+/*********************************************************************
+ *
+ * Function : compile_pattern
+ *
+ * Description : Compiles a host, domain or TAG pattern.
+ *
+ * Parameters :
+ * 1 : pattern = The pattern to compile.
+ * 2 : anchoring = How the regex should be anchored.
+ * Can be either one of NO_ANCHORING,
+ * LEFT_ANCHORED or RIGHT_ANCHORED.
+ * 3 : url = In case of failures, the spec member is
+ * logged and the structure freed.
+ * 4 : regex = Where the compiled regex should be stored.
+ *
+ * Returns : JB_ERR_OK - Success
+ * JB_ERR_MEMORY - Out of memory
+ * JB_ERR_PARSE - Cannot parse regex
+ *
+ *********************************************************************/
+static jb_err compile_pattern(const char *pattern, enum regex_anchoring anchoring,
+ struct url_spec *url, regex_t **regex)
+{
+ int errcode;
+ char rebuf[BUFFER_SIZE];
+ const char *fmt = NULL;
+
+ assert(pattern);
+ assert(strlen(pattern) < sizeof(rebuf) - 2);
+
+ if (pattern[0] == '\0')
+ {
+ *regex = NULL;
+ return JB_ERR_OK;
+ }
+
+ switch (anchoring)
+ {
+ case NO_ANCHORING:
+ fmt = "%s";
+ break;
+ case RIGHT_ANCHORED:
+ fmt = "%s$";
+ break;
+ case LEFT_ANCHORED:
+ fmt = "^%s";
+ break;
+ default:
+ log_error(LOG_LEVEL_FATAL,
+ "Invalid anchoring in compile_pattern %d", anchoring);
+ }
+
+ *regex = zalloc(sizeof(**regex));
+ if (NULL == *regex)
+ {
+ free_url_spec(url);
+ return JB_ERR_MEMORY;
+ }
+
+ snprintf(rebuf, sizeof(rebuf), fmt, pattern);
+
+ errcode = regcomp(*regex, rebuf, (REG_EXTENDED|REG_NOSUB|REG_ICASE));
+
+ if (errcode)
+ {
+ size_t errlen = regerror(errcode, *regex, rebuf, sizeof(rebuf));
+ if (errlen > (sizeof(rebuf) - (size_t)1))
+ {
+ errlen = sizeof(rebuf) - (size_t)1;
+ }
+ rebuf[errlen] = '\0';
+ log_error(LOG_LEVEL_ERROR, "error compiling %s from %s: %s",
+ pattern, url->spec, rebuf);
+ free_url_spec(url);
+
+ return JB_ERR_PARSE;
+ }
+
+ return JB_ERR_OK;
+
+}
+
+
+/*********************************************************************
+ *
+ * Function : compile_url_pattern
+ *
+ * Description : Compiles the three parts of an URL pattern.
+ *
+ * Parameters :
+ * 1 : url = Target url_spec to be filled in.
+ * 2 : buf = The url pattern to compile. Will be messed up.
+ *
+ * Returns : JB_ERR_OK - Success
+ * JB_ERR_MEMORY - Out of memory
+ * JB_ERR_PARSE - Cannot parse regex
+ *
+ *********************************************************************/
+static jb_err compile_url_pattern(struct url_spec *url, char *buf)
+{
+ char *p;
+
+ p = strchr(buf, '/');
+ if (NULL != p)
+ {
+ /*
+ * Only compile the regex if it consists of more than
+ * a single slash, otherwise it wouldn't affect the result.
+ */
+ if (p[1] != '\0')
+ {
+ /*
+ * XXX: does it make sense to compile the slash at the beginning?
+ */
+ jb_err err = compile_pattern(p, LEFT_ANCHORED, url, &url->preg);
+
+ if (JB_ERR_OK != err)
+ {
+ return err;
+ }
+ }
+ *p = '\0';
+ }
+
+ p = strchr(buf, ':');
+ if (NULL != p)
+ {
+ *p++ = '\0';
+ url->port_list = strdup(p);
+ if (NULL == url->port_list)
+ {
+ return JB_ERR_MEMORY;
+ }
+ }
+ else
+ {
+ url->port_list = NULL;
+ }
+
+ if (buf[0] != '\0')
+ {
+ return compile_host_pattern(url, buf);
+ }
+
+ return JB_ERR_OK;
+
+}
+
+
+#ifdef FEATURE_EXTENDED_HOST_PATTERNS
+/*********************************************************************
+ *
+ * Function : compile_host_pattern
+ *
+ * Description : Parses and compiles a host pattern..
+ *
+ * Parameters :
+ * 1 : url = Target url_spec to be filled in.
+ * 2 : host_pattern = Host pattern to compile.
+ *
+ * Returns : JB_ERR_OK - Success
+ * JB_ERR_MEMORY - Out of memory
+ * JB_ERR_PARSE - Cannot parse regex
+ *
+ *********************************************************************/
+static jb_err compile_host_pattern(struct url_spec *url, const char *host_pattern)
+{
+ return compile_pattern(host_pattern, RIGHT_ANCHORED, url, &url->host_regex);
+}
+
+#else
+
+/*********************************************************************
+ *
+ * Function : compile_host_pattern
+ *
+ * Description : Parses and "compiles" an old-school host pattern.
+ *
+ * Parameters :
+ * 1 : url = Target url_spec to be filled in.
+ * 2 : host_pattern = Host pattern to parse.
+ *
+ * Returns : JB_ERR_OK - Success
+ * JB_ERR_MEMORY - Out of memory
+ * JB_ERR_PARSE - Cannot parse regex
+ *
+ *********************************************************************/
+static jb_err compile_host_pattern(struct url_spec *url, const char *host_pattern)
+{
+ char *v[150];
+ size_t size;
+ char *p;
+
+ /*
+ * Parse domain part
+ */
+ if (host_pattern[strlen(host_pattern) - 1] == '.')
+ {
+ url->unanchored |= ANCHOR_RIGHT;
+ }
+ if (host_pattern[0] == '.')
+ {
+ url->unanchored |= ANCHOR_LEFT;
+ }
+
+ /*
+ * Split domain into components
+ */
+ url->dbuffer = strdup(host_pattern);
+ if (NULL == url->dbuffer)
+ {
+ free_url_spec(url);
+ return JB_ERR_MEMORY;
+ }
+
+ /*
+ * Map to lower case
+ */
+ for (p = url->dbuffer; *p ; p++)
+ {
+ *p = (char)tolower((int)(unsigned char)*p);
+ }
+
+ /*
+ * Split the domain name into components
+ */
+ url->dcount = ssplit(url->dbuffer, ".", v, SZ(v), 1, 1);
+
+ if (url->dcount < 0)
+ {
+ free_url_spec(url);
+ return JB_ERR_MEMORY;
+ }
+ else if (url->dcount != 0)
+ {
+ /*
+ * Save a copy of the pointers in dvec
+ */
+ size = (size_t)url->dcount * sizeof(*url->dvec);
+
+ url->dvec = (char **)malloc(size);
+ if (NULL == url->dvec)
+ {
+ free_url_spec(url);
+ return JB_ERR_MEMORY;
+ }
+
+ memcpy(url->dvec, v, size);
+ }
+ /*
+ * else dcount == 0 in which case we needn't do anything,
+ * since dvec will never be accessed and the pattern will
+ * match all domains.
+ */
+ return JB_ERR_OK;
+}
+
+
+/*********************************************************************
+ *
+ * Function : simplematch
+ *
+ * Description : String matching, with a (greedy) '*' wildcard that
+ * stands for zero or more arbitrary characters and
+ * character classes in [], which take both enumerations
+ * and ranges.
+ *
+ * Parameters :
+ * 1 : pattern = pattern for matching
+ * 2 : text = text to be matched
+ *
+ * Returns : 0 if match, else nonzero
+ *
+ *********************************************************************/
+static int simplematch(const char *pattern, const char *text)
+{
+ const unsigned char *pat = (const unsigned char *)pattern;
+ const unsigned char *txt = (const unsigned char *)text;
+ const unsigned char *fallback = pat;
+ int wildcard = 0;
+
+ unsigned char lastchar = 'a';
+ unsigned i;
+ unsigned char charmap[32];
+
+ while (*txt)
+ {
+
+ /* EOF pattern but !EOF text? */
+ if (*pat == '\0')
+ {
+ if (wildcard)
+ {
+ pat = fallback;
+ }
+ else
+ {
+ return 1;
+ }
+ }
+
+ /* '*' in the pattern? */
+ if (*pat == '*')
+ {
+
+ /* The pattern ends afterwards? Speed up the return. */
+ if (*++pat == '\0')
+ {
+ return 0;
+ }
+
+ /* Else, set wildcard mode and remember position after '*' */
+ wildcard = 1;
+ fallback = pat;
+ }
+
+ /* Character range specification? */
+ if (*pat == '[')
+ {
+ memset(charmap, '\0', sizeof(charmap));
+
+ while (*++pat != ']')
+ {
+ if (!*pat)
+ {
+ return 1;
+ }
+ else if (*pat == '-')
+ {
+ if ((*++pat == ']') || *pat == '\0')
+ {
+ return(1);
+ }
+ for (i = lastchar; i <= *pat; i++)
+ {
+ charmap[i / 8] |= (unsigned char)(1 << (i % 8));
+ }
+ }
+ else
+ {
+ charmap[*pat / 8] |= (unsigned char)(1 << (*pat % 8));
+ lastchar = *pat;
+ }
+ }
+ } /* -END- if Character range specification */
+
+
+ /*
+ * Char match, or char range match?
+ */
+ if ( (*pat == *txt)
+ || (*pat == '?')
+ || ((*pat == ']') && (charmap[*txt / 8] & (1 << (*txt % 8)))) )
+ {
+ /*
+ * Sucess: Go ahead
+ */
+ pat++;
+ }
+ else if (!wildcard)
+ {
+ /*
+ * No match && no wildcard: No luck
+ */
+ return 1;
+ }
+ else if (pat != fallback)
+ {
+ /*
+ * Increment text pointer if in char range matching
+ */
+ if (*pat == ']')
+ {
+ txt++;
+ }
+ /*
+ * Wildcard mode && nonmatch beyond fallback: Rewind pattern
+ */
+ pat = fallback;
+ /*
+ * Restart matching from current text pointer
+ */
+ continue;
+ }
+ txt++;
+ }
+
+ /* Cut off extra '*'s */
+ if(*pat == '*') pat++;
+
+ /* If this is the pattern's end, fine! */
+ return(*pat);
+
+}
+
+