wchardatastart = wchardata = (unsigned int *)palloc((sourcedatalength+1)*sizeof(int));
// Based on pg_utf2wchar_with_len from wchar.c
- while (sourcedatalength > 0 && *sourcedata)
+ // Postgresql strings are not zero terminalted
+ while (sourcedatalength > 0)
{
if ((*sourcedata & 0x80) == 0)
{
c1 = *sourcedata++ & 0x1f;
c2 = *sourcedata++ & 0x3f;
*wchardata = (c1 << 6) | c2;
- wchardata++;
+ if (*wchardata < 65536) wchardata++;
sourcedatalength -= 2;
}
else if ((*sourcedata & 0xf0) == 0xe0)
c2 = *sourcedata++ & 0x3f;
c3 = *sourcedata++ & 0x3f;
*wchardata = (c1 << 12) | (c2 << 6) | c3;
- wchardata++;
+ if (*wchardata < 65536) wchardata++;
sourcedatalength -= 3;
}
else if ((*sourcedata & 0xf8) == 0xf0)
c3 = *sourcedata++ & 0x3f;
c4 = *sourcedata++ & 0x3f;
*wchardata = (c1 << 18) | (c2 << 12) | (c3 << 6) | c4;
- wchardata++;
+ if (*wchardata < 65536) wchardata++;
sourcedatalength -= 4;
}
else if ((*sourcedata & 0xfc) == 0xf8)
// table does not extend beyond 4 char long, just skip
if (sourcedatalength < 5) break;
sourcedatalength -= 5;
+ sourcedata += 5;
}
else if ((*sourcedata & 0xfe) == 0xfc)
{
// table does not extend beyond 4 char long, just skip
if (sourcedatalength < 6) break;
sourcedatalength -= 6;
+ sourcedata += 6;
}
else
{
// assume lenngth 1, silently drop bogus characters
sourcedatalength--;
+ sourcedata += 1;
}
}
*wchardata = 0;
str_replace(buffer, &len, &changes, " und ", 5, " ", 1, 0);
str_replace(buffer, &len, &changes, " en ", 4, " ", 1, 0);
str_replace(buffer, &len, &changes, " et ", 4, " ", 1, 0);
- str_replace(buffer, &len, &changes, " e ", 3, " ", 1, 0);
str_replace(buffer, &len, &changes, " y ", 3, " ", 1, 0);
// 'the' (and similar)