feat(ast2ir): 添加左值标识支持以改善表达式处理

- 在 scc_ast2ir_expr 函数中添加 is_lvalue 参数来区分左值和右值表达式
- 更新二元表达式处理逻辑,特别是赋值操作符的处理
- 改进标识符表达式的处理,根据是否为左值决定返回存储位置还是加载值
- 修复哈希比较函数的实现
- 移除调试相关的注释代码

refactor(parser): 优化语法分析器错误处理和控制流

- 移除不必要的错误恢复辅助注释
- 修改表达式解析的控制流程,将直接返回改为使用 break 语句
- 添加语义分析回调,在解析完成后进行标识符查找和验证

refactor(sema): 增强语义分析阶段的符号表管理

- 改进标识符查找逻辑,增加对非变量标识符的检查
- 扩展声明处理范围,包括变量和参数声明的符号表注册
- 为函数声明添加作用域管理

fix(parser): 修正单元测试中的类型定义

- 将 long long 类型定义改为 int 类型,解决测试兼容性问题

refactor(sccf): 重构文件格式定义和构建器实现

- 重命名符号类型枚举值 OBJECT 为 EXTERN
- 重命名段类型枚举值 RELOC 为 RELOCS
- 修正结构体字段命名的一致性问题
- 重新设计 SCCF 构建器的数据结构和API
- 添加符号表、字符串表和重定位表的构建支持

refactor(target): 重命名Windows PE相关类型定义

- 将 scc_winpe_* 类型重命名为 scc_pe_* 以保持命名一致性

chore: 添加 sccf2target 模块用于格式转换

- 创建新的库模块用于 SCCF 到目标格式的转换
- 实现 PE 格式转换的基本功能
- 添加示例程序演示格式转换过程
This commit is contained in:
zzy
2026-03-19 12:11:57 +08:00
parent 5f915ba8d3
commit 02a6c684f1
18 changed files with 575 additions and 101 deletions

View File

@@ -16,10 +16,6 @@
#define sccf_size_t uint64_t
#define sccf_isize_t int64_t
#ifdef __cplusplus
extern "C" {
#endif
/** SCCF魔数 */
#define SCCF_MAGIC "SCCFmt\0\0"
@@ -47,7 +43,7 @@ typedef enum {
SCCF_SYM_TYPE_UNDEF = 0, ///< 未定义
SCCF_SYM_TYPE_FUNC = 1, ///< 函数
SCCF_SYM_TYPE_DATA = 2, ///< 数据
SCCF_SYM_TYPE_OBJECT = 3, ///< 对象
SCCF_SYM_TYPE_EXTERN = 3, ///< 外部符号
} sccf_sym_type_t;
/** 符号绑定类型 */
@@ -73,7 +69,7 @@ typedef enum {
SCCF_SECT_UNINIT_DATA = 4, ///< BSS段(未初始化数据)
SCCF_SECT_SYMTAB = 5, ///< 符号表
SCCF_SECT_STRTAB = 6, ///< 字符串表
SCCF_SECT_RELOC = 7, ///< 重定位表
SCCF_SECT_RELOCS = 7, ///< 重定位表
} sccf_sect_type_t;
/** 重定位类型 */
@@ -99,37 +95,37 @@ typedef struct sccf_header {
* @brief SCCF段
*/
typedef struct {
sccf_byte_t name[8]; ///< 段名称 仅供展示 eg. emoji or ".text"
sccf_enum_t scf_sect_type; ///< 段类型 (内部实际区分的方式)
sccf_size_t size; ///< 段数据实际的大小
sccf_size_t data_size; ///< 段数据的有效数据大小
sccf_size_t addralign; ///< 内存对齐要求 (字节对齐, 如 1,2,4,8,...)
sccf_size_t info; ///< 段信息 (如条目数,链接索引等)
sccf_size_t reserved[2]; ///< 保留
sccf_byte_t name[8]; ///< 段名称 仅供展示 eg. emoji or ".text"
sccf_enum_t sccf_sect_type; ///< 段类型 (内部实际区分的方式)
sccf_size_t size; ///< 段数据实际的大小
sccf_size_t data_size; ///< 段数据的有效数据大小
sccf_size_t addralign; ///< 内存对齐要求 (字节对齐, 如 1,2,4,8,...)
sccf_size_t info; ///< 段信息 (如条目数,链接索引等)
sccf_size_t reserved[2]; ///< 保留
} sccf_sect_header_t;
/**
* @brief SCCF符号表
*/
typedef struct {
sccf_size_t name_offset; ///< 符号名称在字符串表中的偏移量
sccf_enum_t scf_sym_type; ///< 符号类型
sccf_enum_t scf_sym_bind; ///< 符号绑定类型
sccf_enum_t scf_sym_vis; ///< 符号可见性
sccf_enum_t scf_sect_type; ///< 该符号的段类型
sccf_size_t scf_sect_offset; ///< 该符号在段中的偏移量
sccf_size_t scf_sym_size; ///< 该符号符号选中的大小
sccf_size_t name_offset; ///< 符号名称在字符串表中的偏移量
sccf_enum_t sccf_sym_type; ///< 符号类型
sccf_enum_t sccf_sym_bind; ///< 符号绑定类型
sccf_enum_t sccf_sym_vis; ///< 符号可见性
sccf_enum_t sccf_sect_type; ///< 该符号的段类型
sccf_size_t sccf_sect_offset; ///< 该符号在段中的偏移量
sccf_size_t sccf_sym_size; ///< 该符号选中的大小
} sccf_sym_t;
/**
* @brief SCCF重定向条目
*/
typedef struct {
sccf_size_t offset; ///< 在数据段中的偏移量
sccf_size_t sym_idx; ///< 符号索引
sccf_enum_t type; ///< 重定位类型
sccf_enum_t sect_type; ///< 段类型(代码段/数据段)
sccf_isize_t addend; ///< 加数
sccf_enum_t sect_type; ///< 需要重定位的段类型(代码段/数据段)
sccf_size_t sym_idx; ///< 符号索引(重定向指向的符号)
sccf_size_t offset; ///< 在段中的偏移量(用于重定向的地址)
sccf_isize_t addend; ///< 加数(用于获取相对位置时PC的额外值)
} sccf_reloc_t;
/**
@@ -158,8 +154,4 @@ typedef struct {
* };
*/
#ifdef __cplusplus
}
#endif
#endif /* __SCC_FORMAT_H__ */

View File

@@ -4,24 +4,66 @@
#include "sccf_utils.h"
#include <scc_utils.h>
typedef SCC_VEC(sccf_sym_t) sccf_sym_vec_t;
typedef SCC_VEC(sccf_reloc_t) sccf_reloc_vec_t;
typedef struct {
sccf_t sccf;
int aligned;
scc_strpool_t strpool;
scc_hashtable_t str2offset;
sccf_sym_vec_t syms;
sccf_strtab_t strtab;
sccf_reloc_vec_t relocs;
sccf_sym_vec_t symtab;
scc_hashtable_t str2sym;
scc_hashtable_t str2offset;
} sccf_builder_t;
void sccf_builder_init(sccf_builder_t *builder);
usize sccf_builder_add_symbol(sccf_builder_t *builder, const char *name,
sccf_sym_t *sym);
usize sccf_builder_get_symbol_idx(sccf_builder_t *builder, const char *name);
static inline sccf_sym_t *
sccf_builder_get_symbol_unsafe(sccf_builder_t *builder, usize idx) {
return &scc_vec_at(builder->symtab, idx);
}
void sccf_builder_add_reloc(sccf_builder_t *builder, sccf_reloc_t reloc);
/**
* @brief 必须确保参数合法
*
* @param builder
* @param sect_header
* @param sect_data
*/
void sccf_builder_add_section(sccf_builder_t *builder,
sccf_sect_header_t *sect_header,
sccf_sect_data_t *sect_data);
static inline void sccf_builder_add_text_section(sccf_builder_t *builder,
sccf_sect_data_t *sect_data) {
sccf_sect_header_t text_header = {
.name = ".text",
.addralign = 1,
.data_size = scc_vec_size(*sect_data),
.sccf_sect_type = SCCF_SECT_CODE,
.size = scc_vec_size(*sect_data),
.info = 0,
.reserved = {0},
};
sccf_builder_add_section(builder, &text_header, sect_data);
}
static inline void sccf_builder_add_data_section(sccf_builder_t *builder,
sccf_sect_data_t *sect_data) {
sccf_sect_header_t text_header = {
.name = ".data",
.addralign = 1,
.data_size = scc_vec_size(*sect_data),
.sccf_sect_type = SCCF_SECT_DATA,
.size = scc_vec_size(*sect_data),
.info = 0,
.reserved = {0},
};
sccf_builder_add_section(builder, &text_header, sect_data);
}
const sccf_t *sccf_builder_to_sccf(sccf_builder_t *builder);
void sccf_builder_to_buffer(sccf_builder_t *builder, sccf_buffer_t *buffer);
void sccf_builder_to_file(sccf_builder_t *builder, const char *file_path);

View File

@@ -11,7 +11,7 @@ typedef struct {
sccf_vec_t link_sccfs;
scc_strpool_t strpool;
scc_hashtable_t str2offset;
sccf_sym_vec_t syms;
sccf_sym_vec_t symtab;
sccf_reloc_vec_t relocs;
} sccf_linker_t;

View File

@@ -100,7 +100,7 @@ static inline usize sccf_find_sect_by_type(u8 *base, sccf_enum_t type) {
const sccf_header_t *hdr = (const sccf_header_t *)base;
for (usize i = 0; i < (usize)hdr->sect_header_num; ++i) {
sccf_sect_header_t *sh = sccf_sect_header(base, i);
if (sh->scf_sect_type == type)
if (sh->sccf_sect_type == type)
return i;
}
return (usize)hdr->sect_header_num;
@@ -167,6 +167,9 @@ typedef SCC_VEC(u8) sccf_buffer_t;
typedef SCC_VEC(u8) sccf_sect_data_t;
typedef SCC_VEC(sccf_sect_data_t) sccf_sect_data_vec_t;
typedef SCC_VEC(sccf_sect_header_t) sccf_sect_header_vec_t;
typedef SCC_VEC(sccf_sym_t) sccf_sym_vec_t;
typedef SCC_VEC(sccf_reloc_t) sccf_reloc_vec_t;
typedef SCC_VEC(char) sccf_strtab_t;
typedef struct {
sccf_header_t header;
@@ -241,7 +244,7 @@ static inline void sccf_parse(sccf_t *sccf, sccf_buffer_t *buffer, int copied) {
* @param[in] sccf
* @return usize
*/
static inline usize sccf_size(sccf_t *sccf) {
static inline usize sccf_size(const sccf_t *sccf) {
if (scc_vec_size(sccf->sect_datas) != scc_vec_size(sccf->sect_headers) ||
scc_vec_size(sccf->sect_headers) != sccf->header.sect_header_num) {
Panic();
@@ -265,7 +268,7 @@ static inline usize sccf_size(sccf_t *sccf) {
* @param[in] sccf
* @param[out] buffer
*/
static inline void sccf_write(sccf_t *sccf, sccf_buffer_t *buffer) {
static inline void sccf_write(const sccf_t *sccf, sccf_buffer_t *buffer) {
usize size = sccf_size(sccf);
if (scc_vec_size(*buffer) < size) {
scc_vec_realloc(*buffer, size);

View File

@@ -3,12 +3,53 @@
void sccf_builder_init(sccf_builder_t *builder) {
builder->aligned = 64;
sccf_init(&builder->sccf);
scc_strpool_init(&builder->strpool);
scc_hashtable_init(&builder->str2offset,
(scc_hashtable_hash_func_t)scc_strhash32,
(scc_hashtable_equal_func_t)scc_strcmp);
scc_hashtable_init(&builder->str2sym,
(scc_hashtable_hash_func_t)scc_strhash32,
(scc_hashtable_equal_func_t)scc_strcmp);
scc_vec_init(builder->strtab);
scc_vec_init(builder->relocs);
scc_vec_init(builder->syms);
scc_vec_init(builder->symtab);
///< Push null
scc_vec_push(builder->strtab, (char)'\0');
///< Push null
scc_vec_push(builder->symtab, (sccf_sym_t){0});
}
usize sccf_builder_add_symbol(sccf_builder_t *builder, const char *name,
sccf_sym_t *sym) {
usize offset = 0;
offset = (usize)scc_hashtable_get(&builder->str2offset, name);
if (offset == 0) {
offset = scc_vec_size(builder->strtab);
scc_hashtable_set(&builder->str2offset, name, (void *)offset);
while (*name) {
scc_vec_push(builder->strtab, *name);
name++;
}
scc_vec_push(builder->strtab, '\0');
}
sym->name_offset = offset;
usize sym_idx = scc_vec_size(builder->symtab);
offset = (usize)scc_hashtable_get(&builder->str2sym, name);
if (offset == 0) {
scc_hashtable_set(&builder->str2sym, name, (void *)sym_idx);
}
scc_vec_push(builder->symtab, *sym);
return sym_idx;
}
usize sccf_builder_get_symbol_idx(sccf_builder_t *builder, const char *name) {
usize offset = (usize)scc_hashtable_get(&builder->str2sym, name);
return offset;
}
void sccf_builder_add_reloc(sccf_builder_t *builder, sccf_reloc_t reloc) {
scc_vec_push(builder->relocs, reloc);
}
void sccf_builder_add_section(sccf_builder_t *builder,
@@ -20,14 +61,58 @@ void sccf_builder_add_section(sccf_builder_t *builder,
scc_vec_push(builder->sccf.sect_datas, *sect_data);
}
void sccf_builder_to_buffer(sccf_builder_t *builder, sccf_buffer_t *buffer) {
Assert(builder != null && buffer != null);
const sccf_t *sccf_builder_to_sccf(sccf_builder_t *builder) {
// TODO symtab strtab reloc
// sccf_sect_header_t symtab_header;
// sccf_sect_data_t symtab_data;
// sccf_builder_add_section(builder, &symtab_header, &symtab_data);
sccf_sect_header_t sect_header;
if (scc_vec_size(builder->strtab)) {
sect_header = (sccf_sect_header_t){
.name = ".strtab",
.info = 0,
.data_size = scc_vec_size(builder->strtab),
.addralign = 1,
.size = scc_vec_size(builder->strtab),
.sccf_sect_type = SCCF_SECT_STRTAB,
};
// TODO 转换成 u8[]
sccf_builder_add_section(builder, &sect_header,
(void *)&builder->strtab);
}
sccf_write(&builder->sccf, buffer);
if (scc_vec_size(builder->symtab)) {
sect_header = (sccf_sect_header_t){
.name = ".symtab",
.info = 0,
.data_size = scc_vec_size(builder->symtab),
.addralign = 1,
.size = scc_vec_size(builder->symtab),
.sccf_sect_type = SCCF_SECT_SYMTAB,
};
// TODO 转换成 u8[]
sccf_builder_add_section(builder, &sect_header,
(void *)&builder->symtab);
}
if (scc_vec_size(builder->relocs)) {
sect_header = (sccf_sect_header_t){
.name = ".relocs",
.info = 0,
.data_size = scc_vec_size(builder->relocs),
.addralign = 1,
.size = scc_vec_size(builder->relocs),
.sccf_sect_type = SCCF_SECT_RELOCS,
};
// TODO 转换成 u8[]
sccf_builder_add_section(builder, &sect_header,
(void *)&builder->relocs);
}
return &builder->sccf;
}
void sccf_builder_to_buffer(sccf_builder_t *builder, sccf_buffer_t *buffer) {
Assert(builder != null && buffer != null);
sccf_write(sccf_builder_to_sccf(builder), buffer);
}
void sccf_builder_to_file(sccf_builder_t *builder, const char *file_path) {