static rb_encoding*
enc_compatible_latter(VALUE str1, VALUE str2, int idx1, int idx2)
{
int isstr1, isstr2;
rb_encoding *enc1 = rb_enc_from_index(idx1);
rb_encoding *enc2 = rb_enc_from_index(idx2);
isstr2 = RB_TYPE_P(str2, T_STRING);
if (isstr2 && RSTRING_LEN(str2) == 0)
return enc1;
isstr1 = RB_TYPE_P(str1, T_STRING);
if (isstr1 && RSTRING_LEN(str1) == 0)
return (rb_enc_asciicompat(enc1) && rb_enc_str_asciionly_p(str2)) ? enc1 : enc2;
if (!rb_enc_asciicompat(enc1) || !rb_enc_asciicompat(enc2)) {
return 0;
}
/* objects whose encoding is the same of contents */
if (!isstr2 && idx2 == ENCINDEX_US_ASCII)
return enc1;
if (!isstr1 && idx1 == ENCINDEX_US_ASCII)
return enc2;
if (!isstr1) {
VALUE tmp = str1;
int idx0 = idx1;
str1 = str2;
str2 = tmp;
idx1 = idx2;
idx2 = idx0;
idx0 = isstr1;
isstr1 = isstr2;
isstr2 = idx0;
}
if (isstr1) {
int cr1, cr2;
cr1 = rb_enc_str_coderange(str1);
if (isstr2) {
cr2 = rb_enc_str_coderange(str2);
if (cr1 != cr2) {
/* may need to handle ENC_CODERANGE_BROKEN */
if (cr1 == ENC_CODERANGE_7BIT) return enc2;
if (cr2 == ENC_CODERANGE_7BIT) return enc1;
}
if (cr2 == ENC_CODERANGE_7BIT) {
return enc1;
}
}
if (cr1 == ENC_CODERANGE_7BIT)
return enc2;
}
return 0;
}
static VALUE
str_compat_and_valid(VALUE str, rb_encoding *enc)
{
int cr;
str = StringValue(str);
cr = rb_enc_str_coderange(str);
if (cr == ENC_CODERANGE_BROKEN) {
#ifdef PRIsVALUE
rb_raise(rb_eArgError, "replacement must be valid byte sequence '%+"PRIsVALUE"'", str);
#else
str = rb_inspect(str);
rb_raise(rb_eArgError, "replacement must be valid byte sequence '%s'", RSTRING_PTR(str));
RB_GC_GUARD(str);
#endif
}
else if (cr == ENC_CODERANGE_7BIT) {
rb_encoding *e = STR_ENC_GET(str);
if (!rb_enc_asciicompat(enc)) {
rb_raise(rb_eEncCompatError, "incompatible character encodings: %s and %s",
rb_enc_name(enc), rb_enc_name(e));
}
}
else { /* ENC_CODERANGE_VALID */
rb_encoding *e = STR_ENC_GET(str);
if (enc != e) {
rb_raise(rb_eEncCompatError, "incompatible character encodings: %s and %s",
rb_enc_name(enc), rb_enc_name(e));
}
}
return str;
}
static VALUE encoding_spec_rb_enc_str_coderange(VALUE self, VALUE str) {
int coderange = rb_enc_str_coderange(str);
switch(coderange) {
case ENC_CODERANGE_UNKNOWN:
return ID2SYM(rb_intern("coderange_unknown"));
case ENC_CODERANGE_7BIT:
return ID2SYM(rb_intern("coderange_7bit"));
case ENC_CODERANGE_VALID:
return ID2SYM(rb_intern("coderange_valid"));
case ENC_CODERANGE_BROKEN:
return ID2SYM(rb_intern("coderange_broken"));
default:
return ID2SYM(rb_intern("coderange_unrecognized"));
}
}
VALUE native_slot_encode_and_freeze_string(upb_fieldtype_t type, VALUE value) {
rb_encoding* desired_encoding = (type == UPB_TYPE_STRING) ?
kRubyStringUtf8Encoding : kRubyString8bitEncoding;
VALUE desired_encoding_value = rb_enc_from_encoding(desired_encoding);
// Note: this will not duplicate underlying string data unless necessary.
value = rb_str_encode(value, desired_encoding_value, 0, Qnil);
if (type == UPB_TYPE_STRING &&
rb_enc_str_coderange(value) == ENC_CODERANGE_BROKEN) {
rb_raise(rb_eEncodingError, "String is invalid UTF-8");
}
// Ensure the data remains valid. Since we called #encode a moment ago,
// this does not freeze the string the user assigned.
rb_obj_freeze(value);
return value;
}
rb_encoding*
rb_enc_compatible(VALUE str1, VALUE str2)
{
int idx1, idx2;
rb_encoding *enc1, *enc2;
idx1 = rb_enc_get_index(str1);
idx2 = rb_enc_get_index(str2);
if (idx1 < 0 || idx2 < 0)
return 0;
if (idx1 == idx2) {
return rb_enc_from_index(idx1);
}
enc1 = rb_enc_from_index(idx1);
enc2 = rb_enc_from_index(idx2);
if (TYPE(str2) == T_STRING && RSTRING_LEN(str2) == 0)
return (idx1 == ENCINDEX_US_ASCII && rb_enc_asciicompat(enc2)) ? enc2 : enc1;
if (TYPE(str1) == T_STRING && RSTRING_LEN(str1) == 0)
return (idx2 == ENCINDEX_US_ASCII && rb_enc_asciicompat(enc1)) ? enc1 : enc2;
if (!rb_enc_asciicompat(enc1) || !rb_enc_asciicompat(enc2)) {
return 0;
}
/* objects whose encoding is the same of contents */
if (BUILTIN_TYPE(str2) != T_STRING && idx2 == ENCINDEX_US_ASCII)
return enc1;
if (BUILTIN_TYPE(str1) != T_STRING && idx1 == ENCINDEX_US_ASCII)
return enc2;
if (BUILTIN_TYPE(str1) != T_STRING) {
VALUE tmp = str1;
int idx0 = idx1;
str1 = str2;
str2 = tmp;
idx1 = idx2;
idx2 = idx0;
}
if (BUILTIN_TYPE(str1) == T_STRING) {
int cr1, cr2;
cr1 = rb_enc_str_coderange(str1);
if (BUILTIN_TYPE(str2) == T_STRING) {
cr2 = rb_enc_str_coderange(str2);
if (cr1 != cr2) {
/* may need to handle ENC_CODERANGE_BROKEN */
if (cr1 == ENC_CODERANGE_7BIT) return enc2;
if (cr2 == ENC_CODERANGE_7BIT) return enc1;
}
if (cr2 == ENC_CODERANGE_7BIT) {
if (idx1 == ENCINDEX_ASCII) return enc2;
return enc1;
}
}
if (cr1 == ENC_CODERANGE_7BIT)
return enc2;
}
return 0;
}
static VALUE
optimized_unescape(VALUE str, VALUE encoding)
{
long i, len, beg = 0;
VALUE dest = 0;
const char *cstr;
int cr, origenc, encidx = rb_to_encoding_index(encoding);
len = RSTRING_LEN(str);
cstr = RSTRING_PTR(str);
for (i = 0; i < len; ++i) {
char buf[1];
const char c = cstr[i];
int clen = 0;
if (c == '%') {
if (i + 3 > len) break;
if (!ISXDIGIT(cstr[i+1])) continue;
if (!ISXDIGIT(cstr[i+2])) continue;
buf[0] = ((char_to_number(cstr[i+1]) << 4)
| char_to_number(cstr[i+2]));
clen = 2;
}
else if (c == '+') {
buf[0] = ' ';
}
else {
continue;
}
if (!dest) {
dest = rb_str_buf_new(len);
}
rb_str_cat(dest, cstr + beg, i - beg);
i += clen;
beg = i + 1;
rb_str_cat(dest, buf, 1);
}
if (dest) {
rb_str_cat(dest, cstr + beg, len - beg);
preserve_original_state(str, dest);
cr = ENC_CODERANGE_UNKNOWN;
}
else {
dest = rb_str_dup(str);
cr = ENC_CODERANGE(str);
}
origenc = rb_enc_get_index(str);
if (origenc != encidx) {
rb_enc_associate_index(dest, encidx);
if (!ENC_CODERANGE_CLEAN_P(rb_enc_str_coderange(dest))) {
rb_enc_associate_index(dest, origenc);
if (cr != ENC_CODERANGE_UNKNOWN)
ENC_CODERANGE_SET(dest, cr);
}
}
return dest;
}
