luc14n0/glib
1
0
Fork 0
mirror of https://gitlab.gnome.org/GNOME/glib.git synced 2024-12-29 00:46:16 +01:00
Code Issues Packages Projects Releases Wiki Activity

Optimized branching in g_utf8_validate()

Browse Source 
The number of branches and logical operations can be reduced by
never producing a resulting wide character value to check its range.
Instead, individual bytes in the sequence are validated
depending on the branch taken on the basis of preceding bytes.
The syntax given in RFC 3629 is made use of.

https://bugzilla.gnome.org/show_bug.cgi?id=738504
This commit is contained in:
Mikhail Zabaluev 2014-10-14 01:18:57 +03:00 committed by Matthias Clasen
parent 5644ee5083
commit 3188b8ee79
1 changed files with 80 additions and 69 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

125
glib/gutf8.c
View File

@ -1442,20 +1442,18 @@ g_ucs4_to_utf16 (const gunichar *str,
return result; return result;
} }
#define CONTINUATION_CHAR \ #define VALIDATE_BYTE(mask, expect) \
G_STMT_START { \ G_STMT_START { \
if ((*(guchar *)p & 0xc0) != 0x80) /* 10xxxxxx */ \ if (G_UNLIKELY((*(guchar *)p & (mask)) != (expect))) \
goto error; \ goto error; \
val <<= 6; \
val |= (*(guchar *)p) & 0x3f; \
} G_STMT_END } G_STMT_END
/* see IETF RFC 3629 Section 4 */
static const gchar * static const gchar *
fast_validate (const char *str) fast_validate (const char *str)
{ {
gunichar val = 0;
gunichar min = 0;
const gchar *p; const gchar *p;
for (p = str; *p; p++) for (p = str; *p; p++)
@ -1467,45 +1465,52 @@ fast_validate (const char *str)
const gchar *last; const gchar *last;
last = p; last = p;
if ((*(guchar *)p & 0xe0) == 0xc0) /* 110xxxxx */ if (*(guchar *)p < 0xe0) /* 110xxxxx */
{ {
if (G_UNLIKELY ((*(guchar *)p & 0x1e) == 0)) if (G_UNLIKELY ((*(guchar *)p & 0x1e) == 0))
goto error; goto error;
p++;
if (G_UNLIKELY ((*(guchar *)p & 0xc0) != 0x80)) /* 10xxxxxx */
goto error;
} }
else else
{ {
if ((*(guchar *)p & 0xf0) == 0xe0) /* 1110xxxx */ if (*(guchar *)p < 0xf0) /* 1110xxxx */
{ {
min = (1 << 11); switch (*(guchar *)p++ & 0x0f)
val = *(guchar *)p & 0x0f; {
goto TWO_REMAINING; case 0:
VALIDATE_BYTE(0xe0, 0xa0); /* 0xa0 ... 0xbf */
break;
case 0x0d:
VALIDATE_BYTE(0xe0, 0x80); /* 0x80 ... 0x9f */
break;
default:
VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
} }
else if ((*(guchar *)p & 0xf8) == 0xf0) /* 11110xxx */ }
else if (*(guchar *)p < 0xf5) /* 11110xxx excluding out-of-range */
{ {
min = (1 << 16); switch (*(guchar *)p++ & 0x07)
val = *(guchar *)p & 0x07; {
case 0:
VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
if (G_UNLIKELY((*(guchar *)p & 0x30) == 0))
goto error;
break;
case 4:
VALIDATE_BYTE(0xf0, 0x80); /* 0x80 ... 0x8f */
break;
default:
VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
}
p++;
VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
} }
else else
goto error; goto error;
p++;
CONTINUATION_CHAR;
TWO_REMAINING:
p++;
CONTINUATION_CHAR;
p++;
CONTINUATION_CHAR;
if (G_UNLIKELY (val < min))
goto error;
if (G_UNLIKELY (!UNICODE_VALID(val)))
goto error;
} }
p++;
VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
continue; continue;
error: error:
@ -1521,8 +1526,6 @@ fast_validate_len (const char *str,
gssize max_len) gssize max_len)
{ {
gunichar val = 0;
gunichar min = 0;
const gchar *p; const gchar *p;
g_assert (max_len >= 0); g_assert (max_len >= 0);
@ -1536,53 +1539,61 @@ fast_validate_len (const char *str,
const gchar *last; const gchar *last;
last = p; last = p;
if ((*(guchar *)p & 0xe0) == 0xc0) /* 110xxxxx */ if (*(guchar *)p < 0xe0) /* 110xxxxx */
{ {
if (G_UNLIKELY (max_len - (p - str) < 2)) if (G_UNLIKELY (max_len - (p - str) < 2))
goto error; goto error;
if (G_UNLIKELY ((*(guchar *)p & 0x1e) == 0)) if (G_UNLIKELY ((*(guchar *)p & 0x1e) == 0))
goto error; goto error;
p++;
if (G_UNLIKELY ((*(guchar *)p & 0xc0) != 0x80)) /* 10xxxxxx */
goto error;
} }
else else
{ {
if ((*(guchar *)p & 0xf0) == 0xe0) /* 1110xxxx */ if (*(guchar *)p < 0xf0) /* 1110xxxx */
{ {
if (G_UNLIKELY (max_len - (p - str) < 3)) if (G_UNLIKELY (max_len - (p - str) < 3))
goto error; goto error;
min = (1 << 11); switch (*(guchar *)p++ & 0x0f)
val = *(guchar *)p & 0x0f; {
goto TWO_REMAINING; case 0:
VALIDATE_BYTE(0xe0, 0xa0); /* 0xa0 ... 0xbf */
break;
case 0x0d:
VALIDATE_BYTE(0xe0, 0x80); /* 0x80 ... 0x9f */
break;
default:
VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
} }
else if ((*(guchar *)p & 0xf8) == 0xf0) /* 11110xxx */ }
else if (*(guchar *)p < 0xf5) /* 11110xxx excluding out-of-range */
{ {
if (G_UNLIKELY (max_len - (p - str) < 4)) if (G_UNLIKELY (max_len - (p - str) < 4))
goto error; goto error;
min = (1 << 16); switch (*(guchar *)p++ & 0x07)
val = *(guchar *)p & 0x07; {
case 0:
VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
if (G_UNLIKELY((*(guchar *)p & 0x30) == 0))
goto error;
break;
case 4:
VALIDATE_BYTE(0xf0, 0x80); /* 0x80 ... 0x8f */
break;
default:
VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
}
p++;
VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
} }
else else
goto error; goto error;
p++;
CONTINUATION_CHAR;
TWO_REMAINING:
p++;
CONTINUATION_CHAR;
p++;
CONTINUATION_CHAR;
if (G_UNLIKELY (val < min))
goto error;
if (G_UNLIKELY (!UNICODE_VALID(val)))
goto error;
} }
p++;
VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
continue; continue;
error: error: