- Enhanced domain part globbing with infix and prefix asterisk

   matching and optional unanchored operation

diff --git a/filters.c b/filters.c
index e5497c2..2de5e29 100644 (file)
--- a/filters.c
+++ b/filters.c
@@ -1,4 +1,4 @@
-const char filters_rcs[] = "$Id: filters.c,v 1.10 2001/05/29 09:50:24 jongfoster Exp $";
+const char filters_rcs[] = "$Id: filters.c,v 1.11 2001/05/29 11:53:23 oes Exp $";
 /*********************************************************************
  *
  * File        :  $Source: /cvsroot/ijbswa/current/filters.c,v $
@@ -38,6 +38,9 @@ const char filters_rcs[] = "$Id: filters.c,v 1.10 2001/05/29 09:50:24 jongfoster
  *
  * Revisions   :
  *    $Log: filters.c,v $
+ *    Revision 1.11  2001/05/29 11:53:23  oes
+ *    "See why" link added to "blocked" page
+ *
  *    Revision 1.10  2001/05/29 09:50:24  jongfoster
  *    Unified blocklist/imagelist/permissionslist.
  *    File format is still under discussion, but the internal changes
@@ -1045,13 +1048,7 @@ struct url_spec dsplit(char *domain)
 
    memset(ret, '\0', sizeof(*ret));
 
-   if ((p = strrchr(domain, '.')))
-   {
-      if (*(++p) == '\0')
-      {
-         ret->toplevel = 1;
-      }
-   }
+   ret->unanchored = (domain[strlen(domain) - 1] == '.');
 
    ret->dbuf = strdup(domain);
 
@@ -1075,6 +1072,7 @@ struct url_spec dsplit(char *domain)
       memcpy(ret->dvec, v, size);
    }
 
+
    return(*ret);
 
 }
@@ -1085,10 +1083,17 @@ struct url_spec dsplit(char *domain)
  * Function    :  domaincmp
  *
  * Description :  Compare domain names.
- *                domaincmp("a.b.c" , "a.b.c")  => 0 (MATCH)
+ *                domaincmp("a.b.c",  "a.b.c")  => 0 (MATCH)
  *                domaincmp("a*.b.c", "a.b.c")  => 0 (MATCH)
+ *                domaincmp("a*.b.c", "abc.b.c")  => 0 (MATCH)
+ *                domaincmp("a*c.b.c","abbc.b.c")  => 0 (MATCH)
+ *                domaincmp("*a.b.c", "dabc.b.c")  => 0 (MATCH)
  *                domaincmp("b.c"   , "a.b.c")  => 0 (MATCH)
+ *                domaincmp("a.b"   , "a.b.c")  => 1 (DIFF)
+ *                domaincmp("a.b."  , "a.b.c")  => 0 (MATCH)
  *                domaincmp(""      , "a.b.c")  => 0 (MATCH)
+ *                
+ * FIXME: I need a definition!
  *
  * Parameters  :
  *          1  :  pattern = a domain that may contain a '*' as a wildcard.
@@ -1104,27 +1109,29 @@ int domaincmp(struct url_spec *pattern, struct url_spec *fqdn)
    char  *p,   *f;   /* chars    */
 
    pv = pattern->dvec;
-   pn = pattern->dcnt;
-
    fv = fqdn->dvec;
-   fn = fqdn->dcnt;
+   fn = pn = 0;
 
-   while ((pn > 0) && (fn > 0))
+   while (fn < fqdn->dcnt && pn < pattern->dcnt)
    {
-      p = pv[--pn];
-      f = fv[--fn];
+      p = pv[pn];
+      f = fv[fn];
 
-      while (*p && *f && (*p == tolower(*f)))
+      if (trivimatch(p, f))
       {
-         p++, f++;
+         if(pn)
+         {
+            return 1;
+         }
       }
-
-      if ((*p != tolower(*f)) && (*p != '*')) return(1);
+      else
+      {
+         pn++;
+      }
+      fn++;
    }
 
-   if (pn > 0) return(1);
-
-   return(0);
+   return ((pn < pattern->dcnt) || ((fn < fqdn->dcnt) && !pattern->unanchored));
 
 }
 

diff --git a/loaders.c b/loaders.c
index 2eb3007..a00b36d 100644 (file)
--- a/loaders.c
+++ b/loaders.c
@@ -1,4 +1,4 @@
-const char loaders_rcs[] = "$Id: loaders.c,v 1.10 2001/05/29 09:50:24 jongfoster Exp $";
+const char loaders_rcs[] = "$Id: loaders.c,v 1.11 2001/05/29 23:25:24 oes Exp $";
 /*********************************************************************
  *
  * File        :  $Source: /cvsroot/ijbswa/current/loaders.c,v $
@@ -35,6 +35,10 @@ const char loaders_rcs[] = "$Id: loaders.c,v 1.10 2001/05/29 09:50:24 jongfoster
  *
  * Revisions   :
  *    $Log: loaders.c,v $
+ *    Revision 1.11  2001/05/29 23:25:24  oes
+ *
+ *     - load_config_line() and load_permissions_file() now use chomp()
+ *
  *    Revision 1.10  2001/05/29 09:50:24  jongfoster
  *    Unified blocklist/imagelist/permissionslist.
  *    File format is still under discussion, but the internal changes
@@ -418,6 +422,7 @@ static int create_url_spec(struct url_spec * url, char * buf)
    url->dbuf = tmp_url->dbuf;
    url->dcnt = tmp_url->dcnt;
    url->dvec = tmp_url->dvec;
+   url->unanchored = tmp_url->unanchored;
 
    return 0; /* OK */
 }

diff --git a/miscutil.c b/miscutil.c
index 2afa5cf..c3c270c 100644 (file)
--- a/miscutil.c
+++ b/miscutil.c
@@ -1,4 +1,4 @@
-const char miscutil_rcs[] = "$Id: miscutil.c,v 1.2 2001/05/29 09:50:24 jongfoster Exp $";
+const char miscutil_rcs[] = "$Id: miscutil.c,v 1.3 2001/05/29 23:10:09 oes Exp $";
 /*********************************************************************
  *
  * File        :  $Source: /cvsroot/ijbswa/current/miscutil.c,v $
@@ -36,6 +36,12 @@ const char miscutil_rcs[] = "$Id: miscutil.c,v 1.2 2001/05/29 09:50:24 jongfoste
  *
  * Revisions   :
  *    $Log: miscutil.c,v $
+ *    Revision 1.3  2001/05/29 23:10:09  oes
+ *
+ *
+ *     - Introduced chomp()
+ *     - Moved strsav() from showargs to miscutil
+ *
  *    Revision 1.2  2001/05/29 09:50:24  jongfoster
  *    Unified blocklist/imagelist/permissionslist.
  *    File format is still under discussion, but the internal changes
@@ -385,6 +391,86 @@ char *strsav(char *old, const char *text_to_append)
 
 }
 
+
+/*********************************************************************
+ *
+ * Function    :  trivimatch
+ *
+ * Description :  Trivial string matching, with only one metacharacter,
+ *                namely '*', which stands for zero or more arbitrary
+ *                characters.
+ *
+ *                Note: The * is greedy, i.e. it will try to match as
+ *                      much text es possible.
+ *
+ * Parameters  :
+ *          1  :  pattern = pattern for matching
+ *          2  :  text    = text to be matched
+ *
+ * Returns     :  0 if match, else nonzero
+ *
+ *********************************************************************/
+int trivimatch(char *pattern, char *text)
+{
+   char *fallback; 
+   char *pat = pattern;
+   char *txt = text;
+   int wildcard = 0;
+  
+   while (*txt)
+   {
+      /* EOF pattern but !EOF text? */
+      if (*pat == '\0')
+      {
+         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;
+      }
+
+      /* Compare: */
+      if (*pat != *txt)
+      {
+         /* In wildcard mode, just try again */
+         if(wildcard)
+         {
+            /* Without wildcard mode, this is fatal! */
+            pat = fallback;
+         }
+
+         /* Bad luck otherwise */
+         else
+         {
+            return 1;
+         }
+      }
+      /* We had a match, advance */
+      else
+      {
+         pat++;
+      }
+      txt++;
+  }
+
+  if(*pat == '*')  pat++;
+
+  /* Hey, we've made it all the way through! */
+  return(*pat);
+
+}
+
 /*
   Local Variables:
   tab-width: 3

diff --git a/miscutil.h b/miscutil.h
index 8dd55b1..02a54d2 100644 (file)
--- a/miscutil.h
+++ b/miscutil.h
@@ -1,6 +1,6 @@
 #ifndef _MISCUTIL_H
 #define _MISCUTIL_H
-#define MISCUTIL_H_VERSION "$Id: miscutil.h,v 1.2 2001/05/29 09:50:24 jongfoster Exp $"
+#define MISCUTIL_H_VERSION "$Id: miscutil.h,v 1.3 2001/05/29 23:10:09 oes Exp $"
 /*********************************************************************
  *
  * File        :  $Source: /cvsroot/ijbswa/current/miscutil.h,v $
@@ -37,6 +37,12 @@
  *
  * Revisions   :
  *    $Log: miscutil.h,v $
+ *    Revision 1.3  2001/05/29 23:10:09  oes
+ *
+ *
+ *     - Introduced chomp()
+ *     - Moved strsav() from showargs to miscutil
+ *
  *    Revision 1.2  2001/05/29 09:50:24  jongfoster
  *    Unified blocklist/imagelist/permissionslist.
  *    File format is still under discussion, but the internal changes
@@ -81,7 +87,9 @@ extern int strcmpic(const char *s1, const char *s2);
 extern int strncmpic(const char *s1, const char *s2, size_t n);
 
 extern char *strsav(char *old, const char *text_to_append);
+
 extern char *chomp(char *string);
+extern int trivimatch(char *pattern, char *text);
 
 #ifdef __MINGW32__
 extern char *strdup(const char *s);

diff --git a/project.h b/project.h
index 281205b..0cb3aba 100644 (file)
--- a/project.h
+++ b/project.h
@@ -1,6 +1,6 @@
 #ifndef _PROJECT_H
 #define _PROJECT_H
-#define PROJECT_H_VERSION "$Id: project.h,v 1.7 2001/05/29 09:50:24 jongfoster Exp $"
+#define PROJECT_H_VERSION "$Id: project.h,v 1.8 2001/05/29 20:09:15 joergs Exp $"
 /*********************************************************************
  *
  * File        :  $Source: /cvsroot/ijbswa/current/project.h,v $
@@ -36,6 +36,9 @@
  *
  * Revisions   :
  *    $Log: project.h,v $
+ *    Revision 1.8  2001/05/29 20:09:15  joergs
+ *    HTTP_REDIRECT_TEMPLATE fixed.
+ *
  *    Revision 1.7  2001/05/29 09:50:24  jongfoster
  *    Unified blocklist/imagelist/permissionslist.
  *    File format is still under discussion, but the internal changes
@@ -405,7 +408,7 @@ struct url_spec
    char  *dbuf;
    char **dvec;
    int    dcnt;
-   int    toplevel;
+   int    unanchored;
 
    char *path;
    int   pathlen;