void
_logdebuginfo(server_info_t info, int level, const char *ctl, ...)
{
va_list va;
char *str = NULL;
int ipstrpad;
if (level <= debug_level) {
va_start(va, ctl);
vasprintf(&str, ctl, va);
if (str) {
if (info->ipstr == NULL || strcmp(info->target, info->ipstr) == 0) {
_logdebug(level, "%s: %s", info->target, str);
} else {
if ((ipstrpad = 15 - strlen(info->ipstr)) < 0)
ipstrpad = 0;
_logdebug(level, "%s %*.*s(%s): %s", info->target,
ipstrpad, ipstrpad, "",
info->ipstr, str);
}
free(str);
}
va_end(va);
}
}
static int utl_pmx_get_limits(const char *pat, const char *pat_end, const char *txt, int braced,
int32_t *c_beg_ptr, int32_t *c_end_ptr, int32_t *c_esc_ptr)
{
int32_t c_beg = '(';
int32_t c_end = ')';
int32_t c_esc = '\0';
int32_t ch;
_logdebug("BRACE: [%.*s]",pat_end-pat,pat);
if (pat < pat_end) { /* <B()\> <Q""\>*/
pat += utl_pmx_nextch(pat,&c_esc);
if (pat < pat_end) {
c_beg = c_esc; c_esc = '\0';
pat += utl_pmx_nextch(pat,&c_end);
}
if (pat < pat_end) {
pat += utl_pmx_nextch(pat,&c_esc);
}
}
else { /* Just <B> or <Q>, try to infer the braces */
c_beg = '\0';
(void)utl_pmx_nextch(txt,&ch);
if (braced) {
if (ch == '(') {c_beg=ch; c_end=')';}
else if (ch == '[') {c_beg=ch; c_end=']';}
else if (ch == '{') {c_beg=ch; c_end='}';}
else if (ch == '<') {c_beg=ch; c_end='>';}
}
else { // Quoted string
c_esc = '\\';
if (ch == '"') {c_beg=ch; c_end=ch;}
else if (ch == '\'') {c_beg=ch; c_end=ch;}
else if (ch == '`') {c_beg=ch; c_end=ch;}
else if (ch == 0x91) {c_beg=ch; c_end=0x92;} /* ANSI single quotes */
else if (ch == 0x93) {c_beg=ch; c_end=0x94;} /* ANSI double quotes */
else if (ch == 0x2018) {c_beg=ch; c_end=0x2019;} /* Unicode single quotes */
else if (ch == 0x201C) {c_beg=ch; c_end=0x201D;} /* Unicode double quotes */
}
if (c_beg=='\0') { /* Valid both as quoted string or braces */
if (ch == '\xAB') {c_beg=ch; c_end='\xBB';} /* Unicode and ISO-8859-1 "<<" and ">>" */
else if (ch == '\x8B') {c_beg=ch; c_end='\x9B';} /* Unicode and ISO-8859-1 "<" and ">" */
else if (ch == 0x2039) {c_beg=ch; c_end=0x203A;} /* Unicode Single pointing Angle Quotation */
else if (ch == 0x2329) {c_beg=ch; c_end=0x232A;} /* Unicode ANGLE BRACKETS */
else if (ch == 0x27E8) {c_beg=ch; c_end=0x27E9;} /* Unicode MATHEMATICAL ANGLE BRACKETS */
else if (ch == 0x27EA) {c_beg=ch; c_end=0x27EB;} /* Unicode MATHEMATICAL DOUBLE ANGLE BRACKETS */
else return 0;
}
}
_logdebug("open:'%d' close:'%d' esc:'%d'",c_beg,c_end,c_esc);
*c_beg_ptr = c_beg;
*c_end_ptr = c_end;
*c_esc_ptr = c_esc;
return 1;
}
static int utl_pmx_get_delimited(const char *pat, const char *txt,int32_t c_beg, int32_t c_end, int32_t c_esc)
{
int n;
const char *s;
int cnt;
int32_t ch;
s = txt;
n = utl_pmx_nextch(s,&ch);
if (n == 0 || ch != c_beg) return 0;
cnt = 0;
do {
s += n;
n = utl_pmx_nextch(s,&ch);
if (ch == '\0') return 0;
_logdebug("BRACE: '%c' cnt:%d",ch,cnt);
if (ch == c_end) { if (cnt == 0) return (s+n)-txt; else cnt--; }
else if (ch == c_beg) { cnt++; }
else if (ch == c_esc) { s += n; n = utl_pmx_nextch(s,&ch); }
} while (ch);
utl_pmx_set_paterror(pat);
return s-txt;
}
static const char *utl_pmx_alt(const char *pat, const char **txt_ptr)
{
int paren=0;
utl_pmx_state_s *state;
int inv;
const char *ret = utl_emptystring;
while (*pat) {
_logdebug("ALT: %s (%d)",pat,utl_pmx_stack_ptr);
switch (*pat++) {
case '%': if (*pat) pat++; /* works for utf8 as well */
break;
case '(': paren++;
utl_pmx_newcap(NULL);
break;
case ')': if (paren > 0) {
paren--;
break;
}
if (utl_pmx_stack_ptr < 2) {
utl_pmx_set_paterror(pat);
break;
}
// {{ If we are here, we have NOT matched what is in the (...)
state = utl_pmx_state_top();
inv = state->inv;
ret = pat;
if (inv) { /* It's ok, we WANTED to fail */
*txt_ptr = pmxstart(state->cap);
utl_pmx_state_pop();
return ret;
}
if (state->n >= state->min_n) { /* It's ok, we matched enough times */
utl_pmx_capt[state->cap][0] = state->txt;
utl_pmx_capt[state->cap][1] = *txt_ptr;
utl_pmx_state_pop();
return ret;
}
// }}
break;
case '<': while (*pat && *pat != '>') pat++;
while (*pat == '>') pat++;
break;
case '|': if (paren == 0) {
state = utl_pmx_state_top();
*txt_ptr = pmxstart(state->cap);
return pat;
}
break;
}
}
return utl_emptystring;
}
const char *utl_pmx_search(const char *pat, const char *txt, int fromstart)
{
const char *ret=NULL;
utl_pmx_error = NULL;
utl_pmx_case = 1;
if (strncmp(pat,"<utf>",5) == 0) {pat+=5; utl_pmx_utf8=1;}
else if (strncmp(pat,"<iso>",5) == 0) {pat+=5; utl_pmx_utf8=0;}
if (*pat == '^') ret = utl_pmx_match(pat+1,txt);
else while (!(ret = utl_pmx_match(pat,txt)) && *txt && !fromstart) {
txt += utl_pmx_utf8 ? utl_pmx_get_utf8(txt, NULL) : 1;
}
_logdebug("ret: %p",ret);
return ret;
}
static int32_t utl_pmx_iscapt(const char *pat, const char *txt)
{
int32_t len = 0;
uint8_t capnum = 0;
const char *cap;
if ('1' <= *pat && *pat <= '9') {
capnum = *pat - '0';
_logdebug("capt: %d %d",capnum,utl_pmx_capnum);
if (capnum < utl_pmx_capnum) {
cap = pmxstart(capnum);
while (cap < pmxend(capnum) && *cap && (*cap == *txt)) {
len++; txt++; cap++;
}
if (cap < pmxend(capnum)) len = 0;
}
}
return len;
}
static const char *utl_pmx_match(const char *pat, const char *txt)
{
int32_t len;
int32_t ch;
int32_t c1;
int16_t inv =0;
utl_pmx_state_s *state;
utl_pmx_state_reset();
utl_pmx_state_push(pat,txt,1,1,0);
while (*pat) {
_logdebug("[MATCH] %d [%s] [%s]",pmxcount(),pat,txt);
c1 = 0;
switch (*pat) {
case '(' : pat++;
if (*pat == '|') {inv = 1; pat++;}
if (!utl_pmx_state_push(pat,txt,1,1,inv))
utl_pmx_set_paterror(pat);
break;
case '|' : pat = utl_pmx_alt_skip(pat);
break;
case ')' : pat++;
_logdebug(")->%d",utl_pmx_stack_ptr);
if (utl_pmx_stack_ptr < 2) {
utl_pmx_set_paterror(pat-1);
break;
}
/* If we are here, we have matched what is in the (...) */
state = utl_pmx_state_top();
inv = state->inv;
if (inv) { /* we shouldn't have matched it :( */
utl_pmx_state_pop();
utl_pmx_FAIL;
}
utl_pmx_capt[state->cap][1] = txt;
state->n++;
_logdebug("match #%d min:%d max:%d",state->n,state->min_n, state->max_n);
if (state->n < state->max_n) {
utl_pmx_capt[state->cap][0] = txt;
pat = state->pat; /* try to match once more */
}
else {
utl_pmx_capt[state->cap][0] = state->txt;
utl_pmx_state_pop();
}
break;
case '<' : if (!utl_pmx_class(&pat,&txt)) utl_pmx_FAIL;
break;
case '%' : if (pat[1]) len = utl_pmx_nextch(++pat, &c1);
default : if (c1 == 0) len = utl_pmx_nextch(pat, &c1);
len = utl_pmx_nextch(txt, &ch);
if (!utl_pmx_case) {
ch = utl_pmx_fold(ch);
c1 = utl_pmx_fold(c1);
}
if (ch != c1) {
_logdebug("FAIL: %d %d",c1,ch);
utl_pmx_FAIL;
}
txt += len;
pat += len;
break;
fail : pat = utl_pmx_alt(pat, &txt) ; /* search for an alternative */
if (*pat == '\0') utl_pmx_capnum = 0;
break;
}
}
utl_pmx_capt[0][1] = txt;
for (len = utl_pmx_capnum; len < utl_pmx_MAXCAPT; len++) {
utl_pmx_capt[len][0] = utl_pmx_capt[len][1] = NULL;
}
_logdebug("res: %p - %p",utl_pmx_capt[0][0],utl_pmx_capt[0][1]);
return utl_pmx_capt[0][0];
}
static int utl_pmx_class(const char **pat_ptr, const char **txt_ptr)
{
int inv = 0;
const char *pat = *pat_ptr;
const char *txt = *txt_ptr;
const char *pat_end;
int32_t len = 0;
int32_t n = 0;
int32_t min_n = 0;
int32_t max_n = 0;
int32_t ch;
_logdebug("class:[%s][%s]",pat,txt);
pat++; /* skip the '<' */
// {{ Find the end of the pattern
pat_end=pat;
while (*pat_end && *pat_end != '>') pat_end++;
if (pat_end[1] == '>') pat_end++; /* allow just one '>' at the end of a pattern */
//}}
// {{ Get how many times the match has to occur.
// Examples: <2-4l> matches 2,3 or 4 lower case letters
// <-4d> matches from 0 to 4 decimal digits
// <2-s> matches at 2 spaces or more
// <3u> matches exactly 3 upper case letters
while ('0' <= *pat && *pat <= '9')
min_n = (min_n*10) + (*pat++ - '0');
if (*pat == '-') {
pat++;
while ('0'<=*pat && *pat <= '9')
max_n = (max_n*10) + (*pat++ - '0');
if (max_n == 0) max_n = INT32_MAX;
}
if (max_n < min_n) max_n = min_n;
if (max_n == 0) {
switch (*pat) {
case '*' : min_n = 0; max_n = INT32_MAX; pat++; break;
case '+' : min_n = 1; max_n = INT32_MAX; pat++; break;
case '?' : min_n = 0; max_n = 1; pat++; break;
default : min_n = 1; max_n = 1; break;
}
}
// }}
if (*pat == '!') {inv = 1; pat++;}
// {{ This is for handling repetition of patterns between parenthesis:
// Example: '<*>(\\'|<!='>)' <-- single quoted string (\' as escaped)
// <3>(xa|pt) <-- "ptptpt" or "ptxaxa" or "xaxapt" or ...
if (pat[0] == '>' && pat[1] == '(') {
pat += 2;
if (!utl_pmx_state_push(pat,txt,min_n,max_n,inv))
utl_pmx_set_paterror(pat);
*pat_ptr = pat;
return 1;
}
// }}
// {{ Matches a pattern n times
#define utl_W(tst) while ((len = utl_pmx_nextch(txt,&ch)) && ((!tst) == inv) && (n<max_n)) {n++; txt+=len;}
switch (*pat) {
case 'a' : utl_W(( utl_isalpha(ch) )); break;
case 's' : utl_W(( utl_isspace(ch) )); break;
case 'u' : utl_W(( utl_isupper(ch) )); break;
case 'l' : utl_W(( utl_islower(ch) )); break;
case 'd' : utl_W(( utl_isdigit(ch) )); break;
case 'k' : utl_W(( utl_isblank(ch) )); break;
case 'x' : utl_W(( utl_isxdigit(ch))); break;
case 'w' : utl_W(( utl_isalnum(ch) )); break;
case 'c' : utl_W(( utl_iscntrl(ch) )); break;
case 'g' : utl_W(( utl_isgraph(ch) )); break;
case 'p' : utl_W(( utl_ispunct(ch) )); break;
case 'r' : utl_W(( utl_isprint(ch) )); break;
case 'i' : utl_W((ch < 0x80)) ; break;
case '.' : utl_W((ch !='\0' && ch !='\n')) ; break;
case '=' : utl_W(utl_pmx_isin_chars(pat+1,pat_end,ch)) ; break;
case '#' : utl_W(utl_pmx_isin_codes(pat+1,pat_end,ch)) ; break;
case 'N' : utl_W((txt[0]=='\r'
? (txt[1] == '\n'? (len++) : 1)
: (txt[0] == '\n'? 1 : 0) )) ; break;
case 'Q' : utl_W((len=utl_pmx_delimited(pat+1,pat_end,txt, UTL_PMX_QUOTED))); break;
case 'B' : utl_W((len=utl_pmx_delimited(pat+1,pat_end,txt, UTL_PMX_BRACED))); break;
case 'I' : utl_pmx_case = 0; n=min_n; break;
case 'C' : utl_pmx_case = 1; n=min_n; break;
case '$' : if (*txt == '\0') n=min_n; break;
case '>' : utl_pmx_set_paterror(pat); return 0;
case '^' : if (inv) utl_pmx_set_paterror(pat); inv = 0;
utl_W((len=utl_pmx_iscapt(pat+1,txt)));
break;
case ':' : if (utl_pmx_ext)
utl_W((len=utl_pmx_ext(pat+1,txt,len,ch)));
break;
default : ; //utl_pmx_set_paterror(pat);
}
#undef utl_W
// }}
// {{ Advance pattern
while (*pat_end == '>') pat_end++;
*pat_ptr=pat_end;
// }}
/* Not enough matches */
if (n < min_n) return 0;
// {{ Advance matched text
*txt_ptr = txt;
// }}
return 1;
}
static int utl_pmx_get_utf8(const char *txt, int32_t *ch)
{
int len;
uint8_t *s = (uint8_t *)txt;
uint8_t first = *s;
int32_t val;
_logdebug("About to get UTF8: %s in %p",txt,ch);
fsm {
fsmSTART {
if (*s <= 0xC1) { val = *s; len = (*s > 0); fsmGOTO(end); }
if (*s <= 0xDF) { val = *s & 0x1F; len = 2; fsmGOTO(len2); }
if (*s == 0xE0) { val = *s & 0x0F; len = 3; fsmGOTO(len3_0); }
if (*s <= 0xEC) { val = *s & 0x0F; len = 3; fsmGOTO(len3_1); }
if (*s == 0xED) { val = *s & 0x0F; len = 3; fsmGOTO(len3_2); }
if (*s <= 0xEF) { val = *s & 0x0F; len = 3; fsmGOTO(len3_1); }
if (*s == 0xF0) { val = *s & 0x07; len = 4; fsmGOTO(len4_0); }
if (*s <= 0xF3) { val = *s & 0x07; len = 4; fsmGOTO(len4_1); }
if (*s == 0xF4) { val = *s & 0x07; len = 4; fsmGOTO(len4_2); }
fsmGOTO(invalid);
}
fsmSTATE(len4_0) {
s++; if ( *s < 0x90 || 0xbf < *s) fsmGOTO(invalid);
val = (val << 6) | (*s & 0x3F);
fsmGOTO(len3_1);
}
fsmSTATE(len4_1) {
s++; if ( *s < 0x80 || 0xbf < *s) fsmGOTO(invalid);
val = (val << 6) | (*s & 0x3F);
fsmGOTO(len3_1);
}
fsmSTATE(len4_2) {
s++; if ( *s < 0x80 || 0x8f < *s) fsmGOTO(invalid);
val = (val << 6) | (*s & 0x3F);
fsmGOTO(len3_1);
}
fsmSTATE(len3_0) {
s++; if ( *s < 0xA0 || 0xbf < *s) fsmGOTO(invalid);
val = (val << 6) | (*s & 0x3F);
fsmGOTO(len2);
}
fsmSTATE(len3_1) {
s++; if ( *s < 0x80 || 0xbf < *s) fsmGOTO(invalid);
val = (val << 6) | (*s & 0x3F);
fsmGOTO(len2);
}
fsmSTATE(len3_2) {
s++; if ( *s < 0x80 || 0x9f < *s) fsmGOTO(invalid);
val = (val << 6) | (*s & 0x3F);
fsmGOTO(len2);
}
fsmSTATE(len2) {
s++; if ( *s < 0x80 || 0xbf < *s) fsmGOTO(invalid);
val = (val << 6) | (*s & 0x3F);
fsmGOTO(end);
}
fsmSTATE(invalid) {val = first; len = 1;}
fsmSTATE(end) { }
}
if (ch) *ch = val;
return len;
}
