===================================================================
RCS file: /home/cvs/OpenXM/src/util/oxgreph.c,v
retrieving revision 1.1
retrieving revision 1.2
diff -u -p -r1.1 -r1.2
--- OpenXM/src/util/oxgreph.c	2006/03/04 07:48:12	1.1
+++ OpenXM/src/util/oxgreph.c	2006/03/04 10:43:38	1.2
@@ -1,4 +1,4 @@
-/* $OpenXM$ */
+/* $OpenXM: OpenXM/src/util/oxgreph.c,v 1.1 2006/03/04 07:48:12 takayama Exp $ */
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
@@ -11,7 +11,14 @@
    It greps files and outputs an html file to browse the files.
    Keyword should be given in euc or utf-8 depending on the coding system of the files. 
    It does not add meta section  of html file.
+
+   Example:
+   s.txt contains "insubunkai"
+   oxgreph --sjis `cat s.txt` $OpenXM_HOME/doc/cfep/html-ja_JP.sjis/html-jp/*.html >t.html
 */
+
+int Sjis = 0;
+int LeaveTag = 0;
 main(int argc, char *argv[]) {
   int i,m,pid,n, cpos,start;
   char *com;
@@ -23,12 +30,18 @@ main(int argc, char *argv[]) {
   m = 256;
   pid = getpid();
   for (i=1; i<argc; i++) {
-	m += strlen(argv[i])+1;
+    m += strlen(argv[i])+1;
   }
   com = (char *)mymalloc(m);
   sprintf(com,"grep ");
   for (i=1; i<argc; i++) {
-	sprintf(&(com[strlen(com)]),"%s ",argv[i]);
+    if (strcmp(argv[i],"--sjis")==0) {
+      Sjis = 1;
+    }else if (strcmp(argv[i],"--leaveTag")==0) {
+      LeaveTag = 1;
+    }else {
+      sprintf(&(com[strlen(com)]),"%s ",argv[i]);
+    }
   }
   sprintf(workf,"/tmp/tmp-oxgreph-%d.txt",pid);
   sprintf(&(com[strlen(com)]),">%s",workf);
@@ -37,29 +50,35 @@ main(int argc, char *argv[]) {
   printf("<ol>\n");
   fp = fopen(workf,"r");
   if (fp == NULL) {
-	fprintf(stderr,"Open error of the workfile.\n"); exit(1);
+    fprintf(stderr,"Open error of the workfile.\n"); exit(1);
   }
   while (fgets(s,SSIZE-1,fp)) {
-	cpos = -1; n = strlen(s);
-	for (i=0; i< n; i++) {
-	  if (s[i] == ':') {
-		cpos = i; break;
-	  }
-	}
-	if (cpos >= 0) {
-	  s[cpos] = 0; start = 0;
-	  for (i=cpos+1; i<n; i++) {
-        if (s[i] == 0x1b) {
-		  i++;
-		  while (s[i] != 0x1b) i++;
+    cpos = -1; n = strlen(s);
+    for (i=0; i< n; i++) {
+      if (s[i] == ':') {
+        cpos = i; break;
+      }
+    }
+    if (cpos >= 0) {
+      s[cpos] = 0; start = 0;
+      for (i=cpos+1; i<n; i++) {
+        if (Sjis && (((unsigned char)s[i]) >= 0x80)) {
+          i++;
+          continue; /* skip the next byte. */
         }
-		if (s[i] == '<') { start=1; }
-		if ((s[i] == '>') && start) { start=0; s[i] = ' '; }
-		if (start) s[i] = 0;
-		if (s[i] == '\n') s[i] = ' ';
-	  }
-	  printf("<li><a href=\"%s\">%s</a> <pre>%s </pre>\n",s,s,&(s[cpos+1]));
-	}
+        if (s[i] == 0x1b) { /* skip escape sequence */ 
+          i++;
+          while (s[i] != 0x1b) i++;
+        }
+        if (LeaveTag) ;
+        else {
+          if (s[i] == '<') { s[i]='{'; }
+          else if (s[i] == '>') { s[i]='}'; }
+        }
+        if (s[i] == '\n') s[i] = ' ';
+      }
+      printf("<li><a href=\"%s\">%s</a> <pre>%s </pre>\n",s,s,&(s[cpos+1]));
+    }
   }
   printf("</ol>\n");
   sprintf(com,"rm -f %s",workf);