nwc10 jnthn: ASAN expodes on the setting: paste.scsys.co.uk/505142 07:01
nwc10 jnthn: culprit seems to be b6e0e2a3876d30d356d9d2d7d1ec5edbd689eb53 08:05
jnthn nwc10: Hmmm... Thanks. 12:15
dalek arVM: 177c859 | jnthn++ | src/spesh/graph.c:
Revert "Fix leak involving inlining handlers."

This reverts commit b6e0e2a3876d30d356d9d2d7d1ec5edbd689eb53. Turns out this is too early, but the place I'd have expected to move this to already *is* freeing handlers. So will need a litle more analysis.
12:43
jnthn puts that one on the "not so easy" list :) 12:44
dalek arVM: db56f04 | jnthn++ | src/strings/decode_stream.c:
Fix a leak around partial slurping.

That is, reading some of a file, then slurp-rest'ing the rest of it, in Perl 6 terms.
17:34
arVM: fef3b35 | timotimo++ | src/jit/graph.c:
functions with MVMRegister foo don't want _F calling conventions
17:54
arVM: db9c558 | timotimo++ | src/jit/graph.c:
comment on _F not being there
timotimo i could imagine i've gone back and forth on that line a few times already 17:55
jnthn Ah, that one 19:18
jnthn fixed a case like that once...
Guess I didn't look hard enough for others.