bf36412699
removed 'roots', replaced with a registry table
...
- removed cosmoM_addRoot
- removed cosmoM_removeRoot
- renamed cosmoV_register to cosmoV_addGlobals
- added cosmoV_addRegistry
- added cosmoV_getRegistry
- added cosmoV_setProto
- added cosmoV_isValueUserType
2023-09-05 14:35:29 -05:00
6701a63a63
capture freezeGC in CPanic
2023-09-05 02:23:31 -05:00
89be01aaf6
even more error handling refactoring
...
removing all of these useless checks has actually made cosmoV_execute just a
lil bit faster :)
2023-09-04 20:14:53 -05:00
fe136f84b5
removed CERROR
2023-08-30 20:14:03 -05:00
bfdd33e01d
fix vm.collect()
...
we don't freeze the vm on entry to C functions now
2023-08-26 15:03:56 -05:00
c0893b8a14
more refactoring; things seem to work fine
...
all example scripts run fine with GC_STRESS enabled
2023-08-30 12:00:52 -05:00
d30bcace9a
don't freezeGC during GC cycle
2023-08-29 23:32:25 -05:00
6a47c82179
fix more GC bugs
2023-08-29 23:21:52 -05:00
d41126e75f
fix cparse.c gc bug
2023-08-29 23:01:47 -05:00
6126b50941
cosmoV_throw() now resets the vm stack as well
...
also a minor GC bug in cosmoO_newError was fixed.
i'm going to try to phase out cosmoM_freezeGC & friends
since that would cause hell with this new
error handling solution. the only thing still using it is the GC.
2023-08-29 16:48:38 -05:00
7fa7eb8d94
fixed minor memory leak in cparse
...
we keep track of internal values used by the parser by pushing them onto the stack
and popping them off once complete.
2023-08-29 15:27:22 -05:00
0633e87aa6
WIP: removed stale error handling
...
currently, scripts seem to run fine. however I'm a bit worried about stack related issues. maybe i'll need to reset state->top as well? but not entirely sure
2023-08-29 14:07:45 -05:00
75d27afe2c
WIP: major error handling refactoring
...
switching to setjmp instead of the really bad global 'panic' flag
2023-08-28 21:13:00 -05:00
3f39211081
minor refactoring
2023-08-25 23:34:21 -05:00
c5e4305ef8
refactored cobj.c:printObject()
...
uses obj type strings from cosmoO_typeStr
2023-08-25 21:28:41 -05:00
1a78a9ab5f
finally fixed this memory bug
...
we were accidentally tracking frees of stuff that was never
allocated lol
2023-08-25 21:22:10 -05:00
2f0f675159
more debug related refactoring
2023-08-25 20:44:24 -05:00
7c5d2f6b65
minor CTable refactoring
2023-08-25 19:57:16 -05:00
f76f2ffa92
refactoring and cleanup
...
cosmoB_loadOSLib -> cosmoB_loadOS
2023-08-24 23:36:32 -05:00
155e0829fb
minor refactoring
2023-06-03 01:39:35 -05:00
2b3825d258
implemented jump table dispatch
...
- currently only enabled on supported platforms (GNU C Compiler + Clang)
- when enabled, sped up examples/fibtest.cosmo by about 20% (avg of 11.179s prior and 8.799 after)
NOTE: malicious dumps can trivially cause crashes now by having junk function chunks
2023-06-03 01:17:28 -05:00
7bca6927a9
fixed formatting
2023-06-01 22:28:07 -05:00
d3647c161b
added vm.disassemble()
2023-06-01 19:04:12 -05:00
d27d94975e
fixed MSVC support
2023-06-01 22:22:44 -05:00
9537a2c7b8
major main.c refactoring
...
- now takes command line arguments:
Usage: ./bin/cosmo [-clsr] [args]
available options are:
-c <in> <out> compile <in> and dump to <out>
-l <in> load dump from <in>
-s <in...> compile and run <in...> script(s)
-r start the repl
2023-05-28 20:52:12 -05:00
c44dc88972
minor stack fixes
2023-05-28 12:19:52 -05:00
d581e68166
proper error handling for dump edgecases
2023-05-28 12:16:00 -05:00
2271681cec
undef macros
2023-05-28 12:03:49 -05:00
cf18bbbe54
fixed GC bug in cundump.c
2023-05-28 11:55:48 -05:00
3a872fb83f
these values can be defined at the top of the function
2023-05-28 00:08:28 -05:00
4ed1c79b50
formatting changes
2023-05-28 00:03:50 -05:00
635f31863f
cvm.c: added cosmoV_undump
2023-05-27 23:16:47 -05:00
49a7f68239
cdump.c: major refactoring; various fixes
2023-05-27 23:15:58 -05:00
8efecf71a4
cdump.c: fix instruction and line info dumping
2023-05-27 23:15:12 -05:00
395f352c6e
started cundump.c:cosmoD_undump(); minor fixes
2023-05-25 21:12:21 -05:00
65d37838cd
started cdump.c:cosmoD_dump()
2023-05-25 19:41:13 -05:00
3b13ae1624
minor refactoring and typos
2023-05-25 19:40:15 -05:00
d1a16d990c
removed stale COBJ_STREAM usage
2023-05-25 18:34:39 -05:00
8ac8085d20
syntax: 'var'->'let' 'function'->'func'
...
- 'var' has some weird scoping connotations with users of JS. better to change it to 'let', which will decide whether to make the variable a local or a global
- 'func' looks visually appealing lol
- some minor refactoring done in cparse.c
2023-02-09 15:58:25 -06:00
e335fd95d6
minor formatting fixes
2023-02-09 12:52:36 -06:00
b902ac90de
removed CObjStream
2023-02-09 12:42:09 -06:00
6056f8eb5b
added clang-format
2023-02-09 12:32:48 -06:00
7b5825668d
Added boilerplate for CObjStream
2021-03-20 01:44:03 -05:00
d13cc398c8
Added os.system() to the os.* library
2021-03-20 01:02:13 -05:00
55e6453589
Improved cosmoO_equals performance for strings
2021-02-23 11:49:57 -06:00
c83dca2ab2
Added '__equal' metamethod, slightly refactored cosmoO_equal
...
- ISTRING_EQUAL has been added
2021-02-20 12:42:13 -06:00
3890c9dd1e
Refactored cosmoO_equals
...
This sets up room for the '__equal' metamethod to be added
- cosmoO_equals now requires the state to be passed
- cosmoV_equals now requires the state to be passed
- cosmoT_get now requires the state to be passed
2021-02-19 17:04:23 -06:00
40739e9bea
cparse.c now throws an error on illegal statements
2021-02-19 12:47:12 -06:00
Inversion
1eec23035f
Temporary fix for possible bug
2023-11-03 22:55:19 -05:00
Inversion
fec26ac380
Add optional custom error message to assert
2023-11-03 22:55:09 -05:00
