refactor: 独立 desugar pass — match/guard/for/if-let/复合赋值去糖提取

5 种去糖逻辑从 parser.c/expr.c 内联代码提取到 desugar.c:
  desugar_match() — match → let + if-else 链
  desugar_guard() — guard → if !(cond)
  desugar_for()  — for-in-to → var + while
  desugar_if_let() — if let → let + if
  desugar_compound_assign() — +=/-= → assign + binary

parser.c 662→564 行, 新增 desugar.c 109 行, 管道+插值保留在 expr.c

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
This commit is contained in:
2026-06-06 20:20:33 +08:00
parent f3cca30cca
commit eaae0b1c62
4 changed files with 149 additions and 122 deletions
+109
View File
@@ -0,0 +1,109 @@
#include "desugar.h"
#include <stdio.h>
#include <string.h>
// === match → let + if-else 链 ===
AstNode* desugar_match(Parser* p, const Token* match_tok,
AstNode* matched_expr,
AstNode** arm_patterns, bool* arm_is_wildcard,
AstNode** arm_bodies, int arm_count) {
const char* varname = arena_strdup_impl(p->arena, "__match_val", 12);
// 从最后一个分支往前构建 if-else 链
AstNode* result = NULL;
for (int i = arm_count - 1; i >= 0; i--) {
if (arm_is_wildcard[i]) {
AstNode* true_cond = ast_make_literal_bool(p->arena, true, tok_loc(match_tok));
result = ast_make_if(p->arena, true_cond, arm_bodies[i], result, tok_loc(match_tok));
} else {
AstNode* cond = ast_make_binary(p->arena, OP_EQ,
ast_make_ident(p->arena, varname, tok_loc(match_tok)),
arm_patterns[i], tok_loc(match_tok));
result = ast_make_if(p->arena, cond, arm_bodies[i], result, tok_loc(match_tok));
}
}
AstNode* let_stmt = ast_make_let(p->arena, varname, TYPE_UNKNOWN,
false, false, matched_expr, NULL, 0, NULL, 0, tok_loc(match_tok));
AstNode* stmts_arr[2] = { let_stmt, result };
AstNode** stmts = arena_alloc_impl(p->arena, 2 * sizeof(AstNode*));
memcpy(stmts, stmts_arr, 2 * sizeof(AstNode*));
return ast_make_block(p->arena, stmts, 2, tok_loc(match_tok));
}
// === guard → if !(cond) { body } ===
AstNode* desugar_guard(Parser* p, const Token* guard_tok,
AstNode* cond, AstNode* body) {
AstNode* not_cond = ast_make_unary(p->arena, OP_NOT, cond, tok_loc(guard_tok));
return ast_make_if(p->arena, not_cond, body, NULL, tok_loc(guard_tok));
}
// === for i in start to end { body } → { var i = start; while i < end { body; i = i + 1; } } ===
AstNode* desugar_for(Parser* p, const Token* for_tok,
const char* var_name, AstNode* start_expr,
AstNode* end_expr, AstNode* body) {
// var i = start;
AstNode* let_stmt = ast_make_let(p->arena, var_name, TYPE_UNKNOWN,
false, true, start_expr, NULL, 0, NULL, 0, tok_loc(for_tok));
// i < end
AstNode* cond = ast_make_binary(p->arena, OP_LT,
ast_make_ident(p->arena, var_name, tok_loc(for_tok)),
end_expr, tok_loc(for_tok));
// i = i + 1
AstNode* incr = ast_make_assign(p->arena, var_name,
ast_make_binary(p->arena, OP_ADD,
ast_make_ident(p->arena, var_name, tok_loc(for_tok)),
ast_make_literal_i64(p->arena, 1, tok_loc(for_tok)),
tok_loc(for_tok)),
tok_loc(for_tok));
// 增量追加到循环体末尾
AstNode** new_stmts = arena_alloc_impl(p->arena,
(body->as.block.stmt_count + 1) * sizeof(AstNode*));
memcpy(new_stmts, body->as.block.stmts, body->as.block.stmt_count * sizeof(AstNode*));
new_stmts[body->as.block.stmt_count] = incr;
AstNode* new_body = ast_make_block(p->arena, new_stmts,
body->as.block.stmt_count + 1, body->loc);
// while i < end { body; i = i + 1; }
AstNode* while_loop = ast_make_while(p->arena, cond, new_body, tok_loc(for_tok));
AstNode* stmts_arr[2] = { let_stmt, while_loop };
AstNode** stmts = arena_alloc_impl(p->arena, 2 * sizeof(AstNode*));
memcpy(stmts, stmts_arr, 2 * sizeof(AstNode*));
return ast_make_block(p->arena, stmts, 2, tok_loc(for_tok));
}
// === if let pattern = expr { then } else { else } ===
AstNode* desugar_if_let(Parser* p, const Token* if_tok,
AstNode* pattern, AstNode* match_expr,
AstNode* then_block, AstNode* else_block) {
static int iflet_counter = 0;
char vname_buf[32];
snprintf(vname_buf, sizeof(vname_buf), "__iflet_%d", iflet_counter++);
const char* vname = arena_strdup_impl(p->arena, vname_buf, strlen(vname_buf));
AstNode* let_stmt = ast_make_let(p->arena, vname, TYPE_UNKNOWN,
false, false, match_expr, NULL, 0, NULL, 0, tok_loc(if_tok));
AstNode* cond = ast_make_binary(p->arena, OP_EQ,
ast_make_ident(p->arena, vname, tok_loc(if_tok)),
pattern, tok_loc(if_tok));
AstNode* if_stmt = ast_make_if(p->arena, cond, then_block, else_block, tok_loc(if_tok));
AstNode* stmts[2] = { let_stmt, if_stmt };
AstNode** arr = arena_alloc_impl(p->arena, 2 * sizeof(AstNode*));
memcpy(arr, stmts, 2 * sizeof(AstNode*));
return ast_make_block(p->arena, arr, 2, tok_loc(if_tok));
}
// === ident += expr → ident = ident + expr ===
AstNode* desugar_compound_assign(Parser* p, const Token* name_tok,
BinaryOp binop, AstNode* rhs) {
const char* vname = arena_strdup_impl(p->arena, name_tok->start, name_tok->length);
AstNode* lhs_ident = ast_make_ident(p->arena, vname, tok_loc(name_tok));
AstNode* bin_expr = ast_make_binary(p->arena, binop, lhs_ident, rhs, tok_loc(name_tok));
return ast_make_assign(p->arena, vname, bin_expr, tok_loc(name_tok));
}
+30
View File
@@ -0,0 +1,30 @@
#ifndef DESUGAR_H
#define DESUGAR_H
#include "parse_internal.h"
// match expr { pat1 => body1, ... } → let + if-else 链
AstNode* desugar_match(Parser* p, const Token* match_tok,
AstNode* matched_expr,
AstNode** arm_patterns, bool* arm_is_wildcard,
AstNode** arm_bodies, int arm_count);
// guard cond else { body } → if !(cond) { body }
AstNode* desugar_guard(Parser* p, const Token* guard_tok,
AstNode* cond, AstNode* body);
// for i in start to end { body } → { var i = start; while i < end { body; i = i + 1; } }
AstNode* desugar_for(Parser* p, const Token* for_tok,
const char* var_name, AstNode* start_expr,
AstNode* end_expr, AstNode* body);
// if let pattern = expr { then } else { else } → { let __match = expr; if __match == pattern ... }
AstNode* desugar_if_let(Parser* p, const Token* if_tok,
AstNode* pattern, AstNode* match_expr,
AstNode* then_block, AstNode* else_block);
// ident <op>= expr → ident = ident <op> expr
AstNode* desugar_compound_assign(Parser* p, const Token* name_tok,
BinaryOp binop, AstNode* rhs);
#endif
+2 -16
View File
@@ -1,4 +1,5 @@
#include "parse_internal.h"
#include "desugar.h"
#include <stdio.h>
#include <string.h>
@@ -287,22 +288,7 @@ AstNode* parse_expr_prec(Parser* p, Precedence min_prec, ErrorInfo* error) {
else_block = parse_block(p, error);
if (!else_block) return NULL;
}
// 去糖: { let __match = expr; if __match == pattern { then } else { else } }
static int iflet_counter = 0;
char vname_buf[32];
snprintf(vname_buf, sizeof(vname_buf), "__iflet_%d", iflet_counter++);
const char* vname = arena_strdup_impl(p->arena, vname_buf, strlen(vname_buf));
AstNode* let_stmt = ast_make_let(p->arena,
vname, TYPE_UNKNOWN,
false, false, match_expr, NULL, 0, NULL, 0, tok_loc(if_tok));
AstNode* cond = ast_make_binary(p->arena, OP_EQ,
ast_make_ident(p->arena, vname, tok_loc(if_tok)),
pattern, tok_loc(if_tok));
AstNode* if_stmt = ast_make_if(p->arena, cond, then_block, else_block, tok_loc(if_tok));
AstNode* stmts[2] = { let_stmt, if_stmt };
AstNode** arr = arena_alloc_impl(p->arena, 2 * sizeof(AstNode*));
memcpy(arr, stmts, 2 * sizeof(AstNode*));
left = ast_make_block(p->arena, arr, 2, tok_loc(if_tok));
left = desugar_if_let(p, if_tok, pattern, match_expr, then_block, else_block);
} else {
// if-expr: if cond { then } else { else }
AstNode* cond = parse_expr(p, error);
+8 -106
View File
@@ -1,4 +1,5 @@
#include "parse_internal.h"
#include "desugar.h"
#include <stdio.h>
#include <string.h>
#include <stdlib.h>
@@ -48,9 +49,6 @@ static AstNode* parse_match_stmt(Parser* p, ErrorInfo* error) {
if (!expect(p, TOK_LBRACE, error, "match 后缺少 '{'")) return NULL;
// 分配临时变量名
const char* varname = arena_strdup_impl(p->arena, "__match_val", 12);
// 收集所有分支
enum { MAX_ARMS = 64 };
bool arm_is_wildcard[MAX_ARMS];
@@ -65,64 +63,29 @@ static AstNode* parse_match_stmt(Parser* p, ErrorInfo* error) {
error->line = peek(p)->line; error->col = peek(p)->col;
return NULL;
}
if (peek(p)->kind == TOK_UNDERSCORE) {
arm_is_wildcard[arm_count] = true;
arm_pattern[arm_count] = NULL;
advance(p); // 跳过 '_'
advance(p);
} else {
arm_is_wildcard[arm_count] = false;
arm_pattern[arm_count] = parse_expr(p, error);
if (!arm_pattern[arm_count]) return NULL;
}
// 解析 '=>'
if (!expect(p, TOK_MATCH_ARROW, error, "match 分支缺少 '=>'")) return NULL;
// 解析分支体(必须是一个代码块)
arm_body[arm_count] = parse_block(p, error);
if (!arm_body[arm_count]) return NULL;
arm_count++;
// 跳过可选逗号
if (peek(p)->kind == TOK_COMMA) advance(p);
}
if (!expect(p, TOK_RBRACE, error, "缺少 '}'")) return NULL;
if (arm_count == 0) {
error->message = "match 表达式至少需要一个分支";
error->filename = p->filename;
error->line = match_tok->line; error->col = match_tok->col;
return NULL;
}
// 从最后一个分支往前构建 if-else 链(最后一个分支 = 最内层 else)
AstNode* result = NULL;
for (int i = arm_count - 1; i >= 0; i--) {
if (arm_is_wildcard[i]) {
// 通配符分支:if (true) { body } else { result }
AstNode* true_cond = ast_make_literal_bool(p->arena, true, tok_loc(match_tok));
result = ast_make_if(p->arena, true_cond, arm_body[i], result, tok_loc(match_tok));
} else {
// if (__match_val == pattern) { body } else { result }
AstNode* cond = ast_make_binary(p->arena, OP_EQ,
ast_make_ident(p->arena, varname, tok_loc(match_tok)),
arm_pattern[i], tok_loc(match_tok));
result = ast_make_if(p->arena, cond, arm_body[i], result, tok_loc(match_tok));
}
}
// 构建 let __match_val = <matched>;
AstNode* let_stmt = ast_make_let(p->arena, varname, TYPE_UNKNOWN,
false, false, matched, NULL, 0, NULL, 0, tok_loc(match_tok));
// 包装为代码块: { let __match_val = <expr>; <if-else 链> }
AstNode* stmts_arr[2] = { let_stmt, result };
AstNode** stmts = arena_alloc_impl(p->arena, 2 * sizeof(AstNode*));
memcpy(stmts, stmts_arr, 2 * sizeof(AstNode*));
return ast_make_block(p->arena, stmts, 2, tok_loc(match_tok));
return desugar_match(p, match_tok, matched, arm_pattern, arm_is_wildcard, arm_body, arm_count);
}
// === 代码块解析 ===
@@ -201,67 +164,19 @@ AstNode* parse_statement(Parser* p, ErrorInfo* error) {
}
if (t->kind == TOK_FOR) {
advance(p); // 跳过 'for'
// 解析循环变量名
advance(p);
const Token* var_name = expect(p, TOK_IDENT, error, "for 后应为变量名");
if (!var_name) return NULL;
// 解析 'in'
if (!expect(p, TOK_IN, error, "缺少 'in'")) return NULL;
// 解析起始表达式
AstNode* start_expr = parse_expr(p, error);
if (!start_expr) return NULL;
// 解析 'to'
if (!expect(p, TOK_TO, error, "缺少 'to'")) return NULL;
// 解析结束表达式
AstNode* end_expr = parse_expr(p, error);
if (!end_expr) return NULL;
// 解析循环体
AstNode* body = parse_block(p, error);
if (!body) return NULL;
// 脱糖: for i in start to end { body; }
// → { var i = start; while i < end { body; i = i + 1; } }
const char* vname = arena_strdup_impl(p->arena, var_name->start, var_name->length);
// 构建: var i = start;
AstNode* let_stmt = ast_make_let(p->arena, vname, TYPE_UNKNOWN, false, true, start_expr, NULL, 0, NULL, 0, tok_loc(var_name));
// 构建: i < end (while 条件)
AstNode* cond = ast_make_binary(p->arena, OP_LT,
ast_make_ident(p->arena, vname, tok_loc(var_name)),
end_expr, tok_loc(var_name));
// 构建: i = i + 1 (循环增量)
AstNode* incr = ast_make_assign(p->arena, vname,
ast_make_binary(p->arena, OP_ADD,
ast_make_ident(p->arena, vname, tok_loc(var_name)),
ast_make_literal_i64(p->arena, 1, tok_loc(var_name)),
tok_loc(var_name)),
tok_loc(var_name));
// 将增量追加到循环体末尾
AstNode** new_stmts = arena_alloc_impl(p->arena,
(body->as.block.stmt_count + 1) * sizeof(AstNode*));
memcpy(new_stmts, body->as.block.stmts, body->as.block.stmt_count * sizeof(AstNode*));
new_stmts[body->as.block.stmt_count] = incr;
AstNode* new_body = ast_make_block(p->arena, new_stmts,
body->as.block.stmt_count + 1, body->loc);
// 构建: while i < end { ... body ... ; i = i + 1; }
AstNode* while_loop = ast_make_while(p->arena, cond, new_body, tok_loc(t));
// 包装: { var i = start; while i < end { ... } }
AstNode* stmts_arr[2] = { let_stmt, while_loop };
AstNode** stmts = arena_alloc_impl(p->arena, 2 * sizeof(AstNode*));
memcpy(stmts, stmts_arr, 2 * sizeof(AstNode*));
return ast_make_block(p->arena, stmts, 2, tok_loc(t));
return desugar_for(p, t, vname, start_expr, end_expr, body);
}
if (t->kind == TOK_MATCH) {
@@ -269,16 +184,13 @@ AstNode* parse_statement(Parser* p, ErrorInfo* error) {
}
if (t->kind == TOK_GUARD) {
// guard expr else { ... } → if !(expr) { ... }
const Token* guard_tok = advance(p);
AstNode* cond = parse_expr(p, error);
if (!cond) return NULL;
if (!expect(p, TOK_ELSE, error, "guard 缺少 'else'")) return NULL;
AstNode* body = parse_block(p, error);
if (!body) return NULL;
// 去糖: if !cond { body }
AstNode* not_cond = ast_make_unary(p->arena, OP_NOT, cond, tok_loc(guard_tok));
return ast_make_if(p->arena, not_cond, body, NULL, tok_loc(guard_tok));
return desugar_guard(p, guard_tok, cond, body);
}
if (t->kind == TOK_RETURN) {
@@ -333,9 +245,8 @@ AstNode* parse_statement(Parser* p, ErrorInfo* error) {
if (t->kind == TOK_IDENT) {
TokenKind next_kind = (t + 1)->kind;
if (next_kind >= TOK_PLUS_EQ && next_kind <= TOK_SLASH_EQ) {
const Token* name = advance(p); // 消费标识符
const Token* name = advance(p);
TokenKind comp_op = advance(p)->kind;
BinaryOp binop;
switch (comp_op) {
case TOK_PLUS_EQ: binop = OP_ADD; break;
@@ -344,19 +255,10 @@ AstNode* parse_statement(Parser* p, ErrorInfo* error) {
case TOK_SLASH_EQ: binop = OP_DIV; break;
default: break;
}
AstNode* rhs = parse_expr(p, error);
if (!rhs) return NULL;
if (!expect(p, TOK_SEMICOLON, error, "缺少 ';'")) return NULL;
AstNode* lhs_ident = ast_make_ident(p->arena,
arena_strdup_impl(p->arena, name->start, name->length),
tok_loc(name));
AstNode* bin_expr = ast_make_binary(p->arena, binop, lhs_ident, rhs,
tok_loc(name));
return ast_make_assign(p->arena,
arena_strdup_impl(p->arena, name->start, name->length),
bin_expr, tok_loc(name));
return desugar_compound_assign(p, name, binop, rhs);
}
}