Commit Graph

481 Commits

Author SHA1 Message Date
dzaima
10d0cdc2fe •SH 2021-08-13 23:22:37 +03:00
dzaima
5c829a7fd7 fix wrong sortedness check for body indices 2021-08-12 05:11:34 +03:00
dzaima
ef92ddc69e •FromUTF8 2021-08-11 00:22:59 +03:00
Marshall Lochbaum
9a0cfa340a Don't allow First of empty or reshaping empty to non-empty 2021-08-07 20:25:07 -04:00
dzaima
2ab2f35a34 move fmtF to •Glyph 2021-08-04 22:06:33 +03:00
dzaima
65fe3d9aa2 basic namespace formatting 2021-08-04 22:02:21 +03:00
dzaima
2710786bf0 format blocks 2021-08-04 21:47:41 +03:00
dzaima
8934db0ec5 impure function mess 2021-08-04 21:06:33 +03:00
dzaima
39d095c8dc fix •-things being considered pure 2021-08-04 20:32:16 +03:00
dzaima
e9aff609bf more m_arr refactoring cleanup 2021-08-03 20:52:53 +03:00
dzaima
6799e8f078 fix >0⥊0 2021-08-03 16:51:25 +03:00
dzaima
849eeff465 initialize ia on allocation always 2021-08-03 02:57:46 +03:00
dzaima
8d7a31c8e9 store object pointer in slice 2021-08-01 18:59:31 +03:00
dzaima
be10e0d509 better fork_c2 2021-08-01 04:15:00 +03:00
dzaima
65f2cca251 better fork_c1 2021-08-01 04:12:46 +03:00
dzaima
86b3882c5e improve ⊸⟜ 2021-08-01 03:53:48 +03:00
dzaima
e5d876eb80 fix -R 2021-08-01 02:32:47 +03:00
dzaima
a444d4827d decSh doesn't really need to be split 2021-08-01 00:27:00 +03:00
dzaima
46d4acb72b special-case HArrs in withFill 2021-07-31 23:56:57 +03:00
dzaima
74a3a87166 handle arr⊸F and (arr G H) faster 2021-07-31 23:29:21 +03:00
dzaima
749b098f1e use decR less 2021-07-31 22:40:33 +03:00
dzaima
1f5cfbe766 more proper PERF_TEST, )ex 2021-07-31 20:02:56 +03:00
dzaima
c00834cf15 move mm_free call into virtual function so dec is simpler 2021-07-30 04:40:56 +03:00
dzaima
36afe7f533 integrate mode for consistent timing
use with
make t=perftest f='-O3 -DPERF_TEST' c
2021-07-30 02:19:15 +03:00
dzaima
b18157d86b native ⥊
it's horrible but it appears to work
2021-07-27 19:48:22 +03:00
dzaima
25c89c06b1 •GetLine 2021-07-25 20:02:48 +03:00
dzaima
64b3a4ae6a clean up cc.bqn, specialize small arrays 2021-07-22 19:48:45 +03:00
dzaima
9f3ca80509 make it actually run
wow
2021-07-22 04:28:19 +03:00
dzaima
e04cd3c135 make gcc warn less 2021-07-22 03:38:25 +03:00
dzaima
4b8210d1bd apparently i broke 𝕨/𝕩.. 2021-07-22 03:23:49 +03:00
dzaima
0aceb2f8bb correct outdated comments in h.h 2021-07-22 01:24:14 +03:00
dzaima
55f7f8ef5b couple more type fixes 2021-07-21 00:03:45 +03:00
dzaima
5dd7288144 strict prototypes 2021-07-20 22:21:59 +03:00
dzaima
7ee8d642dc trailing newlines 2021-07-20 22:18:37 +03:00
dzaima
d4d4521e52 more fixing of warnings 2021-07-20 22:09:59 +03:00
dzaima
0bcf3b592f more comparison type fixing
and some more / opt to make it not slower
2021-07-20 21:56:35 +03:00
dzaima
99896fbace improving loop counter variable types 2021-07-20 21:32:41 +03:00
dzaima
68c2332fef return on FAIL 2021-07-20 19:58:34 +03:00
dzaima
c2c64f9654 move thrF to a full function
0.3% more instrs with it never being called what
2021-07-20 02:57:18 +03:00
dzaima
78cd23907e make err fails more clear 2021-07-19 23:09:10 +03:00
dzaima
0d5af28043 use mut_copyG in one more place 2021-07-19 19:56:14 +03:00
dzaima
2e2b479d37 shorten mut_init 2021-07-19 19:25:25 +03:00
dzaima
ca264d1332 move init out of mut_to 2021-07-19 19:23:36 +03:00
dzaima
cf10914d63 refactor mut type expanding 2021-07-19 04:54:29 +03:00
dzaima
d8cfac38a2 move mut_pfree out of .c 2021-07-19 04:41:01 +03:00
dzaima
0c04b7e03b move vec_add slow case to .c file 2021-07-19 04:36:04 +03:00
dzaima
708e06c0ea simplify mut usage where type is guaranteed correct 2021-07-19 04:22:18 +03:00
dzaima
9c66cbbb4b move noinline fns out of .h files 2021-07-19 03:48:52 +03:00
dzaima
61e58c7a02 basic monadic arith special-case 2021-07-18 18:12:11 +03:00
dzaima
d8e3f78e6c better dyadic arith messages 2021-07-18 17:24:21 +03:00
dzaima
1719374e0a make ubsan happier, optimize nextBC
still doesn't like f==(f64)(someIntType)f but that's, like, not replaceable
2021-07-17 21:25:11 +03:00
dzaima
43d2946447 fix refcount errors in merge 2021-07-16 23:11:34 +03:00
dzaima
9ae1ab80cf multiple body support 2021-07-16 22:42:22 +03:00
dzaima
c192125a9c fix •file in REPL 2021-07-12 12:10:06 +03:00
dzaima
cdc363c94f initialize rtperf time counters 2021-07-12 01:45:28 +03:00
dzaima
bb463c8f87 fix rtverify failing on usz==u64 2021-07-12 01:06:39 +03:00
dzaima
576876dd32 fix warnings for usz==u64 2021-07-12 00:50:29 +03:00
dzaima
ed691f3dae joined 2buddy 2021-07-12 00:17:59 +03:00
dzaima
8f3f9ba114 don't read mmInfo from RAM 2021-07-11 23:24:13 +03:00
dzaima
2a0b91e7b4 counters instead of allocB 2021-07-11 23:07:07 +03:00
dzaima
1b31c88e4c use mm_heapUsed() instead of allocB 2021-07-11 22:45:08 +03:00
dzaima
9982c62526 move OBJ_COUNTER increment 2021-07-11 22:14:28 +03:00
dzaima
48380be063 separate allocB for mmX 2021-07-11 22:09:26 +03:00
dzaima
1d5e7212fd extract bucket definitions from template 2021-07-11 21:13:41 +03:00
dzaima
edd1f308ce clean up mm macros & globals 2021-07-11 21:04:49 +03:00
Marshall Lochbaum
dde05f5f02 Just use bit sets, not hashing, for complementary selection 2021-07-08 20:56:28 -04:00
Marshall Lochbaum
3546419816 Use the correct complement algorithm for the hash set 2021-07-08 17:54:34 -04:00
Marshall Lochbaum
619b9827de Faster random deal, and sorted subset 2021-07-08 17:33:50 -04:00
dzaima
cce52dd40e random deal & subset 2021-07-07 21:31:48 +03:00
dzaima
fc4d36de07 test for MAP_32BIT to enable JIT 2021-07-06 20:04:14 +03:00
dzaima
aef9702b7b list-of-structs → struct-of-lists for TypeInfo 2021-07-06 19:31:42 +03:00
dzaima
08723a3c80 use TI macros everywhere 2021-07-06 19:22:45 +03:00
dzaima
cd60db1c89 optimize fast allocation path 2021-07-05 02:52:42 +03:00
dzaima
d063e83a8d apparently BSZI is barely used 2021-07-05 01:22:49 +03:00
dzaima
7bc8856cd1 inline LOCU 2021-07-05 01:05:43 +03:00
dzaima
2a49cf8205 update objdump usage 2021-07-04 13:49:27 +03:00
dzaima
eca7699580 more ⊔ changes 2021-07-04 05:31:23 +03:00
dzaima
6942b2d04c add checks for 𝕨<¯1 in ⊔ 2021-07-04 05:05:13 +03:00
dzaima
f5388fbdfa depth check for ⊔ 2021-07-04 05:00:56 +03:00
dzaima
774db7a956 tagged bytecode position 2021-07-04 02:13:52 +03:00
dzaima
223d6ccca0 GS_SET on RET_ only when needed 2021-07-04 00:56:43 +03:00
dzaima
51c41f0766 push out perf/writing stuff from m_nvm 2021-07-03 23:32:56 +03:00
dzaima
72896feb8f merged SET_i+POPS 2021-07-03 18:04:27 +03:00
dzaima
9ee36dae94 fix TypeInfo.arrD1 for c32arr 2021-07-03 17:01:39 +03:00
dzaima
ee3ef4370c waste less on thrF 2021-07-03 02:55:22 +03:00
dzaima
7999683edf skip value checking for builtin loading
perf is reporting an 0.15% increase in instruction count (what?), but that's not gonna stop me
2021-07-03 02:38:09 +03:00
dzaima
f75efbcc26 static harr.c functions 2021-07-03 02:04:31 +03:00
dzaima
ceb967bbd4 return Arr* from m_fillarrp 2021-07-03 01:53:26 +03:00
dzaima
08df9bfe7b arrP_sh* → arr_sh* 2021-07-03 01:38:56 +03:00
dzaima
d49d5a3e31 don't use arr_sh* 2021-07-03 01:37:54 +03:00
dzaima
c0a75327ef return Arr* from mut_fp 2021-07-03 01:25:55 +03:00
dzaima
c52a042eb4 return Arr* from slices 2021-07-03 01:25:55 +03:00
dzaima
daa8294252 return Arr* from m_(i32|c32|f64)arrp 2021-07-03 00:37:00 +03:00
dzaima
49810ce8bc rename mm_allocN 2021-07-03 00:27:20 +03:00
dzaima
43d0d1e043 delay tagging for arrays 2021-07-03 00:22:40 +03:00
dzaima
d8b2ac0268 delay tagging 2021-07-03 00:02:56 +03:00
dzaima
c62369c0c5 initialize objects a bit better
0.29% improvement
2021-07-02 23:07:46 +03:00
dzaima
693f8e0139 CSTACK==0 didn't work anyway 2021-07-02 22:29:16 +03:00
dzaima
3451cc48c2 0 parent scope loading 2021-07-02 22:26:14 +03:00
dzaima
be5c22af37 update gStack only when needed
a tiny 0.13% improvement
2021-07-02 22:19:11 +03:00