int timlex(mystring_t **outstr, unsigned long *outnum, struct protstream *stream)
{
int ch;
char *buff_ptr;
char *buff_end;
unsigned long tmpnum = 0;
unsigned long count=0;
int result = TIMSIEVE_OK;
buff_ptr = buffer; /* ptr into the buffer */
buff_end = buffer + maxscriptsize - 10; /* ptr to end of buffer */
while (1)
{
/* get a character
this may block on a read if there is nothing
in the buffer */
ch=prot_getc(stream);
if (ch==EOF) {
/* Lost connection */
return EOF;
}
switch (lexer_state)
{
case LEXER_STATE_RECOVER:
if (ch == '\n') {
lexer_state=LEXER_STATE_NORMAL;
}
if (ch == '\r')
lexer_state=LEXER_STATE_RECOVER_CR;
break;
case LEXER_STATE_RECOVER_CR:
if (ch == '\n')
lexer_state=LEXER_STATE_NORMAL;
break;
case LEXER_STATE_CR:
if (ch == '\n') {
lexer_state=LEXER_STATE_NORMAL;
return EOL;
}
/* otherwise, life is bad */
ERR_PUSHBACK();
case LEXER_STATE_QSTR:
if (ch == '\"') {
/* End of the string */
if (outstr!=NULL)
{
*outstr = NULL;
result = string_allocate(buff_ptr - buffer, buffer, outstr);
if (result != TIMSIEVE_OK)
ERR_PUSHBACK();
}
/*} */
lexer_state=LEXER_STATE_NORMAL;
return STRING;
}
/* illegal character */
if (ch == '\0'
|| ch == '\r'
|| ch == '\n'
|| 0x7F < ((unsigned char)ch))
{
ERR_PUSHBACK();
}
/* Otherwise, we're appending a character */
if (buff_end <= buff_ptr)
ERR_PUSHBACK(); /* too long! */
if (ch == '\\') {
ch=prot_getc(stream);
if (result != TIMSIEVE_OK)
ERR();
if (ch != '\"' && ch != '\\')
ERR_PUSHBACK();
}
*buff_ptr++ = ch;
break;
case LEXER_STATE_LITERAL:
if (('0' <= ch) && (ch <= '9')) {
unsigned long newcount = count * 10 + (ch - '0');
if (newcount < count)
ERR_PUSHBACK(); /* overflow */
/*
* XXX This should be fatal if non-synchronizing.
*/
count = newcount;
break;
}
if (ch != '+')
ERR_PUSHBACK();
ch=prot_getc(stream);
if (ch != '}')
ERR_PUSHBACK();
ch=prot_getc(stream);
if (ch < 0)
ERR();
if (ch != '\r')
ERR_PUSHBACK();
ch=prot_getc(stream);
if (ch < 0)
ERR();
if (ch != '\n')
ERR_PUSHBACK();
if (count > maxscriptsize) {
/* too big, eat the input */
for(;count > 0;count--) {
if(prot_getc(stream)==EOF)
break;
}
ERR();
}
if (outstr!=NULL)
{
*outstr = NULL;
result = string_allocate(count, NULL, outstr);
if (result != TIMSIEVE_OK)
ERR_PUSHBACK();
}
/* there is a literal string on the wire. let's read it */
if (outstr!=NULL) {
char *it = string_DATAPTR(*outstr),
*end = it + count;
while (it < end) {
*it=prot_getc(stream);
it++;
}
} else {
/* just read the chars and throw them away */
unsigned long lup;
for (lup=0;lup<count;lup++)
(void)prot_getc(stream);
}
lexer_state=LEXER_STATE_NORMAL;
return STRING;
case LEXER_STATE_NUMBER:
if (Uisdigit(ch)) {
unsigned long newcount = tmpnum * 10 + (ch - '0');
if (newcount < tmpnum)
ERR_PUSHBACK(); /* overflow */
tmpnum = newcount;
} else {
lexer_state=LEXER_STATE_NORMAL;
prot_ungetc(ch, stream);
if (outnum) *outnum = tmpnum;
return NUMBER;
}
break;
case LEXER_STATE_NORMAL:
if (Uisalpha(ch)) {
lexer_state=LEXER_STATE_ATOM;
*buff_ptr++ = tolower(ch);
break;
}
if (Uisdigit(ch)) {
lexer_state=LEXER_STATE_NUMBER;
tmpnum = ch -'0';
break;
}
switch (ch) {
case '(':
return '(';
case ')':
return ')';
case ' ':
return ' ';
case '\"':
lexer_state=LEXER_STATE_QSTR;
break;
case '*':
return '*';
case '{':
count = 0;
lexer_state=LEXER_STATE_LITERAL;
break;
case '\r':
lexer_state=LEXER_STATE_CR;
break;
case '\n':
lexer_state=LEXER_STATE_NORMAL;
return EOL;
break;
default:
return ch;
}
break;
case LEXER_STATE_ATOM:
if (!Uisalpha(ch)) {
int token;
buffer[ buff_ptr - buffer] = '\0';
/* We've got the atom. */
token = token_lookup((char *) buffer, (int) (buff_ptr - buffer));
if (token!=-1) {
lexer_state=LEXER_STATE_NORMAL;
prot_ungetc(ch, stream);
return token;
} else
ERR_PUSHBACK();
}
if (buff_end <= buff_ptr)
ERR_PUSHBACK(); /* atom too long */
*buff_ptr++ = tolower(ch);
break;
}
} /* while (1) */
/* never reached */
}
static int parse_rfc822(const char *s, time_t *tp, int dayonly)
{
const char *origs = s;
struct tm tm;
time_t t;
char month[4];
int zone_off = 0;
if (!s)
goto baddate;
memset(&tm, 0, sizeof(tm));
s = skip_fws(s);
if (!s)
goto baddate;
if (Uisalpha(*s)) {
/* Day name -- skip over it */
s++;
if (!Uisalpha(*s))
goto baddate;
s++;
if (!Uisalpha(*s))
goto baddate;
s++;
s = skip_fws(s);
if (!s || *s++ != ',')
goto baddate;
s = skip_fws(s);
if (!s)
goto baddate;
}
if (!Uisdigit(*s))
goto baddate;
tm.tm_mday = *s++ - '0';
if (Uisdigit(*s)) {
tm.tm_mday = tm.tm_mday*10 + *s++ - '0';
}
/* Parse month name */
s = skip_fws(s);
if (!s)
goto baddate;
month[0] = *s++;
if (!Uisalpha(month[0]))
goto baddate;
month[1] = *s++;
if (!Uisalpha(month[1]))
goto baddate;
month[2] = *s++;
if (!Uisalpha(month[2]))
goto baddate;
month[3] = '\0';
for (tm.tm_mon = 0; tm.tm_mon < 12; tm.tm_mon++) {
if (!strcasecmp(month, monthname[tm.tm_mon])) break;
}
if (tm.tm_mon == 12)
goto baddate;
/* Parse year */
s = skip_fws(s);
if (!s || !Uisdigit(*s))
goto baddate;
tm.tm_year = *s++ - '0';
if (!Uisdigit(*s))
goto baddate;
tm.tm_year = tm.tm_year * 10 + *s++ - '0';
if (Uisdigit(*s)) {
if (tm.tm_year < 19)
goto baddate;
tm.tm_year -= 19;
tm.tm_year = tm.tm_year * 10 + *s++ - '0';
if (!Uisdigit(*s))
goto baddate;
tm.tm_year = tm.tm_year * 10 + *s++ - '0';
} else {
if (tm.tm_year < 70) {
/* two-digit year, probably after 2000.
* This patent was overturned, right?
*/
tm.tm_year += 100;
}
}
if (Uisdigit(*s)) {
/* five-digit date */
goto baddate;
}
if (tm.tm_mday > monthdays(tm.tm_year, tm.tm_mon))
goto baddate;
s = skip_fws(s);
if (s && !dayonly) {
/* Parse hour */
if (!s || !Uisdigit(*s))
goto badtime;
tm.tm_hour = *s++ - '0';
if (!Uisdigit(*s))
goto badtime;
tm.tm_hour = tm.tm_hour * 10 + *s++ - '0';
if (!s || *s++ != ':')
goto badtime;
/* Parse min */
if (!s || !Uisdigit(*s))
goto badtime;
tm.tm_min = *s++ - '0';
if (!Uisdigit(*s))
goto badtime;
tm.tm_min = tm.tm_min * 10 + *s++ - '0';
if (*s == ':') {
/* Parse sec */
if (!++s || !Uisdigit(*s))
goto badtime;
tm.tm_sec = *s++ - '0';
if (!Uisdigit(*s))
goto badtime;
tm.tm_sec = tm.tm_sec * 10 + *s++ - '0';
}
s = skip_fws(s);
if (s) {
/* Parse timezone offset */
if (*s == '+' || *s == '-') {
/* Parse numeric offset */
int east = (*s++ == '-');
if (!s || !Uisdigit(*s))
goto badzone;
zone_off = *s++ - '0';
if (!s || !Uisdigit(*s))
goto badzone;
zone_off = zone_off * 10 + *s++ - '0';
if (!s || !Uisdigit(*s))
goto badzone;
zone_off = zone_off * 6 + *s++ - '0';
if (!s || !Uisdigit(*s))
goto badzone;
zone_off = zone_off * 10 + *s++ - '0';
if (east)
zone_off = -zone_off;
}
else if (Uisalpha(*s)) {
char zone[4];
zone[0] = *s++;
if (!Uisalpha(*s)) {
/* Parse military (single-char) zone */
zone[1] = '\0';
lcase(zone);
if (zone[0] < 'j')
zone_off = (zone[0] - 'a' + 1) * 60;
else if (zone[0] == 'j')
goto badzone;
else if (zone[0] <= 'm')
zone_off = (zone[0] - 'a') * 60;
else if (zone[0] < 'z')
zone_off = ('m' - zone[0]) * 60;
else
zone_off = 0;
}
else {
zone[1] = *s++;
if (!Uisalpha(*s)) {
/* Parse UT (universal time) */
zone[2] = '\0';
lcase(zone);
if (strcmp(zone, "ut"))
goto badzone;
zone_off = 0;
}
else {
/* Parse 3-char time zone */
char *p;
zone[2] = *s;
zone[3] = '\0';
lcase(zone);
/* GMT (Greenwich mean time) */
if (!strcmp(zone, "gmt"))
zone_off = 0;
/* US time zone */
else {
p = strchr("aecmpyhb", zone[0]);
if (!p || zone[2] != 't')
goto badzone;
zone_off = (strlen(p) - 12) * 60;
if (zone[1] == 'd')
zone_off += 60;
else if (zone[1] != 's')
goto badzone;
}
}
}
}
else
badzone:
zone_off = 0;
}
}
else
badtime:
tm.tm_hour = 12;
tm.tm_isdst = -1;
if (!dayonly)
t = mkgmtime(&tm);
else {
assert(zone_off == 0);
t = mktime(&tm);
}
if (t >= 0) {
*tp = (t - zone_off * 60);
return s - origs;
}
baddate:
return -1;
}
