|
00:04
d4l3k_ joined
00:06
jsimonet left,
dalek left,
d4l3k_ is now known as dalek,
Geth left,
ZofBot left,
synopsebot left
00:08
greppable6 left,
greppable6 joined
00:12
ChanServ left
00:14
jsimonet joined
00:33
ChanServ joined,
tolkien.freenode.net sets mode: +o ChanServ
00:37
AlexDaniel`` left,
wictory[m] left
00:38
ilmari[m] left
00:40
greppable6 left,
greppable6 joined
00:41
squashable6 left,
benchable6 left,
coverable6 left,
undersightable6 left
01:05
ilmari[m] joined
01:27
shareable6 joined
|
|||
| MasterDuke | re the recent hash table discussion: probablydance.com/2018/05/28/a-new...ash-table/ | 01:29 | |
|
01:31
shareable6 left
01:38
coverable6 joined,
benchable6 joined,
squashable6 joined
|
|||
| samcv | MasterDuke: i wish they made it in C and not C++ | 01:39 | |
|
01:45
ilbot3 left
01:47
AlexDaniel`` joined,
wictory[m] joined,
shareable6 joined
01:54
shareable6 left
01:58
ilbot3 joined
|
|||
| MasterDuke | yeah, would be nice | 02:05 | |
|
03:07
shareable6 joined
03:11
shareable6 left
03:22
buggable left,
huggable left,
ZofBot joined,
huggable joined,
buggable joined
03:23
p6lert left,
p6lert joined,
SourceBaby joined
03:32
shareable6 joined
03:33
shareable6 left,
shareable6 joined
03:35
shareable6 left
03:37
shareable6 joined
03:39
shareable6 left,
shareable6 joined
03:42
shareable6 left,
shareable6 joined
03:46
shareable6 left
03:57
shareable6 joined
04:03
shareable6 left,
shareable6 joined
04:05
shareable6 left
05:13
MasterDuke left
06:25
brrt joined
06:39
domidumont joined
06:45
domidumont left
06:46
domidumont joined
07:56
zakharyas joined
07:58
shareable6 joined
08:19
zakharyas left,
zakharyas joined
08:37
Tison joined
|
|||
| Tison | How does MoarVM dump bytecode? | 08:39 | |
| I have a simple p6 program named main.p6 contains `say "Hello"` | |||
| then execute `perl6 --target=mbc --output=main.moarvm main.p6` | 08:40 | ||
| then execute `moar --dump main.moarvm` | |||
| I can see string "Hello" in main.moarvm, but not in its dump | |||
| confused now ... | |||
|
08:47
brrt left
|
|||
| Tison | also in the dump code, I lost variable names. It seems things like `loc_5_obj` represent them, but isn't it nearly unreadable? | 08:56 | |
|
09:10
undersightable6 joined
|
|||
| Tison | BTW, here is a question on ZhiHu, Chinese Quora, "what do you think of " | 09:48 | |
| "What do you think of MoarVM" | 09:49 | ||
| www.zhihu.com/question/278445209 | |||
| jnthn | It'll be because the string compiles into a Perl 6 Str object, so it's in the serialized object blob | 09:50 | |
| Tison | If you guys have some spare time, why not give a brief talk about MoarVM? Also I'm glad to be delegated to post it. | 09:52 | |
| jnthn: Alright, at least I know why :-) | |||
| jnthn | Tison: I did give one on MoarVM's optimization stuff recently :) | 09:53 | |
| Tison | of course, I'll try to gather so many your slides and talks and give my own answer on it. (Though I really know little about MoarVM) | 09:57 | |
| even integers be serialized? I cannot see literal integers also | 10:01 | ||
| & | 10:03 | ||
|
10:08
Tison left
10:10
domidumont left
10:35
brrt joined
|
|||
| brrt | good hi #moarvm | 10:35 | |
| lizmat | brrt o/ | 10:39 | |
| AlexDaniel | squashable6: next | 10:43 | |
| squashable6 | AlexDaniel, ā š Next SQUASHathon in 1 day and ā23 hours (2018-06-02 UTC-12āUTC+14). See github.com/rakudo/rakudo/wiki/Mont...Squash-Day | ||
| brrt | ohai lizmat | 11:00 | |
|
11:37
zakharyas left
11:57
Tison joined
12:24
brrt left
12:38
zakharyas joined
13:06
brrt joined
13:19
Tison left
|
|||
| brrt | heh, i forgot i had a MVM_JIT_DEBUG flag for validating the position in code | 13:24 | |
| timotimo | medium.com/netflix-techblog/netfli...7ca19d47bb - shiny! | 13:27 | |
|
13:53
AlexDaniel left,
AlexDaniel joined
14:39
domidumont joined
15:03
committable6 joined
15:50
Tison joined
15:55
synopsebot joined,
Geth joined
15:57
tadzik1 joined,
tadzik1 left
16:01
tadzik1 joined
16:02
zakharyas left,
tadzik1 is now known as tadzik
16:07
zakharyas joined
16:27
brrt left
16:30
zakharyas left
17:16
samcv left
17:22
samcv joined
17:36
Tison left
|
|||
| samcv | timotimo: so what's with the debugserver.c doesn't always return a value? that gonna be fixed? | 18:25 | |
| clang complains every time and coverity is also unhappy | |||
| lizmat | #132209 | 18:37 | |
| oops | |||
| timotimo | it's gonna be fixed | 18:39 | |
|
18:45
zakharyas joined
19:02
domidumont left
|
|||
| Geth | MoarVM: e2afa7fbc8 | (Samantha McVey)++ | 3 files Speed up the NFA by using a cached grapheme iterator We use a cached grapheme iterator to access graphemes in the string which should give us a nice speed improvement relative to the number of grapheme accesses the NFA makes in the course of its operation on a single call. |
19:07 | |
| MoarVM: 26ab289617 | (Samantha McVey)++ | src/strings/unicode_ops.c Fix bounds issue in unicode collation (though not possible to trigger) This can't be triggered since we don't have anywhere near 10 graphemes in a collation rule, but change it anyway. Since we increase the value of i immediately after the conditional, that could make i == 10 for an array with only 10 elements and cause it to go past its bounds. |
19:10 | ||
| timotimo | someone want to implement pow_i for the jit real quick ... | 19:15 | |
|
19:15
zakharyas left,
zakharyas joined
19:23
travis-ci joined
|
|||
| travis-ci | MoarVM build passed. Samantha McVey 'Speed up the NFA by using a cached grapheme iterator | 19:23 | |
| travis-ci.org/MoarVM/MoarVM/builds/385844677 github.com/MoarVM/MoarVM/compare/d...afa7fbc870 | |||
|
19:23
travis-ci left
|
|||
| Geth | MoarVM: 4708018084 | (Samantha McVey)++ | src/strings/shiftjis.c Fix potential issue in shiftjis decoder if replacement had synthetic -1 Use a separate variable to denote if the grapheme has been set instead of using a magic value for the grapheme. |
19:26 | |
|
19:33
travis-ci joined
|
|||
| travis-ci | MoarVM build passed. Samantha McVey 'Fix bounds issue in unicode collation (though not possible to trigger) | 19:33 | |
| travis-ci.org/MoarVM/MoarVM/builds/385845821 github.com/MoarVM/MoarVM/compare/e...ab2896170e | |||
|
19:33
travis-ci left
19:56
zakharyas left
20:23
klapperl left
|
|||
| Geth | MoarVM: 586950bf07 | (Samantha McVey)++ | src/strings/unicode_ops.c Fix the number of items in the collation union This doesn't change anything except make the code more clear since the union never overflowed because C allocates the largest item in the union. It does make the code clearer and make much more sense though. Issue detected by coverity. |
20:36 | |
| timotimo | gist.github.com/orlp/3551590 | 21:08 | |
| do we want a better pow impl? | |||
| m: use nqp; say nqp::pow_i(10, 10000) | 21:09 | ||
| camelia | 0 | ||
| timotimo | m: use nqp; say nqp::pow_i(10, 1000) | ||
| camelia | 0 | ||
| timotimo | m: use nqp; say nqp::pow_i(10, 100) | ||
| camelia | 0 | ||
| timotimo | m: use nqp; say nqp::pow_i(10, 10) | ||
| camelia | 10000000000 | ||
| timotimo | it's fine to return 0 for those far-too-high values? | 21:10 | |
| in any case, having special implementations of pow for an exponent of 2 would surely be wise :) | 21:21 | ||
| in the jit, i mean | 21:22 | ||
| dogbert17 | timotimo: trying to --profile rindolf's code doesn't work that well ... | 21:34 | |
|
21:40
klapperl joined
|
|||
| timotimo | oh? | 21:48 | |
| worked well on my end, though i reduced the $N by a lot to do it | |||
| samcv | darn my changes to nfa probably haven't helped at all since the stringe we call it on compiling the setting is a flat string | 22:18 | |
| timotimo | oh, we always flatten strings for the regex engine, right? | 23:12 | |