aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
Diffstat
-rw-r--r--src/environment.c4
-rw-r--r--src/parse/controlflow.c278
-rw-r--r--src/parse/controlflow.h18
-rw-r--r--src/parse/files.c70
-rw-r--r--src/parse/files.h4
-rw-r--r--src/parse/functions.c1
-rw-r--r--src/parse/parse.c647
-rw-r--r--src/parse/parse.h31
-rw-r--r--src/parse/statements.c159
-rw-r--r--src/parse/statements.h12
-rw-r--r--src/parse/typedefs.c204
-rw-r--r--src/parse/typedefs.h11
12 files changed, 762 insertions, 677 deletions
diff --git a/src/environment.c b/src/environment.c
index 620d93c4..4a6407a6 100644
--- a/src/environment.c
+++ b/src/environment.c
@@ -5,7 +5,7 @@
#include "environment.h"
#include "naming.h"
-#include "parse/parse.h"
+#include "parse/files.h"
#include "stdlib/datatypes.h"
#include "stdlib/tables.h"
#include "stdlib/text.h"
@@ -19,7 +19,7 @@ public
type_t *PATH_TYPE_TYPE = NULL;
static type_t *declare_type(env_t *env, const char *def_str) {
- ast_t *ast = parse(def_str);
+ ast_t *ast = parse_file_str(def_str);
if (!ast) errx(1, "Couldn't not parse struct def: %s", def_str);
if (ast->tag != Block) errx(1, "Couldn't not parse struct def: %s", def_str);
ast_list_t *statements = Match(ast, Block)->statements;
diff --git a/src/parse/controlflow.c b/src/parse/controlflow.c
new file mode 100644
index 00000000..e2e8dbc3
--- /dev/null
+++ b/src/parse/controlflow.c
@@ -0,0 +1,278 @@
+// Logic for parsing control flow
+
+#include <stdbool.h>
+#include <string.h>
+
+#include "../ast.h"
+#include "../stdlib/util.h"
+#include "context.h"
+#include "controlflow.h"
+#include "errors.h"
+#include "parse.h"
+#include "statements.h"
+#include "suffixes.h"
+#include "utils.h"
+
+public
+ast_t *parse_block(parse_ctx_t *ctx, const char *pos) {
+ const char *start = pos;
+ spaces(&pos);
+
+ ast_list_t *statements = NULL;
+ if (!indent(ctx, &pos)) {
+ // Inline block
+ spaces(&pos);
+ while (*pos) {
+ spaces(&pos);
+ ast_t *stmt = optional(ctx, &pos, parse_statement);
+ if (!stmt) break;
+ statements = new (ast_list_t, .ast = stmt, .next = statements);
+ spaces(&pos);
+ if (!match(&pos, ";")) break;
+ }
+ } else {
+ goto indented;
+ }
+
+ if (indent(ctx, &pos)) {
+ indented:;
+ int64_t block_indent = get_indent(ctx, pos);
+ whitespace(&pos);
+ while (*pos) {
+ ast_t *stmt = optional(ctx, &pos, parse_statement);
+ if (!stmt) {
+ const char *line_start = pos;
+ if (match_word(&pos, "struct"))
+ parser_err(ctx, line_start, eol(pos), "Struct definitions are only allowed at the top level");
+ else if (match_word(&pos, "enum"))
+ parser_err(ctx, line_start, eol(pos), "Enum definitions are only allowed at the top level");
+ else if (match_word(&pos, "func"))
+ parser_err(ctx, line_start, eol(pos), "Function definitions are only allowed at the top level");
+ else if (match_word(&pos, "use"))
+ parser_err(ctx, line_start, eol(pos), "'use' statements are only allowed at the top level");
+
+ spaces(&pos);
+ if (*pos && *pos != '\r' && *pos != '\n') parser_err(ctx, pos, eol(pos), "I couldn't parse this line");
+ break;
+ }
+ statements = new (ast_list_t, .ast = stmt, .next = statements);
+ whitespace(&pos);
+
+ // Guard against having two valid statements on the same line, separated by spaces (but no newlines):
+ if (!memchr(stmt->end, '\n', (size_t)(pos - stmt->end))) {
+ if (*pos) parser_err(ctx, pos, eol(pos), "I don't know how to parse the rest of this line");
+ pos = stmt->end;
+ break;
+ }
+
+ if (get_indent(ctx, pos) != block_indent) {
+ pos = stmt->end; // backtrack
+ break;
+ }
+ }
+ }
+ REVERSE_LIST(statements);
+ return NewAST(ctx->file, start, pos, Block, .statements = statements);
+}
+
+public
+ast_t *parse_pass(parse_ctx_t *ctx, const char *pos) {
+ const char *start = pos;
+ return match_word(&pos, "pass") ? NewAST(ctx->file, start, pos, Pass) : NULL;
+}
+
+public
+ast_t *parse_defer(parse_ctx_t *ctx, const char *pos) {
+ const char *start = pos;
+ if (!match_word(&pos, "defer")) return NULL;
+ ast_t *body = expect(ctx, start, &pos, parse_block, "I expected a block to be deferred here");
+ return NewAST(ctx->file, start, pos, Defer, .body = body);
+}
+
+public
+ast_t *parse_skip(parse_ctx_t *ctx, const char *pos) {
+ const char *start = pos;
+ if (!match_word(&pos, "continue") && !match_word(&pos, "skip")) return NULL;
+ const char *target;
+ if (match_word(&pos, "for")) target = "for";
+ else if (match_word(&pos, "while")) target = "while";
+ else target = get_id(&pos);
+ ast_t *skip = NewAST(ctx->file, start, pos, Skip, .target = target);
+ skip = parse_optional_conditional_suffix(ctx, skip);
+ return skip;
+}
+
+public
+ast_t *parse_stop(parse_ctx_t *ctx, const char *pos) {
+ const char *start = pos;
+ if (!match_word(&pos, "stop") && !match_word(&pos, "break")) return NULL;
+ const char *target;
+ if (match_word(&pos, "for")) target = "for";
+ else if (match_word(&pos, "while")) target = "while";
+ else target = get_id(&pos);
+ ast_t *stop = NewAST(ctx->file, start, pos, Stop, .target = target);
+ stop = parse_optional_conditional_suffix(ctx, stop);
+ return stop;
+}
+
+public
+ast_t *parse_return(parse_ctx_t *ctx, const char *pos) {
+ const char *start = pos;
+ if (!match_word(&pos, "return")) return NULL;
+ ast_t *value = optional(ctx, &pos, parse_expr);
+ ast_t *ret = NewAST(ctx->file, start, pos, Return, .value = value);
+ ret = parse_optional_conditional_suffix(ctx, ret);
+ return ret;
+}
+
+public
+ast_t *parse_do(parse_ctx_t *ctx, const char *pos) {
+ // do [<indent>] body
+ const char *start = pos;
+ if (!match_word(&pos, "do")) return NULL;
+ ast_t *body = expect(ctx, start, &pos, parse_block, "I expected a body for this 'do'");
+ return NewAST(ctx->file, start, pos, Block, .statements = Match(body, Block)->statements);
+}
+
+public
+ast_t *parse_while(parse_ctx_t *ctx, const char *pos) {
+ // while condition ["do"] [<indent>] body
+ const char *start = pos;
+ if (!match_word(&pos, "while")) return NULL;
+
+ const char *tmp = pos;
+ // Shorthand form: `while when ...`
+ if (match_word(&tmp, "when")) {
+ ast_t *when = expect(ctx, start, &pos, parse_when, "I expected a 'when' block after this");
+ if (!when->__data.When.else_body) when->__data.When.else_body = NewAST(ctx->file, pos, pos, Stop);
+ return NewAST(ctx->file, start, pos, While, .body = when);
+ }
+
+ (void)match_word(&pos, "do"); // Optional 'do'
+
+ ast_t *condition = expect(ctx, start, &pos, parse_expr, "I don't see a viable condition for this 'while'");
+ ast_t *body = expect(ctx, start, &pos, parse_block, "I expected a body for this 'while'");
+ return NewAST(ctx->file, start, pos, While, .condition = condition, .body = body);
+}
+
+public
+ast_t *parse_repeat(parse_ctx_t *ctx, const char *pos) {
+ // repeat [<indent>] body
+ const char *start = pos;
+ if (!match_word(&pos, "repeat")) return NULL;
+ ast_t *body = expect(ctx, start, &pos, parse_block, "I expected a body for this 'repeat'");
+ return NewAST(ctx->file, start, pos, Repeat, .body = body);
+}
+
+public
+ast_t *parse_if(parse_ctx_t *ctx, const char *pos) {
+ // "if" <condition> ["then"] <body> ["else" <body>] | "unless" <condition> <body> ["else" <body>]
+ const char *start = pos;
+ int64_t starting_indent = get_indent(ctx, pos);
+
+ bool unless;
+ if (match_word(&pos, "if")) unless = false;
+ else if (match_word(&pos, "unless")) unless = true;
+ else return NULL;
+
+ ast_t *condition = unless ? NULL : optional(ctx, &pos, parse_declaration);
+ if (!condition) condition = expect(ctx, start, &pos, parse_expr, "I expected to find a condition for this 'if'");
+
+ if (unless) condition = WrapAST(condition, Not, condition);
+
+ (void)match_word(&pos, "then"); // Optional 'then'
+ ast_t *body = expect(ctx, start, &pos, parse_block, "I expected a body for this 'if' statement");
+
+ const char *tmp = pos;
+ whitespace(&tmp);
+ ast_t *else_body = NULL;
+ const char *else_start = pos;
+ if (get_indent(ctx, tmp) == starting_indent && match_word(&tmp, "else")) {
+ pos = tmp;
+ spaces(&pos);
+ else_body = optional(ctx, &pos, parse_if);
+ if (!else_body) else_body = expect(ctx, else_start, &pos, parse_block, "I expected a body for this 'else'");
+ }
+ return NewAST(ctx->file, start, pos, If, .condition = condition, .body = body, .else_body = else_body);
+}
+
+public
+ast_t *parse_when(parse_ctx_t *ctx, const char *pos) {
+ // when <expr> (is var : Tag <body>)* [else <body>]
+ const char *start = pos;
+ int64_t starting_indent = get_indent(ctx, pos);
+
+ if (!match_word(&pos, "when")) return NULL;
+
+ ast_t *subject = optional(ctx, &pos, parse_declaration);
+ if (!subject) subject = expect(ctx, start, &pos, parse_expr, "I expected to find an expression for this 'when'");
+
+ when_clause_t *clauses = NULL;
+ const char *tmp = pos;
+ whitespace(&tmp);
+ while (get_indent(ctx, tmp) == starting_indent && match_word(&tmp, "is")) {
+ pos = tmp;
+ spaces(&pos);
+ ast_t *pattern = expect(ctx, start, &pos, parse_expr, "I expected a pattern to match here");
+ spaces(&pos);
+ when_clause_t *new_clauses = new (when_clause_t, .pattern = pattern, .next = clauses);
+ while (match(&pos, ",")) {
+ pattern = expect(ctx, start, &pos, parse_expr, "I expected a pattern to match here");
+ new_clauses = new (when_clause_t, .pattern = pattern, .next = new_clauses);
+ spaces(&pos);
+ }
+ (void)match_word(&pos, "then"); // Optional 'then'
+ ast_t *body = expect(ctx, start, &pos, parse_block, "I expected a body for this 'when' clause");
+ for (when_clause_t *c = new_clauses; c && c != clauses; c = c->next) {
+ c->body = body;
+ }
+ clauses = new_clauses;
+ tmp = pos;
+ whitespace(&tmp);
+ }
+ REVERSE_LIST(clauses);
+
+ ast_t *else_body = NULL;
+ const char *else_start = pos;
+ if (get_indent(ctx, tmp) == starting_indent && match_word(&tmp, "else")) {
+ pos = tmp;
+ else_body = expect(ctx, else_start, &pos, parse_block, "I expected a body for this 'else'");
+ }
+ return NewAST(ctx->file, start, pos, When, .subject = subject, .clauses = clauses, .else_body = else_body);
+}
+
+public
+ast_t *parse_for(parse_ctx_t *ctx, const char *pos) {
+ // for [k,] v in iter [<indent>] body
+ const char *start = pos;
+ if (!match_word(&pos, "for")) return NULL;
+ int64_t starting_indent = get_indent(ctx, pos);
+ spaces(&pos);
+ ast_list_t *vars = NULL;
+ for (;;) {
+ ast_t *var = optional(ctx, &pos, parse_var);
+ if (var) vars = new (ast_list_t, .ast = var, .next = vars);
+
+ spaces(&pos);
+ if (!match(&pos, ",")) break;
+ }
+
+ spaces(&pos);
+ expect_str(ctx, start, &pos, "in", "I expected an 'in' for this 'for'");
+
+ ast_t *iter = expect(ctx, start, &pos, parse_expr, "I expected an iterable value for this 'for'");
+
+ (void)match_word(&pos, "do"); // Optional 'do'
+
+ ast_t *body = expect(ctx, start, &pos, parse_block, "I expected a body for this 'for'");
+
+ const char *else_start = pos;
+ whitespace(&else_start);
+ ast_t *empty = NULL;
+ if (match_word(&else_start, "else") && get_indent(ctx, else_start) == starting_indent) {
+ pos = else_start;
+ empty = expect(ctx, pos, &pos, parse_block, "I expected a body for this 'else'");
+ }
+ REVERSE_LIST(vars);
+ return NewAST(ctx->file, start, pos, For, .vars = vars, .iter = iter, .body = body, .empty = empty);
+}
diff --git a/src/parse/controlflow.h b/src/parse/controlflow.h
new file mode 100644
index 00000000..2ef093d4
--- /dev/null
+++ b/src/parse/controlflow.h
@@ -0,0 +1,18 @@
+// Logic for parsing control flow
+#pragma once
+
+#include "../ast.h"
+#include "context.h"
+
+ast_t *parse_block(parse_ctx_t *ctx, const char *pos);
+ast_t *parse_defer(parse_ctx_t *ctx, const char *pos);
+ast_t *parse_do(parse_ctx_t *ctx, const char *pos);
+ast_t *parse_for(parse_ctx_t *ctx, const char *pos);
+ast_t *parse_if(parse_ctx_t *ctx, const char *pos);
+ast_t *parse_pass(parse_ctx_t *ctx, const char *pos);
+ast_t *parse_repeat(parse_ctx_t *ctx, const char *pos);
+ast_t *parse_return(parse_ctx_t *ctx, const char *pos);
+ast_t *parse_skip(parse_ctx_t *ctx, const char *pos);
+ast_t *parse_stop(parse_ctx_t *ctx, const char *pos);
+ast_t *parse_when(parse_ctx_t *ctx, const char *pos);
+ast_t *parse_while(parse_ctx_t *ctx, const char *pos);
diff --git a/src/parse/files.c b/src/parse/files.c
index 416adc92..6ff01af1 100644
--- a/src/parse/files.c
+++ b/src/parse/files.c
@@ -14,7 +14,10 @@
#include "files.h"
#include "functions.h"
#include "parse.h"
+#include "statements.h"
#include "text.h"
+#include "typedefs.h"
+#include "types.h"
#include "utils.h"
// The cache of {filename -> parsed AST} will hold at most this many entries:
@@ -28,7 +31,6 @@ static ast_t *parse_top_declaration(parse_ctx_t *ctx, const char *pos) {
return declaration;
}
-public
ast_t *parse_file_body(parse_ctx_t *ctx, const char *pos) {
const char *start = pos;
whitespace(&pos);
@@ -111,3 +113,69 @@ ast_t *parse_file(const char *path, jmp_buf *on_err) {
Table$str_set(&cached, path, ast);
return ast;
}
+
+ast_t *parse_use(parse_ctx_t *ctx, const char *pos) {
+ const char *start = pos;
+
+ ast_t *var = parse_var(ctx, pos);
+ if (var) {
+ pos = var->end;
+ spaces(&pos);
+ if (!match(&pos, ":=")) return NULL;
+ spaces(&pos);
+ }
+
+ if (!match_word(&pos, "use")) return NULL;
+ spaces(&pos);
+ size_t name_len = strcspn(pos, " \t\r\n;");
+ if (name_len < 1) parser_err(ctx, start, pos, "There is no module name here to use");
+ char *name = GC_strndup(pos, name_len);
+ pos += name_len;
+ while (match(&pos, ";"))
+ continue;
+ int what;
+ if (name[0] == '<' || ends_with(name, ".h")) {
+ what = USE_HEADER;
+ } else if (starts_with(name, "-l")) {
+ what = USE_SHARED_OBJECT;
+ } else if (ends_with(name, ".c")) {
+ what = USE_C_CODE;
+ } else if (ends_with(name, ".S") || ends_with(name, ".s")) {
+ what = USE_ASM;
+ } else if (starts_with(name, "./") || starts_with(name, "/") || starts_with(name, "../")
+ || starts_with(name, "~/")) {
+ what = USE_LOCAL;
+ } else {
+ what = USE_MODULE;
+ }
+ return NewAST(ctx->file, start, pos, Use, .var = var, .path = name, .what = what);
+}
+
+ast_t *parse_extern(parse_ctx_t *ctx, const char *pos) {
+ const char *start = pos;
+ if (!match_word(&pos, "extern")) return NULL;
+ spaces(&pos);
+ const char *name = get_id(&pos);
+ spaces(&pos);
+ if (!match(&pos, ":")) parser_err(ctx, start, pos, "I couldn't get a type for this extern");
+ type_ast_t *type = expect(ctx, start, &pos, parse_type, "I couldn't parse the type for this extern");
+ return NewAST(ctx->file, start, pos, Extern, .name = name, .type = type);
+}
+
+public
+ast_t *parse_file_str(const char *str) {
+ file_t *file = spoof_file("<string>", str);
+ parse_ctx_t ctx = {
+ .file = file,
+ .on_err = NULL,
+ };
+
+ const char *pos = file->text;
+ whitespace(&pos);
+ ast_t *ast = parse_file_body(&ctx, pos);
+ pos = ast->end;
+ whitespace(&pos);
+ if (pos < file->text + file->len && *pos != '\0')
+ parser_err(&ctx, pos, pos + strlen(pos), "I couldn't parse this part of the string");
+ return ast;
+}
diff --git a/src/parse/files.h b/src/parse/files.h
index b8dc8726..3073449e 100644
--- a/src/parse/files.h
+++ b/src/parse/files.h
@@ -7,5 +7,9 @@
#include "../ast.h"
#include "context.h"
+ast_t *parse_file_str(const char *str);
ast_t *parse_file(const char *path, jmp_buf *on_err);
+
ast_t *parse_file_body(parse_ctx_t *ctx, const char *pos);
+ast_t *parse_use(parse_ctx_t *ctx, const char *pos);
+ast_t *parse_extern(parse_ctx_t *ctx, const char *pos);
diff --git a/src/parse/functions.c b/src/parse/functions.c
index 7786da7a..4293b43b 100644
--- a/src/parse/functions.c
+++ b/src/parse/functions.c
@@ -11,6 +11,7 @@
#include "../ast.h"
#include "../stdlib/util.h"
#include "context.h"
+#include "controlflow.h"
#include "errors.h"
#include "functions.h"
#include "parse.h"
diff --git a/src/parse/parse.c b/src/parse/parse.c
index a4462cd6..d4e51f1f 100644
--- a/src/parse/parse.c
+++ b/src/parse/parse.c
@@ -1,6 +1,5 @@
// Recursive descent parser for parsing code
-#include <gc.h>
#include <stdbool.h>
#include <string.h>
@@ -9,6 +8,7 @@
#include "binops.h"
#include "containers.h"
#include "context.h"
+#include "controlflow.h"
#include "errors.h"
#include "files.h"
#include "functions.h"
@@ -79,152 +79,6 @@ ast_t *parse_reduction(parse_ctx_t *ctx, const char *pos) {
return NewAST(ctx->file, start, pos, Reduction, .iter = iter, .op = op, .key = key);
}
-ast_t *parse_if(parse_ctx_t *ctx, const char *pos) {
- // "if" <condition> ["then"] <body> ["else" <body>] | "unless" <condition> <body> ["else" <body>]
- const char *start = pos;
- int64_t starting_indent = get_indent(ctx, pos);
-
- bool unless;
- if (match_word(&pos, "if")) unless = false;
- else if (match_word(&pos, "unless")) unless = true;
- else return NULL;
-
- ast_t *condition = unless ? NULL : optional(ctx, &pos, parse_declaration);
- if (!condition) condition = expect(ctx, start, &pos, parse_expr, "I expected to find a condition for this 'if'");
-
- if (unless) condition = WrapAST(condition, Not, condition);
-
- (void)match_word(&pos, "then"); // Optional 'then'
- ast_t *body = expect(ctx, start, &pos, parse_block, "I expected a body for this 'if' statement");
-
- const char *tmp = pos;
- whitespace(&tmp);
- ast_t *else_body = NULL;
- const char *else_start = pos;
- if (get_indent(ctx, tmp) == starting_indent && match_word(&tmp, "else")) {
- pos = tmp;
- spaces(&pos);
- else_body = optional(ctx, &pos, parse_if);
- if (!else_body) else_body = expect(ctx, else_start, &pos, parse_block, "I expected a body for this 'else'");
- }
- return NewAST(ctx->file, start, pos, If, .condition = condition, .body = body, .else_body = else_body);
-}
-
-ast_t *parse_when(parse_ctx_t *ctx, const char *pos) {
- // when <expr> (is var : Tag <body>)* [else <body>]
- const char *start = pos;
- int64_t starting_indent = get_indent(ctx, pos);
-
- if (!match_word(&pos, "when")) return NULL;
-
- ast_t *subject = optional(ctx, &pos, parse_declaration);
- if (!subject) subject = expect(ctx, start, &pos, parse_expr, "I expected to find an expression for this 'when'");
-
- when_clause_t *clauses = NULL;
- const char *tmp = pos;
- whitespace(&tmp);
- while (get_indent(ctx, tmp) == starting_indent && match_word(&tmp, "is")) {
- pos = tmp;
- spaces(&pos);
- ast_t *pattern = expect(ctx, start, &pos, parse_expr, "I expected a pattern to match here");
- spaces(&pos);
- when_clause_t *new_clauses = new (when_clause_t, .pattern = pattern, .next = clauses);
- while (match(&pos, ",")) {
- pattern = expect(ctx, start, &pos, parse_expr, "I expected a pattern to match here");
- new_clauses = new (when_clause_t, .pattern = pattern, .next = new_clauses);
- spaces(&pos);
- }
- (void)match_word(&pos, "then"); // Optional 'then'
- ast_t *body = expect(ctx, start, &pos, parse_block, "I expected a body for this 'when' clause");
- for (when_clause_t *c = new_clauses; c && c != clauses; c = c->next) {
- c->body = body;
- }
- clauses = new_clauses;
- tmp = pos;
- whitespace(&tmp);
- }
- REVERSE_LIST(clauses);
-
- ast_t *else_body = NULL;
- const char *else_start = pos;
- if (get_indent(ctx, tmp) == starting_indent && match_word(&tmp, "else")) {
- pos = tmp;
- else_body = expect(ctx, else_start, &pos, parse_block, "I expected a body for this 'else'");
- }
- return NewAST(ctx->file, start, pos, When, .subject = subject, .clauses = clauses, .else_body = else_body);
-}
-
-ast_t *parse_for(parse_ctx_t *ctx, const char *pos) {
- // for [k,] v in iter [<indent>] body
- const char *start = pos;
- if (!match_word(&pos, "for")) return NULL;
- int64_t starting_indent = get_indent(ctx, pos);
- spaces(&pos);
- ast_list_t *vars = NULL;
- for (;;) {
- ast_t *var = optional(ctx, &pos, parse_var);
- if (var) vars = new (ast_list_t, .ast = var, .next = vars);
-
- spaces(&pos);
- if (!match(&pos, ",")) break;
- }
-
- spaces(&pos);
- expect_str(ctx, start, &pos, "in", "I expected an 'in' for this 'for'");
-
- ast_t *iter = expect(ctx, start, &pos, parse_expr, "I expected an iterable value for this 'for'");
-
- (void)match_word(&pos, "do"); // Optional 'do'
-
- ast_t *body = expect(ctx, start, &pos, parse_block, "I expected a body for this 'for'");
-
- const char *else_start = pos;
- whitespace(&else_start);
- ast_t *empty = NULL;
- if (match_word(&else_start, "else") && get_indent(ctx, else_start) == starting_indent) {
- pos = else_start;
- empty = expect(ctx, pos, &pos, parse_block, "I expected a body for this 'else'");
- }
- REVERSE_LIST(vars);
- return NewAST(ctx->file, start, pos, For, .vars = vars, .iter = iter, .body = body, .empty = empty);
-}
-
-ast_t *parse_do(parse_ctx_t *ctx, const char *pos) {
- // do [<indent>] body
- const char *start = pos;
- if (!match_word(&pos, "do")) return NULL;
- ast_t *body = expect(ctx, start, &pos, parse_block, "I expected a body for this 'do'");
- return NewAST(ctx->file, start, pos, Block, .statements = Match(body, Block)->statements);
-}
-
-ast_t *parse_while(parse_ctx_t *ctx, const char *pos) {
- // while condition ["do"] [<indent>] body
- const char *start = pos;
- if (!match_word(&pos, "while")) return NULL;
-
- const char *tmp = pos;
- // Shorthand form: `while when ...`
- if (match_word(&tmp, "when")) {
- ast_t *when = expect(ctx, start, &pos, parse_when, "I expected a 'when' block after this");
- if (!when->__data.When.else_body) when->__data.When.else_body = NewAST(ctx->file, pos, pos, Stop);
- return NewAST(ctx->file, start, pos, While, .body = when);
- }
-
- (void)match_word(&pos, "do"); // Optional 'do'
-
- ast_t *condition = expect(ctx, start, &pos, parse_expr, "I don't see a viable condition for this 'while'");
- ast_t *body = expect(ctx, start, &pos, parse_block, "I expected a body for this 'while'");
- return NewAST(ctx->file, start, pos, While, .condition = condition, .body = body);
-}
-
-ast_t *parse_repeat(parse_ctx_t *ctx, const char *pos) {
- // repeat [<indent>] body
- const char *start = pos;
- if (!match_word(&pos, "repeat")) return NULL;
- ast_t *body = expect(ctx, start, &pos, parse_block, "I expected a body for this 'repeat'");
- return NewAST(ctx->file, start, pos, Repeat, .body = body);
-}
-
ast_t *parse_heap_alloc(parse_ctx_t *ctx, const char *pos) {
const char *start = pos;
if (!match(&pos, "@")) return NULL;
@@ -298,51 +152,6 @@ ast_t *parse_bool(parse_ctx_t *ctx, const char *pos) {
else return NULL;
}
-ast_t *parse_pass(parse_ctx_t *ctx, const char *pos) {
- const char *start = pos;
- return match_word(&pos, "pass") ? NewAST(ctx->file, start, pos, Pass) : NULL;
-}
-
-ast_t *parse_defer(parse_ctx_t *ctx, const char *pos) {
- const char *start = pos;
- if (!match_word(&pos, "defer")) return NULL;
- ast_t *body = expect(ctx, start, &pos, parse_block, "I expected a block to be deferred here");
- return NewAST(ctx->file, start, pos, Defer, .body = body);
-}
-
-ast_t *parse_skip(parse_ctx_t *ctx, const char *pos) {
- const char *start = pos;
- if (!match_word(&pos, "continue") && !match_word(&pos, "skip")) return NULL;
- const char *target;
- if (match_word(&pos, "for")) target = "for";
- else if (match_word(&pos, "while")) target = "while";
- else target = get_id(&pos);
- ast_t *skip = NewAST(ctx->file, start, pos, Skip, .target = target);
- skip = parse_optional_conditional_suffix(ctx, skip);
- return skip;
-}
-
-ast_t *parse_stop(parse_ctx_t *ctx, const char *pos) {
- const char *start = pos;
- if (!match_word(&pos, "stop") && !match_word(&pos, "break")) return NULL;
- const char *target;
- if (match_word(&pos, "for")) target = "for";
- else if (match_word(&pos, "while")) target = "while";
- else target = get_id(&pos);
- ast_t *stop = NewAST(ctx->file, start, pos, Stop, .target = target);
- stop = parse_optional_conditional_suffix(ctx, stop);
- return stop;
-}
-
-ast_t *parse_return(parse_ctx_t *ctx, const char *pos) {
- const char *start = pos;
- if (!match_word(&pos, "return")) return NULL;
- ast_t *value = optional(ctx, &pos, parse_expr);
- ast_t *ret = NewAST(ctx->file, start, pos, Return, .value = value);
- ret = parse_optional_conditional_suffix(ctx, ret);
- return ret;
-}
-
ast_t *parse_none(parse_ctx_t *ctx, const char *pos) {
const char *start = pos;
if (!match_word(&pos, "none")) return NULL;
@@ -410,111 +219,6 @@ ast_t *parse_term(parse_ctx_t *ctx, const char *pos) {
ast_t *parse_expr(parse_ctx_t *ctx, const char *pos) { return parse_infix_expr(ctx, pos, 0); }
-ast_t *parse_declaration(parse_ctx_t *ctx, const char *pos) {
- const char *start = pos;
- ast_t *var = parse_var(ctx, pos);
- if (!var) return NULL;
- pos = var->end;
- spaces(&pos);
- if (!match(&pos, ":")) return NULL;
- spaces(&pos);
- type_ast_t *type = optional(ctx, &pos, parse_type);
- spaces(&pos);
- ast_t *val = NULL;
- if (match(&pos, "=")) {
- val = optional(ctx, &pos, parse_extended_expr);
- if (!val) {
- if (optional(ctx, &pos, parse_use))
- parser_err(ctx, start, pos, "'use' statements are only allowed at the top level of a file");
- else parser_err(ctx, pos, eol(pos), "This is not a valid expression");
- }
- }
- return NewAST(ctx->file, start, pos, Declare, .var = var, .type = type, .value = val);
-}
-
-ast_t *parse_update(parse_ctx_t *ctx, const char *pos) {
- const char *start = pos;
- ast_t *lhs = optional(ctx, &pos, parse_expr);
- if (!lhs) return NULL;
- spaces(&pos);
- ast_e op;
- if (match(&pos, "+=")) op = PlusUpdate;
- else if (match(&pos, "++=")) op = ConcatUpdate;
- else if (match(&pos, "-=")) op = MinusUpdate;
- else if (match(&pos, "*=")) op = MultiplyUpdate;
- else if (match(&pos, "/=")) op = DivideUpdate;
- else if (match(&pos, "^=")) op = PowerUpdate;
- else if (match(&pos, "<<=")) op = LeftShiftUpdate;
- else if (match(&pos, "<<<=")) op = UnsignedLeftShiftUpdate;
- else if (match(&pos, ">>=")) op = RightShiftUpdate;
- else if (match(&pos, ">>>=")) op = UnsignedRightShiftUpdate;
- else if (match(&pos, "and=")) op = AndUpdate;
- else if (match(&pos, "or=")) op = OrUpdate;
- else if (match(&pos, "xor=")) op = XorUpdate;
- else return NULL;
- ast_t *rhs = expect(ctx, start, &pos, parse_extended_expr, "I expected an expression here");
- return new (ast_t, .file = ctx->file, .start = start, .end = pos, .tag = op, .__data.PlusUpdate.lhs = lhs,
- .__data.PlusUpdate.rhs = rhs);
-}
-
-ast_t *parse_assignment(parse_ctx_t *ctx, const char *pos) {
- const char *start = pos;
- ast_list_t *targets = NULL;
- for (;;) {
- ast_t *lhs = optional(ctx, &pos, parse_term);
- if (!lhs) break;
- targets = new (ast_list_t, .ast = lhs, .next = targets);
- spaces(&pos);
- if (!match(&pos, ",")) break;
- whitespace(&pos);
- }
-
- if (!targets) return NULL;
-
- spaces(&pos);
- if (!match(&pos, "=")) return NULL;
- if (match(&pos, "=")) return NULL; // == comparison
-
- ast_list_t *values = NULL;
- for (;;) {
- ast_t *rhs = optional(ctx, &pos, parse_extended_expr);
- if (!rhs) break;
- values = new (ast_list_t, .ast = rhs, .next = values);
- spaces(&pos);
- if (!match(&pos, ",")) break;
- whitespace(&pos);
- }
-
- REVERSE_LIST(targets);
- REVERSE_LIST(values);
-
- return NewAST(ctx->file, start, pos, Assign, .targets = targets, .values = values);
-}
-
-ast_t *parse_statement(parse_ctx_t *ctx, const char *pos) {
- ast_t *stmt = NULL;
- if ((stmt = parse_declaration(ctx, pos)) || (stmt = parse_doctest(ctx, pos)) || (stmt = parse_assert(ctx, pos)))
- return stmt;
-
- if (!(false || (stmt = parse_update(ctx, pos)) || (stmt = parse_assignment(ctx, pos))))
- stmt = parse_extended_expr(ctx, pos);
-
- for (bool progress = (stmt != NULL); progress;) {
- ast_t *new_stmt;
- progress = false;
- if (stmt->tag == Var) {
- progress = (false || (new_stmt = parse_method_call_suffix(ctx, stmt))
- || (new_stmt = parse_fncall_suffix(ctx, stmt)));
- } else if (stmt->tag == FunctionCall) {
- new_stmt = parse_optional_conditional_suffix(ctx, stmt);
- progress = (new_stmt != stmt);
- }
-
- if (progress) stmt = new_stmt;
- }
- return stmt;
-}
-
ast_t *parse_extended_expr(parse_ctx_t *ctx, const char *pos) {
ast_t *expr = NULL;
@@ -526,354 +230,7 @@ ast_t *parse_extended_expr(parse_ctx_t *ctx, const char *pos) {
return parse_expr(ctx, pos);
}
-ast_t *parse_block(parse_ctx_t *ctx, const char *pos) {
- const char *start = pos;
- spaces(&pos);
-
- ast_list_t *statements = NULL;
- if (!indent(ctx, &pos)) {
- // Inline block
- spaces(&pos);
- while (*pos) {
- spaces(&pos);
- ast_t *stmt = optional(ctx, &pos, parse_statement);
- if (!stmt) break;
- statements = new (ast_list_t, .ast = stmt, .next = statements);
- spaces(&pos);
- if (!match(&pos, ";")) break;
- }
- } else {
- goto indented;
- }
-
- if (indent(ctx, &pos)) {
- indented:;
- int64_t block_indent = get_indent(ctx, pos);
- whitespace(&pos);
- while (*pos) {
- ast_t *stmt = optional(ctx, &pos, parse_statement);
- if (!stmt) {
- const char *line_start = pos;
- if (match_word(&pos, "struct"))
- parser_err(ctx, line_start, eol(pos), "Struct definitions are only allowed at the top level");
- else if (match_word(&pos, "enum"))
- parser_err(ctx, line_start, eol(pos), "Enum definitions are only allowed at the top level");
- else if (match_word(&pos, "func"))
- parser_err(ctx, line_start, eol(pos), "Function definitions are only allowed at the top level");
- else if (match_word(&pos, "use"))
- parser_err(ctx, line_start, eol(pos), "'use' statements are only allowed at the top level");
-
- spaces(&pos);
- if (*pos && *pos != '\r' && *pos != '\n') parser_err(ctx, pos, eol(pos), "I couldn't parse this line");
- break;
- }
- statements = new (ast_list_t, .ast = stmt, .next = statements);
- whitespace(&pos);
-
- // Guard against having two valid statements on the same line, separated by spaces (but no newlines):
- if (!memchr(stmt->end, '\n', (size_t)(pos - stmt->end))) {
- if (*pos) parser_err(ctx, pos, eol(pos), "I don't know how to parse the rest of this line");
- pos = stmt->end;
- break;
- }
-
- if (get_indent(ctx, pos) != block_indent) {
- pos = stmt->end; // backtrack
- break;
- }
- }
- }
- REVERSE_LIST(statements);
- return NewAST(ctx->file, start, pos, Block, .statements = statements);
-}
-
-ast_t *parse_namespace(parse_ctx_t *ctx, const char *pos) {
- const char *start = pos;
- whitespace(&pos);
- int64_t indent = get_indent(ctx, pos);
- ast_list_t *statements = NULL;
- for (;;) {
- const char *next = pos;
- whitespace(&next);
- if (get_indent(ctx, next) != indent) break;
- ast_t *stmt;
- if ((stmt = optional(ctx, &pos, parse_struct_def)) || (stmt = optional(ctx, &pos, parse_func_def))
- || (stmt = optional(ctx, &pos, parse_enum_def)) || (stmt = optional(ctx, &pos, parse_lang_def))
- || (stmt = optional(ctx, &pos, parse_extend)) || (stmt = optional(ctx, &pos, parse_convert_def))
- || (stmt = optional(ctx, &pos, parse_use)) || (stmt = optional(ctx, &pos, parse_extern))
- || (stmt = optional(ctx, &pos, parse_inline_c)) || (stmt = optional(ctx, &pos, parse_declaration))) {
- statements = new (ast_list_t, .ast = stmt, .next = statements);
- pos = stmt->end;
- whitespace(&pos); // TODO: check for newline
- // if (!(space_types & WHITESPACE_NEWLINES)) {
- // pos = stmt->end;
- // break;
- // }
- } else {
- if (get_indent(ctx, next) > indent && next < eol(next))
- parser_err(ctx, next, eol(next), "I couldn't parse this namespace declaration");
- break;
- }
- }
- REVERSE_LIST(statements);
- return NewAST(ctx->file, start, pos, Block, .statements = statements);
-}
-
-ast_t *parse_struct_def(parse_ctx_t *ctx, const char *pos) {
- // struct Foo(...) [: \n body]
- const char *start = pos;
- if (!match_word(&pos, "struct")) return NULL;
-
- int64_t starting_indent = get_indent(ctx, pos);
-
- spaces(&pos);
- const char *name = get_id(&pos);
- if (!name) parser_err(ctx, start, pos, "I expected a name for this struct");
- spaces(&pos);
-
- if (!match(&pos, "(")) parser_err(ctx, pos, pos, "I expected a '(' and a list of fields here");
-
- arg_ast_t *fields = parse_args(ctx, &pos);
-
- whitespace(&pos);
- bool secret = false, external = false, opaque = false;
- if (match(&pos, ";")) { // Extra flags
- whitespace(&pos);
- for (;;) {
- if (match_word(&pos, "secret")) {
- secret = true;
- } else if (match_word(&pos, "extern")) {
- external = true;
- } else if (match_word(&pos, "opaque")) {
- if (fields)
- parser_err(ctx, pos - strlen("opaque"), pos, "A struct can't be opaque if it has fields defined");
- opaque = true;
- } else {
- break;
- }
-
- if (!match_separator(&pos)) break;
- }
- }
-
- expect_closing(ctx, &pos, ")", "I wasn't able to parse the rest of this struct");
-
- ast_t *namespace = NULL;
- const char *ns_pos = pos;
- whitespace(&ns_pos);
- int64_t ns_indent = get_indent(ctx, ns_pos);
- if (ns_indent > starting_indent) {
- pos = ns_pos;
- namespace = optional(ctx, &pos, parse_namespace);
- }
- if (!namespace) namespace = NewAST(ctx->file, pos, pos, Block, .statements = NULL);
- return NewAST(ctx->file, start, pos, StructDef, .name = name, .fields = fields, .namespace = namespace,
- .secret = secret, .external = external, .opaque = opaque);
-}
-
-ast_t *parse_enum_def(parse_ctx_t *ctx, const char *pos) {
- // tagged union: enum Foo(a, b(x:Int,y:Int)=5, ...) [: \n namespace]
- const char *start = pos;
- if (!match_word(&pos, "enum")) return NULL;
- int64_t starting_indent = get_indent(ctx, pos);
- spaces(&pos);
- const char *name = get_id(&pos);
- if (!name) parser_err(ctx, start, pos, "I expected a name for this enum");
- spaces(&pos);
- if (!match(&pos, "(")) return NULL;
-
- tag_ast_t *tags = NULL;
- whitespace(&pos);
- for (;;) {
- spaces(&pos);
- const char *tag_name = get_id(&pos);
- if (!tag_name) break;
-
- spaces(&pos);
- arg_ast_t *fields;
- bool secret = false;
- if (match(&pos, "(")) {
- whitespace(&pos);
- fields = parse_args(ctx, &pos);
- whitespace(&pos);
- if (match(&pos, ";")) { // Extra flags
- whitespace(&pos);
- secret = match_word(&pos, "secret");
- whitespace(&pos);
- }
- expect_closing(ctx, &pos, ")", "I wasn't able to parse the rest of this tagged union member");
- } else {
- fields = NULL;
- }
-
- tags = new (tag_ast_t, .name = tag_name, .fields = fields, .secret = secret, .next = tags);
-
- if (!match_separator(&pos)) break;
- }
-
- whitespace(&pos);
- expect_closing(ctx, &pos, ")", "I wasn't able to parse the rest of this enum definition");
-
- REVERSE_LIST(tags);
-
- if (tags == NULL) parser_err(ctx, start, pos, "This enum does not have any tags!");
-
- ast_t *namespace = NULL;
- const char *ns_pos = pos;
- whitespace(&ns_pos);
- int64_t ns_indent = get_indent(ctx, ns_pos);
- if (ns_indent > starting_indent) {
- pos = ns_pos;
- namespace = optional(ctx, &pos, parse_namespace);
- }
- if (!namespace) namespace = NewAST(ctx->file, pos, pos, Block, .statements = NULL);
-
- return NewAST(ctx->file, start, pos, EnumDef, .name = name, .tags = tags, .namespace = namespace);
-}
-
-ast_t *parse_lang_def(parse_ctx_t *ctx, const char *pos) {
- const char *start = pos;
- // lang Name: [namespace...]
- if (!match_word(&pos, "lang")) return NULL;
- int64_t starting_indent = get_indent(ctx, pos);
- spaces(&pos);
- const char *name = get_id(&pos);
- if (!name) parser_err(ctx, start, pos, "I expected a name for this lang");
- spaces(&pos);
-
- ast_t *namespace = NULL;
- const char *ns_pos = pos;
- whitespace(&ns_pos);
- int64_t ns_indent = get_indent(ctx, ns_pos);
- if (ns_indent > starting_indent) {
- pos = ns_pos;
- namespace = optional(ctx, &pos, parse_namespace);
- }
- if (!namespace) namespace = NewAST(ctx->file, pos, pos, Block, .statements = NULL);
-
- return NewAST(ctx->file, start, pos, LangDef, .name = name, .namespace = namespace);
-}
-
-ast_t *parse_extend(parse_ctx_t *ctx, const char *pos) {
- const char *start = pos;
- // extend Name: body...
- if (!match_word(&pos, "extend")) return NULL;
- int64_t starting_indent = get_indent(ctx, pos);
- spaces(&pos);
- const char *name = get_id(&pos);
- if (!name) parser_err(ctx, start, pos, "I expected a name for this lang");
-
- ast_t *body = NULL;
- const char *ns_pos = pos;
- whitespace(&ns_pos);
- int64_t ns_indent = get_indent(ctx, ns_pos);
- if (ns_indent > starting_indent) {
- pos = ns_pos;
- body = optional(ctx, &pos, parse_namespace);
- }
- if (!body) body = NewAST(ctx->file, pos, pos, Block, .statements = NULL);
-
- return NewAST(ctx->file, start, pos, Extend, .name = name, .body = body);
-}
-
-ast_t *parse_extern(parse_ctx_t *ctx, const char *pos) {
- const char *start = pos;
- if (!match_word(&pos, "extern")) return NULL;
- spaces(&pos);
- const char *name = get_id(&pos);
- spaces(&pos);
- if (!match(&pos, ":")) parser_err(ctx, start, pos, "I couldn't get a type for this extern");
- type_ast_t *type = expect(ctx, start, &pos, parse_type, "I couldn't parse the type for this extern");
- return NewAST(ctx->file, start, pos, Extern, .name = name, .type = type);
-}
-
-ast_t *parse_doctest(parse_ctx_t *ctx, const char *pos) {
- const char *start = pos;
- if (!match(&pos, ">>")) return NULL;
- spaces(&pos);
- ast_t *expr = expect(ctx, start, &pos, parse_statement, "I couldn't parse the expression for this doctest");
- whitespace(&pos);
- ast_t *expected = NULL;
- if (match(&pos, "=")) {
- spaces(&pos);
- expected = expect(ctx, start, &pos, parse_extended_expr, "I couldn't parse the expected expression here");
- } else {
- pos = expr->end;
- }
- return NewAST(ctx->file, start, pos, DocTest, .expr = expr, .expected = expected);
-}
-
-ast_t *parse_assert(parse_ctx_t *ctx, const char *pos) {
- const char *start = pos;
- if (!match_word(&pos, "assert")) return NULL;
- spaces(&pos);
- ast_t *expr = expect(ctx, start, &pos, parse_extended_expr, "I couldn't parse the expression for this assert");
- spaces(&pos);
- ast_t *message = NULL;
- if (match(&pos, ",")) {
- whitespace(&pos);
- message = expect(ctx, start, &pos, parse_extended_expr, "I couldn't parse the error message for this assert");
- } else {
- pos = expr->end;
- }
- return NewAST(ctx->file, start, pos, Assert, .expr = expr, .message = message);
-}
-
-ast_t *parse_use(parse_ctx_t *ctx, const char *pos) {
- const char *start = pos;
-
- ast_t *var = parse_var(ctx, pos);
- if (var) {
- pos = var->end;
- spaces(&pos);
- if (!match(&pos, ":=")) return NULL;
- spaces(&pos);
- }
-
- if (!match_word(&pos, "use")) return NULL;
- spaces(&pos);
- size_t name_len = strcspn(pos, " \t\r\n;");
- if (name_len < 1) parser_err(ctx, start, pos, "There is no module name here to use");
- char *name = GC_strndup(pos, name_len);
- pos += name_len;
- while (match(&pos, ";"))
- continue;
- int what;
- if (name[0] == '<' || ends_with(name, ".h")) {
- what = USE_HEADER;
- } else if (starts_with(name, "-l")) {
- what = USE_SHARED_OBJECT;
- } else if (ends_with(name, ".c")) {
- what = USE_C_CODE;
- } else if (ends_with(name, ".S") || ends_with(name, ".s")) {
- what = USE_ASM;
- } else if (starts_with(name, "./") || starts_with(name, "/") || starts_with(name, "../")
- || starts_with(name, "~/")) {
- what = USE_LOCAL;
- } else {
- what = USE_MODULE;
- }
- return NewAST(ctx->file, start, pos, Use, .var = var, .path = name, .what = what);
-}
-
-ast_t *parse(const char *str) {
- file_t *file = spoof_file("<string>", str);
- parse_ctx_t ctx = {
- .file = file,
- .on_err = NULL,
- };
-
- const char *pos = file->text;
- whitespace(&pos);
- ast_t *ast = parse_file_body(&ctx, pos);
- pos = ast->end;
- whitespace(&pos);
- if (pos < file->text + file->len && *pos != '\0')
- parser_err(&ctx, pos, pos + strlen(pos), "I couldn't parse this part of the string");
- return ast;
-}
-
-ast_t *parse_expression(const char *str) {
+ast_t *parse_expr_str(const char *str) {
file_t *file = spoof_file("<string>", str);
parse_ctx_t ctx = {
.file = file,
diff --git a/src/parse/parse.h b/src/parse/parse.h
index 57d798ce..c7c97f24 100644
--- a/src/parse/parse.h
+++ b/src/parse/parse.h
@@ -1,49 +1,22 @@
-#pragma once
-
// Parsing logic
+#pragma once
#include "../ast.h"
#include "context.h"
-ast_t *parse(const char *str);
-ast_t *parse_expression(const char *str);
+ast_t *parse_expr_str(const char *str);
-ast_t *parse_assignment(parse_ctx_t *ctx, const char *pos);
-ast_t *parse_block(parse_ctx_t *ctx, const char *pos);
ast_t *parse_bool(parse_ctx_t *ctx, const char *pos);
-ast_t *parse_declaration(parse_ctx_t *ctx, const char *pos);
-ast_t *parse_defer(parse_ctx_t *ctx, const char *pos);
-ast_t *parse_do(parse_ctx_t *ctx, const char *pos);
-ast_t *parse_doctest(parse_ctx_t *ctx, const char *pos);
-ast_t *parse_assert(parse_ctx_t *ctx, const char *pos);
-ast_t *parse_enum_def(parse_ctx_t *ctx, const char *pos);
ast_t *parse_expr(parse_ctx_t *ctx, const char *pos);
ast_t *parse_extended_expr(parse_ctx_t *ctx, const char *pos);
-ast_t *parse_extern(parse_ctx_t *ctx, const char *pos);
-ast_t *parse_for(parse_ctx_t *ctx, const char *pos);
ast_t *parse_heap_alloc(parse_ctx_t *ctx, const char *pos);
-ast_t *parse_if(parse_ctx_t *ctx, const char *pos);
-ast_t *parse_lang_def(parse_ctx_t *ctx, const char *pos);
-ast_t *parse_extend(parse_ctx_t *ctx, const char *pos);
-ast_t *parse_namespace(parse_ctx_t *ctx, const char *pos);
ast_t *parse_negative(parse_ctx_t *ctx, const char *pos);
ast_t *parse_not(parse_ctx_t *ctx, const char *pos);
ast_t *parse_none(parse_ctx_t *ctx, const char *pos);
ast_t *parse_parens(parse_ctx_t *ctx, const char *pos);
-ast_t *parse_pass(parse_ctx_t *ctx, const char *pos);
ast_t *parse_reduction(parse_ctx_t *ctx, const char *pos);
-ast_t *parse_repeat(parse_ctx_t *ctx, const char *pos);
-ast_t *parse_return(parse_ctx_t *ctx, const char *pos);
-ast_t *parse_skip(parse_ctx_t *ctx, const char *pos);
ast_t *parse_stack_reference(parse_ctx_t *ctx, const char *pos);
-ast_t *parse_statement(parse_ctx_t *ctx, const char *pos);
-ast_t *parse_stop(parse_ctx_t *ctx, const char *pos);
-ast_t *parse_struct_def(parse_ctx_t *ctx, const char *pos);
ast_t *parse_term(parse_ctx_t *ctx, const char *pos);
ast_t *parse_term_no_suffix(parse_ctx_t *ctx, const char *pos);
-ast_t *parse_update(parse_ctx_t *ctx, const char *pos);
-ast_t *parse_use(parse_ctx_t *ctx, const char *pos);
ast_t *parse_var(parse_ctx_t *ctx, const char *pos);
-ast_t *parse_when(parse_ctx_t *ctx, const char *pos);
-ast_t *parse_while(parse_ctx_t *ctx, const char *pos);
ast_t *parse_deserialize(parse_ctx_t *ctx, const char *pos);
diff --git a/src/parse/statements.c b/src/parse/statements.c
new file mode 100644
index 00000000..c4ab5608
--- /dev/null
+++ b/src/parse/statements.c
@@ -0,0 +1,159 @@
+// Logic for parsing statements
+
+#include <gc.h>
+#include <stdbool.h>
+#include <string.h>
+
+#include "../ast.h"
+#include "../stdlib/util.h"
+#include "context.h"
+#include "errors.h"
+#include "files.h"
+#include "parse.h"
+#include "statements.h"
+#include "suffixes.h"
+#include "types.h"
+#include "utils.h"
+
+public
+ast_t *parse_declaration(parse_ctx_t *ctx, const char *pos) {
+ const char *start = pos;
+ ast_t *var = parse_var(ctx, pos);
+ if (!var) return NULL;
+ pos = var->end;
+ spaces(&pos);
+ if (!match(&pos, ":")) return NULL;
+ spaces(&pos);
+ type_ast_t *type = optional(ctx, &pos, parse_type);
+ spaces(&pos);
+ ast_t *val = NULL;
+ if (match(&pos, "=")) {
+ val = optional(ctx, &pos, parse_extended_expr);
+ if (!val) {
+ if (optional(ctx, &pos, parse_use))
+ parser_err(ctx, start, pos, "'use' statements are only allowed at the top level of a file");
+ else parser_err(ctx, pos, eol(pos), "This is not a valid expression");
+ }
+ }
+ return NewAST(ctx->file, start, pos, Declare, .var = var, .type = type, .value = val);
+}
+
+public
+ast_t *parse_assignment(parse_ctx_t *ctx, const char *pos) {
+ const char *start = pos;
+ ast_list_t *targets = NULL;
+ for (;;) {
+ ast_t *lhs = optional(ctx, &pos, parse_term);
+ if (!lhs) break;
+ targets = new (ast_list_t, .ast = lhs, .next = targets);
+ spaces(&pos);
+ if (!match(&pos, ",")) break;
+ whitespace(&pos);
+ }
+
+ if (!targets) return NULL;
+
+ spaces(&pos);
+ if (!match(&pos, "=")) return NULL;
+ if (match(&pos, "=")) return NULL; // == comparison
+
+ ast_list_t *values = NULL;
+ for (;;) {
+ ast_t *rhs = optional(ctx, &pos, parse_extended_expr);
+ if (!rhs) break;
+ values = new (ast_list_t, .ast = rhs, .next = values);
+ spaces(&pos);
+ if (!match(&pos, ",")) break;
+ whitespace(&pos);
+ }
+
+ REVERSE_LIST(targets);
+ REVERSE_LIST(values);
+
+ return NewAST(ctx->file, start, pos, Assign, .targets = targets, .values = values);
+}
+
+public
+ast_t *parse_update(parse_ctx_t *ctx, const char *pos) {
+ const char *start = pos;
+ ast_t *lhs = optional(ctx, &pos, parse_expr);
+ if (!lhs) return NULL;
+ spaces(&pos);
+ ast_e op;
+ if (match(&pos, "+=")) op = PlusUpdate;
+ else if (match(&pos, "++=")) op = ConcatUpdate;
+ else if (match(&pos, "-=")) op = MinusUpdate;
+ else if (match(&pos, "*=")) op = MultiplyUpdate;
+ else if (match(&pos, "/=")) op = DivideUpdate;
+ else if (match(&pos, "^=")) op = PowerUpdate;
+ else if (match(&pos, "<<=")) op = LeftShiftUpdate;
+ else if (match(&pos, "<<<=")) op = UnsignedLeftShiftUpdate;
+ else if (match(&pos, ">>=")) op = RightShiftUpdate;
+ else if (match(&pos, ">>>=")) op = UnsignedRightShiftUpdate;
+ else if (match(&pos, "and=")) op = AndUpdate;
+ else if (match(&pos, "or=")) op = OrUpdate;
+ else if (match(&pos, "xor=")) op = XorUpdate;
+ else return NULL;
+ ast_t *rhs = expect(ctx, start, &pos, parse_extended_expr, "I expected an expression here");
+ return new (ast_t, .file = ctx->file, .start = start, .end = pos, .tag = op, .__data.PlusUpdate.lhs = lhs,
+ .__data.PlusUpdate.rhs = rhs);
+}
+
+public
+ast_t *parse_doctest(parse_ctx_t *ctx, const char *pos) {
+ const char *start = pos;
+ if (!match(&pos, ">>")) return NULL;
+ spaces(&pos);
+ ast_t *expr = expect(ctx, start, &pos, parse_statement, "I couldn't parse the expression for this doctest");
+ whitespace(&pos);
+ ast_t *expected = NULL;
+ if (match(&pos, "=")) {
+ spaces(&pos);
+ expected = expect(ctx, start, &pos, parse_extended_expr, "I couldn't parse the expected expression here");
+ } else {
+ pos = expr->end;
+ }
+ return NewAST(ctx->file, start, pos, DocTest, .expr = expr, .expected = expected);
+}
+
+public
+ast_t *parse_assert(parse_ctx_t *ctx, const char *pos) {
+ const char *start = pos;
+ if (!match_word(&pos, "assert")) return NULL;
+ spaces(&pos);
+ ast_t *expr = expect(ctx, start, &pos, parse_extended_expr, "I couldn't parse the expression for this assert");
+ spaces(&pos);
+ ast_t *message = NULL;
+ if (match(&pos, ",")) {
+ whitespace(&pos);
+ message = expect(ctx, start, &pos, parse_extended_expr, "I couldn't parse the error message for this assert");
+ } else {
+ pos = expr->end;
+ }
+ return NewAST(ctx->file, start, pos, Assert, .expr = expr, .message = message);
+}
+
+public
+ast_t *parse_statement(parse_ctx_t *ctx, const char *pos) {
+ ast_t *stmt = NULL;
+ if ((stmt = parse_declaration(ctx, pos)) || (stmt = parse_doctest(ctx, pos)) || (stmt = parse_assert(ctx, pos)))
+ return stmt;
+
+ if (!(false || (stmt = parse_update(ctx, pos)) || (stmt = parse_assignment(ctx, pos))))
+ stmt = parse_extended_expr(ctx, pos);
+
+ for (bool progress = (stmt != NULL); progress;) {
+ ast_t *new_stmt;
+ progress = false;
+ if (stmt->tag == Var) {
+ progress = (false || (new_stmt = parse_method_call_suffix(ctx, stmt))
+ || (new_stmt = parse_fncall_suffix(ctx, stmt)));
+ } else if (stmt->tag == FunctionCall) {
+ new_stmt = parse_optional_conditional_suffix(ctx, stmt);
+ progress = (new_stmt != stmt);
+ }
+
+ if (progress) stmt = new_stmt;
+ }
+ return stmt;
+}
diff --git a/src/parse/statements.h b/src/parse/statements.h
new file mode 100644
index 00000000..fbc1af8f
--- /dev/null
+++ b/src/parse/statements.h
@@ -0,0 +1,12 @@
+// Logic for parsing statements
+#pragma once
+
+#include "../ast.h"
+#include "context.h"
+
+ast_t *parse_assignment(parse_ctx_t *ctx, const char *pos);
+ast_t *parse_declaration(parse_ctx_t *ctx, const char *pos);
+ast_t *parse_doctest(parse_ctx_t *ctx, const char *pos);
+ast_t *parse_assert(parse_ctx_t *ctx, const char *pos);
+ast_t *parse_statement(parse_ctx_t *ctx, const char *pos);
+ast_t *parse_update(parse_ctx_t *ctx, const char *pos);
diff --git a/src/parse/typedefs.c b/src/parse/typedefs.c
new file mode 100644
index 00000000..73fe9d7c
--- /dev/null
+++ b/src/parse/typedefs.c
@@ -0,0 +1,204 @@
+// Parsing logic for type definitions
+
+#include <stdbool.h>
+#include <string.h>
+
+#include "../ast.h"
+#include "../stdlib/util.h"
+#include "context.h"
+#include "errors.h"
+#include "files.h"
+#include "functions.h"
+#include "statements.h"
+#include "text.h"
+#include "typedefs.h"
+#include "utils.h"
+
+ast_t *parse_namespace(parse_ctx_t *ctx, const char *pos) {
+ const char *start = pos;
+ whitespace(&pos);
+ int64_t indent = get_indent(ctx, pos);
+ ast_list_t *statements = NULL;
+ for (;;) {
+ const char *next = pos;
+ whitespace(&next);
+ if (get_indent(ctx, next) != indent) break;
+ ast_t *stmt;
+ if ((stmt = optional(ctx, &pos, parse_struct_def)) || (stmt = optional(ctx, &pos, parse_func_def))
+ || (stmt = optional(ctx, &pos, parse_enum_def)) || (stmt = optional(ctx, &pos, parse_lang_def))
+ || (stmt = optional(ctx, &pos, parse_extend)) || (stmt = optional(ctx, &pos, parse_convert_def))
+ || (stmt = optional(ctx, &pos, parse_use)) || (stmt = optional(ctx, &pos, parse_extern))
+ || (stmt = optional(ctx, &pos, parse_inline_c)) || (stmt = optional(ctx, &pos, parse_declaration))) {
+ statements = new (ast_list_t, .ast = stmt, .next = statements);
+ pos = stmt->end;
+ whitespace(&pos); // TODO: check for newline
+ // if (!(space_types & WHITESPACE_NEWLINES)) {
+ // pos = stmt->end;
+ // break;
+ // }
+ } else {
+ if (get_indent(ctx, next) > indent && next < eol(next))
+ parser_err(ctx, next, eol(next), "I couldn't parse this namespace declaration");
+ break;
+ }
+ }
+ REVERSE_LIST(statements);
+ return NewAST(ctx->file, start, pos, Block, .statements = statements);
+}
+
+ast_t *parse_struct_def(parse_ctx_t *ctx, const char *pos) {
+ // struct Foo(...) [: \n body]
+ const char *start = pos;
+ if (!match_word(&pos, "struct")) return NULL;
+
+ int64_t starting_indent = get_indent(ctx, pos);
+
+ spaces(&pos);
+ const char *name = get_id(&pos);
+ if (!name) parser_err(ctx, start, pos, "I expected a name for this struct");
+ spaces(&pos);
+
+ if (!match(&pos, "(")) parser_err(ctx, pos, pos, "I expected a '(' and a list of fields here");
+
+ arg_ast_t *fields = parse_args(ctx, &pos);
+
+ whitespace(&pos);
+ bool secret = false, external = false, opaque = false;
+ if (match(&pos, ";")) { // Extra flags
+ whitespace(&pos);
+ for (;;) {
+ if (match_word(&pos, "secret")) {
+ secret = true;
+ } else if (match_word(&pos, "extern")) {
+ external = true;
+ } else if (match_word(&pos, "opaque")) {
+ if (fields)
+ parser_err(ctx, pos - strlen("opaque"), pos, "A struct can't be opaque if it has fields defined");
+ opaque = true;
+ } else {
+ break;
+ }
+
+ if (!match_separator(&pos)) break;
+ }
+ }
+
+ expect_closing(ctx, &pos, ")", "I wasn't able to parse the rest of this struct");
+
+ ast_t *namespace = NULL;
+ const char *ns_pos = pos;
+ whitespace(&ns_pos);
+ int64_t ns_indent = get_indent(ctx, ns_pos);
+ if (ns_indent > starting_indent) {
+ pos = ns_pos;
+ namespace = optional(ctx, &pos, parse_namespace);
+ }
+ if (!namespace) namespace = NewAST(ctx->file, pos, pos, Block, .statements = NULL);
+ return NewAST(ctx->file, start, pos, StructDef, .name = name, .fields = fields, .namespace = namespace,
+ .secret = secret, .external = external, .opaque = opaque);
+}
+
+ast_t *parse_enum_def(parse_ctx_t *ctx, const char *pos) {
+ // tagged union: enum Foo(a, b(x:Int,y:Int)=5, ...) [: \n namespace]
+ const char *start = pos;
+ if (!match_word(&pos, "enum")) return NULL;
+ int64_t starting_indent = get_indent(ctx, pos);
+ spaces(&pos);
+ const char *name = get_id(&pos);
+ if (!name) parser_err(ctx, start, pos, "I expected a name for this enum");
+ spaces(&pos);
+ if (!match(&pos, "(")) return NULL;
+
+ tag_ast_t *tags = NULL;
+ whitespace(&pos);
+ for (;;) {
+ spaces(&pos);
+ const char *tag_name = get_id(&pos);
+ if (!tag_name) break;
+
+ spaces(&pos);
+ arg_ast_t *fields;
+ bool secret = false;
+ if (match(&pos, "(")) {
+ whitespace(&pos);
+ fields = parse_args(ctx, &pos);
+ whitespace(&pos);
+ if (match(&pos, ";")) { // Extra flags
+ whitespace(&pos);
+ secret = match_word(&pos, "secret");
+ whitespace(&pos);
+ }
+ expect_closing(ctx, &pos, ")", "I wasn't able to parse the rest of this tagged union member");
+ } else {
+ fields = NULL;
+ }
+
+ tags = new (tag_ast_t, .name = tag_name, .fields = fields, .secret = secret, .next = tags);
+
+ if (!match_separator(&pos)) break;
+ }
+
+ whitespace(&pos);
+ expect_closing(ctx, &pos, ")", "I wasn't able to parse the rest of this enum definition");
+
+ REVERSE_LIST(tags);
+
+ if (tags == NULL) parser_err(ctx, start, pos, "This enum does not have any tags!");
+
+ ast_t *namespace = NULL;
+ const char *ns_pos = pos;
+ whitespace(&ns_pos);
+ int64_t ns_indent = get_indent(ctx, ns_pos);
+ if (ns_indent > starting_indent) {
+ pos = ns_pos;
+ namespace = optional(ctx, &pos, parse_namespace);
+ }
+ if (!namespace) namespace = NewAST(ctx->file, pos, pos, Block, .statements = NULL);
+
+ return NewAST(ctx->file, start, pos, EnumDef, .name = name, .tags = tags, .namespace = namespace);
+}
+
+ast_t *parse_lang_def(parse_ctx_t *ctx, const char *pos) {
+ const char *start = pos;
+ // lang Name: [namespace...]
+ if (!match_word(&pos, "lang")) return NULL;
+ int64_t starting_indent = get_indent(ctx, pos);
+ spaces(&pos);
+ const char *name = get_id(&pos);
+ if (!name) parser_err(ctx, start, pos, "I expected a name for this lang");
+ spaces(&pos);
+
+ ast_t *namespace = NULL;
+ const char *ns_pos = pos;
+ whitespace(&ns_pos);
+ int64_t ns_indent = get_indent(ctx, ns_pos);
+ if (ns_indent > starting_indent) {
+ pos = ns_pos;
+ namespace = optional(ctx, &pos, parse_namespace);
+ }
+ if (!namespace) namespace = NewAST(ctx->file, pos, pos, Block, .statements = NULL);
+
+ return NewAST(ctx->file, start, pos, LangDef, .name = name, .namespace = namespace);
+}
+
+ast_t *parse_extend(parse_ctx_t *ctx, const char *pos) {
+ const char *start = pos;
+ // extend Name: body...
+ if (!match_word(&pos, "extend")) return NULL;
+ int64_t starting_indent = get_indent(ctx, pos);
+ spaces(&pos);
+ const char *name = get_id(&pos);
+ if (!name) parser_err(ctx, start, pos, "I expected a name for this lang");
+
+ ast_t *body = NULL;
+ const char *ns_pos = pos;
+ whitespace(&ns_pos);
+ int64_t ns_indent = get_indent(ctx, ns_pos);
+ if (ns_indent > starting_indent) {
+ pos = ns_pos;
+ body = optional(ctx, &pos, parse_namespace);
+ }
+ if (!body) body = NewAST(ctx->file, pos, pos, Block, .statements = NULL);
+
+ return NewAST(ctx->file, start, pos, Extend, .name = name, .body = body);
+}
diff --git a/src/parse/typedefs.h b/src/parse/typedefs.h
new file mode 100644
index 00000000..1a746cff
--- /dev/null
+++ b/src/parse/typedefs.h
@@ -0,0 +1,11 @@
+// Parsing logic for type definitions
+#pragma once
+
+#include "../ast.h"
+#include "context.h"
+
+ast_t *parse_lang_def(parse_ctx_t *ctx, const char *pos);
+ast_t *parse_enum_def(parse_ctx_t *ctx, const char *pos);
+ast_t *parse_struct_def(parse_ctx_t *ctx, const char *pos);
+ast_t *parse_extend(parse_ctx_t *ctx, const char *pos);
+ast_t *parse_namespace(parse_ctx_t *ctx, const char *pos);
generated by cgit v1.2.3 (git 2.25.1) at 2026-03-01 10:56:26 +0000