#include <string.h>
#include <sys/stat.h>
#include <unistd.h>
+#include "u8-mbtouc.h"
static uint8_t *data;
static size_t n;
}
+static bool __attribute__((unused))
+all_utf8(const char *p_)
+{
+ const uint8_t *p = (const uint8_t *) p_;
+ size_t len = strlen ((char *) p);
+ for (size_t ofs = 0, mblen; ofs < len; ofs += mblen)
+ {
+ ucs4_t uc;
+
+ mblen = u8_mbtouc (&uc, p + ofs, len - ofs);
+ if ((uc < 32 && uc != '\n') || uc == 127 || uc == 0xfffd)
+ return false;
+ }
+ return true;
+}
+
static char *
get_string(const char *where)
{
match_byte(0);
match_byte(0);
- if (match_u32 (1))
- match_byte (0);
- else if (match_byte (1))
+ if (match_byte (1))
{
- match_byte (0);
- if (!match_u32 (2))
- match_u32_assert (1);
- match_byte (0);
+ if (!match_byte (0) && !match_byte (1))
+ match_byte_assert (2);
}
- else if (!match_u32(2))
- match_u32_assert (0);
+ else if (!match_byte (2))
+ match_byte_assert (0);
+ match_byte_assert (0);
+ match_byte_assert (0);
+ match_byte_assert (0);
int indx = get_u32();
int n_categories = get_u32();
printf ("<dimension index=\"%d\">\n", indx);
dump_value (stdout, 0);
- /* This byte is usually 0x02 but many other values have been spotted. */
+ /* This byte is usually 0 but many other values have been spotted. */
pos++;
if (!match_byte(0) && !match_byte(1))
match_byte_assert(0);
if (!match_byte(0x40) && !match_byte(0x20) && !match_byte(0x80) && !match_byte(0x10) && !match_byte(0x70))
match_byte_assert(0x50);
- if (!match_byte(0x41))
- match_byte_assert(0x51);
+ match_byte_assert(0x41);
if (!match_u32(0) && !match_u32(1))
match_u32_assert(2);
match_byte_assert(0);
/* OK, this seems really unlikely to be totally correct, but it matches my corpus... */
if (!match_u32(0) && !match_u32(2))
- match_u32_assert(0xfaad);
+ {
+ if (i == 7)
+ match_u32_assert(0xfaad);
+ else
+ match_u32_assert(0);
+ }
if (!match_u32(0) && !match_u32(1) && !match_u32(2))
match_u32_assert(3);
if (version > 1)
{
- /* These seem unlikely to be correct too. */
if (i != 3)
{
if (!match_u32(8))