cell/source/qjs_utf8.c

#include "qjs_utf8.h"
#include "qjs_blob.h"
#include "jsffi.h"
#include <string.h>
#include <stdlib.h>

#include "kim.h"

// Get codepoints from a UTF-8 string
JSC_CCALL(utf8_codepoints,
  const char *str = JS_ToCString(js, argv[0]);
  if (!str) return JS_EXCEPTION;
  
  JSValue arr = JS_NewArray(js);
  int idx = 0;
  
  char *ptr = (char*)str;
  while (*ptr) {
    int codepoint = decode_utf8(&ptr);
    JS_SetPropertyUint32(js, arr, idx++, JS_NewInt32(js, codepoint));
  }
  
  JS_FreeCString(js, str);
  ret = arr;
)

// Create UTF-8 string from codepoints
JSC_CCALL(utf8_from_codepoints,
  int len = JS_ArrayLength(js, argv[0]);
  
  // Allocate buffer (worst case: 4 bytes per codepoint + null)
  char *buffer = malloc(len * 4 + 1);
  char *ptr = buffer;
  
  for (int i = 0; i < len; i++) {
    JSValue val = JS_GetPropertyUint32(js, argv[0], i);
    int codepoint;
    JS_ToInt32(js, &codepoint, val);
    JS_FreeValue(js, val);
    
    encode_utf8(&ptr, codepoint);
  }
  
  *ptr = '\0';
  ret = JS_NewString(js, buffer);
  free(buffer);
)

// Count UTF-8 characters (runes) in a string
JSC_SCALL(utf8_length,
  int count = utf8_count(str);
  ret = JS_NewInt32(js, count);
)

// Validate UTF-8 string
JSC_SCALL(utf8_validate,
  char *ptr = (char*)str;
  int valid = 1;
  
  while (*ptr) {
    int start_pos = ptr - str;
    int codepoint = decode_utf8(&ptr);
    
    // Check for invalid sequences
    if (codepoint < 0 || codepoint > 0x10FFFF || 
        (codepoint >= 0xD800 && codepoint <= 0xDFFF)) {
      valid = 0;
      break;
    }
    
    // Check for overlong encodings
    int bytes_used = ptr - (str + start_pos);
    if ((codepoint <= 0x7F && bytes_used != 1) ||
        (codepoint <= 0x7FF && bytes_used != 2) ||
        (codepoint <= 0xFFFF && bytes_used != 3) ||
        (codepoint <= 0x10FFFF && bytes_used != 4)) {
      valid = 0;
      break;
    }
  }
  
  ret = JS_NewBool(js, valid);
)

// Get byte length of UTF-8 string
JSC_SCALL(utf8_byte_length,
  ret = JS_NewInt32(js, strlen(str));
)

// Encode string to UTF-8 bytes
JSC_SCALL(utf8_encode,
  size_t len = strlen(str);
  ret = js_new_blob_stoned_copy(js, str, len);
)

// Decode UTF-8 bytes to string
JSC_CCALL(utf8_decode,
  size_t len;
  void *data = js_get_blob_data(js, &len, argv[0]);
  if (!data) return JS_ThrowTypeError(js, "Expected blob");
  
  // Create null-terminated string
  char *str = malloc(len + 1);
  memcpy(str, data, len);
  str[len] = '\0';
  
  ret = JS_NewString(js, str);
  free(str);
)

// Slice UTF-8 string by character indices (not byte indices)
JSC_CCALL(utf8_slice,
  const char *str = JS_ToCString(js, argv[0]);
  if (!str) return JS_EXCEPTION;
  
  int start = 0;
  int end = utf8_count(str);
  
  if (argc > 1) JS_ToInt32(js, &start, argv[1]);
  if (argc > 2) JS_ToInt32(js, &end, argv[2]);
  
  // Handle negative indices
  int total = end;
  if (start < 0) start = total + start;
  if (end < 0) end = total + end;
  
  // Clamp values
  if (start < 0) start = 0;
  if (end > total) end = total;
  if (start >= end) {
    JS_FreeCString(js, str);
    return JS_NewString(js, "");
  }
  
  // Find start position
  char *ptr = (char*)str;
  for (int i = 0; i < start && *ptr; i++) {
    decode_utf8(&ptr);
  }
  char *start_ptr = ptr;
  
  // Find end position
  for (int i = start; i < end && *ptr; i++) {
    decode_utf8(&ptr);
  }
  
  // Create substring
  size_t slice_len = ptr - start_ptr;
  char *slice = malloc(slice_len + 1);
  memcpy(slice, start_ptr, slice_len);
  slice[slice_len] = '\0';
  
  ret = JS_NewString(js, slice);
  free(slice);
  JS_FreeCString(js, str);
)

// Get character at index
JSC_CCALL(utf8_char_at,
  const char *str = JS_ToCString(js, argv[0]);
  if (!str) return JS_EXCEPTION;
  
  int index;
  JS_ToInt32(js, &index, argv[1]);
  
  char *ptr = (char*)str;
  int count = 0;
  
  // Skip to index
  while (*ptr && count < index) {
    decode_utf8(&ptr);
    count++;
  }
  
  if (!*ptr || count != index) {
    JS_FreeCString(js, str);
    return JS_UNDEFINED;
  }
  
  // Get the character
  char *char_start = ptr;
  decode_utf8(&ptr);
  
  size_t char_len = ptr - char_start;
  char *result = malloc(char_len + 1);
  memcpy(result, char_start, char_len);
  result[char_len] = '\0';
  
  ret = JS_NewString(js, result);
  free(result);
  JS_FreeCString(js, str);
)

static const JSCFunctionListEntry js_utf8_funcs[] = {
  MIST_FUNC_DEF(utf8, codepoints, 1),
  MIST_FUNC_DEF(utf8, from_codepoints, 1),
  MIST_FUNC_DEF(utf8, length, 1),
  MIST_FUNC_DEF(utf8, validate, 1),
  MIST_FUNC_DEF(utf8, byte_length, 1),
  MIST_FUNC_DEF(utf8, encode, 1),
  MIST_FUNC_DEF(utf8, decode, 1),
  MIST_FUNC_DEF(utf8, slice, 3),
  MIST_FUNC_DEF(utf8, char_at, 2),
};

JSValue js_utf8_use(JSContext *js)
{
  JSValue mod = JS_NewObject(js);
  JS_SetPropertyFunctionList(js, mod, js_utf8_funcs, countof(js_utf8_funcs));
  return mod;
}