Age | Commit message (Collapse) | Author |
|
|
|
TODO: Use mutex in lhs expr and set resolved_type
|
|
Hash map now stores hash of keys to reduce the number of hash
operations.
Positive: faster insert/get.
Negative: more space required (to store usize hash).
|
|
instance
|
|
|
|
|
|
|
|
Fix hash map bug on rorder (invalid size).
Fix string to int/float conversion in tokenizer.
Add more ssa functions
|
|
|
|
|
|
future steps
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Fix issue where not all files are parsed
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
It caused issues when you have parentheses to surround math expression,
for example: ((func() + 34) * 54)
is easier to parse if closure has to begin with 'fn'.
Also removed requirement for semicolons. Semicolons can't even be used
optionally yet.
|
|
|
|
|
|
|
|
already parsed
|
|
|
|
|
|
only join
|
|
|
|
|
|
|
|
allocator
|
|
|
|
Scoped allocator gives us better performance and cleanup code for error
cases is much cleaner
|
|
Scoped allocator gives us better performance and cleanup code for error
cases is much cleaner
|