek joined 06:27 vendethiel joined 06:40 FROGGS joined
dalek arVM/even-moar-jit: df06151 | brrt++ | src/ (4 files):
Fix dynamic array memory corruption

Doubling the size is not always enough. Also minor fixes in exprlist and in the 'compiler', noting that void values do not need to be computed to memory.
06:45
06:51 vendethiel joined 06:57 zakharyas joined 07:01 brrt joined
brrt \o 07:09
07:16 vendethiel joined 07:25 FROGGS joined 07:40 vendethiel joined
dalek arVM/even-moar-jit: cba9b92 | jnthn++ | src/6model/reprs/MultiDimArray.c:
Implement cloning multi-dim arrays.
07:59
MoarVM/even-moar-jit: 53468a9 | jnthn++ | src/ (3 files):
MoarVM/even-moar-jit: Fill out remaining multi-dim array ops.
brrt dalek, that was a merge, obviously 08:00
bad robot
08:00 dalek joined
brrt hmmm 08:05
i'm going to blog some more
FROGGS brrt++ 08:06
brrt but progress this last week was decidely non-impressive
hmm :-) 08:08
08:24 vendethiel joined
brrt brrt-to-the-future.blogspot.nl/2015...eport.html 08:42
hmm 08:47
ping jnthn
09:20 brrt joined
brrt hmm, bbl 09:38
09:46 Ven joined
dalek arVM: 9c9f089 | (Jimmy Zhuo)++ | src/ (23 files):
removed MVM_checked_free_null, and changed to MVM_free
10:03
10:08 brrt joined 10:15 vendethiel joined
brrt netsplit again? 12:03
dalek arVM: b35ff45 | brrt++ | Configure.pl:
Reverse library order for ASAN

Apparantly ASAN complains (on linux, at least) if it isn't the first linked library. So we make it so.
12:14
arVM/even-moar-jit: 9c9f089 | (Jimmy Zhuo)++ | src/ (23 files):
removed MVM_checked_free_null, and changed to MVM_free
arVM/even-moar-jit: b35ff45 | brrt++ | Configure.pl:
Reverse library order for ASAN

Apparantly ASAN complains (on linux, at least) if it isn't the first linked library. So we make it so.
arVM/even-moar-jit: 38fcdaa | brrt++ | / (24 files):
Merge remote-tracking branch 'origin/master' into even-moar-jit
jnthn afternoon, #moarvm 12:16
12:17 rurban joined
brrt \o jnthn 12:19
i was looking for you :-)
still in kiev?
jnthn No, just home to Prague today
brrt long flight?
jnthn 2h 12:20
brrt ah, that's quite reasonable :-)
jnthn But seem to be running a bit of a temperature today, and also got up early for it, so it felt long anyway...
brrt ASAN finds me bugs, bugs, bugs
jnthn ASAN is often right ;)
brrt :-)
jnthn At least, I didn't see a case where it wasn't yet
ugh, Moar build busted on Windows 12:21
Well, MSVC ot be specific
rc\core\frame.c(974) : error C2143: syntax error : missing ';' before 'type'
src\core\frame.c(975) : error C2143: syntax error : missing ';' before 'type'
and a lot more
brrt o.O
hmm
on master?
jnthn aye 12:22
brrt i have not built on windows for quite some time
jnthn Looks like hoelzro is to blame :) 12:23
But his patch was an overall good, so I won't complain too much :)
brrt hmm, asan complains about leaks, some of which i know of (but have decided not to care yet) 12:24
but no SEGV's now, so that's something 12:25
dalek arVM/even-moar-jit: 0130e96 | brrt++ | src/jit/ (2 files):
Fix two SEGV spotted by ASAN
jnthn So bustage... 12:32
jnthn is still fixing
dalek arVM: 500508d | jnthn++ | src/ (6 files):
MSVC build fixes.
12:38
brrt ahah, that old bug again 12:39
nklein.com/2015/06/syntactic-corn-syrup/ this is weird. i literally just proved that 'compiling' lisp is worth ~200 LoC, this guy says lisp is harder to compile than c, wat
12:42 FROGGS joined
dalek arVM/leave: f2a7480 | FROGGS++ | / (9 files):
start to implement exception: return + label

This is about making "LABEL.leave(42)" work.
13:08
brrt bbl 13:09
13:31 colomon joined 14:04 colomon joined 14:29 colomon joined
hoelzro oops =/ 14:34
I made one small Windows-only mistake?
jnthn hoelzro: Well, C89 strictness mistake 14:38
hoelzro: I fixed it
hoelzro ahhhhh
14:38 lizmat joined
hoelzro so Moar builds with C99 on *nix, but C89 on Windows? 14:39
jnthn It's about compiler, not platform.
hoelzro oh, I see
jnthn MSVC doesn't do C99 :/
arnsholt hoelzro: The MS C compiler only implements bits and pieces of C99
jnthn I read the next MSVC will be better at it
hoelzro ah yes, I remember hearing that 14:40
jnthn Which is like, OK, progress, but we'll not be able to really take advantage of it for another decade until the older versions that don't do it can be dropped.
hoelzro heh
14:55 JimmyZ_ joined 15:36 nwc10 joined, psch joined 15:37 dagurval joined, xiaomiao joined 15:39 FROGGS joined 15:54 JimmyZ_ joined 16:00 JimmyZ__ joined 16:30 TEttinger joined 16:35 JimmyZ_ joined 17:12 JimmyZ_ joined 17:51 rurban joined
dalek arVM: 34e8998 | jnthn++ | src/6model/reprs/MultiDimArray.c:
Fix serializing uncomposed MultiDimArray type.
20:09
arVM: 771a1e7 | jnthn++ | src/6model/reprconv.c:
Add missing concreteness checks.
20:16
20:37 synbot6 joined
vendethiel jnthn, brrt, others: I'm not sure it's been linked here (I'm not sure where I got the link, to be honest), but this article is very interesting WRT MOP optimiaztions: stefan-marr.de/papers/pldi-marr-et-...ogramming/ 21:02
PerlJam vendethiel: I think that masak mentioned it the other day. 21:04
vendethiel PerlJam: alright, my bad then :). 21:05
PerlJam (or ... I've seen the link before and in my head it's associated with masak)
vendethiel PerlJam: maybe I saw it in the backlog, maybe from reddit, dunno :) 21:07
21:13 colomon joined 21:27 colomon joined 22:05 synbot6 joined