c64f53a050
Detect duplicate variable declarations correctly
...
Later I'll probably want to make it so that you can explicitly override
the value of a declared variable
2019-03-10 17:02:01 -07:00
8f176543c7
Nested scopes in symbol table
2019-03-10 16:04:20 -07:00
9716b5e55b
Symbol table detects some duplicate symbols
2019-03-08 03:57:32 -08:00
956353cd80
Move rc! macro to util
...
So it can be used anywhere
2019-03-08 01:15:19 -08:00
98db60498a
Add very basic symbol table test shim
2019-03-07 23:51:31 -08:00
7694afc9e2
Add type for talking about symbol paths
...
to symbol table
2019-03-07 20:45:12 -08:00
0bcd7e6f41
Add new_env method
...
This is basically the same as the one on the evaluator and makes use of
the ScopeStack - maybe need to generalize this more?
2019-02-27 02:15:19 -08:00
d515b1658a
Some fixes
2019-02-24 16:24:45 -08:00
e501f4bd10
Various cleanup of comments, stringifying types
2019-02-23 09:59:41 -08:00
5bac01cf20
More boilerplate for apply
2019-02-23 03:55:46 -08:00
0e9b3229e9
Refactor Arrow; add general handle_apply
2019-02-23 03:33:56 -08:00
b709cfd51a
Start adding call
2019-02-23 02:50:11 -08:00
e34295a6f7
Starting on lambda typechecking
2019-02-23 02:45:11 -08:00
8dc34e4b49
Fresh type var
2019-02-23 01:27:32 -08:00
2cc3367666
Unify var-var
2019-02-23 01:20:19 -08:00
452f2ab188
Unify var-const
2019-02-23 01:18:15 -08:00
be175a2b75
Add more infrastructure for unify
2019-02-23 00:59:58 -08:00
00a0de4431
Add ena crate for unification
2019-02-23 00:34:44 -08:00
f041cc17d2
Wrap all Expression nodes in Meta<Expression>
2019-02-21 23:35:18 -08:00
95fe1941a1
Kill some unused items
2019-02-21 18:39:41 -08:00
b35262c444
Rename Node -> Meta
2019-02-21 01:49:15 -08:00
9bb3a2be88
Add type data handle on Node
...
I think the way I want to handle this is a two-step process: first infer and
fill in variables, then unfiy in a separate step. Storing the data in
the AST is handy.
2019-02-21 01:46:27 -08:00
9fa0576547
Rename ExpressionType -> ExpressionKind
2019-02-21 01:26:51 -08:00
6fba0cc5b4
Add variables
2019-02-21 01:17:34 -08:00
a6eb2b4020
Allow type annotations in let expressions
2019-02-20 22:44:45 -08:00
03793e08d3
Typechecking operators
2019-02-20 03:27:46 -08:00
2be55958f4
add Into<String> arg for error constructors
2019-02-20 02:06:58 -08:00
bcf48d0ecb
First tests for typechecking
2019-02-20 01:33:45 -08:00
f0ed63ccf3
Basic if-statement checking
2019-02-19 23:00:41 -08:00
6012bd1087
Variables
2019-02-19 21:41:07 -08:00
df7e74c79d
Types with arguments
2019-02-17 04:31:02 -08:00
abbd02eaef
Use ty! macro
2019-02-17 04:25:38 -08:00
993741e67f
Get rid of typecheck_
2019-02-17 04:08:49 -08:00
fbb7b995b8
Rename mk_type! to ty!
...
Doesn't seem to conflict with the same macro in the parser tests, so should be
ok
2019-02-17 03:38:15 -08:00
9d4f086a04
Put mk_type! in typechecking module
2019-02-17 03:36:12 -08:00
e38ae1c3f1
Fix type to make it compile
2019-02-15 21:11:50 -08:00
d969d573fa
Starting work on values
2019-02-12 21:14:13 -08:00
35da1748f0
Some more type work
2019-02-10 12:21:12 -08:00
5e1799268d
Unification works with bad annotations
2019-02-10 07:32:12 -08:00
42a801d346
Rename Order -> Ordering
2019-02-10 07:06:30 -08:00
a80e1bd706
Type name infra
2019-02-10 07:05:01 -08:00
afd9aa52c5
More infra around unify
2019-02-10 06:53:11 -08:00
5a70784346
Adding unify stub
2019-02-10 06:48:25 -08:00
0dff177e8f
Add more literals kill errors
2019-02-10 05:33:55 -08:00
cf91f74912
Pass through type info to repl
2019-02-10 05:31:58 -08:00
06e9452718
More type infrastructure
...
From here on out, I can start playing with concrete code that attempts
to actually typecheck limited syntactic constructs, and see what I end
up with.
2019-02-10 05:24:11 -08:00
7d3ae36058
AST-walking infrastructure
2019-02-10 04:42:30 -08:00
e8f1f51639
Move (most of) the type definitions back to typechecking module
...
Still need to figure out the macro export thing
2019-02-10 04:30:37 -08:00
170cf349d7
Starting typechecking work again
2019-02-09 00:25:12 -08:00
c0111e30bc
SymbolTable: Add Record type
2019-01-25 00:57:01 -08:00
c225e469ee
Change Record variant representation
2019-01-24 20:47:20 -08:00
1ce06bc0ef
More symbol-table refactoring
2019-01-20 22:32:58 -08:00
10c3a60515
Some symbol-table refactoring
2019-01-20 00:23:08 -08:00
a63dcf91b0
Replace // with `quot`
2019-01-10 20:57:13 -08:00
479a098e0f
Make note to fix parsing bug
2019-01-09 17:37:51 -08:00
1085b528fe
Don't care about case-sensitivity
2019-01-08 02:38:10 -08:00
9b3b5c5541
Token offsets
2019-01-08 02:11:19 -08:00
ab8e24a276
ParseError always has token now
2019-01-08 01:04:46 -08:00
09e2d8579d
Remove all ParseErrors that don't return the failed token
2019-01-08 01:00:40 -08:00
ee7861cbd0
Fix how impl blocks work
...
This gets rid of a token-less parseerror
2019-01-08 00:51:56 -08:00
b88def8a2e
Make error msg better
2019-01-07 16:52:46 -08:00
30676722a3
Transition to edition 2018
2019-01-07 13:00:37 -08:00
801c90aaa7
Reorder parameters in pass functions
2019-01-07 02:43:31 -08:00
fa295aab28
Show location of error in parse error
2019-01-07 01:52:54 -08:00
a0f4abb9a3
Add SourceReference
2019-01-06 01:58:16 -08:00
78b454fb32
Delete this line
2019-01-05 20:41:49 -08:00
5491169d55
Refactor parsing structure
...
TokenHandler should contain all the methods for actually manipulating
tokens, Parser should only contain the recursive descent methods
2019-01-05 20:23:07 -08:00
2b338fd3c9
Move .next() onto token_handler
2019-01-05 18:29:24 -08:00
821f321261
More Node-wrapping of Expression
2019-01-05 18:11:51 -08:00
846eeae04c
More use of Token instead of TokenKind
2019-01-05 17:50:54 -08:00
22f2750853
More use of Token in error messages
2019-01-05 17:40:05 -08:00
f2f8ac7ee8
Make parserrors have token
2019-01-05 17:32:49 -08:00
d0c5dce92b
Use get_kind()
2019-01-05 17:28:35 -08:00
8eda74c9a5
Starting to keep track of locations of errors in file
2019-01-05 17:18:10 -08:00
2efac109ef
Node for TupleLiteral
2019-01-05 16:06:55 -08:00
215e2bbb0d
PrefixOp have Node
2019-01-05 16:02:30 -08:00
2590def3be
More Node-wrapping
2019-01-05 15:54:03 -08:00
879a7de83d
Wrap Expression in Node
2019-01-05 15:47:44 -08:00
4c2e0b8a21
Get rid of old code from old ideas
2019-01-05 15:35:51 -08:00
282c42da3c
Adding Node intermediate type to AST
2019-01-05 01:44:32 -08:00
7ac97ca6e8
Kill ast_visitor
2018-11-26 01:57:44 -08:00
26a8ff307f
Add generic Node type
2018-11-20 03:21:10 -08:00
6be208b51d
Minor fix for parsing error messages
2018-11-20 03:03:08 -08:00
e00948cad9
Add ast_visitor mod
2018-11-17 02:09:16 -08:00
0af6fed505
Clear up some code a bit
2018-11-17 01:10:23 -08:00
1f527f7949
Rename TokenType -> TokenKind
2018-11-16 23:17:34 -08:00
b198984fc5
implement From for Expression-types
2018-11-16 14:06:04 -08:00
58779f8470
Rename method, make sourcemap optional
2018-11-16 12:58:10 -08:00
a0fa50392c
Fix compile error
2018-11-16 12:46:29 -08:00
d357876b16
WIP source map stuff
2018-11-16 04:12:07 -08:00
e42f0c644c
Introduce source map
2018-11-16 03:56:55 -08:00
2ec7bf3b9a
Some initial work on passing token metadata to AST
2018-11-16 03:51:03 -08:00
5147e1a3eb
Handle underscores in identifiers
2018-11-15 16:19:53 -08:00
955c073174
Got typechecker unused errors down to one
2018-11-13 02:39:02 -08:00
7c46a29141
Start adding doc comments
2018-11-11 18:04:44 -08:00
0adc761e72
Kill an unimplemented!()
2018-11-11 02:48:51 -08:00
b2039a7b67
Parameterize Type type over existential/universal
2018-11-10 16:33:42 -08:00
b4c4531e4d
Rename for more concision
2018-11-10 14:11:29 -08:00
2d36ad44d6
Converting over types
...
WIP
2018-11-09 02:50:29 -08:00
21132a369c
Paramaterize Type
2018-11-09 02:05:59 -08:00
ff0294c56e
Typechecking shouldn't fail yet
2018-11-09 02:02:08 -08:00
e39356c0e5
Even more type work
2018-11-09 00:21:34 -08:00
d44bb02d61
Even more types
2018-11-08 20:30:17 -08:00
9056e9b0e1
More type work2
2018-11-08 02:29:54 -08:00
e9b90412ce
More type work
2018-11-08 02:12:01 -08:00
65c47c20fc
Change name of monad in which type inference happens
2018-11-07 17:01:07 -08:00
fab3fb8ec2
More basic types + test
2018-11-07 16:39:32 -08:00
0d5ccd21fe
TConst
2018-11-07 15:39:40 -08:00
69b7b9f528
Print out types to REPL
2018-11-07 13:44:28 -08:00
9a09f40222
More typing work
2018-11-07 03:39:31 -08:00
020819550b
More typechecking infrastructure
2018-11-06 16:47:34 -08:00
15f9dbe7a6
Typechecking infrastructure
2018-11-06 13:44:52 -08:00
836bed1207
Added janky map to prelude
2018-11-06 03:02:32 -08:00
cee5b085d5
Simpler syntax for single param in lambdas
...
This kind of implies that I might want -> for function types after all,
instead of :
2018-11-06 02:58:57 -08:00
837a55c718
Test for nested function call
2018-11-06 02:42:28 -08:00
f4f89b39b6
Handle nested function calls
2018-11-06 02:40:10 -08:00
c6b4ed7ee4
Basic lambdas
2018-11-06 01:19:16 -08:00
be425860af
Starting on lambdas
2018-11-05 21:13:31 -08:00
17e88b33f2
Eval test doesn't need to be a macro
...
Can be a fn
2018-11-05 21:07:06 -08:00
47f7eb1ef6
Make prelude be separate file
2018-11-05 20:55:03 -08:00
72d0cfe466
More macro test consolidation
2018-11-05 20:52:18 -08:00
cea2f63b44
Use macros to make types more concise
2018-11-05 20:12:10 -08:00
eec315dd58
Get rid of exprstatement! macro
...
For shorter exst! one
2018-11-05 19:58:55 -08:00
1e9aa91c5d
More concise test macros
2018-11-05 19:57:11 -08:00
9813609ad7
Minor test refactoring
2018-11-05 19:17:53 -08:00
5953d9d815
type annotations on lambdas
2018-11-05 19:10:34 -08:00
a74e09c761
Change lambda syntax
2018-11-05 18:51:01 -08:00
ad53d4394b
Get rid of println
2018-11-05 14:52:51 -08:00
151246e1c5
Test for pattern-matching
2018-11-05 14:11:49 -08:00
77d2826918
Pattern-match on structured objects
2018-11-05 14:01:14 -08:00
1bd48ed5db
Fix problem with parsing commas
...
I should probably rethink how delimited block expressions like if-blocks
(and eventually for-blocks) work
2018-11-05 13:07:08 -08:00
c394b81746
More pattern-matching
2018-11-05 04:02:04 -08:00
ec29077247
More tuple-matching
...
Also discovered parser bug
2018-11-05 03:41:03 -08:00
62043ac2d1
Starting on pattern-matching tuples
...
Lots of duplicated code here
2018-11-05 03:17:03 -08:00
bada386979
More work on subpattern matching
2018-11-03 12:53:09 -07:00
e71d404071
Finished this refactor
2018-11-02 19:54:04 -07:00
cab4702bd6
Refactoring matching - WIP
...
doesn't work yet
2018-11-01 02:43:47 -07:00
ec5a9d457e
String patterns
2018-10-31 01:45:16 -07:00
bfbc1580aa
Make tag optional
2018-10-30 23:36:55 -07:00
2d6c9010b9
More work here
2018-10-30 18:53:34 -07:00
f4ff92302f
Use subpattern abstraction
2018-10-30 18:46:06 -07:00
e88ed97b06
Add subpattern struct
2018-10-30 18:39:25 -07:00
b8df09e956
Change eval strategy to use conditional sigil
2018-10-29 01:50:43 -07:00
d7f0147a4f
Add conditional target placeholder expr
2018-10-28 12:45:45 -07:00
f883512882
New abstraction layer in Schala-lang parser
...
Just for manipulating tokens
2018-10-21 16:33:21 -07:00
37070a6b3e
Move pass chain generation from macro to codegen
2018-10-20 18:00:05 -07:00
d7baf065fb
Changing what method to call to start parsing
2018-10-20 15:41:09 -07:00
6b42f8b8de
Change how parsing works
2018-10-20 14:27:00 -07:00
7de536ade0
Install failure crate
2018-10-20 11:17:18 -07:00
4679a9fc7f
Remove compiler warnings
2018-10-20 00:55:37 -07:00
c25354b2c7
Get rid of typechecking code (for now)
...
I'm tired of seeing the errors. See branch last_commit_with_typechecking
2018-10-20 00:41:56 -07:00
5f8b842bf2
Delete newline
2018-10-20 00:22:41 -07:00
fef66e345b
Subpattern field
2018-10-19 17:43:22 -07:00
e57d33eae7
More work on more patterns
...
-need to convert guard into a possibly-empty vec
2018-10-19 17:27:06 -07:00
dca9ad06c3
Handle HalfExpr closer to correct
2018-10-19 11:02:10 -07:00
354148c5ba
rename codegen -> chala-lang-codegen
2018-10-19 09:57:35 -07:00
6219a06d6f
Converted all parser methods to use the annotation
2018-10-19 02:56:11 -07:00
3b20b40eb7
Proc macro generated code for parsing seems to work
2018-10-19 02:45:35 -07:00
4ecf63c54d
Okay the proc_macro is actually doing something
...
At the cost of breaking code
2018-10-19 02:36:23 -07:00
3d00667caf
Add test for ignored pattern
2018-10-18 15:55:24 -07:00
4b9c7e38dd
Rename TypeName -> TypeIdentifier
2018-10-18 13:27:09 -07:00
dff204069f
Starting to implement Ignored pattern
2018-10-18 01:54:36 -07:00
f2282f0101
case_match_expression split out into its own method
2018-10-18 01:49:42 -07:00
40ccea8c05
Separate assign_expression method
2018-10-18 01:46:30 -07:00
cae6f2f768
Rename schala-codegen -> schala-repl-codegen
2018-10-18 01:09:29 -07:00
1be6991f55
Making eval expression method a bit less complex
...
by splitting it into submethods
2018-10-17 20:46:16 -07:00
1b60bd38ff
Add codegen crate for schala-lang
2018-10-17 15:29:32 -07:00
3b20b9e209
Put schala-lang crates into a subdirectory
2018-10-17 14:51:48 -07:00
de0e150536
Fix if-block parsing to handle newlines
2018-10-17 13:44:29 -07:00
baf51fb147
Boolean patterns
2018-10-17 12:43:09 -07:00
dc9e493fa1
Handle more patterns at reduce_ast level
2018-10-16 17:18:03 -07:00
d57a8045a9
Rename test helper
2018-10-16 04:11:18 -07:00
50d5176b45
Fix bug add test
2018-10-16 04:10:28 -07:00
501eaeee87
Implement numeric pattern matching
2018-10-16 03:54:08 -07:00
8619c94217
Start handling numeric patterns
...
Still need to add eval support for this
2018-10-16 01:38:41 -07:00
9927a6b1fd
Implement custom interpreter directives - and a wtf?
...
See the comment about &mut self vs &self
2018-10-15 20:29:23 -07:00
abe2db25b2
full if matching working with basic patterns
2018-10-15 19:54:17 -07:00
a99a5f10a4
Indicate index explicitly in SymbolTable debug
2018-10-15 19:37:02 -07:00
80eb703f5e
Finally got get_doc hookup in codegen macro working
2018-09-29 01:20:31 -07:00
4fccff5e27
Working on improved proc_macro handling
2018-09-27 04:07:42 -07:00
5d4505241a
get rid of completed todo
2018-09-22 00:26:38 -07:00
f67793308e
Part of the work for a doc handler
2018-09-22 00:24:27 -07:00
693766fa59
Proc macros are stable now
2018-09-21 19:46:31 -07:00
1abbe2e448
Add guard to Alternative
...
The semantics are:
-if tag is Some(_), assume the condition is a constructor,
and compare tags
- if guard is Some(_), evaluate true/false *after* having
applied any bound variables
With this, I can technically get rid of bare conditionals now, since
they are the same as an Alternative with a None tag
2018-08-27 12:45:08 -07:00
065bdd6bda
Starting custom operators
...
Can now parse custom operators. Maybe I want to make it so that you
have to put explicit backticks if you define a custom operator,
currently you can just do: fn +() { .. }
2018-08-24 16:49:59 -07:00
e125e8b440
Add spaceship operator for getting an ord
2018-08-24 16:29:28 -07:00
8565c7dfb3
Some work on reduced ast pattern
2018-08-24 16:04:18 -07:00
f885d5dfb6
Remove type alias
2018-08-22 23:22:08 -07:00
b85725125c
Start using HalfExp
2018-08-22 16:41:31 -07:00
2d961d6402
Fix other pattern parsing bugs
2018-08-21 20:02:10 -07:00
fa7b6ce96b
Handle negatives in patterns correctly
2018-08-21 19:57:45 -07:00
1d5e5aa735
Some type renaming in builtins
...
Builtins will remain entirely separate from the actual type
representation, whatever that ends up being
2018-08-19 22:00:20 -07:00
2c298c7247
Add warning for undefined operator
...
In practice this will probably always not typecheck, but it's a valid
parse
2018-08-19 21:40:30 -07:00
f00fee0e37
Rename StateStack -> ScopeStack
2018-08-19 21:31:45 -07:00
0d13b5e3bc
Preliminary support for binops in if-discriminators
...
The BNF grammar is a bit more liberal than any successfully-compiled
schala program should be, in that it allows things like `if x < is
pattern`. It's okay if that parses successfully and then is an error at
typechecking.
2018-08-19 21:25:07 -07:00
98f597f00a
Implement comparison operators correctly
2018-08-19 21:11:43 -07:00
fb71881409
Refactor binop parsing
2018-08-19 20:33:50 -07:00
d1c3b4a81b
Starting on halfexprs / binops
2018-08-19 18:44:54 -07:00
f9181b5786
use expr_or_block where appropriate
2018-08-19 15:58:31 -07:00
0e914cf057
Error message for parsing guards
2018-08-19 15:12:34 -07:00
greg