PARROT_WARN_UNUSED_RESULT
INTVAL
encoding_equal(PARROT_INTERP, ARGIN(const STRING *lhs), ARGIN(const STRING *rhs))
{
ASSERT_ARGS(encoding_equal)
String_iter l_iter, r_iter;
const UINTVAL len = STRING_length(lhs);
if (len != STRING_length(rhs))
return 0;
if (len == 0)
return 1;
if (lhs == rhs)
return 1;
if (lhs->hashval && rhs->hashval && lhs->hashval != rhs->hashval)
return 0;
if (lhs->encoding == rhs->encoding)
return memcmp(lhs->strstart, rhs->strstart, STRING_byte_length(lhs)) == 0;
STRING_ITER_INIT(interp, &l_iter);
STRING_ITER_INIT(interp, &r_iter);
while (l_iter.charpos < len) {
const UINTVAL cl = STRING_iter_get_and_advance(interp, lhs, &l_iter);
const UINTVAL cr = STRING_iter_get_and_advance(interp, rhs, &r_iter);
if (cl != cr)
return 0;
}
return 1;
}
PARROT_WARN_UNUSED_RESULT
INTVAL
encoding_rindex(PARROT_INTERP, ARGIN(const STRING *src),
ARGIN(const STRING *search), INTVAL offset)
{
ASSERT_ARGS(encoding_rindex)
String_iter search_iter, search_start, start;
const UINTVAL len = search->strlen;
UINTVAL c0;
INTVAL skip;
if (offset < 0
|| len == 0
|| src->strlen < len)
return -1;
skip = src->strlen - len;
if (offset < skip)
skip = offset;
STRING_ITER_INIT(interp, &start);
STRING_iter_skip(interp, src, &start, skip);
STRING_ITER_INIT(interp, &search_start);
c0 = STRING_iter_get_and_advance(interp, search, &search_start);
while (1) {
UINTVAL c1 = STRING_iter_get(interp, src, &start, 0);
if (c1 == c0) {
UINTVAL c2;
String_iter iter = start;
STRING_iter_skip(interp, src, &iter, 1);
search_iter = search_start;
do {
if (search_iter.charpos >= len)
return start.charpos;
c1 = STRING_iter_get_and_advance(interp, src, &iter);
c2 = STRING_iter_get_and_advance(interp, search, &search_iter);
} while (c1 == c2);
}
if (start.charpos == 0)
break;
STRING_iter_skip(interp, src, &start, -1);
}
return -1;
}
PARROT_CANNOT_RETURN_NULL
PARROT_WARN_UNUSED_RESULT
STRING *
encoding_to_encoding(PARROT_INTERP, ARGIN(const STRING *src),
ARGIN(const STR_VTABLE *encoding), double avg_bytes)
{
ASSERT_ARGS(encoding_to_encoding)
STRING *result;
String_iter src_iter, dest_iter;
UINTVAL src_len, alloc_bytes;
UINTVAL max_bytes = encoding->max_bytes_per_codepoint;
if (src->encoding == encoding)
return Parrot_str_clone(interp, src);
src_len = src->strlen;
result = Parrot_gc_new_string_header(interp, 0);
result->encoding = encoding;
result->strlen = src_len;
if (!src_len)
return result;
alloc_bytes = (UINTVAL)(src_len * avg_bytes);
if (alloc_bytes < max_bytes)
alloc_bytes = max_bytes;
Parrot_gc_allocate_string_storage(interp, result, alloc_bytes);
result->bufused = alloc_bytes;
STRING_ITER_INIT(interp, &src_iter);
STRING_ITER_INIT(interp, &dest_iter);
while (src_iter.charpos < src_len) {
const UINTVAL c = STRING_iter_get_and_advance(interp, src, &src_iter);
const UINTVAL needed = dest_iter.bytepos + max_bytes;
if (needed > result->bufused) {
alloc_bytes = src_len - src_iter.charpos;
alloc_bytes = (UINTVAL)(alloc_bytes * avg_bytes);
alloc_bytes += needed;
Parrot_gc_reallocate_string_storage(interp, result, alloc_bytes);
result->bufused = alloc_bytes;
}
STRING_iter_set_and_advance(interp, result, &dest_iter, c);
}
result->bufused = dest_iter.bytepos;
return result;
}
PARROT_WARN_UNUSED_RESULT
INTVAL
encoding_find_cclass(PARROT_INTERP, INTVAL flags, ARGIN(const STRING *src),
UINTVAL offset, UINTVAL count)
{
ASSERT_ARGS(encoding_find_cclass)
String_iter iter;
UINTVAL codepoint;
UINTVAL end = offset + count;
static UINTVAL last_char_offset;
static String_iter cached_iter;
static STRING *last_string = 0;
if (last_string == src && offset > last_char_offset) {
iter = cached_iter;
STRING_iter_skip(interp, src, &iter, offset - last_char_offset);
}
else if (last_string == src && offset == last_char_offset) {
iter = cached_iter;
}
else {
STRING_ITER_INIT(interp, &iter);
STRING_iter_skip(interp, src, &iter, offset);
}
end = src->strlen < end ? src->strlen : end;
while (iter.charpos < end) {
codepoint = STRING_iter_get_and_advance(interp, src, &iter);
if (codepoint >= 256) {
if (u_iscclass(interp, codepoint, flags))
goto return_and_cache;
}
else {
if (Parrot_iso_8859_1_typetable[codepoint] & flags)
goto return_and_cache;
}
}
return end;
return_and_cache:
if (iter.charpos > 128) {
last_char_offset = iter.charpos;
cached_iter = iter;
last_string = (STRING*)PTR2INTVAL(src);
}
return iter.charpos - 1;
}
PARROT_WARN_UNUSED_RESULT
INTVAL
encoding_compare(PARROT_INTERP, ARGIN(const STRING *lhs), ARGIN(const STRING *rhs))
{
ASSERT_ARGS(encoding_compare)
String_iter l_iter, r_iter;
const UINTVAL l_len = STRING_length(lhs);
const UINTVAL r_len = STRING_length(rhs);
UINTVAL min_len;
if (r_len == 0)
return l_len != 0;
if (l_len == 0)
return -1;
STRING_ITER_INIT(interp, &l_iter);
STRING_ITER_INIT(interp, &r_iter);
min_len = l_len > r_len ? r_len : l_len;
while (l_iter.charpos < min_len) {
const UINTVAL cl = STRING_iter_get_and_advance(interp, lhs, &l_iter);
const UINTVAL cr = STRING_iter_get_and_advance(interp, rhs, &r_iter);
if (cl != cr)
return cl < cr ? -1 : 1;
}
if (l_len < r_len)
return -1;
if (l_len > r_len)
return 1;
return 0;
}
PARROT_CANNOT_RETURN_NULL
static STRING *
ascii_to_encoding(PARROT_INTERP, ARGIN(const STRING *src))
{
ASSERT_ARGS(ascii_to_encoding)
STRING *dest;
if (STRING_max_bytes_per_codepoint(src) == 1) {
unsigned char * const src_buf = (unsigned char *)src->strstart;
UINTVAL offs;
for (offs = 0; offs < src->strlen; ++offs) {
UINTVAL c = src_buf[offs];
if (c >= 0x80)
Parrot_ex_throw_from_c_args(interp, NULL, EXCEPTION_LOSSY_CONVERSION,
"lossy conversion to ascii");
}
dest = Parrot_str_clone(interp, src);
dest->encoding = Parrot_ascii_encoding_ptr;
}
else {
String_iter iter;
unsigned char *p;
const UINTVAL len = src->strlen;
dest = Parrot_str_new_init(interp, NULL, len,
Parrot_ascii_encoding_ptr, 0);
p = (unsigned char *)dest->strstart;
STRING_ITER_INIT(interp, &iter);
while (iter.charpos < len) {
const UINTVAL c = STRING_iter_get_and_advance(interp, src, &iter);
if (c >= 0x80)
Parrot_ex_throw_from_c_args(interp, NULL, EXCEPTION_LOSSY_CONVERSION,
"can't convert unicode string to ascii");
*p++ = c;
}
dest->bufused = len;
dest->strlen = len;
}
return dest;
}
PARROT_WARN_UNUSED_RESULT
PARROT_CANNOT_RETURN_NULL
static STRING *
ucs4_to_encoding(PARROT_INTERP, ARGIN(const STRING *src))
{
ASSERT_ARGS(ucs4_to_encoding)
const UINTVAL len = src->strlen;
UINTVAL i;
STRING *res;
utf32_t *ptr;
if (src->encoding == Parrot_ucs4_encoding_ptr)
return Parrot_str_copy(interp, src);
res = Parrot_str_new_init(interp, NULL, len * 4,
Parrot_ucs4_encoding_ptr, 0);
ptr = (utf32_t *)res->strstart;
if (STRING_max_bytes_per_codepoint(src) == 1) {
const unsigned char *s = (unsigned char *)src->strstart;
for (i = 0; i < len; i++) {
ptr[i] = s[i];
}
}
else {
String_iter iter;
STRING_ITER_INIT(interp, &iter);
while (iter.charpos < len) {
i = iter.charpos;
ptr[i] = STRING_iter_get_and_advance(interp, src, &iter);
}
}
res->strlen = len;
res->bufused = len * 4;
return res;
}
PARROT_WARN_UNUSED_RESULT
size_t
encoding_hash(PARROT_INTERP, ARGIN(const STRING *src), size_t hashval)
{
ASSERT_ARGS(encoding_hash)
DECL_CONST_CAST;
STRING * const s = PARROT_const_cast(STRING *, src);
String_iter iter;
STRING_ITER_INIT(interp, &iter);
while (iter.charpos < s->strlen) {
const UINTVAL c = STRING_iter_get_and_advance(interp, s, &iter);
hashval += hashval << 5;
hashval += c;
}
s->hashval = hashval;
return hashval;
}
PARROT_WARN_UNUSED_RESULT
INTVAL
encoding_find_not_cclass(PARROT_INTERP, INTVAL flags, ARGIN(const STRING *src),
UINTVAL offset, UINTVAL count)
{
ASSERT_ARGS(encoding_find_not_cclass)
String_iter iter;
UINTVAL codepoint;
UINTVAL end = offset + count;
int bit;
static UINTVAL last_char_offset;
static String_iter cached_iter;
static STRING *last_string = 0;
if (offset > src->strlen) {
/* XXX: Throw in this case? */
return offset + count;
}
if (last_string == src && offset > last_char_offset) {
iter = cached_iter;
STRING_iter_skip(interp, src, &iter, offset - last_char_offset);
}
else if (last_string == src && offset == last_char_offset) {
iter = cached_iter;
}
else {
STRING_ITER_INIT(interp, &iter);
if (offset)
STRING_iter_skip(interp, src, &iter, offset);
}
end = src->strlen < end ? src->strlen : end;
if (flags == enum_cclass_any)
return end;
while (iter.charpos < end) {
codepoint = STRING_iter_get_and_advance(interp, src, &iter);
if (codepoint >= 256) {
for (bit = enum_cclass_uppercase;
bit <= enum_cclass_word ; bit <<= 1) {
if ((bit & flags) && !u_iscclass(interp, codepoint, bit))
goto return_and_cache;
}
}
else {
if (!(Parrot_iso_8859_1_typetable[codepoint] & flags))
goto return_and_cache;
}
}
return end;
return_and_cache:
if (iter.charpos > 128) {
last_char_offset = iter.charpos;
cached_iter = iter;
last_string = (STRING*)PTR2INTVAL(src);
}
return iter.charpos - 1;
}