|
github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm Set by AlexDaniel on 12 June 2018. |
|||
|
00:02
Ven`` joined
00:03
p6bannerbot sets mode: +v Ven``
01:31
kiwi_60 joined,
p6bannerbot sets mode: +v kiwi_60
01:32
kiwi_60 left
03:32
leont left
07:40
evalable6 left
07:44
evalable6 joined
07:45
p6bannerbot sets mode: +v evalable6
07:47
brrt joined
07:48
p6bannerbot sets mode: +v brrt
|
|||
| brrt | \o | 07:50 | |
| hmm | 08:24 | ||
| strictly and for consistency, I should probably know when I'm trying emit a store-of-num as a store-of-int | 08:25 | ||
| but. I can also get away with not knowing | |||
| now question is; is that realistic | |||
| the problem is this: | 08:26 | ||
| - I move things like 'const_n64' to use a new 'const_num' operator | |||
| - the const_num operator yields a 'num' type value | 08:27 | ||
| - the 'store' operator doesn't work on 'num' values, so there is a 'store_num' operator introduced as wlel | 08:28 | ||
| so far goes the theory | |||
| the practical implementation is that 'num' type values are mapped to xmm registers | |||
| so that 'store_num' maps to 'movsd [addr], xmm1', 'load_num' to 'movsd xmm0, [addr]' etc | 08:31 | ||
| now, i need to figure out the proper operator type for the automatially assigned store.... | 08:32 | ||
| oh, hmm | 09:10 | ||
| local_types and lexical_tyeps are MVMuint16, but all actual values fit in MVMuint8; what's more, MVM_spesh_get_reg_type will give you an MVMint8 | 09:11 | ||
| Geth | MoarVM/jit-expr-float: 7 commits pushed by (Bart Wiegmans)++
|
09:53 | |
| brrt | hang on.... | 09:55 | |
| Geth | MoarVM: 45cf901307 | (Bart Wiegmans)++ | build/Makefile.in [JIT] Compile minilua with only the needed flags Issue #999 seems to be avoidable mostly by restricting minilua to compile with just the libraries that it needs, rather than all that MoarVM brings in. Or thats' my theory at least. Removes the always-recompile behaviour, which was annoying me. |
10:09 | |
| MoarVM/jit-expr-float: 6 commits pushed by (Bart Wiegmans)++ | 10:10 | ||
|
10:28
travis-ci joined,
p6bannerbot sets mode: +v travis-ci
|
|||
| travis-ci | MoarVM build failed. Bart Wiegmans '[JIT] Compile minilua with only the needed flags | 10:28 | |
| travis-ci.org/MoarVM/MoarVM/builds/475937707 github.com/MoarVM/MoarVM/compare/4...cf90130766 | |||
|
10:28
travis-ci left
|
|||
| brrt | :-( | 10:32 | |
| I don't understand... | 10:35 | ||
| looks like I"m going to have to get a VM then | 10:39 | ||
|
11:07
brrt left
|
|||
| Geth | MoarVM: 2ee7596dd3 | (Stefan Seifert)++ | src/jit/graph.c Fix call to memcpy with NULL *src in lego JIT compiler MVM_platform_cpu_count does not have any arguments, so there's nothing to copy. |
11:07 | |
| nine | dogbert11: this fixes the null pointer runtime error in src/jit/graph.c:39 | 11:08 | |
|
11:28
travis-ci joined,
p6bannerbot sets mode: +v travis-ci
|
|||
| travis-ci | MoarVM build failed. Stefan Seifert 'Fix call to memcpy with NULL *src in lego JIT compiler | 11:28 | |
| travis-ci.org/MoarVM/MoarVM/builds/475947940 github.com/MoarVM/MoarVM/compare/4...e7596dd394 | |||
|
11:28
travis-ci left
11:55
domidumont joined,
p6bannerbot sets mode: +v domidumont
13:06
Ven`` left
13:16
patrickb joined
13:17
p6bannerbot sets mode: +v patrickb
14:22
lucasb joined,
p6bannerbot sets mode: +v lucasb
14:29
leont joined
14:30
p6bannerbot sets mode: +v leont
14:56
leont left
15:11
leont joined
15:12
p6bannerbot sets mode: +v leont
15:19
Ven`` joined
15:20
p6bannerbot sets mode: +v Ven``
|
|||
| dogbert11 | nine++ thx | 15:58 | |
| .seen samcv | |||
| yoleaux | I saw samcv 15:52Z in #perl6-dev: <samcv> so i guess the answer is yet :) | ||
|
15:58
Ven`` left
|
|||
| robertle | how does moar determine how much memory to use? is there a limit like the -Xmx on the jvm? | 15:59 | |
| dogbert11 | .tell samcv time permitting would it be possible for you to take a look at M#1030 ? | ||
| yoleaux | dogbert11: I'll pass your message to samcv. | ||
| synopsebot | M#1030 [open]: github.com/MoarVM/MoarVM/issues/1030 [⚠ blocker ⚠] Valgrind errors when running t/spec/S32-io/utf16.t | ||
|
17:03
brrt joined
17:04
p6bannerbot sets mode: +v brrt
|
|||
| brrt | ohai #moarvm | 17:04 | |
|
17:05
domidumont left
|
|||
| brrt | gcc, as recent as 4.8, is a bit more strange than you might believe | 17:05 | |
| gcc -lm foo.c -o foo # works on newer gcc, clang, etc | 17:06 | ||
| gcc foo.c -lm -o foo # required on older gcc | |||
| anyway.... all is now well again :-) | 17:10 | ||
| samcv | dogbert11, let me take a look now | 17:14 | |
| yoleaux | 15:59Z <dogbert11> samcv: time permitting would it be possible for you to take a look at M#1030 ? | ||
| synopsebot | M#1030 [open]: github.com/MoarVM/MoarVM/issues/1030 [⚠ blocker ⚠] Valgrind errors when running t/spec/S32-io/utf16.t | ||
|
17:18
huggable left
|
|||
| Geth | MoarVM: 583a2de8f7 | (Bart Wiegmans)++ | build/Makefile.in [JIT] LDLIBS should go after the source files, before output I had no idea that this used to be a requirement. But this caused travis to fail and be angry. |
17:18 | |
| MoarVM: 60c152f6f6 | (Samantha McVey)++ | tools/release.sh Use GNU type archive to fix archive extract on AIX tar This should hopefully fix issue #1005, so the next tar archive we create will extract fine with AIX. I'm guessing the latest version of GNU tar changes something regarding the format so AIX can no longer support it. Some background on the weirdness of the TAR format can be found here: ftp.gnu.org/old-gnu/Manuals/tar/htm...r_117.html |
17:19 | ||
| dogbert11 | ++samcv | 17:24 | |
| Geth | MoarVM: 07de92199f | (Samantha McVey)++ | src/6model/reprs/VMArray.c Fix compiler warnings on some printf specifiers Use cross platform PRI macros instead of %llu and %lld |
||
|
17:30
Ven`` joined
17:31
p6bannerbot sets mode: +v Ven``
17:39
ZofBot left,
buggable left
17:41
travis-ci joined,
p6bannerbot sets mode: +v travis-ci
|
|||
| travis-ci | MoarVM build passed. Bart Wiegmans '[JIT] LDLIBS should go after the source files, before output | 17:41 | |
| travis-ci.org/MoarVM/MoarVM/builds/476031006 github.com/MoarVM/MoarVM/compare/2...3a2de8f72c | |||
|
17:41
travis-ci left
|
|||
| brrt | yay | 17:44 | |
|
17:51
travis-ci joined,
p6bannerbot sets mode: +v travis-ci
|
|||
| travis-ci | MoarVM build passed. Samantha McVey 'Fix compiler warnings on some printf specifiers | 17:51 | |
| travis-ci.org/MoarVM/MoarVM/builds/476032668 github.com/MoarVM/MoarVM/compare/6...de92199f2d | |||
|
17:51
travis-ci left
18:11
lucasb left
18:12
zakharyas joined,
p6bannerbot sets mode: +v zakharyas
18:16
p6bannerbot left,
p6bannerbot joined,
moon.freenode.net sets mode: +o p6bannerbot,
domidumont joined
18:17
p6bannerbot sets mode: +v domidumont
18:24
leont left
18:26
lizmat left
18:30
lizmat joined,
p6bannerbot sets mode: +v lizmat
|
|||
| Geth | MoarVM/jit-expr-float: 90ac623c47 | (Bart Wiegmans)++ | 2 files [JIT] Split type check from expr type resolution Needs to implement the actual type check, but I can forget about that for a while until more pressing issues are resolved. |
18:38 | |
MoarVM/jit-expr-float: 7 commits pushed by (Bart Wiegmans)++
|
|||
|
18:40
brrt left
18:43
Ven` joined,
p6bannerbot sets mode: +v Ven`
18:51
leont joined
18:52
p6bannerbot sets mode: +v leont
19:39
Ven`` left
19:40
Ven` left
19:44
Ven`` joined
19:45
p6bannerbot sets mode: +v Ven``
19:55
domidumont left,
Ven`` left
20:29
zakharyas left
20:42
Ven`` joined
20:43
p6bannerbot sets mode: +v Ven``
20:44
patrickb left
20:56
Ven`` left
21:07
Ven`` joined
21:08
p6bannerbot sets mode: +v Ven``
21:14
Ven`` left
21:22
Ven`` joined,
p6bannerbot sets mode: +v Ven``
21:33
AlexDaniel left
21:35
AlexDaniel joined,
p6bannerbot sets mode: +v AlexDaniel
22:00
leont left
22:06
vendethiel- joined
22:07
p6bannerbot sets mode: +v vendethiel-
22:34
leont joined,
p6bannerbot sets mode: +v leont
|
|||