README.md

@@ -1,4 +1,3 @@
 # Akhamoth
 
-An attempt at making something very much like Elixir, but with syntax I prefer
+statically typed functional language
-and static typing as a mandatory part of the language from the beginning.

pages/syntax_reference.md

@@ -1,66 +0,0 @@
-# Syntax Reference
-
-In Akhamoth, everything is an expression.
-
-## Literals
-
-### Identifiers
-
-Identifiers are composed of ASCII letters, digits, and underscores and may not
-begin with a digit. The preferred naming convention is `UpperCamelCase` for type
-names, `snake_case` for variables, functions, and modules, and
-`SCREAMING_SNAKE_CASE` for global constants. Identifiers, beginning with an
-underscore are typically intentionally unused variables and will silence
-warnings about this from the compiler.
-
-> #### Unicode Identifiers {: .info}
->
-> For version 1.0 of Akhamoth, the goal is only to support ASCII identifiers in
-> order to make parsing easier. In the future however, it would be good to look
-> at implementing [UAX #31](https://www.unicode.org/reports/tr31/) unicode
-> identifiers.
-
-### Atoms
-
-Atoms have the exact same syntax as identifiers, but are prefixed with `:`, e.g.
-`:akhamoth`.
-
-### Strings
-
-A String literal consists of `"` followed by any number of other characters and
-then another `"`. Currently there is no support for character escapes of any
-kind. This will be rectified before version 1.0.
-
-### Integers
-
-Integer literals must begin with a digit, followed by any number of digits and
-underscores. Underscores are intended to be used for grouping digits in long
-numbers, e.g. `1_000_000_000`.
-
-## Expressions
-
-### Operators
-
-The following is a list of all operators in Akhamoth, ordered from highest
-precedence to lowest, along with their associativity:
-
-Operator | Associativity
--------- | -------------
-`.`      | left
-`-`      | unary
-`*` `/`  | left
-`+` `-`  | left
-`\|>`    | left
-`==`     | requires parens
-`..`     | requires parens
-`=>`     | right
-`->`     | right
-`=`      | right
-
-### Function Calls
-
-## AST
-
-The design of Akhamoth's AST is essentially the same as [Elixir's][1].
-
-[1]: https://hexdocs.pm/elixir/main/syntax-reference.html#the-elixir-ast

rebar.config

@@ -1,12 +1,3 @@
 {erl_opts, [debug_info]}.
 {project_plugins, [rebar3_ex_doc]}.
 {deps, []}.
-{ex_doc, [
-    {extras, [
-        {"README.md", #{title => <<"Overview">>}},
-        "pages/syntax_reference.md"
-    ]},
-    {main, "README.md"},
-    {source_url, "https://git.wires.systems/wires/akhamoth"},
-    {api_reference, false}
-]}.

src/akh_lexer.erl

@@ -1,64 +1,54 @@
 -module(akh_lexer).
--moduledoc """
+-include("akhamoth.hrl").
-This module contains functions for tokenizing Akhamoth source code.
+-export([
-""".
+    new/1,
+    next/1
+]).
 
--export([new/1, next/1]).
+-type token_kind() ::
+    binary()
+    | '{'
+    | '}'
+    | '['
+    | ']'
+    | '('
+    | ')'
+    | ','
+    | ';'
+    | ':'
+    | '.'
+    | '->'
+    | '+'.
 
--define(is_digit(C), C >= $0, C =< $9).
+-type token() :: {token_kind(), non_neg_integer() | inserted}.
--define(is_id_start(C), C >= $a, C =< $z; C >= $A, C =< $Z; C =:= $_).
 
--doc """
+-define(is_id(C), (C >= $A andalso C =< $Z); (C >= $a andalso C =< $z); C =:= $_).
-A token in the input stream
-""".
--type token() :: any().
 
 -record(lexer, {
     source :: binary(),
     offset = 0 :: non_neg_integer()
 }).
 
-%%% exports
-
--doc """
-Initializes a lexer to tokenize the given binary.
-""".
 -spec new(binary()) -> #lexer{}.
 new(Source) -> #lexer{source = Source}.
 
--doc """
-Attempts to get the next token in the input.
-""".
 -spec next(#lexer{}) -> none | {ok, token(), #lexer{}}.
-next(#lexer{source = <<C, _/bytes>>} = Lx) when ?is_id_start(C) ->
+next(#lexer{source = <<$\n, Rest/bytes>>, offset = Offset} = Lexer) ->
-    lex_id(Lx, 1);
+    next(Lexer#lexer{source = Rest, offset = Offset + 1});
-next(#lexer{source = <<C, _/bytes>>} = Lx) when ?is_digit(C) ->
+next(#lexer{source = <<C, _/bytes>>} = Lexer) when ?is_id(C) ->
-    lex_number(Lx, 1, C - $0);
+    lex_id(Lexer, 1);
 next(#lexer{source = <<>>}) ->
     none.
 
-%%% local functions
+-spec lex_id(#lexer{}, pos_integer()) -> {ok, token(), #lexer{}}.
-
+lex_id(#lexer{source = Source, offset = Offset} = Lexer, Len) when Len < byte_size(Source) ->
-lex_id(#lexer{source = Source, offset = Offset} = Lx, Len) when Len < byte_size(Source) ->
     C = binary:at(Source, Len),
     if
-        ?is_id_start(C); ?is_digit(C) ->
+        ?is_id(C) ->
-            lex_id(Lx, Len + 1);
+            lex_id(Lexer, Len + 1);
         true ->
             {Id, Rest} = split_binary(Source, Len),
-            {ok, {id, Offset, Id}, Lx#lexer{source = Rest, offset = Offset + Len}}
+            {ok, {Id, Offset}, Lexer#lexer{source = Rest, offset = Offset + Len}}
     end;
-lex_id(#lexer{source = Source, offset = Offset} = Lx, Len) ->
+lex_id(#lexer{source = Source, offset = Offset} = Lexer, Len) ->
-    {ok, {id, Offset, Source}, Lx#lexer{source = <<>>, offset = Offset + Len}}.
+    {ok, {Source, Offset}, Lexer#lexer{source = <<>>, offset = Offset + Len}}.
-
-lex_number(#lexer{source = Source, offset = Offset} = Lx, Len, Acc) when Len < byte_size(Source) ->
-    C = binary:at(Source, Len),
-    if
-        ?is_digit(C) ->
-            lex_number(Lx, Len + 1, Acc * 10 + C - $0);
-        true ->
-            {_, Rest} = split_binary(Source, Len),
-            {ok, {number, Offset, Acc}, Lx#lexer{source = Rest, offset = Offset + Len}}
-    end;
-lex_number(#lexer{offset = Offset} = Lx, Len, Acc) ->
-    {ok, {number, Offset, Acc}, Lx#lexer{source = <<>>, offset = Offset + Len}}.

src/akh_source_map.erl

@@ -0,0 +1,21 @@
+-module(akh_source_map).
+-include("akh_source_map.hrl").
+-export([
+    empty/0,
+    insert/3,
+    get_loc/2
+]).
+
+-spec empty() -> source_map().
+empty() -> gb_trees:empty().
+
+-spec insert(Offset :: non_neg_integer(), Line :: pos_integer(), source_map()) ->
+    source_map().
+insert(Offset, Line, SourceMap) -> gb_trees:insert(Offset, Line, SourceMap).
+
+-spec get_loc(Offset :: non_neg_integer(), source_map()) -> #loc{}.
+get_loc(Offset, SourceMap) ->
+    case gb_trees:smaller(Offset, SourceMap) of
+        {Start, Line} -> #loc{line = Line, col = Offset - Start - 1};
+        none -> #loc{line = 0, col = Offset}
+    end.

src/akh_source_map.hrl

@@ -0,0 +1,5 @@
+-type source_map() :: gb_trees:tree(non_neg_integer(), pos_integer()).
+
+-type span() :: {non_neg_integer(), pos_integer()}.
+
+-record(loc, {line :: pos_integer(), col :: non_neg_integer()}).

src/akhamoth.erl

@@ -1,17 +1,14 @@
 -module(akhamoth).
 
--export([compile_file/1, compile_binary/1, collect/2]).
+-export([compile/1, collect/2]).
 
-compile_file(Path) ->
+compile(Path) ->
-    {ok, Source} = file:read_file(Path),
+    {ok, Src} = file:read_file(Path),
-    compile_binary(Source).
+    Lexer = akh_lexer:new(Src),
+    collect(Lexer, []).
 
-compile_binary(Source) ->
+collect(Lexer, Acc) ->
-    Lx = akh_lexer:new(Source),
+    case akh_lexer:next(Lexer) of
-    collect(Lx, []).
-
-collect(Lx, Acc) ->
-    case akh_lexer:next(Lx) of
         none -> Acc;
         {ok, T, L} -> collect(L, [T | Acc])
     end.