Pavel R.
|
072d2c37f0
|
Fixed possibility to initialize empty strings (and booleans)
|
3 years ago |
annelin
|
73de201165
|
added new (optional) types to lparser.c (no idea what does it affect)
|
3 years ago |
annelin
|
be65484feb
|
Removed RAVI_DEFER_STATEMENT define (there are no reasons to disable it)
|
3 years ago |
Pavel R.
|
5359ca1d5f
|
- Added boolean type
- Types now can be optional (example: string?)
- Non-optional types now are more strict (e.g. you can not pass nil to function that expects string)
|
3 years ago |
Dibyendu Majumdar
|
e3d8203bd9
|
Refactor
|
3 years ago |
Dibyendu Majumdar
|
063a55604e
|
Merge branch 'new_type_info' into type-maps
|
3 years ago |
Dibyendu Majumdar
|
658f04c3d8
|
issue #215 adapt fix by XmiliaH from pull request
|
3 years ago |
Dibyendu Majumdar
|
8fd3a1bbab
|
issue #214 since we haven't implemented boolean type annotation fully, the parser will no longer recognize the :boolean annotation.
|
3 years ago |
Dibyendu Majumdar
|
b5afdfaa46
|
issue #215 The top level check v->ravi_type != vartype is not correct as it means we miss out checking scenario where v->k == VINDEXED.
|
3 years ago |
Dibyendu Majumdar
|
88ccaf34a0
|
Reformat
|
3 years ago |
XmiliaH
|
b4359b9391
|
Cleanup and Fixes
|
3 years ago |
XmiliaH
|
7789f4f32d
|
Finish type maps
|
3 years ago |
XmiliaH
|
6af3d804a4
|
First part for type maps
|
3 years ago |
XmiliaH
|
4cce67ec13
|
Improve type deduction
|
3 years ago |
Dibyendu Majumdar
|
77cd6b9527
|
Update copyright notices
|
3 years ago |
Dibyendu Majumdar
|
3b09a22e96
|
issue #204 Code reorg
|
3 years ago |
Dibyendu Majumdar
|
bf2f55eece
|
issue #163 Do not generate TAILCALL opcode if a defer closure is in scope; back-ported from Lua 5.4.
|
4 years ago |
Dibyendu Majumdar
|
062550b84d
|
issue #196 backport generational GC from Lua 5.4
|
4 years ago |
Dibyendu Majumdar
|
b8af71583d
|
issue #163 'defer' implementation now controlled via RAVI_DEFER_STATEMENT macro. This allows testing of performance with and without the 'defer' statement
|
4 years ago |
Dibyendu Majumdar
|
76accd1930
|
Update copyright notices
|
4 years ago |
Dibyendu Majumdar
|
51eeb4e8c2
|
issue #163 simplify defer statement and fix memory leak
|
4 years ago |
Dibyendu Majumdar
|
edce55b997
|
issue #163 Base must be protected as luaF_close can now reallocate stack due to invocation of deferred functions
|
4 years ago |
Dibyendu Majumdar
|
9a0ced9620
|
issue #163 First attempt to implement a Go like defer statement
|
4 years ago |
Dibyendu Majumdar
|
31f4e178ee
|
issue #157 more opcode renaming
|
6 years ago |
Dibyendu Majumdar
|
fc6929aa58
|
issue #157 more opcode renaming
|
6 years ago |
Dibyendu Majumdar
|
d4d679bb3c
|
issue #156
|
6 years ago |
Dibyendu Majumdar
|
9f3043dcc9
|
reduce compiler noise (warnings)
|
6 years ago |
Dibyendu Majumdar
|
471280ca3a
|
tech debt compiler warnings
|
6 years ago |
Dibyendu Majumdar
|
a3fdf76103
|
issue #98 refactoring and start on value types
|
6 years ago |
Dibyendu Majumdar
|
f9793dc527
|
issue #136 added support for user defined type names to have '.' in them
|
6 years ago |
Dibyendu Majumdar
|
64a6a2372b
|
issue #146 Currently Ravi's code generator assumes that #operator returns an integer value, but if user has set a metamethod for __len the operator# might return a non-integer value. This issue came to light when testing lpeg, which redefines #operator to return userdata objects. We now try to only expect integer values when # is used in fornum loops, or applied to table or array. Additionally unless the target is an array, we also emit a TOINT opcode to validate at runtime that the value in integer type
|
6 years ago |
Dibyendu Majumdar
|
b2c6c8ebd4
|
fix a bug in issue #136 implementation and add experimental support for computed gotos in the VM
|
7 years ago |
Dibyendu Majumdar
|
c96483431f
|
issue #136 more type checking
|
7 years ago |
Dibyendu Majumdar
|
1e23abbbff
|
issue #136 more type checking
|
7 years ago |
Dibyendu Majumdar
|
daf2b8bfe9
|
issue #136 wip
|
7 years ago |
Dibyendu Majumdar
|
5e03430d7d
|
issue #136 additional type assertions
|
7 years ago |
Dibyendu Majumdar
|
ac064238bd
|
remove unused code
|
7 years ago |
Dibyendu Majumdar
|
77e2f6268d
|
issue #119
|
7 years ago |
Dibyendu Majumdar
|
0f089b6b92
|
issue #119
|
7 years ago |
Dibyendu Majumdar
|
46a2282f3f
|
issue #111 Lua 5.3.3 bug fixes
|
8 years ago |
Dibyendu Majumdar
|
cd0150e08e
|
issue #100
|
8 years ago |
Dibyendu Majumdar
|
e91538b54b
|
issue #89 add support for 32-bit numbers
|
8 years ago |
Dibyendu Majumdar
|
2213f866a6
|
issue #86 start new attempt to merge 5.3.3
|
8 years ago |
Dibyendu Majumdar
|
25a5fb3c9e
|
issue #86 merge reverted due to bug - to be fixed in the branch first
|
8 years ago |
Dibyendu Majumdar
|
86a2adb4f7
|
issue #86: merge Lua 5.3.3
|
8 years ago |
Dibyendu Majumdar
|
5dc495720f
|
the previous change related to module sharing between functions appears to have introduced memory leak - caught by asan; also apply fixes discussed in Lua mailing list
|
8 years ago |
Dibyendu Majumdar
|
c47f132349
|
use new approach to locate OP_NEWTABLE when checking types of assignments to local variables
|
8 years ago |
Dibyendu Majumdar
|
762ab319f5
|
issue #76 type assertions
|
9 years ago |
Dibyendu Majumdar
|
b23e0eab16
|
issue #76 type assertions
|
9 years ago |
Dibyendu Majumdar
|
aa453e0aae
|
issue #76: work on type assertion
|
9 years ago |