mirror/guile
mirror/guile
1
Fork 0
mirror of https://git.savannah.gnu.org/git/guile.git synced 2025-04-30 03:40:34 +02:00
Code Activity

optimize scm_from_utf8_stringn

Browse source 
* libguile/strings.c (decoding_error): Factor out of scm_from_stringn,
  properly handling errno.
  (scm_from_stringn): Adapt.
  (scm_from_utf8_stringn): Inline the conversion here, to avoid going
  through iconv.
This commit is contained in:
Andy Wingo 2011-10-25 18:18:39 +02:00
parent 25d50a051d
commit 41d1d984ae
1 changed files with 93 additions and 14 deletions
Download patch file Download diff file Expand all files Collapse all files

107
libguile/strings.c
View file

@ -1446,6 +1446,23 @@ scm_decoding_error (const char *subr, int err, const char *message, SCM port)
/* String conversion to/from C. */
static void
decoding_error (const char *func_name, int errno_save,
const char *str, size_t len)
{
/* Raise an error and pass the raw C string as a bytevector to the `throw'
handler. */
SCM bv;
signed char *buf;
buf = scm_gc_malloc_pointerless (len, "bytevector");
memcpy (buf, str, len);
bv = scm_c_take_gc_bytevector (buf, len, SCM_BOOL_F);
scm_decoding_error (func_name, errno_save,
"input locale conversion error", bv);
}
SCM
scm_from_stringn (const char *str, size_t len, const char *encoding,
scm_t_string_failed_conversion_handler handler)
@ -1481,19 +1498,7 @@ scm_from_stringn (const char *str, size_t len, const char *encoding,
NULL, &u32len);
if (SCM_UNLIKELY (u32 == NULL))
{
/* Raise an error and pass the raw C string as a bytevector to the `throw'
handler. */
SCM bv;
signed char *buf;
buf = scm_gc_malloc_pointerless (len, "bytevector");
memcpy (buf, str, len);
bv = scm_c_take_gc_bytevector (buf, len, SCM_BOOL_F);
scm_decoding_error (__func__, errno,
"input locale conversion error", bv);
}
decoding_error (__func__, errno, str, len);
i = 0;
while (i < u32len)
@ -1567,7 +1572,81 @@ scm_from_utf8_string (const char *str)
SCM
scm_from_utf8_stringn (const char *str, size_t len)
{
return scm_from_stringn (str, len, "UTF-8", SCM_FAILED_CONVERSION_ERROR);
size_t i, char_len;
const scm_t_uint8 *ustr = (const scm_t_uint8 *) str;
int ascii = 1, narrow = 1;
SCM res;
if (len == (size_t) -1)
len = strlen (str);
i = 0;
char_len = 0;
while (i < len)
{
if (ustr[i] <= 127)
{
char_len++;
i++;
}
else
{
ucs4_t c;
int nbytes;
ascii = 0;
nbytes = u8_mbtouc (&c, ustr + i, len - i);
if (nbytes < 0)
/* Bad UTF-8. */
decoding_error (__func__, errno, str, len);
if (c > 255)
narrow = 0;
char_len++;
i += nbytes;
}
}
if (ascii)
{
char *dst;
res = scm_i_make_string (char_len, &dst, 0);
memcpy (dst, str, len);
}
else if (narrow)
{
char *dst;
size_t j;
ucs4_t c;
res = scm_i_make_string (char_len, &dst, 0);
for (i = 0, j = 0; i < len; i++, j++)
{
i += u8_mbtouc_unsafe (&c, ustr + i, len - i);
dst[j] = (signed char) c;
}
}
else
{
scm_t_wchar *dst;
size_t j;
ucs4_t c;
res = scm_i_make_wide_string (char_len, &dst, 0);
for (i = 0, j = 0; i < len; i++, j++)
{
i += u8_mbtouc_unsafe (&c, ustr + i, len - i);
dst[j] = c;
}
}
return res;
}
SCM