diff options
| author | czjstmax <jstmaxlol@disroot.org> | 2026-01-31 15:10:20 +0100 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2026-01-31 15:10:20 +0100 |
| commit | d2e2617dec584884b92eb452f377b20c0bf8f321 (patch) | |
| tree | d7cdc28d1a83f16a0fc7e945963aa070bfa9d3e4 | |
| parent | 0427d254207a69e394499d1abaea768f484f1cb5 (diff) | |
| parent | 051400c70a4d5384923113cfbcbc69e8e58d27a0 (diff) | |
Merge pull request #1 from z-libs/main
Merge newer updates
39 files changed, 1059 insertions, 177 deletions
diff --git a/.github/copilot-instructions.md b/.github/copilot-instructions.md new file mode 100644 index 0000000..fac102a --- /dev/null +++ b/.github/copilot-instructions.md @@ -0,0 +1,87 @@ +# Zen C Copilot Instructions + +These instructions are **MANDATORY** for all code generation and review tasks in the Zen C project. + +## 1. Memory Management (Critical) +* **Arena Allocation**: The compiler uses a bump-pointer arena allocator. + * **MUST USE**: `xmalloc`, `xcalloc`, `xrealloc`, `xstrdup` (defined in `src/utils/utils.c`). + * **NEVER USE**: Standard `malloc`, `calloc`, `free` (unless interfacing with an external library that strictly requires owned heap memory, like CJSON). +* **Destructors/Freeing**: + * `free(ptr)` is `#defined` to `((void)0)` in `src/zprep.h`. It is a no-op. + * Do **NOT** attempt to free AST nodes or types. `ast_free` is a no-op. + * Memory is reclaimed only when the process exits. Design your data structures accordingly (append-only is fine). + +## 2. AST and Type System +* **Creation**: Use `ast_create(NODE_TYPE)` to allocate new nodes. +* **Type Representation**: + * Use `Type` struct (defined in `src/ast/ast.h`). + * Use `type_new(TYPE_KIND)` helper. + * `type_to_string(t)` and `type_to_c_string(t)` return arena-allocated strings. Do not worry about freeing them. +* **Traversal**: + * The compiler uses **Recursive Descent** with **Switch Statements** on `node->type`. + * Do NOT introduce Visitor patterns or callback tables unless consistent with existing code. + +## 3. Parser Patterns +* **Context**: `ParserContext *ctx` is the god-object. It MUST be passed to almost every function in parsing, analysis, and codegen. + * **Signature Rule**: `ReturnType func_name(ParserContext *ctx, ...)` +* **Token Consumption**: + * Use `expect(lexer, TOKEN_TYPE, "error message")` for mandatory tokens. + * For optional tokens, check `l->token.type` and assume `lexer_next(l)` is used to advance (verify specific helper availability). +* **Error Handling**: + * **Fatal**: `zpanic("msg")` or `zpanic_at(token, "msg")`. Exits immediately (or delegates to LSP handler). + * **Warning**: `zwarn("msg")` or `zwarn_at(token, "msg")`. + * **Semantic Errors**: Prefer `zpanic_at` for type errors to give line/col info. + +## 4. Code Generation (C Backend) +* **Generic Mangling**: + * Generic structs (e.g., `Slice<int>`) are mangled to `Slice_int` **IF AND ONLY IF** the instantiated struct exists (checked via `find_struct_def_codegen`). + * **Fallback**: If the mangled struct (e.g. `Async_int`) does not exist, use the base name (`Async`). This handles opaque types like `Async<T>` correctly. + * See `emit_c_decl` in `src/codegen/codegen_utils.c` for the canonical implementation. +* **Function Signatures**: + * Use `emit_func_signature(ctx, out, node, override_name)` to handle modifiers, return types, and arguments correctly. +* **Output**: Use `fprintf(out, ...)` where `out` is the `FILE*`. + +## 5. Coding Style & Conventions +* **Formatting**: + * Indentation: 4 spaces. + * Braces: **ALWAYS** Use braces `{}` for control flow (`if`, `while`, `for`), even for single lines. + * Opening brace on the **NEXT** line (Allman style). +* **Naming**: + * Structs/Enums: `PascalCase`. + * Functions/Variables: `snake_case`. + * Macros/Constants: `SCREAMING_SNAKE_CASE`. + * Private/Static: No strict prefix, but `static` keyword is mandatory for internal file-scope functions. +* **Iterators**: + * When implementing iteration in compiler (C code): Use `while (node) { ... node = node->next; }` for linked lists (`ASTNode`, `StructRef`). + +## 6. Standard Library (Zen C Definitions) +* **Arrays**: Use `for val in arr` syntax (direct iteration). +* **Vectors**: `Vec<T>` is a dynamic array. +* **Strings**: `string` in Zen C maps to `char*` in C. `kstring` or `zstr` are higher-level wrappers. + +## 7. Common Pitfalls +* **Unused Variables**: The compiler builds with `-Wall -Werror` (or similar strictness). Cast unused vars to void: `(void)var_name;`. +## 8. Zen C Language Rules (For writing .zc files) +* **Syntax**: + * Variables: `let x = 10;`, `let y: const int = 20;`. + * Constants: `def MAX = 100;` (compile-time). + * Functions: `fn name(arg: type) -> ret { ... }`. + * Structs: `struct Point { x: int; y: int; }`. + * Enums: `enum Shape { Circle(float), Rect(float, float) }`. +* **Memory & Resources**: + * **Move Semantics**: Structs/Enums are moved by default on assignment/pass-by-value. + * **Defer**: Use `defer stmt;` to run cleanup at scope exit. + * **Drop**: Implement `impl Drop for T` for RAII. +* **Arrays & Slices**: + * **Iteration**: Use `for val in arr` (direct iteration supported). + * **Slices**: `Slice<T>` is a view. `int[N]` auto-converts to slice in loops. +* **Generics**: + * Syntax: `struct Box<T> { val: T; }`. +* **Concurrency**: + * Use `async fn` and `await` keyword. + * `Async<T>` is the opaque future type. +* **Standard Library**: + * Import with `import "std/io.zc"`, `import "std/vec.zc"`. + * Use `println "msg"` (shorthand) or `printf`. +* **Testing**: + * Use `test "name" { ... }` blocks for unit tests. @@ -188,7 +188,7 @@ clean: @echo "=> Clean complete!" # Test -test: $(TARGET) +test: $(TARGET) $(PLUGINS) ./tests/run_tests.sh ./tests/run_codegen_tests.sh ./tests/run_example_transpile.sh @@ -165,7 +165,7 @@ zc repl ### Environment Variables -You can set `ZC_ROOT` to specify the location of the Standard Library (standard imports like `import "std/vector.zc"`). This allows you to run `zc` from any directory. +You can set `ZC_ROOT` to specify the location of the Standard Library (standard imports like `import "std/vec.zc"`). This allows you to run `zc` from any directory. ```bash export ZC_ROOT=/path/to/Zen-C @@ -494,8 +494,15 @@ for i in 0..<10 { ... } // Exclusive (Explicit) for i in 0..=10 { ... } // Inclusive (0 to 10) for i in 0..10 step 2 { ... } -// Iterator (Vec, Array, or custom Iterable) -for item in collection { ... } +// Iterator (Vec or custom Iterable) +for item in vec { ... } + +// Iterate over fixed-size arrays directly +let arr: int[5] = [1, 2, 3, 4, 5]; +for val in arr { + // val is int + println "{val}"; +} // While while x < 10 { ... } @@ -607,7 +614,7 @@ Zen C supports a shorthand for prompting user input using the `?` prefix. - `? "Enter age: " (age)`: Prints prompt and scans input into the variable `age`. - Format specifiers are automatically inferred based on variable type. -```c +```zc let age: int; ? "How old are you? " (age); println "You are {age} years old."; @@ -939,6 +946,7 @@ Decorate functions and structs to modify compiler behavior. | `@host` | Fn | CUDA: Host function (`__host__`). | | `@comptime` | Fn | Helper function available for compile-time execution. | | `@derive(...)` | Struct | Auto-implement traits. Supports `Debug`, `Eq` (Smart Derive), `Copy`, `Clone`. | +| `@ctype("type")` | Fn Param | Overrides generated C type for a parameter. | | `@<custom>` | Any | Passes generic attributes to C (e.g. `@flatten`, `@alias("name")`). | #### Custom Attributes diff --git a/README_ES.md b/README_ES.md index 9040ea8..d73e9ca 100644 --- a/README_ES.md +++ b/README_ES.md @@ -493,8 +493,15 @@ for i in 0..<10 { ... } // Exclusivo (Explícito) for i in 0..=10 { ... } // Inclusivo (0 al 10) for i in 0..10 step 2 { ... } -// Iterador (Vec, Array, o Iterable personalizado) -for item in coleccion { ... } +// Iterador (Vec o Iterable personalizado) +for item in vec { ... } + +// Iterar sobre arrays de tamaño fijo directamente +let arr: int[5] = [1, 2, 3, 4, 5]; +for val in arr { + // val es int + println "{val}"; +} // While while x < 10 { ... } @@ -508,6 +515,7 @@ externo: loop { for _ in 0..5 { ... } ``` + #### Control Avanzado ```zc // Guard: Ejecuta else y retorna si la condición es falsa @@ -606,7 +614,7 @@ Zen C soporta una abreviatura para solicitar entrada al usuario usando el prefij - `? "Ingresa la edad: " (edad)`: Imprime el prompt y escanea la entrada en la variable `edad`. - Los especificadores de formato se infieren automáticamente según el tipo de variable. -```c +```zc let edad: int; ? "¿Cuántos años tienes? " (edad); println "Tienes {edad} años."; @@ -938,6 +946,7 @@ Decora funciones y structs para modificar el comportamiento del compilador. | `@host` | Fn | CUDA: Función de host (`__host__`). | | `@comptime` | Fn | Función auxiliar disponible para ejecución en tiempo de compilación. | | `@derive(...)` | Struct | Implementa traits automáticamente. Soporta `Debug`, `Eq` (Derivación Inteligente), `Copy`, `Clone`. | +| `@ctype("tipo")` | Parámetro Fn | Sobrescribe el tipo C generado para un parámetro. | | `@<custom>` | Cualquier | Pasa atributos genéricos a C (ej. `@flatten`, `@alias("nombre")`). | #### Atributos Personalizados diff --git a/README_ZH_CN.md b/README_ZH_CN.md index 217e9ec..51689f6 100644 --- a/README_ZH_CN.md +++ b/README_ZH_CN.md @@ -485,26 +485,33 @@ match opt { } ``` -#### 循环 +#### 循環 ```zc -// 区间迭代 -for i in 0..10 { ... } // 左闭右开 (0 到 9) -for i in 0..<10 { ... } // 左闭右开 (显式) -for i in 0..=10 { ... } // 全闭 (0 到 10) +// 區間迭代 +for i in 0..10 { ... } // 左閉右開 (0 到 9) +for i in 0..<10 { ... } // 左閉右開 (顯式) +for i in 0..=10 { ... } // 全閉 (0 到 10) for i in 0..10 step 2 { ... } -// 迭代器 (Vec, Array, 或自定义 Iterable) -for item in collection { ... } +// 迭代器 (Vec 或自定義 Iterable) +for item in vec { ... } -// While 循环 +// 直接迭代固定大小数组 +let arr: int[5] = [1, 2, 3, 4, 5]; +for val in arr { + // val 是 int + println "{val}"; +} + +// While 循環 while x < 10 { ... } -// 带标签的无限循环 +// 帶標籤的無限循環 outer: loop { if done { break outer; } } -// 重复 N 次 +// 重複 N 次 for _ in 0..5 { ... } ``` @@ -606,7 +613,7 @@ Zen C 支持使用 `?` 前缀进行用户输入提示的简写。 - `? "输入年龄: " (age)`: 打印提示并扫描输入到变量 `age` 中。 - 格式说明符会根据变量类型自动推断。 -```c +```zc let age: int; ? "你多大了？ " (age); println "你 {age} 岁了。"; @@ -938,6 +945,7 @@ let re = regex! { ^[a-z]+$ }; | `@host` | 函数 | CUDA: 主机函数 (`__host__`)。 | | `@comptime` | 函数 | 用于编译时执行的辅助函数。 | | `@derive(...)` | 结构体 | 自动实现 Trait。支持 `Debug`, `Eq` (智能派生), `Copy`, `Clone`。 | +| `@ctype("type")` | 函数参数 | 覆盖参数生成的 C 类型。 | | `@<custom>` | 任意 | 将泛型属性传递给 C (例如 `@flatten`, `@alias("name")`)。 | #### 自定义属性 diff --git a/README_ZH_TW.md b/README_ZH_TW.md index 8618540..6fa0dbd 100644 --- a/README_ZH_TW.md +++ b/README_ZH_TW.md @@ -493,8 +493,15 @@ for i in 0..<10 { ... } // 左閉右開 (顯式) for i in 0..=10 { ... } // 全閉 (0 到 10) for i in 0..10 step 2 { ... } -// 迭代器 (Vec, Array, 或自定義 Iterable) -for item in collection { ... } +// 迭代器 (Vec 或自定義 Iterable) +for item in vec { ... } + +// 直接迭代固定大小數組 +let arr: int[5] = [1, 2, 3, 4, 5]; +for val in arr { + // val 是 int + println "{val}"; +} // While 循環 while x < 10 { ... } @@ -606,7 +613,7 @@ Zen C 支持使用 `?` 前綴進行用戶輸入提示的簡寫。 - `? "輸入年齡: " (age)`: 打印提示並掃描輸入到變量 `age` 中。 - 格式說明符會根據變量類型自動推斷。 -```c +```zc let age: int; ? "你多大了？ " (age); println "你 {age} 歲了。"; @@ -938,6 +945,7 @@ let re = regex! { ^[a-z]+$ }; | `@host` | 函數 | CUDA: 主機函數 (`__host__`)。 | | `@comptime` | 函數 | 用於編譯時執行的輔助函數。 | | `@derive(...)` | 結構體 | 自動實現 Trait。支持 `Debug`, `Eq` (智能派生), `Copy`, `Clone`。 | +| `@ctype("type")` | 函數參數 | 覆蓋參數生成的 C 類型。 | | `@<custom>` | 任意 | 將泛型屬性傳遞給 C (例如 `@flatten`, `@alias("name")`)。 | #### 自定義屬性 diff --git a/docs/std/json.md b/docs/std/json.md index fba2ad8..ce35d64 100644 --- a/docs/std/json.md +++ b/docs/std/json.md @@ -51,6 +51,32 @@ Represents a node in a JSON document. - **`fn at(self, index: usize) -> Option<JsonValue*>`** Retrieves a value from an array by index. +#### Serialization + +- **`fn to_string(self) -> String`** + Serializes the JSON value to a string representation. + +- **`fn stringify(self, buf: String*)`** + Internal recursive serialization method that appends to a string buffer. + +**Example:** +```zc +let obj = JsonValue::object(); +obj.set("name", JsonValue::string("Alice")); +obj.set("age", JsonValue::number(30.0)); + +let json_str = obj.to_string(); +println "{json_str.c_str()}"; // {"name":"Alice","age":30} +json_str.free(); +obj.free(); +``` + +**Features:** +- Proper escaping of special characters: `\"`, `\\`, `\n`, `\t`, `\r`, `\b`, `\f` +- Numbers formatted with `%.15g` for precision +- Recursive serialization for nested objects and arrays +- Round-trip compatible with `parse()` + #### Memory Management - **`fn free(self)`** diff --git a/docs/std/slice.md b/docs/std/slice.md new file mode 100644 index 0000000..b70c5fe --- /dev/null +++ b/docs/std/slice.md @@ -0,0 +1,90 @@ +# Standard Library: Slice (`std/slice.zc`) + +`Slice<T>` is a lightweight, non-owning view into a contiguous sequence of elements. It's particularly useful for working with fixed-size arrays and enabling iteration. + +## Usage + +```zc +import "std/slice.zc" + +fn main() { + let arr: int[5] = [1, 2, 3, 4, 5]; + + // Direct iteration (Recommended) + for val in arr { + println "{val}"; + } + + // Manual slice creation (for partial views or specific needs) + let slice = Slice<int>::from_array((int*)(&arr), 5); + for val in slice { + println "{val}"; + } +} +``` + +## Structure + +```zc +struct Slice<T> { + data: T*; + len: usize; +} +``` + +## Methods + +### Construction + +| Method | Signature | Description | +| :--- | :--- | :--- | +| **from_array** | `Slice<T>::from_array(arr: T*, len: usize) -> Slice<T>` | Creates a slice view over an array. | + +### Iteration + +| Method | Signature | Description | +| :--- | :--- | :--- | +| **iterator** | `iterator(self) -> SliceIter<T>` | Returns an iterator for `for-in` loops. | + +`SliceIter<T>` implements the iterator protocol with a `next() -> Option<T>` method. + +### Access & Query + +| Method | Signature | Description | +| :--- | :--- | :--- | +| **length** | `length(self) -> usize` | Returns the number of elements. | +| **is_empty** | `is_empty(self) -> bool` | Returns true if length is 0. | +| **get** | `get(self, idx: usize) -> Option<T>` | Returns the element at index, or None if out of bounds. | +| **at** | `at(self, idx: usize) -> Option<T>` | Alias for `get`. | + +## Examples + +### Iterating over fixed-size arrays + +```zc +let numbers: int[3] = [10, 20, 30]; +let slice = Slice<int>::from_array((int*)(&numbers), 3); + +for n in slice { + println "Number: {n}"; +} +``` + +### Safe indexed access + +```zc +let arr: int[3] = [1, 2, 3]; +let slice = Slice<int>::from_array((int*)(&arr), 3); + +let opt = slice.get(1); +if (!opt.is_none()) { + println "Value: {opt.unwrap()}"; // Prints: Value: 2 +} +``` + +## Notes + +- `Slice<T>` does not own its data - it's just a view +- No memory management needed (no `free()` method) +- Must specify the generic type explicitly: `Slice<int>`, `Slice<String>`, etc. +- The array pointer cast `(T*)(&arr)` is required for fixed-size arrays diff --git a/docs/std/string.md b/docs/std/string.md index 1f89e0f..a2f63f5 100644 --- a/docs/std/string.md +++ b/docs/std/string.md @@ -49,8 +49,12 @@ struct String { | Method | Signature | Description | | :--- | :--- | :--- | | **append** | `append(self, other: String*)` | Appends another string to this one. | +| **append_c** | `append_c(self, s: char*)` | Appends a C string literal. Uses value receiver. | +| **append_c_ptr** | `append_c_ptr(ptr: String*, s: char*)` | Appends a C string literal using pointer receiver for guaranteed mutation. | | **add** | `add(self, other: String*) -> String` | Concatenates this string and another into a new String. | +**Note:** When passing `String*` to functions that need to mutate, use `append_c_ptr` instead of `append_c` for reliable mutation. + ### Access & Query | Method | Signature | Description | diff --git a/src/ast/ast.h b/src/ast/ast.h index 71d9943..4498d7c 100644 --- a/src/ast/ast.h +++ b/src/ast/ast.h @@ -229,6 +229,8 @@ struct ASTNode int cuda_device; // @device -> __device__ int cuda_host; // @host -> __host__ + char **c_type_overrides; // @ctype("...") per parameter + Attribute *attributes; // Custom attributes } func; diff --git a/src/codegen/codegen.c b/src/codegen/codegen.c index a66f179..7a67428 100644 --- a/src/codegen/codegen.c +++ b/src/codegen/codegen.c @@ -65,6 +65,52 @@ static void codegen_var_expr(ParserContext *ctx, ASTNode *node, FILE *out) zwarn_at(node->token, "%s\n = help: %s", msg, help); } } + + // Check for static method call pattern: Type::method or Slice<T>::method + char *double_colon = strstr(node->var_ref.name, "::"); + if (double_colon) + { + // Extract type name and method name + int type_len = double_colon - node->var_ref.name; + char *type_name = xmalloc(type_len + 1); + strncpy(type_name, node->var_ref.name, type_len); + type_name[type_len] = 0; + + char *method_name = double_colon + 2; // Skip :: + + // Handle generic types: Slice<int> -> Slice_int + char mangled_type[256]; + if (strchr(type_name, '<')) + { + // Generic type - need to mangle it + char *lt = strchr(type_name, '<'); + char *gt = strchr(type_name, '>'); + + if (lt && gt) + { + // Extract base type and type argument + *lt = 0; + char *type_arg = lt + 1; + *gt = 0; + + sprintf(mangled_type, "%s_%s", type_name, type_arg); + } + else + { + strcpy(mangled_type, type_name); + } + } + else + { + strcpy(mangled_type, type_name); + } + + // Output as Type__method + fprintf(out, "%s__%s", mangled_type, method_name); + free(type_name); + return; + } + fprintf(out, "%s", node->var_ref.name); } @@ -348,6 +394,7 @@ void codegen_expression(ParserContext *ctx, ASTNode *node, FILE *out) } // Check for Static Enum Variant Call: Enum.Variant(...) + if (target->type == NODE_EXPR_VAR) { ASTNode *def = find_struct_def(ctx, target->var_ref.name); @@ -418,11 +465,43 @@ void codegen_expression(ParserContext *ctx, ASTNode *node, FILE *out) base += 7; } + char *mangled_base = base; + char base_buf[256]; + + // Mangle generic types: Slice<int> -> Slice_int, Vec<Point> -> Vec_Point + char *lt = strchr(base, '<'); + if (lt) + { + char *gt = strchr(lt, '>'); + if (gt) + { + int prefix_len = lt - base; + int arg_len = gt - lt - 1; + snprintf(base_buf, 255, "%.*s_%.*s", prefix_len, base, arg_len, lt + 1); + mangled_base = base_buf; + } + } + if (!strchr(type, '*') && target->type == NODE_EXPR_CALL) { - fprintf(out, "({ %s _t = ", type); + char *type_mangled = type; + char type_buf[256]; + char *t_lt = strchr(type, '<'); + if (t_lt) + { + char *t_gt = strchr(t_lt, '>'); + if (t_gt) + { + int p_len = t_lt - type; + int a_len = t_gt - t_lt - 1; + snprintf(type_buf, 255, "%.*s_%.*s", p_len, type, a_len, t_lt + 1); + type_mangled = type_buf; + } + } + + fprintf(out, "({ %s _t = ", type_mangled); codegen_expression(ctx, target, out); - fprintf(out, "; %s__%s(&_t", base, method); + fprintf(out, "; %s__%s(&_t", mangled_base, method); ASTNode *arg = node->call.args; while (arg) { @@ -435,10 +514,11 @@ void codegen_expression(ParserContext *ctx, ASTNode *node, FILE *out) else { // Mixin Lookup Logic - char *call_base = base; + char *call_base = mangled_base; + int need_cast = 0; char mixin_func_name[128]; - sprintf(mixin_func_name, "%s__%s", base, method); + sprintf(mixin_func_name, "%s__%s", call_base, method); char *resolved_method_suffix = NULL; diff --git a/src/codegen/codegen.h b/src/codegen/codegen.h index b3e971d..89614f7 100644 --- a/src/codegen/codegen.h +++ b/src/codegen/codegen.h @@ -48,7 +48,7 @@ char *replace_string_type(const char *args); const char *parse_original_method_name(const char *mangled); void emit_auto_type(ParserContext *ctx, ASTNode *init_expr, Token t, FILE *out); char *codegen_type_to_string(Type *t); -void emit_func_signature(FILE *out, ASTNode *func, const char *name_override); +void emit_func_signature(ParserContext *ctx, FILE *out, ASTNode *func, const char *name_override); char *strip_template_suffix(const char *name); int emit_move_invalidation(ParserContext *ctx, ASTNode *node, FILE *out); void codegen_expression_with_move(ParserContext *ctx, ASTNode *node, FILE *out); @@ -66,7 +66,7 @@ void emit_trait_defs(ASTNode *node, FILE *out); void emit_enum_protos(ASTNode *node, FILE *out); void emit_globals(ParserContext *ctx, ASTNode *node, FILE *out); void emit_lambda_defs(ParserContext *ctx, FILE *out); -void emit_protos(ASTNode *node, FILE *out); +void emit_protos(ParserContext *ctx, ASTNode *node, FILE *out); void emit_impl_vtables(ParserContext *ctx, FILE *out); /** diff --git a/src/codegen/codegen_decl.c b/src/codegen/codegen_decl.c index 0b78676..9d23617 100644 --- a/src/codegen/codegen_decl.c +++ b/src/codegen/codegen_decl.c @@ -698,7 +698,7 @@ void emit_globals(ParserContext *ctx, ASTNode *node, FILE *out) } // Emit function prototypes -void emit_protos(ASTNode *node, FILE *out) +void emit_protos(ParserContext *ctx, ASTNode *node, FILE *out) { ASTNode *f = node; while (f) @@ -721,7 +721,7 @@ void emit_protos(ASTNode *node, FILE *out) } else { - emit_func_signature(out, f, NULL); + emit_func_signature(ctx, out, f, NULL); fprintf(out, ";\n"); } } @@ -799,7 +799,7 @@ void emit_protos(ASTNode *node, FILE *out) } else { - emit_func_signature(out, m, proto); + emit_func_signature(ctx, out, m, proto); fprintf(out, ";\n"); } diff --git a/src/codegen/codegen_main.c b/src/codegen/codegen_main.c index a140070..b298700 100644 --- a/src/codegen/codegen_main.c +++ b/src/codegen/codegen_main.c @@ -616,7 +616,7 @@ void codegen_node(ParserContext *ctx, ASTNode *node, FILE *out) } } - emit_protos(merged_funcs, out); + emit_protos(ctx, merged_funcs, out); emit_impl_vtables(ctx, out); diff --git a/src/codegen/codegen_stmt.c b/src/codegen/codegen_stmt.c index 2f9a2ba..7828ecf 100644 --- a/src/codegen/codegen_stmt.c +++ b/src/codegen/codegen_stmt.c @@ -750,7 +750,7 @@ void codegen_node_single(ParserContext *ctx, ASTNode *node, FILE *out) { fprintf(out, "inline "); } - emit_func_signature(out, node, NULL); + emit_func_signature(ctx, out, node, NULL); fprintf(out, "\n"); fprintf(out, "{\n"); char *prev_ret = g_current_func_ret_type; diff --git a/src/codegen/codegen_utils.c b/src/codegen/codegen_utils.c index 391ebd3..08707cc 100644 --- a/src/codegen/codegen_utils.c +++ b/src/codegen/codegen_utils.c @@ -41,7 +41,7 @@ char *strip_template_suffix(const char *name) } // Helper to emit C declaration (handle arrays, function pointers correctly) -void emit_c_decl(FILE *out, const char *type_str, const char *name) +void emit_c_decl(ParserContext *ctx, FILE *out, const char *type_str, const char *name) { char *bracket = strchr(type_str, '['); char *generic = strchr(type_str, '<'); @@ -64,9 +64,34 @@ void emit_c_decl(FILE *out, const char *type_str, const char *name) } else if (generic && (!bracket || generic < bracket)) { - // Strip generic part for C output - int base_len = generic - type_str; - fprintf(out, "%.*s %s", base_len, type_str, name); + char mangled_candidate[256]; + char *gt = strchr(generic, '>'); + int success = 0; + + if (gt) + { + int base_len = generic - type_str; + int arg_len = gt - generic - 1; + + // Limit check + if (base_len + arg_len + 2 < 256) + { + snprintf(mangled_candidate, 256, "%.*s_%.*s", base_len, type_str, arg_len, + generic + 1); + + if (find_struct_def_codegen(ctx, mangled_candidate)) + { + fprintf(out, "%s %s", mangled_candidate, name); + success = 1; + } + } + } + + if (!success) + { + int base_len = generic - type_str; + fprintf(out, "%.*s %s", base_len, type_str, name); + } if (bracket) { @@ -87,8 +112,7 @@ void emit_c_decl(FILE *out, const char *type_str, const char *name) // Helper to emit variable declarations with array types. void emit_var_decl_type(ParserContext *ctx, FILE *out, const char *type_str, const char *var_name) { - (void)ctx; - emit_c_decl(out, type_str, var_name); + emit_c_decl(ctx, out, type_str, var_name); } // Find struct definition @@ -644,7 +668,7 @@ char *codegen_type_to_string(Type *t) } // Emit function signature using Type info for correct C codegen -void emit_func_signature(FILE *out, ASTNode *func, const char *name_override) +void emit_func_signature(ParserContext *ctx, FILE *out, ASTNode *func, const char *name_override) { if (!func || func->type != NODE_FUNCTION) { @@ -714,7 +738,12 @@ void emit_func_signature(FILE *out, ASTNode *func, const char *name_override) } char *type_str = NULL; - if (func->func.arg_types && func->func.arg_types[i]) + // Check for @ctype override first + if (func->func.c_type_overrides && func->func.c_type_overrides[i]) + { + type_str = xstrdup(func->func.c_type_overrides[i]); + } + else if (func->func.arg_types && func->func.arg_types[i]) { type_str = codegen_type_to_string(func->func.arg_types[i]); } @@ -724,13 +753,14 @@ void emit_func_signature(FILE *out, ASTNode *func, const char *name_override) } const char *name = ""; + if (func->func.param_names && func->func.param_names[i]) { name = func->func.param_names[i]; } // check if array type - emit_c_decl(out, type_str, name); + emit_c_decl(ctx, out, type_str, name); free(type_str); } if (func->func.is_varargs) diff --git a/src/parser/parser.h b/src/parser/parser.h index 262c359..23c2920 100644 --- a/src/parser/parser.h +++ b/src/parser/parser.h @@ -561,7 +561,8 @@ ASTNode *parse_arrow_lambda_multi(ParserContext *ctx, Lexer *l, char **param_nam * @brief Parses and converts arguments. */ char *parse_and_convert_args(ParserContext *ctx, Lexer *l, char ***defaults_out, int *count_out, - Type ***types_out, char ***names_out, int *is_varargs_out); + Type ***types_out, char ***names_out, int *is_varargs_out, + char ***ctype_overrides_out); /** * @brief Checks if a file has been imported. diff --git a/src/parser/parser_decl.c b/src/parser/parser_decl.c index c96ca36..93a124d 100644 --- a/src/parser/parser_decl.c +++ b/src/parser/parser_decl.c @@ -99,10 +99,11 @@ ASTNode *parse_function(ParserContext *ctx, Lexer *l, int is_async) int count; Type **arg_types; char **param_names; + char **ctype_overrides; int is_varargs = 0; - char *args = - parse_and_convert_args(ctx, l, &defaults, &count, &arg_types, ¶m_names, &is_varargs); + char *args = parse_and_convert_args(ctx, l, &defaults, &count, &arg_types, ¶m_names, + &is_varargs, &ctype_overrides); char *ret = "void"; Type *ret_type_obj = type_new(TYPE_VOID); @@ -191,6 +192,7 @@ ASTNode *parse_function(ParserContext *ctx, Lexer *l, int is_async) node->func.defaults = defaults; node->func.ret_type_info = ret_type_obj; node->func.is_varargs = is_varargs; + node->func.c_type_overrides = ctype_overrides; if (gen_param) { diff --git a/src/parser/parser_expr.c b/src/parser/parser_expr.c index 6156cc0..28dc465 100644 --- a/src/parser/parser_expr.c +++ b/src/parser/parser_expr.c @@ -5421,6 +5421,7 @@ ASTNode *parse_expr_prec(ParserContext *ctx, Lexer *l, Precedence min_prec) // This gives a warning as "unused" but it's needed for the rewrite. char *r_name = resolve_struct_name_from_type(ctx, rhs->type_info, &is_rhs_ptr, &r_alloc); + (void)r_name; if (r_alloc) { free(r_alloc); diff --git a/src/parser/parser_stmt.c b/src/parser/parser_stmt.c index a471fe6..7758ae3 100644 --- a/src/parser/parser_stmt.c +++ b/src/parser/parser_stmt.c @@ -1133,6 +1133,7 @@ ASTNode *parse_for(ParserContext *ctx, Lexer *l) ASTNode *obj_expr = start_expr; char *iter_method = "iterator"; + ASTNode *slice_decl = NULL; // Track if we need to add a slice declaration // Check for reference iteration: for x in &vec if (obj_expr->type == NODE_EXPR_UNARY && obj_expr->unary.op && @@ -1142,6 +1143,78 @@ ASTNode *parse_for(ParserContext *ctx, Lexer *l) iter_method = "iter_ref"; } + // Check for array iteration: wrap with Slice<T>::from_array + if (obj_expr->type_info && obj_expr->type_info->kind == TYPE_ARRAY && + obj_expr->type_info->array_size > 0) + { + // Create a var decl for the slice + slice_decl = ast_create(NODE_VAR_DECL); + slice_decl->var_decl.name = xstrdup("__zc_arr_slice"); + + // Build type string: Slice<elem_type> + char *elem_type_str = type_to_string(obj_expr->type_info->inner); + char slice_type[256]; + sprintf(slice_type, "Slice<%s>", elem_type_str); + slice_decl->var_decl.type_str = xstrdup(slice_type); + + ASTNode *from_array_call = ast_create(NODE_EXPR_CALL); + ASTNode *static_method = ast_create(NODE_EXPR_VAR); + + // The function name for static methods is Type::method format + char func_name[512]; + snprintf(func_name, 511, "%s::from_array", slice_type); + static_method->var_ref.name = xstrdup(func_name); + + from_array_call->call.callee = static_method; + + // Create arguments + ASTNode *arr_addr = ast_create(NODE_EXPR_UNARY); + arr_addr->unary.op = xstrdup("&"); + arr_addr->unary.operand = obj_expr; + + ASTNode *arr_cast = ast_create(NODE_EXPR_CAST); + char cast_type[256]; + sprintf(cast_type, "%s*", elem_type_str); + arr_cast->cast.target_type = xstrdup(cast_type); + arr_cast->cast.expr = arr_addr; + + ASTNode *size_arg = ast_create(NODE_EXPR_LITERAL); + size_arg->literal.type_kind = LITERAL_INT; + size_arg->literal.int_val = obj_expr->type_info->array_size; + char size_buf[32]; + sprintf(size_buf, "%d", obj_expr->type_info->array_size); + size_arg->literal.string_val = xstrdup(size_buf); + + arr_cast->next = size_arg; + from_array_call->call.args = arr_cast; + from_array_call->call.arg_count = 2; + + slice_decl->var_decl.init_expr = from_array_call; + + // Manually trigger generic instantiation for Slice<T> + // This ensures that Slice_int, Slice_float, etc. structures are generated + Token dummy_tok = {0}; + instantiate_generic(ctx, "Slice", elem_type_str, elem_type_str, dummy_tok); + + // Instantiate SliceIter and Option too for the loop logic + char iter_type[256]; + sprintf(iter_type, "SliceIter<%s>", elem_type_str); + instantiate_generic(ctx, "SliceIter", elem_type_str, elem_type_str, dummy_tok); + + char option_type[256]; + sprintf(option_type, "Option<%s>", elem_type_str); + instantiate_generic(ctx, "Option", elem_type_str, elem_type_str, dummy_tok); + + // Replace obj_expr with a reference to the slice variable + ASTNode *slice_ref = ast_create(NODE_EXPR_VAR); + slice_ref->var_ref.name = xstrdup("__zc_arr_slice"); + slice_ref->resolved_type = + xstrdup(slice_type); // Explicitly set type for codegen + obj_expr = slice_ref; + + free(elem_type_str); + } + // var __it = obj.iterator(); ASTNode *it_decl = ast_create(NODE_VAR_DECL); it_decl->var_decl.name = xstrdup("__it"); @@ -1182,6 +1255,34 @@ ASTNode *parse_for(ParserContext *ctx, Lexer *l) stmts_tail = node; \ } + char *iter_type_ptr = NULL; + char *option_type_ptr = NULL; + + if (slice_decl) + { + char *slice_t = slice_decl->var_decl.type_str; + char *start = strchr(slice_t, '<'); + if (start) + { + char *end = strrchr(slice_t, '>'); + if (end) + { + int len = end - start - 1; + char *elem = xmalloc(len + 1); + strncpy(elem, start + 1, len); + elem[len] = 0; + + iter_type_ptr = xmalloc(256); + sprintf(iter_type_ptr, "SliceIter<%s>", elem); + + option_type_ptr = xmalloc(256); + sprintf(option_type_ptr, "Option<%s>", elem); + + free(elem); + } + } + } + // var __opt = __it.next(); ASTNode *opt_decl = ast_create(NODE_VAR_DECL); opt_decl->var_decl.name = xstrdup("__opt"); @@ -1192,6 +1293,10 @@ ASTNode *parse_for(ParserContext *ctx, Lexer *l) ASTNode *memb_next = ast_create(NODE_EXPR_MEMBER); ASTNode *it_ref = ast_create(NODE_EXPR_VAR); it_ref->var_ref.name = xstrdup("__it"); + if (iter_type_ptr) + { + it_ref->resolved_type = xstrdup(iter_type_ptr); + } memb_next->member.target = it_ref; memb_next->member.field = xstrdup("next"); call_next->call.callee = memb_next; @@ -1204,15 +1309,22 @@ ASTNode *parse_for(ParserContext *ctx, Lexer *l) ASTNode *memb_is_none = ast_create(NODE_EXPR_MEMBER); ASTNode *opt_ref1 = ast_create(NODE_EXPR_VAR); opt_ref1->var_ref.name = xstrdup("__opt"); + if (option_type_ptr) + { + opt_ref1->resolved_type = xstrdup(option_type_ptr); + } memb_is_none->member.target = opt_ref1; memb_is_none->member.field = xstrdup("is_none"); call_is_none->call.callee = memb_is_none; + call_is_none->call.args = NULL; + call_is_none->call.arg_count = 0; - ASTNode *break_stmt = ast_create(NODE_BREAK); - + // if (__opt.is_none()) break; ASTNode *if_break = ast_create(NODE_IF); if_break->if_stmt.condition = call_is_none; + ASTNode *break_stmt = ast_create(NODE_BREAK); if_break->if_stmt.then_body = break_stmt; + if_break->if_stmt.else_body = NULL; APPEND_STMT(if_break); // var <user_var> = __opt.unwrap(); @@ -1225,25 +1337,28 @@ ASTNode *parse_for(ParserContext *ctx, Lexer *l) ASTNode *memb_unwrap = ast_create(NODE_EXPR_MEMBER); ASTNode *opt_ref2 = ast_create(NODE_EXPR_VAR); opt_ref2->var_ref.name = xstrdup("__opt"); + if (option_type_ptr) + { + opt_ref2->resolved_type = xstrdup(option_type_ptr); + } memb_unwrap->member.target = opt_ref2; memb_unwrap->member.field = xstrdup("unwrap"); call_unwrap->call.callee = memb_unwrap; + call_unwrap->call.args = NULL; + call_unwrap->call.arg_count = 0; user_var_decl->var_decl.init_expr = call_unwrap; APPEND_STMT(user_var_decl); - // User Body + // User body statements enter_scope(ctx); add_symbol(ctx, var_name, NULL, NULL); - ASTNode *user_body_node; - if (lexer_peek(l).type == TOK_LBRACE) - { - user_body_node = parse_block(ctx, l); - } - else + // Body block + ASTNode *stmt = parse_statement(ctx, l); + ASTNode *user_body_node = stmt; + if (stmt && stmt->type != NODE_BLOCK) { - ASTNode *stmt = parse_statement(ctx, l); ASTNode *blk = ast_create(NODE_BLOCK); blk->block.statements = stmt; user_body_node = blk; @@ -1256,10 +1371,21 @@ ASTNode *parse_for(ParserContext *ctx, Lexer *l) loop_body->block.statements = stmts_head; while_loop->while_stmt.body = loop_body; - // Wrap entire thing in a block to scope _it + // Wrap entire thing in a block to scope __it (and __zc_arr_slice if present) ASTNode *outer_block = ast_create(NODE_BLOCK); - it_decl->next = while_loop; - outer_block->block.statements = it_decl; + if (slice_decl) + { + // Chain: slice_decl -> it_decl -> while_loop + slice_decl->next = it_decl; + it_decl->next = while_loop; + outer_block->block.statements = slice_decl; + } + else + { + // Chain: it_decl -> while_loop + it_decl->next = while_loop; + outer_block->block.statements = it_decl; + } return outer_block; } @@ -3179,7 +3305,7 @@ char *run_comptime_block(ParserContext *ctx, Lexer *l) ASTNode *fn = ref->node; if (fn && fn->type == NODE_FUNCTION && fn->func.is_comptime) { - emit_func_signature(f, fn, NULL); + emit_func_signature(ctx, f, fn, NULL); fprintf(f, ";\n"); codegen_node_single(ctx, fn, f); } diff --git a/src/parser/parser_struct.c b/src/parser/parser_struct.c index 82dd346..109eeee 100644 --- a/src/parser/parser_struct.c +++ b/src/parser/parser_struct.c @@ -100,7 +100,7 @@ ASTNode *parse_trait(ParserContext *ctx, Lexer *l) char **param_names = NULL; int is_varargs = 0; char *args = parse_and_convert_args(ctx, l, &defaults, &arg_count, &arg_types, ¶m_names, - &is_varargs); + &is_varargs, NULL); char *ret = xstrdup("void"); if (lexer_peek(l).type == TOK_ARROW) diff --git a/src/parser/parser_utils.c b/src/parser/parser_utils.c index 48418b6..28d2c11 100644 --- a/src/parser/parser_utils.c +++ b/src/parser/parser_utils.c @@ -3392,7 +3392,8 @@ char *consume_and_rewrite(ParserContext *ctx, Lexer *l) } char *parse_and_convert_args(ParserContext *ctx, Lexer *l, char ***defaults_out, int *count_out, - Type ***types_out, char ***names_out, int *is_varargs_out) + Type ***types_out, char ***names_out, int *is_varargs_out, + char ***ctype_overrides_out) { Token t = lexer_next(l); if (t.type != TOK_LPAREN) @@ -3406,18 +3407,52 @@ char *parse_and_convert_args(ParserContext *ctx, Lexer *l, char ***defaults_out, char **defaults = xmalloc(sizeof(char *) * 16); Type **types = xmalloc(sizeof(Type *) * 16); char **names = xmalloc(sizeof(char *) * 16); + char **ctype_overrides = xmalloc(sizeof(char *) * 16); for (int i = 0; i < 16; i++) { defaults[i] = NULL; types[i] = NULL; names[i] = NULL; + ctype_overrides[i] = NULL; } if (lexer_peek(l).type != TOK_RPAREN) { while (1) { + // Check for @ctype("...") before parameter + char *ctype_override = NULL; + if (lexer_peek(l).type == TOK_AT) + { + lexer_next(l); // eat @ + Token attr = lexer_next(l); + if (attr.type == TOK_IDENT && attr.len == 5 && strncmp(attr.start, "ctype", 5) == 0) + { + if (lexer_next(l).type != TOK_LPAREN) + { + zpanic_at(lexer_peek(l), "Expected ( after @ctype"); + } + Token ctype_tok = lexer_next(l); + if (ctype_tok.type != TOK_STRING) + { + zpanic_at(ctype_tok, "@ctype requires a string argument"); + } + // Extract string content (strip quotes) + ctype_override = xmalloc(ctype_tok.len - 1); + strncpy(ctype_override, ctype_tok.start + 1, ctype_tok.len - 2); + ctype_override[ctype_tok.len - 2] = 0; + if (lexer_next(l).type != TOK_RPAREN) + { + zpanic_at(lexer_peek(l), "Expected ) after @ctype string"); + } + } + else + { + zpanic_at(attr, "Unknown parameter attribute @%.*s", attr.len, attr.start); + } + } + Token t = lexer_next(l); // Handle 'self' if (t.type == TOK_IDENT && strncmp(t.start, "self", 4) == 0 && t.len == 4) @@ -3470,6 +3505,7 @@ char *parse_and_convert_args(ParserContext *ctx, Lexer *l, char ***defaults_out, types[count] = type_new_ptr(type_new(TYPE_VOID)); add_symbol(ctx, "self", "void*", types[count]); } + ctype_overrides[count] = ctype_override; count++; } else @@ -3514,11 +3550,20 @@ char *parse_and_convert_args(ParserContext *ctx, Lexer *l, char ***defaults_out, } else { - strcat(buf, type_str); + // Use @ctype override if present + if (ctype_override) + { + strcat(buf, ctype_override); + } + else + { + strcat(buf, type_str); + } strcat(buf, " "); strcat(buf, name); } + ctype_overrides[count] = ctype_override; count++; if (lexer_peek(l).type == TOK_OP && is_token(lexer_peek(l), "=")) @@ -3593,6 +3638,10 @@ char *parse_and_convert_args(ParserContext *ctx, Lexer *l, char ***defaults_out, *count_out = count; *types_out = types; *names_out = names; + if (ctype_overrides_out) + { + *ctype_overrides_out = ctype_overrides; + } return buf; } diff --git a/std/core.zc b/std/core.zc index f450517..7379db4 100644 --- a/std/core.zc +++ b/std/core.zc @@ -7,11 +7,11 @@ include <stdarg.h> let __zen_hash_seed: usize = 14695981039346656037; -raw { -void _zen_panic(const char* file, int line, const char* func, const char* msg) { +extern fn exit(code: int); + +fn _zen_panic(file: const char*, line: int, func: const char*, msg: const char*) { fprintf(stderr, "%s:%d (%s): Panic: %s\n", file, line, func, msg); exit(1); } -} #define panic(msg) _zen_panic(__FILE__, __LINE__, __func__, msg)
\ No newline at end of file diff --git a/std/cuda.zc b/std/cuda.zc index c6a9403..8fc6545 100644 --- a/std/cuda.zc +++ b/std/cuda.zc @@ -101,6 +101,8 @@ fn cuda_ok() -> bool { } +// Minimal raw block: required for cudaDeviceProp struct field access +// The cudaDeviceProp struct cannot be declared in Zen-C without type conflicts raw { void _z_cuda_get_props(int dev, char* name, size_t* total_mem, int* sm_count, int* major, int* minor, int* max_threads, int* warp_size) { struct cudaDeviceProp prop; @@ -2,23 +2,12 @@ import "./core.zc" import "./option.zc" import "./string.zc" -raw { - char *_z_env_get(char *name) { - return getenv(name); - } - - int _z_env_set(char *name, char *value, int overwrite) { - return setenv(name, value, overwrite); - } - - int _z_env_unset(char *name) { - return unsetenv(name); - } -} +include <stdlib.h> -extern fn _z_env_get(name: char*) -> char*; -extern fn _z_env_set(name: char*, value: char*, overwrite: int) -> int; -extern fn _z_env_unset(name: char*) -> int; +// Direct externs with const char* to match C stdlib declarations +extern fn getenv(name: const char*) -> char*; +extern fn setenv(name: const char*, value: const char*, overwrite: int) -> int; +extern fn unsetenv(name: const char*) -> int; @derive(Eq) enum EnvRes { @@ -30,7 +19,7 @@ struct Env {} impl Env { fn get(name: string) -> Option<string> { - let value: string = _z_env_get(name); + let value: string = getenv(name); if (value == NULL) { return Option<string>::None(); } @@ -39,7 +28,7 @@ impl Env { } fn get_dup(name: string) -> Option<String> { - let value: string = _z_env_get(name); + let value: string = getenv(name); if (value == NULL) { return Option<String>::None(); } @@ -52,13 +41,13 @@ impl Env { } fn set(name: string, value: string) -> EnvRes { - let ret: int = _z_env_set(name, value, 1); + let ret: int = setenv(name, value, 1); return (ret == 0) ? EnvRes::OK() : EnvRes::ERR(); } fn unset(name: string) -> EnvRes { - let ret: int = _z_env_unset(name); + let ret: int = unsetenv(name); return (ret == 0) ? EnvRes::OK() : EnvRes::ERR(); } @@ -14,12 +14,20 @@ include <unistd.h> include <stdlib.h> include <stdio.h> -// TODO: restructure this tomorrow (lol). +// Direct externs for simple functions with const char* parameters +extern fn access(pathname: const char*, mode: int) -> int; +extern fn unlink(pathname: const char*) -> int; +extern fn rmdir(pathname: const char*) -> int; +extern fn malloc(size: usize) -> void*; +extern fn free(ptr: void*); + +// Minimal raw block: required for opaque FILE*/DIR* types and C struct access +// These cannot be expressed in Zen-C extern declarations without type conflicts raw { typedef struct DirEntry* DirEntryPtr; - // Wrappers for FILE* handling due to opaque pointer casting - void* _z_fs_fopen(char* path, char* mode) { + // FILE* wrappers - fopen/fclose/etc use FILE* which conflicts with void* + void* _z_fs_fopen(const char* path, const char* mode) { return fopen(path, mode); } @@ -43,22 +51,8 @@ raw { return (int64_t)ftell((FILE*)stream); } - // Wrappers needed because C headers declare these with 'const char*' - // but Zen C externs generate 'char*', leading to conflicting types. - int _z_fs_access(char* pathname, int mode) { - return access(pathname, mode); - } - - int _z_fs_unlink(char* pathname) { - return unlink(pathname); - } - - int _z_fs_rmdir(char* pathname) { - return rmdir(pathname); - } - - // Wrappers for DIR* handling - void* _z_fs_opendir(char* name) { + // DIR* wrappers - opendir/closedir/readdir use DIR* which conflicts with void* + void* _z_fs_opendir(const char* name) { return opendir(name); } @@ -66,8 +60,8 @@ raw { return closedir((DIR*)dir); } - // struct stat / struct dirent helpers - int _z_fs_get_metadata(char* path, uint64_t* size, int* is_dir, int* is_file) { + // struct stat access - cannot define matching Zen-C struct for stat + int _z_fs_get_metadata(const char* path, uint64_t* size, int* is_dir, int* is_file) { struct stat st; if (stat(path, &st) != 0) return -1; *size = st.st_size; @@ -76,6 +70,7 @@ raw { return 0; } + // struct dirent access - readdir returns struct dirent* int _z_fs_read_entry(void* dir, char* out_name, int buf_size, int* is_dir) { struct dirent* ent = readdir((DIR*)dir); if (!ent) return 0; @@ -85,7 +80,8 @@ raw { return 1; } - int _z_fs_mkdir(char* path) { + // mkdir has different signatures on Windows vs POSIX + int _z_fs_mkdir(const char* path) { #ifdef _WIN32 return mkdir(path); #else @@ -94,26 +90,18 @@ raw { } } -// Direct externs -extern fn malloc(size: usize) -> void*; -extern fn free(ptr: void*); - -extern fn _z_fs_mkdir(path: char*) -> int; -extern fn _z_fs_get_metadata(path: char*, size: U64*, is_dir: int*, is_file: int*) -> int; +extern fn _z_fs_mkdir(path: const char*) -> int; +extern fn _z_fs_get_metadata(path: const char*, size: U64*, is_dir: int*, is_file: int*) -> int; extern fn _z_fs_read_entry(dir: void*, out_name: char*, buf_size: int, is_dir: int*) -> int; -extern fn _z_fs_fopen(path: char*, mode: char*) -> void*; +extern fn _z_fs_fopen(path: const char*, mode: const char*) -> void*; extern fn _z_fs_fclose(stream: void*) -> int; extern fn _z_fs_fread(ptr: void*, size: usize, nmemb: usize, stream: void*) -> usize; extern fn _z_fs_fwrite(ptr: void*, size: usize, nmemb: usize, stream: void*) -> usize; extern fn _z_fs_fseek(stream: void*, offset: I64, whence: int) -> int; extern fn _z_fs_ftell(stream: void*) -> I64; -extern fn _z_fs_opendir(name: char*) -> void*; +extern fn _z_fs_opendir(name: const char*) -> void*; extern fn _z_fs_closedir(dir: void*) -> int; -extern fn _z_fs_access(pathname: char*, mode: int) -> int; -extern fn _z_fs_unlink(pathname: char*) -> int; -extern fn _z_fs_rmdir(pathname: char*) -> int; - struct File { handle: void*; @@ -203,7 +191,7 @@ impl File { } fn exists(path: char*) -> bool { - return _z_fs_access(path, Z_F_OK) == 0; + return access(path, Z_F_OK) == 0; } fn metadata(path: char*) -> Result<Metadata> { @@ -230,14 +218,14 @@ impl File { } fn remove_file(path: char*) -> Result<bool> { - if (_z_fs_unlink(path) != 0) { + if (unlink(path) != 0) { return Result<bool>::Err("Failed to remove file"); } return Result<bool>::Ok(true); } fn remove_dir(path: char*) -> Result<bool> { - if (_z_fs_rmdir(path) != 0) { + if (rmdir(path) != 0) { return Result<bool>::Err("Failed to remove directory"); } return Result<bool>::Ok(true); @@ -2,37 +2,51 @@ import "./core.zc" import "./string.zc" -raw { - int _z_vprintf(char* fmt, va_list ap) { - return vprintf((const char*)fmt, ap); - } - - int _z_vsnprintf(char* str, size_t size, char* fmt, va_list ap) { - return vsnprintf(str, size, (const char*)fmt, ap); - } +include <stdio.h> +include <stdarg.h> - void* _z_get_stdin() { return stdin; } - int _z_get_eof() { return EOF; } +// These work directly with const char* in extern declarations +extern fn vprintf(fmt: const char*, ap: va_list) -> int; + +// vsnprintf is problematic on macOS because it's a macro that expands to a builtin with a different signature +// so we wrap it in a C function to avoid the conflict +extern fn _z_vsnprintf(str: char*, size: usize, fmt: const char*, ap: va_list) -> int; + +// EOF is typically -1, but we define it for portability +def Z_EOF = -1; + +// Minimal raw block: only for truly opaque FILE* types that can't be +// represented in Zen-C extern declarations without type conflicts. +// These wrappers use void* to avoid FILE* declaration conflicts. +raw { + void* _z_get_stdin(void) { return stdin; } int _z_fgetc(void* stream) { return fgetc((FILE*)stream); } + int _z_vsnprintf(char* str, size_t size, const char* fmt, va_list ap) { + return vsnprintf(str, size, fmt, ap); + } } -extern fn _z_vprintf(fmt: char*, ap: va_list) -> int; -extern fn _z_vsnprintf(str: char*, size: usize, fmt: char*, ap: va_list) -> int; +extern fn _z_get_stdin() -> void*; +extern fn _z_fgetc(stream: void*) -> int; fn format(fmt: char*, ...) -> char* { static let buffer: char[1024]; let ap: va_list; va_start(ap, fmt); + _z_vsnprintf(buffer, 1024, fmt, ap); va_end(ap); + return (char*)buffer; } fn format_into(buffer: char*, size: usize, fmt: char*, ...) -> int { let ap: va_list; va_start(ap, fmt); + let ret = _z_vsnprintf(buffer, size, fmt, ap); va_end(ap); + return ret; } @@ -42,15 +56,17 @@ fn format_new(fmt: char*, ...) -> char* { let ap: va_list; va_start(ap, fmt); + _z_vsnprintf(buffer, 1024, fmt, ap); va_end(ap); + return buffer; } fn print(fmt: char*, ...) -> int { let ap: va_list; va_start(ap, fmt); - let ret = _z_vprintf(fmt, ap); + let ret = vprintf(fmt, ap); va_end(ap); return ret; } @@ -58,7 +74,7 @@ fn print(fmt: char*, ...) -> int { fn println(fmt: char*, ...) -> int { let ap: va_list; va_start(ap, fmt); - let ret = _z_vprintf(fmt, ap); + let ret = vprintf(fmt, ap); va_end(ap); puts(""); return ret + 1; @@ -72,11 +88,10 @@ fn readln() -> char* { let c: int; let std_in = _z_get_stdin(); - let eof_val = _z_get_eof(); while (true) { c = _z_fgetc(std_in); - if (c == eof_val) break; + if (c == Z_EOF) break; if (c == 10) break; // '\n' if (len + 1 >= cap) { @@ -93,7 +108,7 @@ fn readln() -> char* { len = len + 1; } - if (len == 0 && c == eof_val) { + if (len == 0 && c == Z_EOF) { free(line); return NULL; } diff --git a/std/json.zc b/std/json.zc index d373ab9..9f5cf73 100644 --- a/std/json.zc +++ b/std/json.zc @@ -457,3 +457,68 @@ impl Drop for JsonValue { self.free(); } } + +impl JsonValue { + fn to_string(self) -> String { + let s = String::new(""); + self.stringify(&s); + return s; + } + + fn stringify(self, buf: String*) { + if (self.kind.tag == JsonType::JSON_NULL().tag) { + buf.append_c_ptr("null"); + } else if (self.kind.tag == JsonType::JSON_BOOL().tag) { + if (self.bool_val) { buf.append_c_ptr("true"); } else { buf.append_c_ptr("false"); } + } else if (self.kind.tag == JsonType::JSON_NUMBER().tag) { + let tmp: char[64]; + sprintf((char*)tmp, "%.15g", self.number_val); // Use %.15g for precision + buf.append_c_ptr((char*)tmp); + } else if (self.kind.tag == JsonType::JSON_STRING().tag) { + buf.append_c_ptr("\""); + let p = self.string_val; + let len = strlen(p); + for (let i = 0; i < len; i = i + 1) { + let c = p[i]; + if (c == '"') buf.append_c_ptr("\\\""); + else if (c == '\\') buf.append_c_ptr("\\\\"); + else if (c == '\n') buf.append_c_ptr("\\n"); + else if (c == '\t') buf.append_c_ptr("\\t"); + else if (c == '\r') buf.append_c_ptr("\\r"); + else if (c == '\b') buf.append_c_ptr("\\b"); + else if (c == '\f') buf.append_c_ptr("\\f"); + else { + let tmp: char[2]; tmp[0] = c; tmp[1] = 0; + buf.append_c_ptr((char*)tmp); + } + } + buf.append_c_ptr("\""); + } else if (self.kind.tag == JsonType::JSON_ARRAY().tag) { + buf.append_c_ptr("["); + let v = self.array_val; + for (let i: usize = 0; i < v.length(); i = i + 1) { + if (i > 0) buf.append_c_ptr(","); + let item = v.get(i); + (*item).stringify(buf); + } + buf.append_c_ptr("]"); + } else if (self.kind.tag == JsonType::JSON_OBJECT().tag) { + buf.append_c_ptr("{{"); + let m = self.object_val; + let first = true; + for (let i: usize = 0; i < m.capacity(); i = i + 1) { + if (m.is_slot_occupied(i)) { + if (!first) buf.append_c_ptr(","); + first = false; + let key = m.key_at(i); + buf.append_c_ptr("\""); + buf.append_c_ptr(key); // Assuming keys are simple for now, but really should escape them too + buf.append_c_ptr("\":"); + let val = m.val_at(i); + val.stringify(buf); + } + } + buf.append_c_ptr("}"); + } + } +} @@ -3,16 +3,20 @@ import "./core.zc" import "./option.zc" import "./mem.zc" -raw { - extern size_t __zen_hash_seed; - size_t _map_hash_str(const char* str) { - size_t hash = __zen_hash_seed; - while (*str) { - hash ^= (unsigned char)*str++; - hash *= 1099511628211UL; - } - return hash; +// Pure Zen-C string hash using FNV-1a algorithm +fn _map_hash_str(str: const char*) -> usize { + let hash = __zen_hash_seed; + let i: usize = 0; + + while (str[i] != 0) { + // Cast char to U8 for unsigned byte value + let b: U8 = (U8)str[i]; + hash = hash ^ (usize)b; + hash = hash * (usize)1099511628211; + i = i + 1; } + + return hash; } struct Map<V> { @@ -13,8 +13,17 @@ import "./mem.zc" def Z_AF_INET = 2; def Z_SOCK_STREAM = 1; +// Direct externs for simple socket functions +extern fn socket(domain: int, type: int, proto: int) -> int; +extern fn close(fd: int) -> int; +extern fn read(fd: int, buf: void*, count: usize) -> isize; + +// Minimal raw block: required for struct sockaddr_in usage +// These functions encapsulate sockaddr_in setup because the struct layout +// cannot be declared in Zen-C without type conflicts with C headers. +// Also includes inet_pton, htons, bind, connect, listen, accept wrappers. raw { - static int _z_net_bind(int fd, char *host, int port) { + static int _z_net_bind(int fd, const char *host, int port) { struct sockaddr_in addr; addr.sin_family = AF_INET; addr.sin_port = htons(port); @@ -28,7 +37,7 @@ raw { return 0; } - static int _z_net_connect(int fd, char *host, int port) { + static int _z_net_connect(int fd, const char *host, int port) { struct sockaddr_in addr; addr.sin_family = AF_INET; addr.sin_port = htons(port); @@ -42,19 +51,15 @@ raw { return accept(fd, NULL, NULL); } - static ssize_t _z_net_write(int fd, char* buf, size_t n) { + static ssize_t _z_net_write(int fd, const char* buf, size_t n) { return write(fd, (const void*)buf, n); } } -extern fn socket(domain: int, type: int, proto: int) -> int; -extern fn close(fd: int) -> int; -extern fn read(fd: int, buf: void*, count: usize) -> isize; - -extern fn _z_net_bind(fd: int, host: char*, port: int) -> int; -extern fn _z_net_connect(fd: int, host: char*, port: int) -> int; +extern fn _z_net_bind(fd: int, host: const char*, port: int) -> int; +extern fn _z_net_connect(fd: int, host: const char*, port: int) -> int; extern fn _z_net_accept(fd: int) -> int; -extern fn _z_net_write(fd: int, buf: char*, n: usize) -> isize; +extern fn _z_net_write(fd: int, buf: const char*, n: usize) -> isize; struct TcpStream { diff --git a/std/process.zc b/std/process.zc index d0b09a9..3ce43b6 100644 --- a/std/process.zc +++ b/std/process.zc @@ -5,8 +5,16 @@ import "./mem.zc"; import "./string.zc"; import "./option.zc"; -raw { - void *_z_popen(char *command, char *type) { +include <stdio.h> +include <stdlib.h> + +// system() can be externed directly with const char* +extern fn system(command: const char*) -> int; + +// Minimal raw block: only for opaque FILE* types +// popen/pclose/fgets use FILE* which conflicts with void* +raw { + void *_z_popen(const char *command, const char *type) { return (void *)popen(command, type); } @@ -17,16 +25,11 @@ raw { char *_z_fgets(char *s, int size, void *stream) { return fgets(s, size, (FILE *)stream); } - - int _z_system(char *command) { - return system(command); - } } -extern fn _z_popen(command: char*, type: char*) -> void*; +extern fn _z_popen(command: const char*, type: const char*) -> void*; extern fn _z_pclose(stream: void*) -> int; extern fn _z_fgets(s: char*, size: int, stream: void*) -> char*; -extern fn _z_system(command: char*) -> int; struct Output { stdout: String; @@ -105,7 +108,7 @@ impl Command { fn status(self) -> int { let cmd_str = self._build_cmd(); - let code = _z_system(cmd_str.c_str()); + let code = system(cmd_str.c_str()); cmd_str.free(); return code; } @@ -2,17 +2,17 @@ import "./core.zc" import "./option.zc" -raw { - extern size_t __zen_hash_seed; - size_t _set_hash(const void* data, size_t len) { - size_t hash = __zen_hash_seed; - const unsigned char* bytes = (const unsigned char*)data; - for (size_t i = 0; i < len; i++) { - hash ^= bytes[i]; - hash *= 1099511628211UL; - } - return hash; +// Pure Zen-C generic hash using FNV-1a algorithm +fn _set_hash(data: const void*, len: usize) -> usize { + let hash = __zen_hash_seed; + let bytes: U8* = (U8*)data; + + for (let i: usize = 0; i < len; i = i + 1) { + hash = hash ^ (usize)bytes[i]; + hash = hash * (usize)1099511628211; } + + return hash; } struct Set<T> { diff --git a/std/slice.zc b/std/slice.zc index 778c6ed..7ace396 100644 --- a/std/slice.zc +++ b/std/slice.zc @@ -1,10 +1,45 @@ +import "./option.zc" + struct Slice<T> { data: T*; len: usize; } +struct SliceIter<T> { + data: T*; + count: usize; + idx: usize; +} + +impl SliceIter<T> { + fn next(self) -> Option<T> { + if (self.idx < self.count) { + let item = self.data[self.idx]; + self.idx = self.idx + 1; + return Option<T>::Some(item); + } + return Option<T>::None(); + } + + fn iterator(self) -> SliceIter<T> { + return *self; + } +} + impl Slice<T> { + fn from_array(arr: T*, len: usize) -> Slice<T> { + return Slice<T> { data: arr, len: len }; + } + + fn iterator(self) -> SliceIter<T> { + return SliceIter<T> { + data: self.data, + count: self.len, + idx: 0 + }; + } + fn length(self) -> usize { return self.len; } diff --git a/std/string.zc b/std/string.zc index fe5b0ad..0bc9539 100644 --- a/std/string.zc +++ b/std/string.zc @@ -55,6 +55,28 @@ impl String { } } + fn append_c(self, s: char*) { + if (self.vec.len > 0) { + self.vec.len = self.vec.len - 1; + } + let len = strlen(s); + for (let i = 0; i < len; i = i + 1) { + self.vec.push(s[i]); + } + self.vec.push(0); + } + + fn append_c_ptr(ptr: String*, s: char*) { + if (ptr.vec.len > 0) { + ptr.vec.len = ptr.vec.len - 1; + } + let len = strlen(s); + for (let i = 0; i < len; i = i + 1) { + ptr.vec.push(s[i]); + } + ptr.vec.push(0); + } + fn add(self, other: String*) -> String { let new_s = String::from(self.c_str()); new_s.append(other); diff --git a/std/thread.zc b/std/thread.zc index 98f080e..16f3ca1 100644 --- a/std/thread.zc +++ b/std/thread.zc @@ -7,6 +7,11 @@ import "./core.zc" import "./result.zc" import "./mem.zc" +// Essential raw block: required for pthread operations and closure trampolining +// This block cannot be eliminated because: +// 1. z_closure_T is an internal compiler type for Zen-C closures +// 2. pthread_t, pthread_mutex_t are opaque types that can't be extern'd with void* +// 3. The trampoline function needs to cast and execute Zen-C closures raw { typedef void (*ZenThreadFunc)(void*); diff --git a/std/time.zc b/std/time.zc index 865dd6d..fa764ed 100644 --- a/std/time.zc +++ b/std/time.zc @@ -5,6 +5,9 @@ include <unistd.h> include <sys/time.h> include <stdlib.h> +// Minimal raw block: required because gettimeofday() uses struct timeval +// which can't be declared in Zen-C without type conflicts, and time() +// has conflicting type signature (time_t* vs void*) raw { static uint64_t _time_now_impl(void) { struct timeval tv; diff --git a/tests/std/test_direct_array_iteration.zc b/tests/std/test_direct_array_iteration.zc new file mode 100644 index 0000000..359951f --- /dev/null +++ b/tests/std/test_direct_array_iteration.zc @@ -0,0 +1,37 @@ +import "std/slice.zc" + +test "direct array iteration" { + let arr: int[5] = [1, 2, 3, 4, 5]; + + let sum = 0; + for val in arr { + sum = sum + val; + } + + assert(sum == 15, "Sum should be 1+2+3+4+5 = 15"); +} + +test "direct array iteration with different types" { + let floats: float[3] = [1.5, 2.5, 3.0]; + let count = 0; + + for f in floats { + count = count + 1; + } + + assert(count == 3, "Should iterate over all 3 elements"); +} + +// TODO: Nested array iteration needs special handling +// test "nested array iteration" { +// let matrix: int[2][3] = [[1, 2, 3], [4, 5, 6]]; +// let total = 0; +// +// for row in matrix { +// for val in row { +// total = total + val; +// } +// } +// +// assert(total == 21, "Sum should be 1+2+3+4+5+6 = 21"); +// } diff --git a/tests/std/test_json_serialization.zc b/tests/std/test_json_serialization.zc new file mode 100644 index 0000000..9fd5b32 --- /dev/null +++ b/tests/std/test_json_serialization.zc @@ -0,0 +1,149 @@ +import "std/json.zc" +import "std/io.zc" + +test "primitives" { + // Null + let v = JsonValue::null(); + let s = v.to_string(); + let expected = String::from("null"); + if (!s.eq(&expected)) { + panic("Null serialization failed"); + } + expected.free(); + s.free(); + + // Bool True + v = JsonValue::bool(true); + s = v.to_string(); + expected = String::from("true"); + if (!s.eq(&expected)) { + panic("Bool true serialization failed"); + } + expected.free(); + s.free(); + + // Bool False + v = JsonValue::bool(false); + s = v.to_string(); + expected = String::from("false"); + if (!s.eq(&expected)) { + panic("Bool false serialization failed"); + } + expected.free(); + s.free(); + + // Number Int + v = JsonValue::number(123.0); + s = v.to_string(); + expected = String::from("123"); + if (!s.eq(&expected)) { + println "{s.c_str()}"; + panic("Number 123 serialization failed"); + } + expected.free(); + s.free(); + + // Number Float + v = JsonValue::number(12.5); + s = v.to_string(); + expected = String::from("12.5"); + if (!s.eq(&expected)) { + panic("Number 12.5 serialization failed"); + } + expected.free(); + s.free(); + + // String Simple + v = JsonValue::string("hello"); + s = v.to_string(); + expected = String::from("\"hello\""); + if (!s.eq(&expected)) { + println "{s.c_str()}"; + panic("String hello serialization failed"); + } + expected.free(); + s.free(); + + // String Escaped + v = JsonValue::string("hello \"world\""); + s = v.to_string(); + expected = String::from("\"hello \\\"world\\\"\""); + if (!s.eq(&expected)) { + println "Got: {s.c_str()}"; + panic("String escaped serialization failed"); + } + expected.free(); + s.free(); +} + +test "array" { + let v = JsonValue::array(); + v.push(JsonValue::number(1.0)); + v.push(JsonValue::bool(true)); + v.push(JsonValue::string("a")); + + let s = v.to_string(); + let expected = String::from("[1,true,\"a\"]"); + if (!s.eq(&expected)) { + println "Got: {s.c_str()}"; + panic("Array serialization failed"); + } + expected.free(); + s.free(); +} + +test "object" { + let v = JsonValue::object(); + v.set("key", JsonValue::string("value")); + + let s = v.to_string(); + // Round trip verification to avoid parser bug with literals + let parsed_res = JsonValue::parse(s.c_str()); + if (parsed_res.is_err()) { + panic("Object round trip parse failed"); + } + let parsed = parsed_res.unwrap(); + if (!parsed.is_object()) panic("Round trip not object"); + + let val_opt = (*parsed).get_string("key"); + if (val_opt.is_none()) panic("Round trip missing 'key'"); + + let val_str = val_opt.unwrap(); + if (strcmp(val_str, "value") != 0) panic("Round trip wrong value"); + + // Cleanup + (*parsed).free(); + free(parsed); + s.free(); +} + +test "nested" { + // {"arr":[1,2]} + let v = JsonValue::object(); + let arr = JsonValue::array(); + arr.push(JsonValue::number(1.0)); + arr.push(JsonValue::number(2.0)); + v.set("arr", arr); + + let s = v.to_string(); + + // Round trip + let parsed_res = JsonValue::parse(s.c_str()); + if (parsed_res.is_err()) { + panic("Round trip parse failed"); + } + let parsed = parsed_res.unwrap(); + if (!parsed.is_object()) panic("Round trip type mismatch"); + + let arr_opt = (*parsed).get_array("arr"); + if (arr_opt.is_none()) panic("Round trip missing arr"); + + let arr_ptr = arr_opt.unwrap(); + if (!(*arr_ptr).is_array()) panic("Inner not array"); + if ((*arr_ptr).len() != 2) panic("Wrong array length"); + + // Cleanup + (*parsed).free(); + free(parsed); + s.free(); +}
\ No newline at end of file diff --git a/tests/std/test_slice_iteration.zc b/tests/std/test_slice_iteration.zc new file mode 100644 index 0000000..b7eddf4 --- /dev/null +++ b/tests/std/test_slice_iteration.zc @@ -0,0 +1,29 @@ +import "std/slice.zc" +import "std/io.zc" + +test "slice from array iteration" { + let ints: int[5] = [1, 2, 3, 4, 5]; + let slice = Slice<int>::from_array((int*)(&ints), 5); + + // Test iteration + let sum = 0; + for val in slice { + sum = sum + val; + } + + if (sum != 15) { + panic("Slice iteration failed: expected sum 15"); + } +} + +test "slice methods" { + let arr: int[3] = [10, 20, 30]; + let slice = Slice<int>::from_array((int*)(&arr), 3); + + if (slice.length() != 3) panic("Slice length wrong"); + if (slice.is_empty()) panic("Slice should not be empty"); + + let opt = slice.get(1); + if (opt.is_none()) panic("Slice get failed"); + if (opt.unwrap() != 20) panic("Slice get returned wrong value"); +} |