Perl 6 - the future is here, just unevenly distributed

IRC log for #moarvm, 2017-05-16

| Channels | #moarvm index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:31 lizmat_ joined #moarvm
01:49 ilbot3 joined #moarvm
01:49 Topic for #moarvm is now https://github.com/moarvm/moarvm | IRC logs at  http://irclog.perlgeek.de/moarvm/today
05:06 geekosaur joined #moarvm
05:41 brrt joined #moarvm
05:42 domidumont joined #moarvm
05:51 domidumont joined #moarvm
06:39 brrt joined #moarvm
07:30 brrt joined #moarvm
07:38 brrt joined #moarvm
07:43 domidumont joined #moarvm
07:46 zakharyas joined #moarvm
07:46 domidumont1 joined #moarvm
07:47 zakharyas joined #moarvm
08:07 brrt joined #moarvm
08:08 lizmat_ And a new Perl 6 Weekly hits the Net: https://p6weekly.wordpress.com/2017/05/16/2017-20-crossing-the-alps/
08:08 brrt \o lizmat_
08:13 zakharyas joined #moarvm
09:26 brrt joined #moarvm
09:34 Geth ¦ MoarVM/even-moar-jit: 45bf6e94eb | (Bart Wiegmans)++ | 3 files
09:34 Geth ¦ MoarVM/even-moar-jit: Disable expr compiling for throwish/invokish ops
09:34 Geth ¦ MoarVM/even-moar-jit:
09:34 Geth ¦ MoarVM/even-moar-jit: These ops need guarding, and that is not yet implemented.
09:34 Geth ¦ MoarVM/even-moar-jit: This fixes CORE.setting compilation for now.
09:34 Geth ¦ MoarVM/even-moar-jit: review: https://github.com/MoarVM/MoarVM/commit/45bf6e94eb
09:34 Geth ¦ MoarVM/even-moar-jit: 9bc0ed2b9c | (Bart Wiegmans)++ | 2 files
09:35 Geth ¦ MoarVM/even-moar-jit: Split jit-dump from jit-bisect
09:35 Geth ¦ MoarVM/even-moar-jit:
09:35 Geth ¦ MoarVM/even-moar-jit: The jit-bisect debug tool was somewhat unreliable and frankly did
09:35 Geth ¦ MoarVM/even-moar-jit: too many things at once. I've split the code to get a specific JIT
09:35 Geth ¦ MoarVM/even-moar-jit: dump into its own script and refactored the bisecting code. Stuff
09:35 Geth ¦ MoarVM/even-moar-jit: is generally more reliable now.
09:35 Geth ¦ MoarVM/even-moar-jit: review: https://github.com/MoarVM/MoarVM/commit/9bc0ed2b9c
09:46 brrt i'm still figuring out what is broken and why...
10:01 zakharyas joined #moarvm
11:41 praisethemoon joined #moarvm
12:54 Geth ¦ MoarVM/even-moar-jit: 3f5211cb8c | (Bart Wiegmans)++ | src/jit/linear_scan.c
12:54 Geth ¦ MoarVM/even-moar-jit: tile->args is not tile->values
12:54 Geth ¦ MoarVM/even-moar-jit:
12:54 Geth ¦ MoarVM/even-moar-jit: Also, values and register_spec are +1 biased compared to refs because
12:54 Geth ¦ MoarVM/even-moar-jit: they include the output register which refs does not. It might be a
12:54 Geth ¦ MoarVM/even-moar-jit: good idea to change that (have tile->refs[0] == tile->node).
12:54 Geth ¦ MoarVM/even-moar-jit: review: https://github.com/MoarVM/MoarVM/commit/3f5211cb8c
12:54 Geth ¦ MoarVM/even-moar-jit: e0b99d02bd | (Bart Wiegmans)++ | src/jit/linear_scan.c
12:54 Geth ¦ MoarVM/even-moar-jit: Invoke prepare_arglist_and_call for CALL tile
12:54 Geth ¦ MoarVM/even-moar-jit:
12:54 Geth ¦ MoarVM/even-moar-jit: It could potentially happen that the register used by CALL would be
12:54 Geth ¦ MoarVM/even-moar-jit: allocated after the ARGLIST was processed, e.g. when the associated
12:54 Geth ¦ MoarVM/even-moar-jit: value is spilled and loaded by a synthetic tile. This could then
12:55 Geth ¦ MoarVM/even-moar-jit: overwrite the register picked by prepare_arglist_and_call for
12:55 Geth ¦ MoarVM/even-moar-jit: resolving the conflict between the ARGLIST and CALL registers.
12:55 Geth ¦ MoarVM/even-moar-jit: Now that prepare_arglist_and_call is executed for the CALL
12:55 Geth ¦ MoarVM/even-moar-jit: tile this can no longer happen.
12:55 Geth ¦ MoarVM/even-moar-jit: review: https://github.com/MoarVM/MoarVM/commit/e0b99d02bd
12:55 Geth ¦ MoarVM/even-moar-jit: 83e4976189 | (Bart Wiegmans)++ | 8 files
12:55 Geth ¦ MoarVM/even-moar-jit: Wrap invokish and throwish with a guard tile
12:55 Geth ¦ MoarVM/even-moar-jit:
12:55 Geth ¦ MoarVM/even-moar-jit: Because this means that control flow can be interrupted and resumed
12:55 Geth ¦ MoarVM/even-moar-jit: right after the compiled node, the expression assembler flushes all
12:55 Geth ¦ MoarVM/even-moar-jit: active values after the guard, forcing them to be reloaded after.
12:55 Geth ¦ MoarVM/even-moar-jit:
12:55 Geth ¦ MoarVM/even-moar-jit: The pre-throwish pseudotile (which is responsible for loading the JIT
12:55 Geth ¦ MoarVM/even-moar-jit: <…commit message has 7 more lines…>
12:55 Geth ¦ MoarVM/even-moar-jit: review: https://github.com/MoarVM/MoarVM/commit/83e4976189
12:55 brrt ^ the last commit unblocks decont and similar invokish / throwish ops
12:56 brrt can be improved (i want to do 'optimistic' flushing, i.e only flush when necessary) but yeah
13:04 nine So settings compilation now works?
13:17 brrt joined #moarvm
13:17 brrt yes
13:18 brrt i intend to keep it working, but with ever larger pieces of code JITted :-)
13:18 brrt nine++ nice blog post on the PTS
13:21 nine Thanks :)
13:21 nine brrt: that sounds as it's about time people joined the effort?
13:23 brrt i'm guessing we're very near that point, yes
13:25 Zoffix \o/
14:01 zakharyas joined #moarvm
15:06 zakharyas joined #moarvm
15:25 brrt joined #moarvm
17:21 zakharyas joined #moarvm
17:21 AlexDaniel joined #moarvm
17:23 samcv joined #moarvm
17:36 praisethemoon joined #moarvm
17:36 praisethemoon joined #moarvm
18:29 japhb joined #moarvm
18:49 domidumont joined #moarvm
18:57 Ven joined #moarvm
19:33 domidumont joined #moarvm
19:42 Ven_ joined #moarvm
19:52 robertle joined #moarvm
19:58 Ven joined #moarvm
20:01 zakharyas joined #moarvm
20:27 Ven_ joined #moarvm
20:48 committable6 joined #moarvm
20:48 Ven_ joined #moarvm
22:19 MasterDuke_ joined #moarvm

| Channels | #moarvm index | Today | | Search | Google Search | Plain-Text | summary