added reverse match, used defines for match modes, fixed a argument parsing bug that would break on the second line of input if using non-simple matching

1 files changed, 45 insertions, 30 deletions

diff --git a/urimatch.c b/urimatch.c
index a90fb68..91177f2 100644
--- a/urimatch.c
+++ b/urimatch.c
@@ -6,29 +6,43 @@
 
 #define LINE_LENGTH 1024
 
-int match(char negate,char *part,char *arg) {
- if(negate) {
-  if(part == 0) return 1;//we found that the part isn't here!
- } else {
-  if(part) {
-   if(!fnmatch(arg,part,FNM_NOESCAPE)) return 1;
-  }
+#define MATCH_UNEXIST 0
+#define MATCH_PATTERN 1
+#define MATCH_REVERSE -1
+
+// return 1 if the match and rule should have the main function print this URL.
+int match(char rule,char *part,char *arg) {
+ switch(rule) {
+  case MATCH_UNEXIST:
+   if(part == 0) return 1;
+   break;
+  case MATCH_REVERSE:
+   if(part) {
+     if(!fnmatch(arg,part,FNM_NOESCAPE)) return 0;
+     else return 1;
+   }
+   break;
+  case MATCH_PATTERN:
+   if(part) if(!fnmatch(arg,part,FNM_NOESCAPE)) return 1;
+   break;
+  default:
+   fprintf(stderr,"oh god. what the hell happened to get here?\n");
+   break;
  }
  return 0;
 }
 
 int main(int argc,char *argv[]) {
- int i;
+ int i,j;;
  int ret=1;
  struct uri u;
- char negate=0;
+ char rule=MATCH_PATTERN;
  char *line=malloc(LINE_LENGTH);
  char copy[LINE_LENGTH];
  if(argc < 2) {
-  printf("usage: urimatch [-][n][s|u|k|d|D|P|p|q|f] [string]\n");
+  printf("usage: urimatch [-][n][r][s|u|k|d|P|p|q|f] [string]\n");
   printf("scheme://username:password@domain:port/path?query_string#fragment_id\n");
   printf("s://u:k@d:P/p?q#f\n");
-  printf("The D flag is special. it matches its argument against the last bytes of the input url's domain.\n");
   printf("This allows matching of subdomains, like `echo epoch.ano | urimatch -D ano` would match.\n");
   printf("the 'n' flag can be put before any of the other flags to check for a missing.\n");
   return 1;
@@ -40,30 +54,31 @@ int main(int argc,char *argv[]) {
   memset(&u,0,sizeof(u));
   urifromline(&u,line);
   //use the character in argv[1] to match stdin against argv[2]. if match print whole line.
-  for(i=1;i<argc;i+=2) {
-   if(negate) {i--;}//we didn't really need to go that far.
-   negate=0;
-   if(argv[i][0] == '-') argv[i]++;
-   if(argv[i][0] == 'n') {argv[i]++; negate=1; }//heh.
+  for(i=1;i<argc;i+=2) {//for each line, we're reparsing the arguments...
+   rule=MATCH_PATTERN;//default
+   j=0;
    switch(argv[i][0]) {
-    case 's': if(match(negate,u.scheme,argv[i+1]))       { printf("%s\n",copy); ret=0;} break;
-    case 'u': if(match(negate,u.username,argv[i+1]))     { printf("%s\n",copy); ret=0;} break;
-    case 'k': if(match(negate,u.password,argv[i+1]))     { printf("%s\n",copy); ret=0;} break;
-    case 'd': if(match(negate,u.domain,argv[i+1]))       { printf("%s\n",copy); ret=0;} break;
-    case 'P': if(match(negate,u.port,argv[i+1]))         { printf("%s\n",copy); ret=0;} break;
-    case 'p': if(match(negate,u.path,argv[i+1]))         { printf("%s\n",copy); ret=0;} break;
-    case 'q': if(match(negate,u.query_string,argv[i+1])) { printf("%s\n",copy); ret=0;} break;
-    case 'f': if(match(negate,u.fragment_id,argv[i+1]))  { printf("%s\n",copy); ret=0;} break;
-    case 'D': //not sure how to look for a missing one of these. it'd be like d.
-     if(u.domain && argv[i+1] && strlen(u.domain) >= strlen(argv[i+1]) && !strncmp(u.domain+strlen(u.domain)-strlen(argv[i+1]),argv[i+1],strlen(argv[i+1]))) {
-      printf("%s\n",copy);
-      ret=0;
-     }
-     break;
+     case '-': j=1; rule=MATCH_PATTERN; break;//so we can't be modifying these! how the fuck?
+     case 'n': j=1; rule=MATCH_UNEXIST; break;
+     case 'r': j=1; rule=MATCH_REVERSE; break;
+     default: break;//handled by the next switch
+   }
+   switch(argv[i][j]) {
+    case 's': if(match(rule,u.scheme,argv[i+1]))       { printf("%s\n",copy); ret=0;} break;
+    case 'u': if(match(rule,u.username,argv[i+1]))     { printf("%s\n",copy); ret=0;} break;
+    case 'k': if(match(rule,u.password,argv[i+1]))     { printf("%s\n",copy); ret=0;} break;
+    case 'd': if(match(rule,u.domain,argv[i+1]))       { printf("%s\n",copy); ret=0;} break;
+    case 'P': if(match(rule,u.port,argv[i+1]))         { printf("%s\n",copy); ret=0;} break;
+    case 'p': if(match(rule,u.path,argv[i+1]))         { printf("%s\n",copy); ret=0;} break;
+    case 'q': if(match(rule,u.query_string,argv[i+1])) { printf("%s\n",copy); ret=0;} break;
+    case 'f': if(match(rule,u.fragment_id,argv[i+1]))  { printf("%s\n",copy); ret=0;} break;
+    // TODO: after a while, get rid of 'D' completely.
+    case 'D': fprintf(stderr,"the 'D' flag has been removed.\ninstead of: D 'host'\nuse: d '*host'\n");
     default:
      printf("unknown url part letter! '%c'\n",argv[i][0]);
      return ret;
    }
+   if(rule == MATCH_UNEXIST) {i--;}//UNEXIST doesn't consume the argument after it. so, go backwards 1. we're about to go ahead 2.
   }
  }
  return ret;