/*
* Mark character usage in *f based on the hexadecimal bitmap found in
* string s. A two-digit offset separated by a colon gives the initial
* character code. We have no way of knowing how many times each character
* is used or how many strings get created when showing the characters so
* we just estimate two usages per character and one string per pair of
* usages.
*/
static void
includechars(fontdesctype *f, char *s)
{
int b, c, d;
int l = strlen(s);
while (l > 0 && (s[l-1] == '\n' || s[l-1] == '\r'))
s[--l] = 0;
if (!ISXDIGIT(s[0]) || !ISXDIGIT(s[1]) || s[2]!=':'
|| strspn(s+3,"0123456789ABCDEFabcdef") < l-3) {
fprintf(stderr, "%s\n", s);
error("Bad syntax in included font usage table");
return;
}
c = (xdig[(int)(s[0])] << 4) + xdig[(int)(s[1])];
s += 2;
while (*++s) {
d = xdig[(int)*s];
for (b=8; b!=0; b>>=1) {
if ((d&b)!=0) {
pagecost ++;
(void) prescanchar(&f->chardesc[c]);
}
if (++c==256) return;
}
}
}
char *url_unescape_len(const char *string, int *olen) {
int alloc = (int)strlen(string)+1;
char *ns = (char*) xmalloc(alloc);
unsigned char in;
int strindex=0;
long hex;
while(--alloc > 0) {
in = *string;
if(('%' == in) && ISXDIGIT(string[1]) && ISXDIGIT(string[2])) {
char hexstr[3]; // '%XX'
hexstr[0] = string[1];
hexstr[1] = string[2];
hexstr[2] = 0;
hex = strtol(hexstr, NULL, 16);
in = (unsigned char)hex; /* hex is always < 256 */
string+=2;
alloc-=2;
}
ns[strindex++] = in;
string++;
}
ns[strindex]=0;
if(olen) *olen = strindex;
return ns;
}
/*
* Curl_urldecode() URL decodes the given string.
*
* Optionally detects control characters (byte codes lower than 32) in the
* data and rejects such data.
*
* Returns a pointer to a malloced string in *ostring with length given in
* *olen. If length == 0, the length is assumed to be strlen(string).
*
*/
CURLcode Curl_urldecode(struct SessionHandle *data,
const char *string, size_t length,
char **ostring, size_t *olen,
bool reject_ctrl)
{
size_t alloc = (length?length:strlen(string))+1;
char *ns = malloc(alloc);
unsigned char in;
size_t strindex=0;
unsigned long hex;
CURLcode res;
if(!ns)
return CURLE_OUT_OF_MEMORY;
while(--alloc > 0) {
in = *string;
if(('%' == in) && (alloc > 2) &&
ISXDIGIT(string[1]) && ISXDIGIT(string[2])) {
/* this is two hexadecimal digits following a '%' */
char hexstr[3];
char *ptr;
hexstr[0] = string[1];
hexstr[1] = string[2];
hexstr[2] = 0;
hex = strtoul(hexstr, &ptr, 16);
in = curlx_ultouc(hex); /* this long is never bigger than 255 anyway */
res = Curl_convert_from_network(data, &in, 1);
if(res) {
/* Curl_convert_from_network calls failf if unsuccessful */
free(ns);
return res;
}
string+=2;
alloc-=2;
}
if(reject_ctrl && (in < 0x20)) {
free(ns);
return CURLE_URL_MALFORMAT;
}
ns[strindex++] = in;
string++;
}
ns[strindex]=0; /* terminate it */
if(olen)
/* store output size */
*olen = strindex;
if(ostring)
/* store output string */
*ostring = ns;
return CURLE_OK;
}
bfd_size_type
compute_build_id_size (const char *style)
{
if (streq (style, "md5") || streq (style, "uuid"))
return 128 / 8;
if (streq (style, "sha1"))
return 160 / 8;
if (strneq (style, "0x", 2))
{
bfd_size_type size = 0;
/* ID is in string form (hex). Count the bytes. */
const char *id = style + 2;
do
{
if (ISXDIGIT (id[0]) && ISXDIGIT (id[1]))
{
++size;
id += 2;
}
else if (*id == '-' || *id == ':')
++id;
else
{
size = 0;
break;
}
} while (*id != '\0');
return size;
}
return 0;
}
/*
* Unescapes the given URL escaped string of given length. Returns a
* pointer to a malloced string with length given in *olen.
* If length == 0, the length is assumed to be strlen(string).
* If olen == NULL, no output length is stored.
*/
char *curl_easy_unescape(CURL *handle, const char *string, int length,
int *olen)
{
int alloc = (length?length:(int)strlen(string))+1;
char *ns = malloc(alloc);
unsigned char in;
int strindex=0;
long hex;
#ifndef CURL_DOES_CONVERSIONS
/* avoid compiler warnings */
(void)handle;
#endif
if( !ns )
return NULL;
while(--alloc > 0) {
in = *string;
if(('%' == in) && ISXDIGIT(string[1]) && ISXDIGIT(string[2])) {
/* this is two hexadecimal digits following a '%' */
char hexstr[3];
char *ptr;
hexstr[0] = string[1];
hexstr[1] = string[2];
hexstr[2] = 0;
hex = strtol(hexstr, &ptr, 16);
in = (unsigned char)hex; /* this long is never bigger than 255 anyway */
#ifdef CURL_DOES_CONVERSIONS
/* escape sequences are always in ASCII so convert them on non-ASCII hosts */
if(!handle ||
(Curl_convert_from_network(handle, &in, 1) != CURLE_OK)) {
/* Curl_convert_from_network calls failf if unsuccessful */
free(ns);
return NULL;
}
#endif /* CURL_DOES_CONVERSIONS */
string+=2;
alloc-=2;
}
ns[strindex++] = in;
string++;
}
ns[strindex]=0; /* terminate it */
if(olen)
/* store output size */
*olen = strindex;
return ns;
}
/*
* Unescapes the given URL escaped string of given length. Returns a
* pointer to a malloced string with length given in *olen.
* If length == 0, the length is assumed to be strlen(string).
* If olen == NULL, no output length is stored.
*/
char *curl_easy_unescape(CURL *handle, const char *string, int length,
int *olen)
{
int alloc = (length?length:(int)strlen(string))+1;
char *ns = malloc(alloc);
unsigned char in;
int strindex=0;
unsigned long hex;
CURLcode res;
if(!ns)
return NULL;
while(--alloc > 0) {
in = *string;
if(('%' == in) && ISXDIGIT(string[1]) && ISXDIGIT(string[2])) {
/* this is two hexadecimal digits following a '%' */
char hexstr[3];
char *ptr;
hexstr[0] = string[1];
hexstr[1] = string[2];
hexstr[2] = 0;
hex = strtoul(hexstr, &ptr, 16);
in = curlx_ultouc(hex); /* this long is never bigger than 255 anyway */
res = Curl_convert_from_network(handle, &in, 1);
if(res) {
/* Curl_convert_from_network calls failf if unsuccessful */
free(ns);
return NULL;
}
string+=2;
alloc-=2;
}
ns[strindex++] = in;
string++;
}
ns[strindex]=0; /* terminate it */
if(olen)
/* store output size */
*olen = strindex;
return ns;
}
/*-
*-----------------------------------------------------------------------
* CondCvtArg --
* Convert the given number into a double. If the number begins
* with 0x, it is interpreted as a hexadecimal integer
* and converted to a double from there. All other strings just have
* strtod called on them.
*
* Results:
* Sets 'value' to double value of string.
* Returns true if the string was a valid number, false o.w.
*
* Side Effects:
* Can change 'value' even if string is not a valid number.
*-----------------------------------------------------------------------
*/
static bool
CondCvtArg(const char *str, double *value)
{
if (*str == '0' && str[1] == 'x') {
long i;
for (str += 2, i = 0; *str; str++) {
int x;
if (ISDIGIT(*str))
x = *str - '0';
else if (ISXDIGIT(*str))
x = 10 + *str - (ISUPPER(*str) ? 'A' : 'a');
else
return false;
i = (i << 4) + x;
}
*value = (double) i;
return true;
}
else {
char *eptr;
*value = strtod(str, &eptr);
return *eptr == '\0';
}
}
static char *
convert_escape (struct type *type, const char *dest_charset,
char *p, char *limit, struct obstack *output)
{
/* Skip the backslash. */
ADVANCE;
switch (*p)
{
case '\\':
obstack_1grow (output, '\\');
++p;
break;
case 'x':
ADVANCE;
if (!ISXDIGIT (*p))
error (_("\\x used with no following hex digits."));
p = convert_hex (type, p, limit, output);
break;
case '0':
case '1':
case '2':
case '3':
case '4':
case '5':
case '6':
case '7':
p = convert_octal (type, p, limit, output);
break;
case 'u':
case 'U':
{
int length = *p == 'u' ? 4 : 8;
ADVANCE;
if (!ISXDIGIT (*p))
error (_("\\u used with no following hex digits"));
p = convert_ucn (p, limit, dest_charset, output, length);
}
}
return p;
}
static char *
_util_unescape(const char *string, int length)
{
int alloc = (length?length:(int)strlen(string))+1;
char *ns = malloc(alloc);
unsigned char in;
int strindex=0;
unsigned long hex;
if( !ns )
return NULL;
while(--alloc > 0) {
in = *string;
if(('%' == in) && ISXDIGIT(string[1]) && ISXDIGIT(string[2])) {
/* this is two hexadecimal digits following a '%' */
char hexstr[3];
char *ptr;
hexstr[0] = string[1];
hexstr[1] = string[2];
hexstr[2] = 0;
hex = strtoul(hexstr, &ptr, 16);
in = (unsigned char)(hex & (unsigned long) 0xFF);
// in = ultouc(hex); /* this long is never bigger than 255 anyway */
string+=2;
alloc-=2;
}
ns[strindex++] = in;
string++;
}
ns[strindex]=0; /* terminate it */
return ns;
}
void decode_string (char* s)
{
char *p = s;
for (; *s; s++, p++)
{
if (*s != '%')
*p = *s;
else
{
/* Do nothing if at the end of the string, or if the chars
are not hex-digits. */
if (!*(s + 1) || !*(s + 2)
|| !(ISXDIGIT (*(s + 1)) && ISXDIGIT (*(s + 2))))
{
*p = *s;
continue;
}
*p = (ASC2HEXD (*(s + 1)) << 4) + ASC2HEXD (*(s + 2));
s += 2;
}
}
*p = '\0';
}
/*
* -------------------------------------------------------
* Check if string is a Hex number
*
* Return:
* TRUE - identifier
* FALSE - otherwise
* -------------------------------------------------------
*/
int cli_is_hex(char *p)
{
if (('+' == *p) || ('-' == *p))
p++;
if (('0' == *p) && ('X' == toupper(*(p + 1))))
{
p = p + 2;
}
while (*p && ISXDIGIT(*p))
p++;
return ((*p) ? FALSE : TRUE);
}
static char *
convert_hex (struct type *type, char *p,
char *limit, struct obstack *output)
{
unsigned long value = 0;
while (p < limit && ISXDIGIT (*p))
{
value = 16 * value + host_hex_value (*p);
++p;
}
emit_numeric_character (type, value, output);
return p;
}
static char *
convert_ucn (char *p, char *limit, const char *dest_charset,
struct obstack *output, int length)
{
unsigned long result = 0;
gdb_byte data[4];
int i;
for (i = 0; i < length && p < limit && ISXDIGIT (*p); ++i, ++p)
result = (result << 4) + host_hex_value (*p);
for (i = 3; i >= 0; --i)
{
data[i] = result & 0xff;
result >>= 8;
}
convert_between_encodings ("UTF-32BE", dest_charset, data,
4, 4, output, translit_none);
return p;
}
/*
* static int
* inet_net_pton_ipv4(src, dst, size)
* convert IPv4 network number from presentation to network format.
* accepts hex octets, hex strings, decimal octets, and /CIDR.
* "size" is in bytes and describes "dst".
* return:
* number of bits, either imputed classfully or specified with /CIDR,
* or -1 if some failure occurred (check errno). ENOENT means it was
* not an IPv4 network specification.
* note:
* network byte order assumed. this means 192.5.5.240/28 has
* 0b11110000 in its fourth octet.
* note:
* On Windows we store the error in the thread errno, not
* in the winsock error code. This is to avoid loosing the
* actual last winsock error. So use macro ERRNO to fetch the
* errno this funtion sets when returning (-1), not SOCKERRNO.
* author:
* Paul Vixie (ISC), June 1996
*/
static int
inet_net_pton_ipv4(const char *src, unsigned char *dst, size_t size)
{
static const char xdigits[] = "0123456789abcdef";
static const char digits[] = "0123456789";
int n, ch, tmp = 0, dirty, bits;
const unsigned char *odst = dst;
ch = *src++;
if (ch == '0' && (src[0] == 'x' || src[0] == 'X')
&& ISXDIGIT(src[1])) {
/* Hexadecimal: Eat nybble string. */
if (!size)
goto emsgsize;
dirty = 0;
src++; /* skip x or X. */
while ((ch = *src++) != '\0' && ISXDIGIT(ch)) {
if (ISUPPER(ch))
ch = tolower(ch);
n = (int)(strchr(xdigits, ch) - xdigits);
if (dirty == 0)
tmp = n;
else
tmp = (tmp << 4) | n;
if (++dirty == 2) {
if (!size--)
goto emsgsize;
*dst++ = (unsigned char) tmp;
dirty = 0;
}
}
if (dirty) { /* Odd trailing nybble? */
if (!size--)
goto emsgsize;
*dst++ = (unsigned char) (tmp << 4);
}
} else if (ISDIGIT(ch)) {
/* Decimal: eat dotted digit string. */
for (;;) {
tmp = 0;
do {
n = (int)(strchr(digits, ch) - digits);
tmp *= 10;
tmp += n;
if (tmp > 255)
goto enoent;
} while ((ch = *src++) != '\0' &&
ISDIGIT(ch));
if (!size--)
goto emsgsize;
*dst++ = (unsigned char) tmp;
if (ch == '\0' || ch == '/')
break;
if (ch != '.')
goto enoent;
ch = *src++;
if (!ISDIGIT(ch))
goto enoent;
}
} else
goto enoent;
bits = -1;
if (ch == '/' &&
ISDIGIT(src[0]) && dst > odst) {
/* CIDR width specifier. Nothing can follow it. */
ch = *src++; /* Skip over the /. */
bits = 0;
do {
n = (int)(strchr(digits, ch) - digits);
bits *= 10;
bits += n;
} while ((ch = *src++) != '\0' && ISDIGIT(ch));
if (ch != '\0')
goto enoent;
if (bits > 32)
goto emsgsize;
}
/* Firey death and destruction unless we prefetched EOS. */
if (ch != '\0')
goto enoent;
/* If nothing was written to the destination, we found no address. */
if (dst == odst)
goto enoent;
/* If no CIDR spec was given, infer width from net class. */
if (bits == -1) {
if (*odst >= 240) /* Class E */
bits = 32;
else if (*odst >= 224) /* Class D */
bits = 8;
else if (*odst >= 192) /* Class C */
bits = 24;
else if (*odst >= 128) /* Class B */
bits = 16;
else /* Class A */
bits = 8;
/* If imputed mask is narrower than specified octets, widen. */
if (bits < ((dst - odst) * 8))
bits = (int)(dst - odst) * 8;
/*
* If there are no additional bits specified for a class D
* address adjust bits to 4.
*/
if (bits == 8 && *odst == 224)
bits = 4;
}
/* Extend network to cover the actual mask. */
while (bits > ((dst - odst) * 8)) {
if (!size--)
goto emsgsize;
*dst++ = '\0';
}
return (bits);
enoent:
SET_ERRNO(ENOENT);
return (-1);
emsgsize:
SET_ERRNO(EMSGSIZE);
return (-1);
}
int
is_valid_ipv6_address (const char *str, const char *end)
{
/* Use lower-case for these to avoid clash with system headers. */
enum {
ns_inaddrsz = 4,
ns_in6addrsz = 16,
ns_int16sz = 2
};
const char *curtok;
int tp;
const char *colonp;
int saw_xdigit;
unsigned int val;
tp = 0;
colonp = NULL;
if (str == end)
return 0;
/* Leading :: requires some special handling. */
if (*str == ':')
{
++str;
if (str == end || *str != ':')
return 0;
}
curtok = str;
saw_xdigit = 0;
val = 0;
while (str < end)
{
int ch = *str++;
/* if ch is a number, add it to val. */
if (ISXDIGIT (ch))
{
val <<= 4;
val |= XDIGIT_TO_NUM (ch);
if (val > 0xffff)
return 0;
saw_xdigit = 1;
continue;
}
/* if ch is a colon ... */
if (ch == ':')
{
curtok = str;
if (saw_xdigit == 0)
{
if (colonp != NULL)
return 0;
colonp = str + tp;
continue;
}
else if (str == end)
return 0;
if (tp > ns_in6addrsz - ns_int16sz)
return 0;
tp += ns_int16sz;
saw_xdigit = 0;
val = 0;
continue;
}
/* if ch is a dot ... */
if (ch == '.' && (tp <= ns_in6addrsz - ns_inaddrsz)
&& is_valid_ipv4_address (curtok, end) == 1)
{
tp += ns_inaddrsz;
saw_xdigit = 0;
break;
}
return 0;
}
if (saw_xdigit == 1)
{
if (tp > ns_in6addrsz - ns_int16sz)
return 0;
tp += ns_int16sz;
}
if (colonp != NULL)
{
if (tp == ns_in6addrsz)
return 0;
tp = ns_in6addrsz;
}
if (tp != ns_in6addrsz)
return 0;
return 1;
}
/* Match pattern "p" against the a virtually-joined string consisting
* of "text" and any strings in array "a". */
static int dowild(const uchar *p, const uchar *text, const uchar*const *a)
{
uchar p_ch;
#ifdef WILD_TEST_ITERATIONS
wildmatch_iteration_count++;
#endif
for ( ; (p_ch = *p) != '\0'; text++, p++) {
int matched, special;
uchar t_ch, prev_ch;
while ((t_ch = *text) == '\0') {
if (*a == NULL) {
if (p_ch != '*')
return ABORT_ALL;
break;
}
text = *a++;
}
if (force_lower_case && ISUPPER(t_ch))
t_ch = tolower(t_ch);
switch (p_ch) {
case '\\':
/* Literal match with following character. Note that the test
* in "default" handles the p[1] == '\0' failure case. */
p_ch = *++p;
/* FALLTHROUGH */
default:
if (t_ch != p_ch)
return FALSE;
continue;
case '?':
/* Match anything but '/'. */
if (t_ch == '/')
return FALSE;
continue;
case '*':
if (*++p == '*') {
while (*++p == '*') {}
special = TRUE;
} else
special = FALSE;
if (*p == '\0') {
/* Trailing "**" matches everything. Trailing "*" matches
* only if there are no more slash characters. */
if (!special) {
do {
if (strchr((char*)text, '/') != NULL)
return FALSE;
} while ((text = *a++) != NULL);
}
return TRUE;
}
while (1) {
if (t_ch == '\0') {
if ((text = *a++) == NULL)
break;
t_ch = *text;
continue;
}
if ((matched = dowild(p, text, a)) != FALSE) {
if (!special || matched != ABORT_TO_STARSTAR)
return matched;
} else if (!special && t_ch == '/')
return ABORT_TO_STARSTAR;
t_ch = *++text;
}
return ABORT_ALL;
case '[':
p_ch = *++p;
#ifdef NEGATE_CLASS2
if (p_ch == NEGATE_CLASS2)
p_ch = NEGATE_CLASS;
#endif
/* Assign literal TRUE/FALSE because of "matched" comparison. */
special = p_ch == NEGATE_CLASS? TRUE : FALSE;
if (special) {
/* Inverted character class. */
p_ch = *++p;
}
prev_ch = 0;
matched = FALSE;
do {
if (!p_ch)
return ABORT_ALL;
if (p_ch == '\\') {
p_ch = *++p;
if (!p_ch)
return ABORT_ALL;
if (t_ch == p_ch)
matched = TRUE;
} else if (p_ch == '-' && prev_ch && p[1] && p[1] != ']') {
p_ch = *++p;
if (p_ch == '\\') {
p_ch = *++p;
if (!p_ch)
return ABORT_ALL;
}
if (t_ch <= p_ch && t_ch >= prev_ch)
matched = TRUE;
p_ch = 0; /* This makes "prev_ch" get set to 0. */
} else if (p_ch == '[' && p[1] == ':') {
const uchar *s;
int i;
for (s = p += 2; (p_ch = *p) && p_ch != ']'; p++) {}
if (!p_ch)
return ABORT_ALL;
i = p - s - 1;
if (i < 0 || p[-1] != ':') {
/* Didn't find ":]", so treat like a normal set. */
p = s - 2;
p_ch = '[';
if (t_ch == p_ch)
matched = TRUE;
continue;
}
if (CC_EQ(s,i, "alnum")) {
if (ISALNUM(t_ch))
matched = TRUE;
} else if (CC_EQ(s,i, "alpha")) {
if (ISALPHA(t_ch))
matched = TRUE;
} else if (CC_EQ(s,i, "blank")) {
if (ISBLANK(t_ch))
matched = TRUE;
} else if (CC_EQ(s,i, "cntrl")) {
if (ISCNTRL(t_ch))
matched = TRUE;
} else if (CC_EQ(s,i, "digit")) {
if (ISDIGIT(t_ch))
matched = TRUE;
} else if (CC_EQ(s,i, "graph")) {
if (ISGRAPH(t_ch))
matched = TRUE;
} else if (CC_EQ(s,i, "lower")) {
if (ISLOWER(t_ch))
matched = TRUE;
} else if (CC_EQ(s,i, "print")) {
if (ISPRINT(t_ch))
matched = TRUE;
} else if (CC_EQ(s,i, "punct")) {
if (ISPUNCT(t_ch))
matched = TRUE;
} else if (CC_EQ(s,i, "space")) {
if (ISSPACE(t_ch))
matched = TRUE;
} else if (CC_EQ(s,i, "upper")) {
if (ISUPPER(t_ch))
matched = TRUE;
} else if (CC_EQ(s,i, "xdigit")) {
if (ISXDIGIT(t_ch))
matched = TRUE;
} else /* malformed [:class:] string */
return ABORT_ALL;
p_ch = 0; /* This makes "prev_ch" get set to 0. */
} else if (t_ch == p_ch)
matched = TRUE;
} while (prev_ch = p_ch, (p_ch = *++p) != ']');
if (matched == special || t_ch == '/')
return FALSE;
continue;
}
}
do {
if (*text)
return FALSE;
} while ((text = *a++) != NULL);
return TRUE;
}
bfd_vma
bfd_scan_vma (const char *string, const char **end, int base)
{
bfd_vma value;
bfd_vma cutoff;
unsigned int cutlim;
int overflow;
/* Let the host do it if possible. */
if (sizeof (bfd_vma) <= sizeof (unsigned long))
return strtoul (string, (char **) end, base);
if (sizeof (bfd_vma) <= sizeof (unsigned long long))
return strtoull (string, (char **) end, base);
if (base == 0)
{
if (string[0] == '0')
{
if ((string[1] == 'x') || (string[1] == 'X'))
base = 16;
else
base = 8;
}
}
if ((base < 2) || (base > 36))
base = 10;
if (base == 16
&& string[0] == '0'
&& (string[1] == 'x' || string[1] == 'X')
&& ISXDIGIT (string[2]))
{
string += 2;
}
cutoff = (~ (bfd_vma) 0) / (bfd_vma) base;
cutlim = (~ (bfd_vma) 0) % (bfd_vma) base;
value = 0;
overflow = 0;
while (1)
{
unsigned int digit;
digit = *string;
if (ISDIGIT (digit))
digit = digit - '0';
else if (ISALPHA (digit))
digit = TOUPPER (digit) - 'A' + 10;
else
break;
if (digit >= (unsigned int) base)
break;
if (value > cutoff || (value == cutoff && digit > cutlim))
overflow = 1;
value = value * base + digit;
++string;
}
if (overflow)
value = ~ (bfd_vma) 0;
if (end != NULL)
*end = string;
return value;
}
static int isxdigit(int c)
{
return (ISXDIGIT(c));
}
static int
internal_function
internal_fnmatch (const char *pattern, const char *string, int no_leading_period, int flags)
{
register const char *p = pattern, *n = string;
register unsigned char c;
/* Note that this evaluates C many times. */
# ifdef _LIBC
# define FOLD(c) ((flags & FNM_CASEFOLD) ? tolower (c) : (c))
# else
# define FOLD(c) ((flags & FNM_CASEFOLD) && ISUPPER (c) ? tolower (c) : (c))
# endif
while ((c = *p++) != '\0')
{
c = FOLD (c);
switch (c)
{
case '?':
if (*n == '\0')
return FNM_NOMATCH;
else if (*n == '/' && (flags & FNM_FILE_NAME))
return FNM_NOMATCH;
else if (*n == '.' && no_leading_period
&& (n == string
|| (n[-1] == '/' && (flags & FNM_FILE_NAME))))
return FNM_NOMATCH;
break;
case '\\':
if (!(flags & FNM_NOESCAPE))
{
c = *p++;
if (c == '\0')
/* Trailing \ loses. */
return FNM_NOMATCH;
c = FOLD (c);
}
if (FOLD ((unsigned char) *n) != c)
return FNM_NOMATCH;
break;
case '*':
if (*n == '.' && no_leading_period
&& (n == string
|| (n[-1] == '/' && (flags & FNM_FILE_NAME))))
return FNM_NOMATCH;
for (c = *p++; c == '?' || c == '*'; c = *p++)
{
if (*n == '/' && (flags & FNM_FILE_NAME))
/* A slash does not match a wildcard under FNM_FILE_NAME. */
return FNM_NOMATCH;
else if (c == '?')
{
/* A ? needs to match one character. */
if (*n == '\0')
/* There isn't another character; no match. */
return FNM_NOMATCH;
else
/* One character of the string is consumed in matching
this ? wildcard, so *??? won't match if there are
less than three characters. */
++n;
}
}
if (c == '\0')
/* The wildcard(s) is/are the last element of the pattern.
If the name is a file name and contains another slash
this does mean it cannot match. */
return ((flags & FNM_FILE_NAME) && strchr (n, '/') != NULL
? FNM_NOMATCH : 0);
else
{
const char *endp;
endp = __strchrnul (n, (flags & FNM_FILE_NAME) ? '/' : '\0');
if (c == '[')
{
int flags2 = ((flags & FNM_FILE_NAME)
? flags : (flags & ~FNM_PERIOD));
for (--p; n < endp; ++n)
if (internal_fnmatch (p, n,
(no_leading_period
&& (n == string
|| (n[-1] == '/'
&& (flags
& FNM_FILE_NAME)))),
flags2)
== 0)
return 0;
}
else if (c == '/' && (flags & FNM_FILE_NAME))
{
while (*n != '\0' && *n != '/')
++n;
if (*n == '/'
&& (internal_fnmatch (p, n + 1, flags & FNM_PERIOD,
flags) == 0))
return 0;
}
else
{
int flags2 = ((flags & FNM_FILE_NAME)
? flags : (flags & ~FNM_PERIOD));
if (c == '\\' && !(flags & FNM_NOESCAPE))
c = *p;
c = FOLD (c);
for (--p; n < endp; ++n)
if (FOLD ((unsigned char) *n) == c
&& (internal_fnmatch (p, n,
(no_leading_period
&& (n == string
|| (n[-1] == '/'
&& (flags
& FNM_FILE_NAME)))),
flags2) == 0))
return 0;
}
}
/* If we come here no match is possible with the wildcard. */
return FNM_NOMATCH;
case '[':
{
/* Nonzero if the sense of the character class is inverted. */
static int posixly_correct;
register int not;
char cold;
if (posixly_correct == 0)
posixly_correct = getenv ("POSIXLY_CORRECT") != NULL ? 1 : -1;
if (*n == '\0')
return FNM_NOMATCH;
if (*n == '.' && no_leading_period && (n == string
|| (n[-1] == '/'
&& (flags
& FNM_FILE_NAME))))
return FNM_NOMATCH;
if (*n == '/' && (flags & FNM_FILE_NAME))
/* `/' cannot be matched. */
return FNM_NOMATCH;
not = (*p == '!' || (posixly_correct < 0 && *p == '^'));
if (not)
++p;
c = *p++;
for (;;)
{
unsigned char fn = FOLD ((unsigned char) *n);
if (!(flags & FNM_NOESCAPE) && c == '\\')
{
if (*p == '\0')
return FNM_NOMATCH;
c = FOLD ((unsigned char) *p);
++p;
if (c == fn)
goto matched;
}
else if (c == '[' && *p == ':')
{
/* Leave room for the null. */
char str[CHAR_CLASS_MAX_LENGTH + 1];
size_t c1 = 0;
# if defined _LIBC || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H)
wctype_t wt;
# endif
const char *startp = p;
for (;;)
{
if (c1 > CHAR_CLASS_MAX_LENGTH)
/* The name is too long and therefore the pattern
is ill-formed. */
return FNM_NOMATCH;
c = *++p;
if (c == ':' && p[1] == ']')
{
p += 2;
break;
}
if (c < 'a' || c >= 'z')
{
/* This cannot possibly be a character class name.
Match it as a normal range. */
p = startp;
c = '[';
goto normal_bracket;
}
str[c1++] = c;
}
str[c1] = '\0';
# if defined _LIBC || (defined HAVE_WCTYPE_H && defined HAVE_WCHAR_H)
wt = IS_CHAR_CLASS (str);
if (wt == 0)
/* Invalid character class name. */
return FNM_NOMATCH;
if (__iswctype (__btowc ((unsigned char) *n), wt))
goto matched;
# else
if ((STREQ (str, "alnum") && ISALNUM ((unsigned char) *n))
|| (STREQ (str, "alpha") && ISALPHA ((unsigned char) *n))
|| (STREQ (str, "blank") && ISBLANK ((unsigned char) *n))
|| (STREQ (str, "cntrl") && ISCNTRL ((unsigned char) *n))
|| (STREQ (str, "digit") && ISDIGIT ((unsigned char) *n))
|| (STREQ (str, "graph") && ISGRAPH ((unsigned char) *n))
|| (STREQ (str, "lower") && ISLOWER ((unsigned char) *n))
|| (STREQ (str, "print") && ISPRINT ((unsigned char) *n))
|| (STREQ (str, "punct") && ISPUNCT ((unsigned char) *n))
|| (STREQ (str, "space") && ISSPACE ((unsigned char) *n))
|| (STREQ (str, "upper") && ISUPPER ((unsigned char) *n))
|| (STREQ (str, "xdigit") && ISXDIGIT ((unsigned char) *n)))
goto matched;
# endif
}
else if (c == '\0')
/* [ (unterminated) loses. */
return FNM_NOMATCH;
else
{
normal_bracket:
if (FOLD (c) == fn)
goto matched;
cold = c;
c = *p++;
if (c == '-' && *p != ']')
{
/* It is a range. */
unsigned char cend = *p++;
if (!(flags & FNM_NOESCAPE) && cend == '\\')
cend = *p++;
if (cend == '\0')
return FNM_NOMATCH;
if (cold <= fn && fn <= FOLD (cend))
goto matched;
c = *p++;
}
}
if (c == ']')
break;
}
if (!not)
return FNM_NOMATCH;
break;
matched:
/* Skip the rest of the [...] that already matched. */
while (c != ']')
{
if (c == '\0')
/* [... (unterminated) loses. */
return FNM_NOMATCH;
c = *p++;
if (!(flags & FNM_NOESCAPE) && c == '\\')
{
if (*p == '\0')
return FNM_NOMATCH;
/* XXX 1003.2d11 is unclear if this is right. */
++p;
}
else if (c == '[' && *p == ':')
{
do
if (*++p == '\0')
return FNM_NOMATCH;
while (*p != ':' || p[1] == ']');
p += 2;
c = *p;
}
}
if (not)
return FNM_NOMATCH;
}
break;
default:
if (c != FOLD ((unsigned char) *n))
return FNM_NOMATCH;
}
++n;
}
if (*n == '\0')
return 0;
if ((flags & FNM_LEADING_DIR) && *n == '/')
/* The FNM_LEADING_DIR flag says that "foo*" matches "foobar/frobozz". */
return 0;
return FNM_NOMATCH;
# undef FOLD
}
bfd_boolean
generate_build_id (bfd *abfd,
const char *style,
checksum_fn checksum_contents,
unsigned char *id_bits,
int size ATTRIBUTE_UNUSED)
{
if (streq (style, "md5"))
{
struct md5_ctx ctx;
md5_init_ctx (&ctx);
if (!(*checksum_contents) (abfd, (sum_fn) &md5_process_bytes, &ctx))
return FALSE;
md5_finish_ctx (&ctx, id_bits);
}
else if (streq (style, "sha1"))
{
struct sha1_ctx ctx;
sha1_init_ctx (&ctx);
if (!(*checksum_contents) (abfd, (sum_fn) &sha1_process_bytes, &ctx))
return FALSE;
sha1_finish_ctx (&ctx, id_bits);
}
else if (streq (style, "uuid"))
{
#ifndef __MINGW32__
int n;
int fd = open ("/dev/urandom", O_RDONLY);
if (fd < 0)
return FALSE;
n = read (fd, id_bits, size);
close (fd);
if (n < size)
return FALSE;
#else /* __MINGW32__ */
typedef RPC_STATUS (RPC_ENTRY * UuidCreateFn) (UUID *);
UUID uuid;
UuidCreateFn uuid_create = 0;
HMODULE rpc_library = LoadLibrary ("rpcrt4.dll");
if (!rpc_library)
return FALSE;
uuid_create = (UuidCreateFn) (void (WINAPI *)(void)) GetProcAddress (rpc_library, "UuidCreate");
if (!uuid_create)
{
FreeLibrary (rpc_library);
return FALSE;
}
if (uuid_create (&uuid) != RPC_S_OK)
{
FreeLibrary (rpc_library);
return FALSE;
}
FreeLibrary (rpc_library);
memcpy (id_bits, &uuid,
(size_t) size < sizeof (UUID) ? (size_t) size : sizeof (UUID));
#endif /* __MINGW32__ */
}
else if (strneq (style, "0x", 2))
{
/* ID is in string form (hex). Convert to bits. */
const char *id = style + 2;
size_t n = 0;
do
{
if (ISXDIGIT (id[0]) && ISXDIGIT (id[1]))
{
id_bits[n] = read_hex (*id++) << 4;
id_bits[n++] |= read_hex (*id++);
}
else if (*id == '-' || *id == ':')
++id;
else
abort (); /* Should have been validated earlier. */
}
while (*id != '\0');
}
else
abort (); /* Should have been validated earlier. */
return TRUE;
}
static int
internal_function
FCT (const CHAR *pattern, const CHAR *string, const CHAR *string_end,
bool no_leading_period, int flags)
{
register const CHAR *p = pattern, *n = string;
register UCHAR c;
#ifdef _LIBC
# if WIDE_CHAR_VERSION
const char *collseq = (const char *)
_NL_CURRENT(LC_COLLATE, _NL_COLLATE_COLLSEQWC);
# else
const UCHAR *collseq = (const UCHAR *)
_NL_CURRENT(LC_COLLATE, _NL_COLLATE_COLLSEQMB);
# endif
#endif
while ((c = *p++) != L_('\0'))
{
bool new_no_leading_period = false;
c = FOLD (c);
switch (c)
{
case L_('?'):
if (__builtin_expect (flags & FNM_EXTMATCH, 0) && *p == '(')
{
int res;
res = EXT (c, p, n, string_end, no_leading_period,
flags);
if (res != -1)
return res;
}
if (n == string_end)
return FNM_NOMATCH;
else if (*n == L_('/') && (flags & FNM_FILE_NAME))
return FNM_NOMATCH;
else if (*n == L_('.') && no_leading_period)
return FNM_NOMATCH;
break;
case L_('\\'):
if (!(flags & FNM_NOESCAPE))
{
c = *p++;
if (c == L_('\0'))
/* Trailing \ loses. */
return FNM_NOMATCH;
c = FOLD (c);
}
if (n == string_end || FOLD ((UCHAR) *n) != c)
return FNM_NOMATCH;
break;
case L_('*'):
if (__builtin_expect (flags & FNM_EXTMATCH, 0) && *p == '(')
{
int res;
res = EXT (c, p, n, string_end, no_leading_period,
flags);
if (res != -1)
return res;
}
if (n != string_end && *n == L_('.') && no_leading_period)
return FNM_NOMATCH;
for (c = *p++; c == L_('?') || c == L_('*'); c = *p++)
{
if (*p == L_('(') && (flags & FNM_EXTMATCH) != 0)
{
const CHAR *endp = END (p);
if (endp != p)
{
/* This is a pattern. Skip over it. */
p = endp;
continue;
}
}
if (c == L_('?'))
{
/* A ? needs to match one character. */
if (n == string_end)
/* There isn't another character; no match. */
return FNM_NOMATCH;
else if (*n == L_('/')
&& __builtin_expect (flags & FNM_FILE_NAME, 0))
/* A slash does not match a wildcard under
FNM_FILE_NAME. */
return FNM_NOMATCH;
else
/* One character of the string is consumed in matching
this ? wildcard, so *??? won't match if there are
less than three characters. */
++n;
}
}
if (c == L_('\0'))
/* The wildcard(s) is/are the last element of the pattern.
If the name is a file name and contains another slash
this means it cannot match, unless the FNM_LEADING_DIR
flag is set. */
{
int result = (flags & FNM_FILE_NAME) == 0 ? 0 : FNM_NOMATCH;
if (flags & FNM_FILE_NAME)
{
if (flags & FNM_LEADING_DIR)
result = 0;
else
{
if (MEMCHR (n, L_('/'), string_end - n) == NULL)
result = 0;
}
}
return result;
}
else
{
const CHAR *endp;
endp = MEMCHR (n, (flags & FNM_FILE_NAME) ? L_('/') : L_('\0'),
string_end - n);
if (endp == NULL)
endp = string_end;
if (c == L_('[')
|| (__builtin_expect (flags & FNM_EXTMATCH, 0) != 0
&& (c == L_('@') || c == L_('+') || c == L_('!'))
&& *p == L_('(')))
{
int flags2 = ((flags & FNM_FILE_NAME)
? flags : (flags & ~FNM_PERIOD));
bool no_leading_period2 = no_leading_period;
for (--p; n < endp; ++n, no_leading_period2 = false)
if (FCT (p, n, string_end, no_leading_period2, flags2)
== 0)
return 0;
}
else if (c == L_('/') && (flags & FNM_FILE_NAME))
{
while (n < string_end && *n != L_('/'))
++n;
if (n < string_end && *n == L_('/')
&& (FCT (p, n + 1, string_end, flags & FNM_PERIOD, flags)
== 0))
return 0;
}
else
{
int flags2 = ((flags & FNM_FILE_NAME)
? flags : (flags & ~FNM_PERIOD));
int no_leading_period2 = no_leading_period;
if (c == L_('\\') && !(flags & FNM_NOESCAPE))
c = *p;
c = FOLD (c);
for (--p; n < endp; ++n, no_leading_period2 = false)
if (FOLD ((UCHAR) *n) == c
&& (FCT (p, n, string_end, no_leading_period2, flags2)
== 0))
return 0;
}
}
/* If we come here no match is possible with the wildcard. */
return FNM_NOMATCH;
case L_('['):
{
/* Nonzero if the sense of the character class is inverted. */
register bool not;
CHAR cold;
UCHAR fn;
if (posixly_correct == 0)
posixly_correct = getenv ("POSIXLY_CORRECT") != NULL ? 1 : -1;
if (n == string_end)
return FNM_NOMATCH;
if (*n == L_('.') && no_leading_period)
return FNM_NOMATCH;
if (*n == L_('/') && (flags & FNM_FILE_NAME))
/* `/' cannot be matched. */
return FNM_NOMATCH;
not = (*p == L_('!') || (posixly_correct < 0 && *p == L_('^')));
if (not)
++p;
fn = FOLD ((UCHAR) *n);
c = *p++;
for (;;)
{
if (!(flags & FNM_NOESCAPE) && c == L_('\\'))
{
if (*p == L_('\0'))
return FNM_NOMATCH;
c = FOLD ((UCHAR) *p);
++p;
if (c == fn)
goto matched;
}
else if (c == L_('[') && *p == L_(':'))
{
/* Leave room for the null. */
CHAR str[CHAR_CLASS_MAX_LENGTH + 1];
size_t c1 = 0;
#if defined _LIBC || WIDE_CHAR_SUPPORT
wctype_t wt;
#endif
const CHAR *startp = p;
for (;;)
{
if (c1 == CHAR_CLASS_MAX_LENGTH)
/* The name is too long and therefore the pattern
is ill-formed. */
return FNM_NOMATCH;
c = *++p;
if (c == L_(':') && p[1] == L_(']'))
{
p += 2;
break;
}
if (c < L_('a') || c >= L_('z'))
{
/* This cannot possibly be a character class name.
Match it as a normal range. */
p = startp;
c = L_('[');
goto normal_bracket;
}
str[c1++] = c;
}
str[c1] = L_('\0');
#if defined _LIBC || WIDE_CHAR_SUPPORT
wt = IS_CHAR_CLASS (str);
if (wt == 0)
/* Invalid character class name. */
return FNM_NOMATCH;
# if defined _LIBC && ! WIDE_CHAR_VERSION
/* The following code is glibc specific but does
there a good job in speeding up the code since
we can avoid the btowc() call. */
if (_ISCTYPE ((UCHAR) *n, wt))
goto matched;
# else
if (ISWCTYPE (BTOWC ((UCHAR) *n), wt))
goto matched;
# endif
#else
if ((STREQ (str, L_("alnum")) && ISALNUM ((UCHAR) *n))
|| (STREQ (str, L_("alpha")) && ISALPHA ((UCHAR) *n))
|| (STREQ (str, L_("blank")) && ISBLANK ((UCHAR) *n))
|| (STREQ (str, L_("cntrl")) && ISCNTRL ((UCHAR) *n))
|| (STREQ (str, L_("digit")) && ISDIGIT ((UCHAR) *n))
|| (STREQ (str, L_("graph")) && ISGRAPH ((UCHAR) *n))
|| (STREQ (str, L_("lower")) && ISLOWER ((UCHAR) *n))
|| (STREQ (str, L_("print")) && ISPRINT ((UCHAR) *n))
|| (STREQ (str, L_("punct")) && ISPUNCT ((UCHAR) *n))
|| (STREQ (str, L_("space")) && ISSPACE ((UCHAR) *n))
|| (STREQ (str, L_("upper")) && ISUPPER ((UCHAR) *n))
|| (STREQ (str, L_("xdigit")) && ISXDIGIT ((UCHAR) *n)))
goto matched;
#endif
c = *p++;
}
#ifdef _LIBC
else if (c == L_('[') && *p == L_('='))
{
UCHAR str[1];
uint32_t nrules =
_NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
const CHAR *startp = p;
c = *++p;
if (c == L_('\0'))
{
p = startp;
c = L_('[');
goto normal_bracket;
}
str[0] = c;
c = *++p;
if (c != L_('=') || p[1] != L_(']'))
{
p = startp;
c = L_('[');
goto normal_bracket;
}
p += 2;
if (nrules == 0)
{
if ((UCHAR) *n == str[0])
goto matched;
}
else
{
const int32_t *table;
# if WIDE_CHAR_VERSION
const int32_t *weights;
const int32_t *extra;
# else
const unsigned char *weights;
const unsigned char *extra;
# endif
const int32_t *indirect;
int32_t idx;
const UCHAR *cp = (const UCHAR *) str;
/* This #include defines a local function! */
# if WIDE_CHAR_VERSION
# include <locale/weightwc.h>
# else
# include <locale/weight.h>
# endif
# if WIDE_CHAR_VERSION
table = (const int32_t *)
_NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEWC);
weights = (const int32_t *)
_NL_CURRENT (LC_COLLATE, _NL_COLLATE_WEIGHTWC);
extra = (const int32_t *)
_NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAWC);
indirect = (const int32_t *)
_NL_CURRENT (LC_COLLATE, _NL_COLLATE_INDIRECTWC);
# else
table = (const int32_t *)
_NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
weights = (const unsigned char *)
_NL_CURRENT (LC_COLLATE, _NL_COLLATE_WEIGHTMB);
extra = (const unsigned char *)
_NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
indirect = (const int32_t *)
_NL_CURRENT (LC_COLLATE, _NL_COLLATE_INDIRECTMB);
# endif
idx = findidx (&cp);
if (idx != 0)
{
/* We found a table entry. Now see whether the
character we are currently at has the same
equivalance class value. */
int len = weights[idx];
int32_t idx2;
const UCHAR *np = (const UCHAR *) n;
idx2 = findidx (&np);
if (idx2 != 0 && len == weights[idx2])
{
int cnt = 0;
while (cnt < len
&& (weights[idx + 1 + cnt]
== weights[idx2 + 1 + cnt]))
++cnt;
if (cnt == len)
goto matched;
}
}
}
c = *p++;
}
#endif
else if (c == L_('\0'))
/* [ (unterminated) loses. */
return FNM_NOMATCH;
else
{
bool is_range = false;
#ifdef _LIBC
bool is_seqval = false;
if (c == L_('[') && *p == L_('.'))
{
uint32_t nrules =
_NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
const CHAR *startp = p;
size_t c1 = 0;
while (1)
{
c = *++p;
if (c == L_('.') && p[1] == L_(']'))
{
p += 2;
break;
}
if (c == '\0')
return FNM_NOMATCH;
++c1;
}
/* We have to handling the symbols differently in
ranges since then the collation sequence is
important. */
is_range = *p == L_('-') && p[1] != L_('\0');
if (nrules == 0)
{
/* There are no names defined in the collation
data. Therefore we only accept the trivial
names consisting of the character itself. */
if (c1 != 1)
return FNM_NOMATCH;
if (!is_range && *n == startp[1])
goto matched;
cold = startp[1];
c = *p++;
}
else
{
int32_t table_size;
const int32_t *symb_table;
# ifdef WIDE_CHAR_VERSION
char str[c1];
size_t strcnt;
# else
# define str (startp + 1)
# endif
const unsigned char *extra;
int32_t idx;
int32_t elem;
int32_t second;
int32_t hash;
# ifdef WIDE_CHAR_VERSION
/* We have to convert the name to a single-byte
string. This is possible since the names
consist of ASCII characters and the internal
representation is UCS4. */
for (strcnt = 0; strcnt < c1; ++strcnt)
str[strcnt] = startp[1 + strcnt];
# endif
table_size =
_NL_CURRENT_WORD (LC_COLLATE,
_NL_COLLATE_SYMB_HASH_SIZEMB);
symb_table = (const int32_t *)
_NL_CURRENT (LC_COLLATE,
_NL_COLLATE_SYMB_TABLEMB);
extra = (const unsigned char *)
_NL_CURRENT (LC_COLLATE,
_NL_COLLATE_SYMB_EXTRAMB);
/* Locate the character in the hashing table. */
hash = elem_hash (str, c1);
idx = 0;
elem = hash % table_size;
second = hash % (table_size - 2);
while (symb_table[2 * elem] != 0)
{
/* First compare the hashing value. */
if (symb_table[2 * elem] == hash
&& c1 == extra[symb_table[2 * elem + 1]]
&& memcmp (str,
&extra[symb_table[2 * elem + 1]
+ 1], c1) == 0)
{
/* Yep, this is the entry. */
idx = symb_table[2 * elem + 1];
idx += 1 + extra[idx];
break;
}
/* Next entry. */
elem += second;
}
if (symb_table[2 * elem] != 0)
{
/* Compare the byte sequence but only if
this is not part of a range. */
# ifdef WIDE_CHAR_VERSION
int32_t *wextra;
idx += 1 + extra[idx];
/* Adjust for the alignment. */
idx = (idx + 3) & ~3;
wextra = (int32_t *) &extra[idx + 4];
# endif
if (! is_range)
{
# ifdef WIDE_CHAR_VERSION
for (c1 = 0;
(int32_t) c1 < wextra[idx];
++c1)
if (n[c1] != wextra[1 + c1])
break;
if ((int32_t) c1 == wextra[idx])
goto matched;
# else
for (c1 = 0; c1 < extra[idx]; ++c1)
if (n[c1] != extra[1 + c1])
break;
if (c1 == extra[idx])
goto matched;
# endif
}
/* Get the collation sequence value. */
is_seqval = true;
# ifdef WIDE_CHAR_VERSION
cold = wextra[1 + wextra[idx]];
# else
/* Adjust for the alignment. */
idx += 1 + extra[idx];
idx = (idx + 3) & ~4;
cold = *((int32_t *) &extra[idx]);
# endif
c = *p++;
}
else if (c1 == 1)
{
/* No valid character. Match it as a
single byte. */
if (!is_range && *n == str[0])
goto matched;
cold = str[0];
c = *p++;
}
else
return FNM_NOMATCH;
}
}
else
# undef str
#endif
{
c = FOLD (c);
normal_bracket:
/* We have to handling the symbols differently in
ranges since then the collation sequence is
important. */
is_range = (*p == L_('-') && p[1] != L_('\0')
&& p[1] != L_(']'));
if (!is_range && c == fn)
goto matched;
cold = c;
c = *p++;
}
if (c == L_('-') && *p != L_(']'))
{
#if _LIBC
/* We have to find the collation sequence
value for C. Collation sequence is nothing
we can regularly access. The sequence
value is defined by the order in which the
definitions of the collation values for the
various characters appear in the source
file. A strange concept, nowhere
documented. */
uint32_t fcollseq;
uint32_t lcollseq;
UCHAR cend = *p++;
# ifdef WIDE_CHAR_VERSION
/* Search in the `names' array for the characters. */
fcollseq = __collseq_table_lookup (collseq, fn);
if (fcollseq == ~((uint32_t) 0))
/* XXX We don't know anything about the character
we are supposed to match. This means we are
failing. */
goto range_not_matched;
if (is_seqval)
lcollseq = cold;
else
lcollseq = __collseq_table_lookup (collseq, cold);
# else
fcollseq = collseq[fn];
lcollseq = is_seqval ? cold : collseq[(UCHAR) cold];
# endif
is_seqval = false;
if (cend == L_('[') && *p == L_('.'))
{
uint32_t nrules =
_NL_CURRENT_WORD (LC_COLLATE,
_NL_COLLATE_NRULES);
const CHAR *startp = p;
size_t c1 = 0;
while (1)
{
c = *++p;
if (c == L_('.') && p[1] == L_(']'))
{
p += 2;
break;
}
if (c == '\0')
return FNM_NOMATCH;
++c1;
}
if (nrules == 0)
{
/* There are no names defined in the
collation data. Therefore we only
accept the trivial names consisting
of the character itself. */
if (c1 != 1)
return FNM_NOMATCH;
cend = startp[1];
}
else
{
int32_t table_size;
const int32_t *symb_table;
# ifdef WIDE_CHAR_VERSION
char str[c1];
size_t strcnt;
# else
# define str (startp + 1)
# endif
const unsigned char *extra;
int32_t idx;
int32_t elem;
int32_t second;
int32_t hash;
# ifdef WIDE_CHAR_VERSION
/* We have to convert the name to a single-byte
string. This is possible since the names
consist of ASCII characters and the internal
representation is UCS4. */
for (strcnt = 0; strcnt < c1; ++strcnt)
str[strcnt] = startp[1 + strcnt];
# endif
table_size =
_NL_CURRENT_WORD (LC_COLLATE,
_NL_COLLATE_SYMB_HASH_SIZEMB);
symb_table = (const int32_t *)
_NL_CURRENT (LC_COLLATE,
_NL_COLLATE_SYMB_TABLEMB);
extra = (const unsigned char *)
_NL_CURRENT (LC_COLLATE,
_NL_COLLATE_SYMB_EXTRAMB);
/* Locate the character in the hashing
table. */
hash = elem_hash (str, c1);
idx = 0;
elem = hash % table_size;
second = hash % (table_size - 2);
while (symb_table[2 * elem] != 0)
{
/* First compare the hashing value. */
if (symb_table[2 * elem] == hash
&& (c1
== extra[symb_table[2 * elem + 1]])
&& memcmp (str,
&extra[symb_table[2 * elem + 1]
+ 1], c1) == 0)
{
/* Yep, this is the entry. */
idx = symb_table[2 * elem + 1];
idx += 1 + extra[idx];
break;
}
/* Next entry. */
elem += second;
}
if (symb_table[2 * elem] != 0)
{
/* Compare the byte sequence but only if
this is not part of a range. */
# ifdef WIDE_CHAR_VERSION
int32_t *wextra;
idx += 1 + extra[idx];
/* Adjust for the alignment. */
idx = (idx + 3) & ~4;
wextra = (int32_t *) &extra[idx + 4];
# endif
/* Get the collation sequence value. */
is_seqval = true;
# ifdef WIDE_CHAR_VERSION
cend = wextra[1 + wextra[idx]];
# else
/* Adjust for the alignment. */
idx += 1 + extra[idx];
idx = (idx + 3) & ~4;
cend = *((int32_t *) &extra[idx]);
# endif
}
else if (symb_table[2 * elem] != 0 && c1 == 1)
{
cend = str[0];
c = *p++;
}
else
return FNM_NOMATCH;
}
# undef str
}
else
{
if (!(flags & FNM_NOESCAPE) && cend == L_('\\'))
cend = *p++;
if (cend == L_('\0'))
return FNM_NOMATCH;
cend = FOLD (cend);
}
/* XXX It is not entirely clear to me how to handle
characters which are not mentioned in the
collation specification. */
if (
# ifdef WIDE_CHAR_VERSION
lcollseq == 0xffffffff ||
# endif
lcollseq <= fcollseq)
{
/* We have to look at the upper bound. */
uint32_t hcollseq;
if (is_seqval)
hcollseq = cend;
else
{
# ifdef WIDE_CHAR_VERSION
hcollseq =
__collseq_table_lookup (collseq, cend);
if (hcollseq == ~((uint32_t) 0))
{
/* Hum, no information about the upper
bound. The matching succeeds if the
lower bound is matched exactly. */
if (lcollseq != fcollseq)
goto range_not_matched;
goto matched;
}
# else
hcollseq = collseq[cend];
# endif
}
if (lcollseq <= hcollseq && fcollseq <= hcollseq)
goto matched;
}
# ifdef WIDE_CHAR_VERSION
range_not_matched:
# endif
#else
/* We use a boring value comparison of the character
values. This is better than comparing using
`strcoll' since the latter would have surprising
and sometimes fatal consequences. */
UCHAR cend = *p++;
if (!(flags & FNM_NOESCAPE) && cend == L_('\\'))
cend = *p++;
if (cend == L_('\0'))
return FNM_NOMATCH;
/* It is a range. */
if (cold <= fn && fn <= cend)
goto matched;
#endif
c = *p++;
}
}
if (c == L_(']'))
break;
}
if (!not)
return FNM_NOMATCH;
break;
matched:
/* Skip the rest of the [...] that already matched. */
do
{
ignore_next:
c = *p++;
if (c == L_('\0'))
/* [... (unterminated) loses. */
return FNM_NOMATCH;
if (!(flags & FNM_NOESCAPE) && c == L_('\\'))
{
if (*p == L_('\0'))
return FNM_NOMATCH;
/* XXX 1003.2d11 is unclear if this is right. */
++p;
}
else if (c == L_('[') && *p == L_(':'))
{
int c1 = 0;
const CHAR *startp = p;
while (1)
{
c = *++p;
if (++c1 == CHAR_CLASS_MAX_LENGTH)
return FNM_NOMATCH;
if (*p == L_(':') && p[1] == L_(']'))
break;
if (c < L_('a') || c >= L_('z'))
{
p = startp;
goto ignore_next;
}
}
p += 2;
c = *p++;
}
else if (c == L_('[') && *p == L_('='))
{
c = *++p;
if (c == L_('\0'))
return FNM_NOMATCH;
c = *++p;
if (c != L_('=') || p[1] != L_(']'))
return FNM_NOMATCH;
p += 2;
c = *p++;
}
else if (c == L_('[') && *p == L_('.'))
{
++p;
while (1)
{
c = *++p;
if (c == '\0')
return FNM_NOMATCH;
if (*p == L_('.') && p[1] == L_(']'))
break;
}
p += 2;
c = *p++;
}
}
while (c != L_(']'));
if (not)
return FNM_NOMATCH;
}
break;
case L_('+'):
case L_('@'):
case L_('!'):
if (__builtin_expect (flags & FNM_EXTMATCH, 0) && *p == '(')
{
int res;
res = EXT (c, p, n, string_end, no_leading_period, flags);
if (res != -1)
return res;
}
goto normal_match;
case L_('/'):
if (NO_LEADING_PERIOD (flags))
{
if (n == string_end || c != (UCHAR) *n)
return FNM_NOMATCH;
new_no_leading_period = true;
break;
}
/* FALLTHROUGH */
default:
normal_match:
if (n == string_end || c != FOLD ((UCHAR) *n))
return FNM_NOMATCH;
}
no_leading_period = new_no_leading_period;
++n;
}
if (n == string_end)
return 0;
if ((flags & FNM_LEADING_DIR) && n != string_end && *n == L_('/'))
/* The FNM_LEADING_DIR flag says that "foo*" matches "foobar/frobozz". */
return 0;
return FNM_NOMATCH;
}
/* Match pattern "p" against "text" */
static int dowild(const uchar *p, const uchar *text, unsigned int flags)
{
uchar p_ch;
const uchar *pattern = p;
int pattern_needs_leading_period =
(flags & WM_PERIOD) && (text[0] == '.');
for ( ; (p_ch = *p) != '\0'; text++, p++) {
int matched, match_slash, negated;
uchar t_ch, prev_ch;
if ((t_ch = *text) == '\0' && p_ch != '*')
return WM_ABORT_ALL;
if ((flags & WM_CASEFOLD) && ISUPPER(t_ch))
t_ch = tolower(t_ch);
if ((flags & WM_CASEFOLD) && ISUPPER(p_ch))
p_ch = tolower(p_ch);
switch (p_ch) {
case '\\':
if (!(flags & WM_NOESCAPE))
/* Literal match with following character. Note that the test
* in "default" handles the p[1] == '\0' failure case. */
p_ch = *++p;
/* FALLTHROUGH */
default:
if (t_ch != p_ch)
return WM_NOMATCH;
if ((flags & WM_PERIOD) &&
(flags & WM_PATHNAME) && t_ch == '/' &&
text[1] == '.' && p[1] != '.')
return WM_NOMATCH;
/* If we needed a leading period, we've matched it. */
pattern_needs_leading_period = 0;
continue;
case '?':
/* Match anything but '/'. */
if ((flags & WM_PATHNAME) && t_ch == '/')
return WM_NOMATCH;
continue;
case '*':
if (*++p == '*') {
const uchar *prev_p = p - 2;
while (*++p == '*') {}
if (!(flags & WM_PATHNAME))
/* without WM_PATHNAME, '*' == '**' */
match_slash = 1;
else if ((prev_p < pattern || *prev_p == '/') &&
(*p == '\0' || *p == '/' ||
(p[0] == '\\' && p[1] == '/'))) {
/*
* Assuming we already match 'foo/' and are at
* <star star slash>, just assume it matches
* nothing and go ahead match the rest of the
* pattern with the remaining string. This
* helps make foo/<*><*>/bar (<> because
* otherwise it breaks C comment syntax) match
* both foo/bar and foo/a/bar.
*/
if (p[0] == '/' &&
dowild(p + 1, text, flags) == WM_MATCH)
return WM_MATCH;
match_slash = 1;
} else
return WM_ABORT_MALFORMED;
} else
/* without WM_PATHNAME, '*' == '**' */
match_slash = flags & WM_PATHNAME ? 0 : 1;
if (*p == '\0') {
/* If we needed a leading period in the pattern but only
* found stars, we didn't match. */
if (pattern_needs_leading_period)
return WM_NOMATCH;
/* Trailing "**" matches everything. Trailing "*" matches
* only if there are no more slash characters. */
if (!match_slash) {
if (strchr((char*)text, '/') != NULL)
return WM_NOMATCH;
if ((flags & WM_PERIOD) &&
strstr((const char*)text, "/.") != NULL)
return WM_NOMATCH;
}
return WM_MATCH;
} else if (!match_slash && *p == '/') {
/*
* _one_ asterisk followed by a slash
* with WM_PATHNAME matches the next
* directory
*/
const char *slash = strchr((char*)text, '/');
if (!slash)
return WM_NOMATCH;
text = (const uchar*)slash;
/* the slash is consumed by the top-level for loop */
break;
}
while (1) {
if (t_ch == '\0')
break;
/*
* Try to advance faster when an asterisk is
* followed by a literal. We know in this case
* that the the string before the literal
* must belong to "*".
* If match_slash is false, do not look past
* the first slash as it cannot belong to '*'.
*/
if (!is_glob_special(*p)) {
p_ch = *p;
if ((flags & WM_CASEFOLD) && ISUPPER(p_ch))
p_ch = tolower(p_ch);
while ((t_ch = *text) != '\0' &&
(match_slash || t_ch != '/')) {
if ((flags & WM_CASEFOLD) && ISUPPER(t_ch))
t_ch = tolower(t_ch);
if (t_ch == p_ch)
break;
text++;
}
if (t_ch != p_ch)
return WM_NOMATCH;
}
if (pattern_needs_leading_period)
return WM_NOMATCH;
if ((matched = dowild(p, text, flags)) != WM_NOMATCH) {
if (!match_slash || matched != WM_ABORT_TO_STARSTAR)
return matched;
} else if (!match_slash && t_ch == '/')
return WM_ABORT_TO_STARSTAR;
else if ((flags & WM_PERIOD) &&
(flags & WM_PATHNAME) && t_ch == '/' &&
text[1] == '.' && p[1] != '.')
return WM_NOMATCH;
t_ch = *++text;
}
return WM_ABORT_ALL;
case '[':
p_ch = *++p;
#ifdef NEGATE_CLASS2
if (p_ch == NEGATE_CLASS2)
p_ch = NEGATE_CLASS;
#endif
/* Assign literal 1/0 because of "matched" comparison. */
negated = p_ch == NEGATE_CLASS ? 1 : 0;
if (negated) {
/* Inverted character class. */
p_ch = *++p;
}
prev_ch = 0;
matched = 0;
do {
if (!p_ch)
return WM_ABORT_ALL;
if (p_ch == '\\') {
p_ch = *++p;
if (!p_ch)
return WM_ABORT_ALL;
if (t_ch == p_ch)
matched = 1;
} else if (p_ch == '-' && prev_ch && p[1] && p[1] != ']') {
p_ch = *++p;
if (p_ch == '\\') {
p_ch = *++p;
if (!p_ch)
return WM_ABORT_ALL;
}
if (t_ch <= p_ch && t_ch >= prev_ch)
matched = 1;
else if ((flags & WM_CASEFOLD) && ISLOWER(t_ch)) {
uchar t_ch_upper = toupper(t_ch);
if (t_ch_upper <= p_ch && t_ch_upper >= prev_ch)
matched = 1;
}
p_ch = 0; /* This makes "prev_ch" get set to 0. */
} else if (p_ch == '[' && p[1] == ':') {
const uchar *s;
int i;
for (s = p += 2; (p_ch = *p) && p_ch != ']'; p++) {} /*SHARED ITERATOR*/
if (!p_ch)
return WM_ABORT_ALL;
i = p - s - 1;
if (i < 0 || p[-1] != ':') {
/* Didn't find ":]", so treat like a normal set. */
p = s - 2;
p_ch = '[';
if (t_ch == p_ch)
matched = 1;
continue;
}
if (CC_EQ(s,i, "alnum")) {
if (ISALNUM(t_ch))
matched = 1;
} else if (CC_EQ(s,i, "alpha")) {
if (ISALPHA(t_ch))
matched = 1;
} else if (CC_EQ(s,i, "blank")) {
if (ISBLANK(t_ch))
matched = 1;
} else if (CC_EQ(s,i, "cntrl")) {
if (ISCNTRL(t_ch))
matched = 1;
} else if (CC_EQ(s,i, "digit")) {
if (ISDIGIT(t_ch))
matched = 1;
} else if (CC_EQ(s,i, "graph")) {
if (ISGRAPH(t_ch))
matched = 1;
} else if (CC_EQ(s,i, "lower")) {
if (ISLOWER(t_ch))
matched = 1;
} else if (CC_EQ(s,i, "print")) {
if (ISPRINT(t_ch))
matched = 1;
} else if (CC_EQ(s,i, "punct")) {
if (ISPUNCT(t_ch))
matched = 1;
} else if (CC_EQ(s,i, "space")) {
if (ISSPACE(t_ch))
matched = 1;
} else if (CC_EQ(s,i, "upper")) {
if (ISUPPER(t_ch))
matched = 1;
else if ((flags & WM_CASEFOLD) && ISLOWER(t_ch))
matched = 1;
} else if (CC_EQ(s,i, "xdigit")) {
if (ISXDIGIT(t_ch))
matched = 1;
} else /* malformed [:class:] string */
return WM_ABORT_ALL;
p_ch = 0; /* This makes "prev_ch" get set to 0. */
} else if (t_ch == p_ch)
matched = 1;
} while (prev_ch = p_ch, (p_ch = *++p) != ']');
if (matched == negated ||
((flags & WM_PATHNAME) && t_ch == '/'))
return WM_NOMATCH;
if ((flags & WM_PATHNAME) && t_ch == '/' &&
(flags & WM_PERIOD) && (text[1] == '.') && (p[1] != '.'))
return WM_NOMATCH;
continue;
}
}
return *text ? WM_NOMATCH : WM_MATCH;
}
static VALUE
optimized_unescape(VALUE str, VALUE encoding)
{
long i, len, beg = 0;
VALUE dest = 0;
const char *cstr;
int cr, origenc, encidx = rb_to_encoding_index(encoding);
len = RSTRING_LEN(str);
cstr = RSTRING_PTR(str);
for (i = 0; i < len; ++i) {
char buf[1];
const char c = cstr[i];
int clen = 0;
if (c == '%') {
if (i + 3 > len) break;
if (!ISXDIGIT(cstr[i+1])) continue;
if (!ISXDIGIT(cstr[i+2])) continue;
buf[0] = ((char_to_number(cstr[i+1]) << 4)
| char_to_number(cstr[i+2]));
clen = 2;
}
else if (c == '+') {
buf[0] = ' ';
}
else {
continue;
}
if (!dest) {
dest = rb_str_buf_new(len);
}
rb_str_cat(dest, cstr + beg, i - beg);
i += clen;
beg = i + 1;
rb_str_cat(dest, buf, 1);
}
if (dest) {
rb_str_cat(dest, cstr + beg, len - beg);
preserve_original_state(str, dest);
cr = ENC_CODERANGE_UNKNOWN;
}
else {
dest = rb_str_dup(str);
cr = ENC_CODERANGE(str);
}
origenc = rb_enc_get_index(str);
if (origenc != encidx) {
rb_enc_associate_index(dest, encidx);
if (!ENC_CODERANGE_CLEAN_P(rb_enc_str_coderange(dest))) {
rb_enc_associate_index(dest, origenc);
if (cr != ENC_CODERANGE_UNKNOWN)
ENC_CODERANGE_SET(dest, cr);
}
}
return dest;
}
struct symbol *parser( )
{
register struct symbol *psym;
register char *pline;
struct varent *pvar;
struct strent *pstr;
char *cp, *plc, *pn;
long lnc;
int i;
long double tem;
/* reference for old Whitesmiths compiler: */
/*
*extern FILE *stdout;
*/
pline = interl; /* get current location in command string */
/* If at beginning of string, must ask for more input */
if( pline == line )
{
if( maccnt > 0 )
{
--maccnt;
cp = maclin;
plc = pline;
while( (*plc++ = *cp++) != 0 )
;
goto mstart;
}
if( takptr < 0 )
{ /* no take file active: prompt keyboard input */
printf("* ");
}
/* Various ways of typing in a command line. */
/*
* Old Whitesmiths call to print "*" immediately
* use RT11 .GTLIN to get command string
* from command file or terminal
*/
/*
* fflush(stdout);
* gtlin(line);
*/
zgets( line, TRUE ); /* keyboard input for other systems: */
mstart:
uposs = 1; /* unary operators possible at start of line */
}
ignore:
/* Skip over spaces */
while( *pline == ' ' )
++pline;
/* unary minus after operator */
if( uposs && (*pline == '-') )
{
psym = &oprtbl[2]; /* UMINUS */
++pline;
goto pdon3;
}
/* COMP */
/*
if( uposs && (*pline == '~') )
{
psym = &oprtbl[3];
++pline;
goto pdon3;
}
*/
if( uposs && (*pline == '+') ) /* ignore leading plus sign */
{
++pline;
goto ignore;
}
/* end of null terminated input */
if( (*pline == '\n') || (*pline == '\0') || (*pline == '\r') )
{
pline = line;
goto endlin;
}
if( *pline == ';' )
{
++pline;
endlin:
psym = &oprtbl[1]; /* EOL */
goto pdon2;
}
/* parser() */
/* Test for numeric input */
if( (ISDIGIT(*pline)) || (*pline == '.') )
{
lnc = 0; /* initialize numeric input to zero */
qnc = 0.0L;
if( *pline == '0' )
{ /* leading "0" may mean octal or hex radix */
++pline;
if( *pline == '.' )
goto decimal; /* 0.ddd */
/* leading "0x" means hexadecimal radix */
if( (*pline == 'x') || (*pline == 'X') )
{
++pline;
while( ISXDIGIT(*pline) )
{
i = *pline++ & 0xff;
if( i >= 'a' )
i -= 047;
if( i >= 'A' )
i -= 07;
i -= 060;
lnc = (lnc << 4) + i;
qnc = lnc;
}
goto numdon;
}
else
{
while( ISOCTAL( *pline ) )
{
i = ((*pline++) & 0xff) - 060;
lnc = (lnc << 3) + i;
qnc = lnc;
}
goto numdon;
}
}
else
{
/* no leading "0" means decimal radix */
/******/
decimal:
pn = number;
while( (ISDIGIT(*pline)) || (*pline == '.') )
*pn++ = *pline++;
/* get possible exponent field */
if( (*pline == 'e') || (*pline == 'E') )
*pn++ = *pline++;
else
goto numcvt;
if( (*pline == '-') || (*pline == '+') )
*pn++ = *pline++;
while( ISDIGIT(*pline) )
*pn++ = *pline++;
numcvt:
*pn++ = ' ';
*pn++ = 0;
#if NE == 6
asctoe64( number, &qnc );
#else
asctoe113( number, &qnc );
#endif
/* sscanf( number, "%le", &nc ); */
}
/* output the number */
numdon:
/* search the symbol table of constants */
pvar = &contbl[0];
for( i=0; i<NCONST; i++ )
{
if( (pvar->attrib & BUSY) == 0 )
goto confnd;
tem = *pvar->value;
if( tem == qnc )
{
psym = (struct symbol *)pvar;
goto pdon2;
}
++pvar;
}
printf( "no room for constant\n" );
psym = (struct symbol *)&contbl[0];
goto pdon2;
confnd:
pvar->spel= contbl[0].spel;
pvar->attrib = CONST | BUSY;
*pvar->value = qnc;
psym = (struct symbol *)pvar;
goto pdon2;
}
/* check for operators */
psym = &oprtbl[3];
for( i=0; i<NOPR; i++ )
{
if( *pline == *(psym->spel) )
goto pdon1;
++psym;
}
/* if quoted, it is a string variable */
if( *pline == '"' )
{
/* find an empty slot for the string */
pstr = strtbl; /* string table */
for( i=0; i<NSTRNG-1; i++ )
{
if( (pstr->attrib & BUSY) == 0 )
goto fndstr;
++pstr;
}
printf( "No room for string\n" );
pstr->attrib |= ILLEG;
psym = (struct symbol *)pstr;
goto pdon0;
fndstr:
pstr->attrib |= BUSY;
plc = pstr->string;
++pline;
for( i=0; i<39; i++ )
{
*plc++ = *pline;
if( (*pline == '\n') || (*pline == '\0') || (*pline == '\r') )
{
illstr:
pstr = &strtbl[NSTRNG-1];
pstr->attrib |= ILLEG;
printf( "Missing string terminator\n" );
psym = (struct symbol *)pstr;
goto pdon0;
}
if( *pline++ == '"' )
goto finstr;
}
goto illstr; /* no terminator found */
finstr:
--plc;
*plc = '\0';
psym = (struct symbol *)pstr;
goto pdon2;
}
/* If none of the above, search function and symbol tables: */
/* copy character string to array lc[] */
plc = &lc[0];
while( ISALPHA(*pline) )
{
/* convert to lower case characters */
if( ISUPPER( *pline ) )
*pline += 040;
*plc++ = *pline++;
}
*plc = 0; /* Null terminate the output string */
/* parser() */
psym = (struct symbol *)menstk[menptr]; /* function table */
plc = &lc[0];
cp = psym->spel;
do
{
if( strcmp( plc, cp ) == 0 )
goto pdon3; /* following unary minus is possible */
++psym;
cp = psym->spel;
}
while( *cp != '\0' );
psym = (struct symbol *)&indtbl[0]; /* indirect symbol table */
plc = &lc[0];
cp = psym->spel;
do
{
if( strcmp( plc, cp ) == 0 )
goto pdon2;
++psym;
cp = psym->spel;
}
while( *cp != '\0' );
pdon0:
pline = line; /* scrub line if illegal symbol */
goto pdon2;
pdon1:
++pline;
if( (psym->attrib & 0xf) == RPAREN )
pdon2: uposs = 0;
else
pdon3: uposs = 1;
interl = pline;
return( psym );
} /* end of parser */
static VALUE
optimized_unescape_html(VALUE str)
{
enum {UNICODE_MAX = 0x10ffff};
rb_encoding *enc = rb_enc_get(str);
unsigned long charlimit = (strcasecmp(rb_enc_name(enc), "UTF-8") == 0 ? UNICODE_MAX :
strcasecmp(rb_enc_name(enc), "ISO-8859-1") == 0 ? 256 :
128);
long i, len, beg = 0;
size_t clen, plen;
int overflow;
const char *cstr;
char buf[6];
VALUE dest = 0;
len = RSTRING_LEN(str);
cstr = RSTRING_PTR(str);
for (i = 0; i < len; i++) {
unsigned long cc;
char c = cstr[i];
if (c != '&') continue;
plen = i - beg;
if (++i >= len) break;
c = (unsigned char)cstr[i];
switch (c) {
case 'a':
++i;
if (len - i >= 4 && memcmp(&cstr[i], "pos;", 4) == 0) {
c = '\'';
i += 3;
}
else if (len - i >= 3 && memcmp(&cstr[i], "mp;", 3) == 0) {
c = '&';
i += 2;
}
else continue;
break;
case 'q':
++i;
if (len - i >= 4 && memcmp(&cstr[i], "uot;", 4) == 0) {
c = '"';
i += 3;
}
else continue;
break;
case 'g':
++i;
if (len - i >= 2 && memcmp(&cstr[i], "t;", 2) == 0) {
c = '>';
i += 1;
}
else continue;
break;
case 'l':
++i;
if (len - i >= 2 && memcmp(&cstr[i], "t;", 2) == 0) {
c = '<';
i += 1;
}
else continue;
break;
case '#':
if (len - ++i >= 2 && ISDIGIT(cstr[i])) {
cc = ruby_scan_digits(&cstr[i], len-i, 10, &clen, &overflow);
}
else if ((cstr[i] == 'x' || cstr[i] == 'X') && len - ++i >= 2 && ISXDIGIT(cstr[i])) {
cc = ruby_scan_digits(&cstr[i], len-i, 16, &clen, &overflow);
}
else continue;
i += clen;
if (overflow || cc >= charlimit || cstr[i] != ';') continue;
if (!dest) {
dest = rb_str_buf_new(len);
}
rb_str_cat(dest, cstr + beg, plen);
if (charlimit > 256) {
rb_str_cat(dest, buf, rb_enc_mbcput((OnigCodePoint)cc, buf, enc));
}
else {
c = (unsigned char)cc;
rb_str_cat(dest, &c, 1);
}
beg = i + 1;
continue;
default:
--i;
continue;
}
if (!dest) {
dest = rb_str_buf_new(len);
}
rb_str_cat(dest, cstr + beg, plen);
rb_str_cat(dest, &c, 1);
beg = i + 1;
}
if (dest) {
rb_str_cat(dest, cstr + beg, len - beg);
preserve_original_state(str, dest);
return dest;
}
else {
return rb_str_dup(str);
}
}
static int loop(const unsigned char *pattern, const unsigned char *string)
{
loop_state state = CURLFNM_LOOP_DEFAULT;
unsigned char *p = (unsigned char *)pattern;
unsigned char *s = (unsigned char *)string;
unsigned char charset[CURLFNM_CHSET_SIZE] = { 0 };
int rc = 0;
for(;;) {
switch(state) {
case CURLFNM_LOOP_DEFAULT:
if(*p == '*') {
while(*(p+1) == '*') /* eliminate multiple stars */
p++;
if(*s == '\0' && *(p+1) == '\0')
return CURL_FNMATCH_MATCH;
rc = loop(p + 1, s); /* *.txt matches .txt <=> .txt matches .txt */
if(rc == CURL_FNMATCH_MATCH)
return CURL_FNMATCH_MATCH;
if(*s) /* let the star eat up one character */
s++;
else
return CURL_FNMATCH_NOMATCH;
}
else if(*p == '?') {
if(ISPRINT(*s)) {
s++;
p++;
}
else if(*s == '\0')
return CURL_FNMATCH_NOMATCH;
else
return CURL_FNMATCH_FAIL; /* cannot deal with other character */
}
else if(*p == '\0') {
if(*s == '\0')
return CURL_FNMATCH_MATCH;
else
return CURL_FNMATCH_NOMATCH;
}
else if(*p == '\\') {
state = CURLFNM_LOOP_BACKSLASH;
p++;
}
else if(*p == '[') {
unsigned char *pp = p+1; /* cannot handle with pointer to register */
if(setcharset(&pp, charset)) {
int found = FALSE;
if(charset[(unsigned int)*s])
found = TRUE;
else if(charset[CURLFNM_ALNUM])
found = ISALNUM(*s);
else if(charset[CURLFNM_ALPHA])
found = ISALPHA(*s);
else if(charset[CURLFNM_DIGIT])
found = ISDIGIT(*s);
else if(charset[CURLFNM_XDIGIT])
found = ISXDIGIT(*s);
else if(charset[CURLFNM_PRINT])
found = ISPRINT(*s);
else if(charset[CURLFNM_SPACE])
found = ISSPACE(*s);
else if(charset[CURLFNM_UPPER])
found = ISUPPER(*s);
else if(charset[CURLFNM_LOWER])
found = ISLOWER(*s);
else if(charset[CURLFNM_BLANK])
found = ISBLANK(*s);
else if(charset[CURLFNM_GRAPH])
found = ISGRAPH(*s);
if(charset[CURLFNM_NEGATE])
found = !found;
if(found) {
p = pp+1;
s++;
memset(charset, 0, CURLFNM_CHSET_SIZE);
}
else
return CURL_FNMATCH_NOMATCH;
}
else
return CURL_FNMATCH_FAIL;
}
else {
if(*p++ != *s++)
return CURL_FNMATCH_NOMATCH;
}
break;
case CURLFNM_LOOP_BACKSLASH:
if(ISPRINT(*p)) {
if(*p++ == *s++)
state = CURLFNM_LOOP_DEFAULT;
else
return CURL_FNMATCH_NOMATCH;
}
else
return CURL_FNMATCH_FAIL;
break;
}
}
}
/*
* ctpl_input_stream_read_number_internal:
* @type: which kind of number match (float, int or both)
*
* see ctpl_input_stream_read_number()
*/
static gboolean
ctpl_input_stream_read_number_internal (CtplInputStream *stream,
gint type,
CtplValue *value,
GError **error)
{
gboolean have_mantissa = FALSE;
gboolean have_exponent = FALSE;
gboolean have_exponent_delim = FALSE;
gboolean have_sign = FALSE;
gboolean have_dot = FALSE;
GString *gstring;
GError *err = NULL;
gboolean in_number = TRUE;
gint base = 10;
#define ISSIGN(c) ((c) == '+' || (c) == '-')
#define ISDIGIT(c) ((c) >= '0' && (c) <= '9')
#define ISBDIGIT(c) ((c) == '0' || (c) == '1')
#define ISODIGIT(c) ((c) >= '0' && (c) <= '7')
#define ISXDIGIT(c) (ISDIGIT (c) || \
((c) >= 'a' && (c) <= 'f') || \
((c) >= 'A' && (c) <= 'F'))
gstring = g_string_new ("");
while (in_number && ! err) {
gchar buf[3];
gssize buf_len;
buf_len = ctpl_input_stream_peek (stream, buf, sizeof (buf), &err);
if (! err) {
gchar c = (buf_len > 0) ? buf[0] : CTPL_EOF;
switch (c) {
case '.':
if (! have_dot && ! have_exponent_delim && (type & READ_FLOAT)) {
g_string_append_c (gstring, c);
have_dot = TRUE;
type &= READ_FLOAT;
} else {
in_number = FALSE;
}
break;
case '+':
case '-':
if (! have_sign && (! have_mantissa ||
(have_exponent_delim && ! have_exponent)) &&
/* ISDIGIT() is fine here even though we probably don't know the
* base yet because the default base is 10 and the exponent or
* power are also in base 10 */
buf_len > 1 && ISDIGIT (buf[1])) {
g_string_append_c (gstring, c);
have_sign = TRUE;
} else {
in_number = FALSE;
}
break;
case 'e':
case 'E':
if (base < 15) {
if (have_mantissa && ! have_exponent_delim &&
(type & READ_FLOAT) && base == 10 &&
((buf_len > 1 && ISDIGIT (buf[1])) ||
(buf_len > 2 && ISSIGN (buf[1]) && ISDIGIT (buf[2])))) {
have_exponent_delim = TRUE;
have_sign = FALSE;
type &= READ_FLOAT;
g_string_append_c (gstring, 'e');
} else {
in_number = FALSE;
}
break;
}
/* Fallthrough */
case 'b':
case 'B':
case 'a':
case 'A':
case 'c':
case 'C':
case 'd':
case 'D':
case 'f':
case 'F':
if (base < 16 || have_exponent_delim /* exponent is decimal */) {
in_number = FALSE;
break;
}
/* Fallthrough */
case '8':
case '9':
if (base < 10) {
in_number = FALSE;
break;
}
case '2':
case '3':
case '4':
case '5':
case '6':
case '7':
if (base < 8) {
in_number = FALSE;
break;
}
/* Fallthrough */
case '0':
if (! have_mantissa && buf_len > 2) {
gboolean is_start = TRUE;
if ((type & READ_INT) &&
(buf[1] == 'b' || buf[1] == 'B') && ISBDIGIT (buf[2])) {
type &= READ_INT;
base = 2;
} else if ((type & READ_INT) &&
(buf[1] == 'o' || buf[1] == 'O') && ISODIGIT (buf[2])) {
type &= READ_INT;
base = 8;
} else if ((buf[1] == 'x' || buf[1] == 'X') && ISXDIGIT (buf[2])) {
/* needed for floating-points */
g_string_append_c (gstring, c);
g_string_append_c (gstring, buf[1]);
base = 16;
} else {
is_start = FALSE;
}
if (is_start) {
/* eat the character we just handled. no need to check the error
* since the data is already cached -- so no error can happen */
ctpl_input_stream_get_c (stream, NULL);
break;
}
}
/* Fallthrough */
case '1':
g_string_append_c (gstring, c);
if (! have_exponent_delim) {
have_mantissa = TRUE;
} else {
have_exponent = TRUE;
}
break;
case 'p':
case 'P':
if (have_mantissa && ! have_exponent_delim &&
(type & READ_FLOAT) && base == 16 &&
((buf_len > 1 && ISDIGIT (buf[1])) ||
(buf_len > 2 && ISSIGN (buf[1]) && ISDIGIT (buf[2])))) {
have_exponent_delim = TRUE;
have_sign = FALSE;
type &= READ_FLOAT;
g_string_append_c (gstring, 'p');
} else {
in_number = FALSE;
}
break;
default:
in_number = FALSE;
}
if (in_number) {
ctpl_input_stream_get_c (stream, &err); /* eat character */
}
}
}
if (! err) {
if (! have_mantissa) {
ctpl_input_stream_set_error (stream, &err, CTPL_IO_ERROR,
CTPL_IO_ERROR_INVALID_NUMBER,
"Missing mantissa in numeric constant");
} else {
gchar *nptr = gstring->str;
gchar *endptr;
gdouble dblval = 0.0;
glong longval = 0;
gint errno_save = errno;
errno = 0;
if (type & READ_INT) {
longval = strtol (nptr, &endptr, base);
} else {
dblval = g_ascii_strtod (nptr, &endptr);
}
if (! endptr || *endptr != 0) {
ctpl_input_stream_set_error (stream, &err, CTPL_IO_ERROR,
CTPL_IO_ERROR_INVALID_NUMBER,
"Invalid base %d numeric constant \"%s\"",
base, nptr);
} else if (errno == ERANGE) {
ctpl_input_stream_set_error (stream, &err, CTPL_IO_ERROR,
CTPL_IO_ERROR_RANGE,
"Overflow in numeric constant conversion");
} else {
if (type & READ_INT) {
ctpl_value_set_int (value, longval);
} else {
ctpl_value_set_float (value, dblval);
}
}
errno = errno_save;
}
}
if (err) {
g_propagate_error (error, err);
}
g_string_free (gstring, TRUE);
#undef ISSIGN
#undef ISDIGIT
#undef ISBDIGIT
#undef ISODIGIT
#undef ISXDIGIT
return ! err;
}
