01:00 tokuhirom_h joined 01:23 ilbot3 joined 01:41 tokuhirom joined 02:47 ilbot3 joined 02:54 vendethiel joined 03:43 tokuhirom joined 05:43 tokuhirom joined 06:25 tokuhirom joined 08:02 FROGGS joined 08:25 tokuhirom joined 09:44 mj41 joined 09:51 vendethiel joined 10:08 kjs_ joined 10:35 ShimmerFairy joined 10:39 leont joined 11:34 FROGGS_ joined 12:27 tokuhirom joined 13:20 kjs_ joined 14:28 tokuhirom joined 15:21 kjs_ joined 16:04 vendethiel joined 16:24 tokuhirom joined 16:42 BinGOs joined 17:29 TimToady joined 18:04 vendethiel joined 18:12 kjs_ joined 18:27 tokuhirom joined
dalek arVM/even-moar-jit: 44b9a55 | brrt++ | 3rdparty/dynasm:
Update dynasm to bugfixed version
18:49
arVM/even-moar-jit: 183e63a | brrt++ | src/jit/core.expr:
Add getf and setf expression macros

Names are shamelessly taken from LISP. Doesn't seem to introduce new errors, which is something at least. Simplifies many templates tremendously.
19:01 tokuhirom_h joined 19:54 vendethiel joined 20:04 kjs_ joined
nwc10 jnthn: IIRC you said that one of the advantages of updating the bytecode to be able to store literal strings as either UTF-8 or ISO-8859-1, was that for the latter one knows it has to be NFG already, so no need to process it 20:50
that's no longer true now is it, for any string containing "\r\n" ? 20:51
if so, would the simplest fix be for the serialiser to always serialise strings containing \r\n as UTF-8, so that they go through the NFG input routines, even if the string is in the ISO-8859-1 range
21:47 kjs_ joined 22:26 zakharyas joined 22:29 tokuhirom joined 22:45 kjs_ joined