cjit: Some standard headers, builtins

packages/cjit/src/includes/builtin.h

@@ -0,0 +1,31 @@
+#pragma once
+
+#define abs(value) __builtin_abs(value)
+#define byteswap(value) __builtin_byteswap(value)
+#define ceil(value) __builtin_ceil(value)
+#define clz(value) __builtin_clz(value)
+#define containerof(ptr, type, member) __builtin_containerof(ptr, type, member)
+#define cos(value) __builtin_cos(value)
+#define ctz(value) __builtin_ctz(value)
+#define embedfile(path) __builtin_embedfile(path)
+#define exp(value) __builtin_exp(value)
+#define exp2(value) __builtin_exp2(value)
+#define floor(value) __builtin_floor(value)
+#define frameaddress() __builtin_frameaddress()
+#define log(value) __builtin_log(value)
+#define log10(value) __builtin_log10(value)
+#define log2(value) __builtin_log2(value)
+#define max(...) __builtin_max(__VA_ARGS__)
+#define memcpy(dest, src, count) __builtin_memcpy(dest, src, count)
+#define memmove(dest, src, count) __builtin_memmove(dest, src, count)
+#define memset(value) __builtin_memset(value)
+#define min(...) __builtin_min(__VA_ARGS__)
+#define popcount(value) __builtin_popcount(value)
+#define returnaddress() __builtin_returnaddress()
+#define round(value) __builtin_round(value)
+#define sin(value) __builtin_sin(value)
+#define sqrt(value) __builtin_sqrt(value)
+#define tan(value) __builtin_tan(value)
+#define trunc(value) __builtin_trunc(value)
+#define typename(type) __builtin_typename(type)
+#define typeof(...) __builtin_typeof(__VA_ARGS__)

packages/cjit/src/includes/stdalign.h

@@ -0,0 +1,7 @@
+#pragma once
+
+#define alignas _Alignas
+#define alignof _Alignof
+
+#define __alignas_is_defined 1
+#define __alignof_is_defined 1

packages/cjit/src/includes/stdbool.h

@@ -0,0 +1,7 @@
+#pragma once
+
+#define bool _Bool
+
+#define true 1
+#define false 0
+#define __bool_true_false_are_defined 1

packages/cjit/src/includes/stddef.h

@@ -0,0 +1,9 @@
+#pragma once
+
+typedef long ptrdiff_t;
+typedef long max_align_t;
+typedef unsigned long size_t;
+
+#define NULL ((void *)0)
+
+#define offsetof(type, member) __builtin_offsetof(type, member)

packages/cjit/src/includes/stdint.h

@@ -0,0 +1,118 @@
+#pragma once
+
+typedef signed char int8_t;
+typedef short int16_t;
+typedef int int32_t;
+typedef long int64_t;
+
+typedef int8_t int_fast8_t;
+typedef int16_t int_fast16_t;
+typedef int32_t int_fast32_t;
+typedef int64_t int_fast64_t;
+
+typedef int8_t int_least8_t;
+typedef int16_t int_least16_t;
+typedef int32_t int_least32_t;
+typedef int64_t int_least64_t;
+
+typedef int64_t intmax_t;
+typedef int64_t intptr_t;
+
+typedef unsigned char uint8_t;
+typedef unsigned short uint16_t;
+typedef unsigned int uint32_t;
+typedef unsigned long uint64_t;
+
+typedef uint8_t uint_fast8_t;
+typedef uint16_t uint_fast16_t;
+typedef uint32_t uint_fast32_t;
+typedef uint64_t uint_fast64_t;
+
+typedef uint8_t uint_least8_t;
+typedef uint16_t uint_least16_t;
+typedef uint32_t uint_least32_t;
+typedef uint64_t uint_least64_t;
+
+typedef uint64_t uintmax_t;
+typedef uint64_t uintptr_t;
+
+#define INT8_MIN (-0x80)
+#define INT16_MIN (-0x8000)
+#define INT32_MIN (-0x80000000)
+#define INT64_MIN (-0x8000000000000000L)
+
+#define INT_FAST8_MIN INT8_MIN
+#define INT_FAST16_MIN INT16_MIN
+#define INT_FAST32_MIN INT32_MIN
+#define INT_FAST64_MIN INT64_MIN
+
+#define INT_LEAST8_MIN INT8_MIN
+#define INT_LEAST16_MIN INT16_MIN
+#define INT_LEAST32_MIN INT32_MIN
+#define INT_LEAST64_MIN INT64_MIN
+
+#define INTPTR_MIN INT64_MIN
+#define INTMAX_MIN INT64_MIN
+
+#define INT8_MAX (0x7F)
+#define INT16_MAX (0x7FFF)
+#define INT32_MAX (0x7FFFFFFF)
+#define INT64_MAX (0x7FFFFFFFFFFFFFFFL)
+
+#define INT_FAST8_MAX INT8_MAX
+#define INT_FAST16_MAX INT16_MAX
+#define INT_FAST32_MAX INT32_MAX
+#define INT_FAST64_MAX INT64_MAX
+
+#define INT_LEAST8_MAX INT8_MAX
+#define INT_LEAST16_MAX INT16_MAX
+#define INT_LEAST32_MAX INT32_MAX
+#define INT_LEAST64_MAX INT64_MAX
+
+#define INTPTR_MAX INT64_MAX
+#define INTMAX_MAX INT64_MAX
+
+#define UINT8_MAX (0xFF)
+#define UINT16_MAX (0xFFFF)
+#define UINT32_MAX (0xFFFFFFFFU)
+#define UINT64_MAX (0xFFFFFFFFFFFFFFFFUL)
+
+#define UINT_FAST8_MAX UINT8_MAX
+#define UINT_FAST16_MAX UINT16_MAX
+#define UINT_FAST32_MAX UINT32_MAX
+#define UINT_FAST64_MAX UINT64_MAX
+
+#define UINT_LEAST8_MAX UINT8_MAX
+#define UINT_LEAST16_MAX UINT16_MAX
+#define UINT_LEAST32_MAX UINT32_MAX
+#define UINT_LEAST64_MAX UINT64_MAX
+
+#define UINTPTR_MAX UINT64_MAX
+#define UINTMAX_MAX UINT64_MAX
+
+#define INT8_C(value) value
+#define INT16_C(value) value
+#define INT32_C(value) value
+#define INT64_C(value) value ## L
+
+#define INTMAX_C(value) value ## L
+
+#define UINT8_C(value) value
+#define UINT16_C(value) value
+#define UINT32_C(value) value ## U
+#define UINT64_C(value) value ## UL
+
+#define UINTMAX_C(value) value ## UL
+
+#define PTRDIFF_MIN (-0x8000000000000000L)
+#define PTRDIFF_MAX (0x7FFFFFFFFFFFFFFFL)
+
+#define SIZE_MAX (0xFFFFFFFFFFFFFFFFUL)
+
+#define WINT_MIN (-0x80000000)
+#define WINT_MAX (0x7FFFFFFF)
+
+#define WCHAR_MIN (0)
+#define WCHAR_MAX (0x10FFFF)
+
+#endif

packages/cjit/src/includes/stdnoreturn.h

@@ -0,0 +1,3 @@
+#pragma once
+
+#define noreturn _Noreturn

packages/cjit/src/tokens.zig

@@ -1,6 +1,7 @@
-pub const Constant = @import("tokens/Constant.zig");
-pub const Keyword = @import("tokens/Keyword.zig");
-pub const Punctuator = @import("tokens/Punctuator.zig");
-pub const Token = @import("tokens/Token.zig");
+pub const Builtin = @import("tokens/Builtin.zig").Builtin;
+pub const Constant = @import("tokens/Constant.zig").Constant;
+pub const Keyword = @import("tokens/Keyword.zig").Keyword;
+pub const Punctuator = @import("tokens/Punctuator.zig").Punctuator;
+pub const Token = @import("tokens/Token.zig").Token;
 pub const Tokenizer = @import("tokens/Tokenizer.zig");
 pub const Utf8Iterator = @import("tokens/Utf8Iterator.zig");

packages/cjit/src/tokens/Builtin.zig

@@ -0,0 +1,156 @@
+const std = @import("std");
+
+pub const Builtin = enum {
+    /// Usage: `__builtin_abs(value)`
+    ///
+    /// `value` can be any integer or real type. Equivalent to
+    /// `value < 0 ? -value : value`. Noop for unsigned integer types.
+    __builtin_abs,
+    /// Usage: `__builtin_byteswap(value)`
+    ///
+    /// `value` can be any integer type.
+    __builtin_byteswap,
+    /// Usage: `__builtin_ceil(value)`
+    ///
+    /// `value` can be any real type.
+    __builtin_ceil,
+    /// Usage: `__builtin_clz(value)`
+    ///
+    /// Count leading zeroes. `value` can be any integer type. The return type
+    /// is `int`.
+    __builtin_clz,
+    /// Usage `__builtin_containerof(ptr, type, member)`
+    ///
+    /// `ptr` must be an pointer to a struct or a union. `type` must be a type.
+    /// `member` must be an identifier. Given `ptr` is a pointer to a given
+    /// member of `type`, returns a pointer to the entire container.
+    __builtin_containerof,
+    /// Usage: `__builtin_cos(value)`
+    ///
+    /// `value` can be any real type.
+    __builtin_cos,
+    /// Usage: `__builtin_ctz(value)`
+    ///
+    /// Count trailing zeroes. `value` can be any integer type. The return type
+    /// is `int`.
+    __builtin_ctz,
+    /// Usage: `__builtin_embedfile(path)`
+    ///
+    /// `path` must be a string literal. The return type is `const char *`. The
+    /// data is null-terminated.
+    __builtin_embedfile,
+    /// Usage `__builtin_exp(value)`
+    ///
+    /// `value` can be any real type. Calculates e^value.
+    __builtin_exp,
+    /// Usage `__builtin_exp2(value)`
+    ///
+    /// `value` can be any real type. Calculates 2^value.
+    __builtin_exp2,
+    /// Usage: `__bultin_floor(value)`
+    ///
+    /// `value` can be any real type.
+    __builtin_floor,
+    /// Usage: `__builtin_frameaddress()`
+    ///
+    /// Returns the value of base pointer. The return type is equivalent to
+    /// `uintptr_t`.
+    __builtin_frameaddress,
+    /// Usage: `__builtin_log(value)`
+    ///
+    /// `value` can be any real type. Calculates natural logarithm (base e).
+    __builtin_log,
+    /// Usage: `__builtin_log10(value)`
+    ///
+    /// `value` can be any real type. Calculates base 10 logarithm.
+    __builtin_log10,
+    /// Usage: `__builtin_log2(value)`
+    ///
+    /// `value` can be any real type. Calculates base 2 logarithm.
+    __builtin_log2,
+    /// Usage: `__builtin_max(...)`
+    ///
+    /// The arguments can be any integer or real types. NaN values are ignored.
+    __builtin_max,
+    /// Usage: `__builtin_memcpy(dest, src, count)`
+    ///
+    /// `dest` and `src` must be pointers. The pointers are reinterpreted as
+    /// pointers to `char`. `count` is coerced to the equivalent of `size_t`.
+    /// `dest` must be a non-const pointer. The regions must not overlap.
+    __builtin_memcpy,
+    /// Usage: `__builtin_memmove(dest, src, count)`
+    ///
+    /// `dest` and `src` must be pointers. The pointers are reinterpreted as
+    /// pointers to `char`. `count` is coerced to the equivalent of `size_t`.
+    /// `dest` must be a non-const pointer. The regions may overlap.
+    __builtin_memmove,
+    /// Usage: `__builtin_memset(dest, ch, count)`
+    ///
+    /// `dest` must be a pointer. The pointer is reinterpreted as pointer to
+    /// `char`. `ch` is cast to `unsigned char`. `dest` must be a non-const
+    // pointer.
+    __builtin_memset,
+    /// Usage: `__builtin_min(...)`
+    ///
+    /// The arguments can be any integer or real types. NaN values are ignored.
+    __builtin_min,
+    /// Usage: `__builtin_offsetof(type, member)`
+    ///
+    /// `type` must be a type. `member` must be an identifier.
+    __builtin_offsetof,
+    /// Usage: `__builtin_popcount(value)`
+    ///
+    /// `value` can be any integer type. The return type is `int`.
+    __builtin_popcount,
+    /// Usage: `__builtin_returnaddress()`
+    ///
+    /// Returns the address of the instruction to run after current function
+    /// returns. The return type is equivalent to `uintptr_t`.
+    __builtin_returnaddress,
+    /// Usage: `__builtin_round(value)`
+    ///
+    /// `value` can be any real type.
+    __builtin_round,
+    /// Usage: `__builtin_sin(value)`
+    ///
+    /// `value` can be any real type.
+    __builtin_sin,
+    /// Usage: `__builtin_sqrt(value)`
+    ///
+    /// `value` can be any real type.
+    __builtin_sqrt,
+    /// Usage: `__builtin_tan(value)`
+    ///
+    /// `value` can be any real type.
+    __builtin_tan,
+    /// Usage: `__builtin_trunc(value)`
+    ///
+    /// `value` can be any real type.
+    __builtin_trunc,
+    /// Usage: `__builtin_typename(type)`
+    ///
+    /// `type` must be a type. The return type is `const char *`.
+    __builtin_typename,
+    /// Usage: `__builtin_typeof(...)`
+    __builtin_typeof,
+
+    pub const map: std.StaticStringMap(Builtin) = blk: {
+        const fields = @typeInfo(Builtin).@"enum".fields;
+
+        var kvs_list: [fields.len]struct { []const u8, Builtin } = undefined;
+        for (fields, 0..) |field, i| {
+            kvs_list[i] = .{ field.name, @field(Builtin, field.name) };
+        }
+
+        break :blk .initComptime(kvs_list);
+    };
+
+    pub fn isBuiltin(identifier: []const u8) ?Builtin {
+        if (std.mem.startsWith(u8, identifier, "__builtin_")) {
+            @branchHint(.unlikely);
+            return map.get(identifier);
+        } else {
+            return null;
+        }
+    }
+};

packages/cjit/src/tokens/Punctuator.zig

@@ -57,14 +57,14 @@ pub const Punctuator = enum(u32) {
     pub const line_continuation_crlf = strToInt3("\\\r\n");
 };
 
-pub fn strToInt1(str: *const u8[1]) u32 {
+fn strToInt1(str: *const u8[1]) u32 {
     return @as(u8, @bitCast(str.*));
 }
 
-pub fn strToInt2(str: *const u8[2]) u32 {
+fn strToInt2(str: *const u8[2]) u32 {
     return @as(u16, @bitCast(str.*));
 }
 
-pub fn strToInt3(str: *const u8[3]) u32 {
+fn strToInt3(str: *const u8[3]) u32 {
     return @as(u24, @bitCast(str.*));
 }

packages/cjit/src/tokens/Token.zig

@@ -1,14 +1,16 @@
 const std = @import("std");
 
-pub const Constant = @import("Constant.zig");
-pub const Keyword = @import("Keyword.zig");
-pub const Punctuator = @import("Punctuator.zig");
+const Builtin = @import("Builtin.zig").Builtin;
+const Constant = @import("Constant.zig").Constant;
+const Keyword = @import("Keyword.zig").Keyword;
+const Punctuator = @import("Punctuator.zig").Punctuator;
 
 pub const Token = union(enum) {
-    keyword: Keyword,
-    identifier: []const u8,
+    builtin: Builtin,
     constant: Constant,
+    keyword: Keyword,
+    punctuator: Punctuator,
+    identifier: []const u8,
     string_literal: [:0]const u8,
     wide_string_literal: [:0]const u32,
-    punctuator: Punctuator,
 };

packages/cjit/src/tokens/Tokenizer.zig

@@ -1,6 +1,7 @@
 const std = @import("std");
 const Self = @This();
 
+const Builtin = @import("Builtin.zig").Builtin;
 const Keyword = @import("Keyword.zig").Keyword;
 const Punctuator = @import("Punctuator.zig").Punctuator;
 const Token = @import("Token.zig").Token;
@@ -84,10 +85,13 @@ pub fn nextToken(self: *Self, arena_allocator: std.mem.Allocator) !?Token {
 
             const identifier = self.str[identifier_start..self.it.ptr];
 
+            // TODO Preprocessor
+
             if (Keyword.isKeyword(identifier)) |keyword| {
                 return .{ .keyword = keyword };
+            } else if (Builtin.isBuiltin(identifier)) |builtin| {
+                return .{ .builtin = builtin };
             } else {
-                // TODO Preprocessor
                 return .{ .identifier = arena_allocator.dupe(u8, identifier) };
             }
         },

packages/cjit/src/x86_64.zig

@@ -4,7 +4,6 @@ const tokens = @import("tokens.zig");
 const types = @import("types.zig");
 
 const Location = Runtime.Location;
-const Punctuator = tokens.Punctuator;
 const Runtime = @import("Runtime.zig");
 const StackValue = @import("StackValue.zig");
 const Type = types.Type;