try to better handle non-ascii file names

src/utils/pathut.cpp

@@ -1,5 +1,5 @@
 #ifndef lint
-static char rcsid[] = "@(#$Id: pathut.cpp,v 1.9 2006-02-02 08:58:11 dockes Exp $ (C) 2004 J.F.Dockes";
+static char rcsid[] = "@(#$Id: pathut.cpp,v 1.10 2006-03-29 11:18:15 dockes Exp $ (C) 2004 J.F.Dockes";
 #endif
 /*
  *   This program is free software; you can redistribute it and/or modify
@@ -189,6 +189,41 @@ list<std::string> path_dirglob(const std::string &dir,
     return res;
 }
 
+std::string url_encode(const std::string url, string::size_type offs)
+{
+    string out = url.substr(0, offs);
+    const char *cp = url.c_str();
+    for (string::size_type i = offs; i < url.size(); i++) {
+	int c;
+	char *h = "0123456789ABCDEF";
+	c = cp[i];
+	if(c <= 0x1f || 
+	   c >= 0x7f || 
+	   c == '<' ||
+	   c == '>' ||
+	   c == ' ' ||
+	   c == '\t'||
+	   c == '"' ||
+	   c == '#' ||
+	   c == '%' ||
+	   c == '{' ||
+	   c == '}' ||
+	   c == '|' ||
+	   c == '\\' ||
+	   c == '^' ||
+	   c == '~'||
+	   c == '[' ||
+	   c == ']' ||
+	   c == '`') {
+	    out += '%';
+	    out += h[(c >> 4) & 0xf];
+	    out += h[c & 0xf];
+	} else {
+	    out += char(c);
+	}
+    }
+    return out;
+}
 
 #else // TEST_PATHUT
 

src/utils/pathut.h

@@ -16,7 +16,7 @@
  */
 #ifndef _PATHUT_H_INCLUDED_
 #define _PATHUT_H_INCLUDED_
-/* @(#$Id: pathut.h,v 1.7 2006-03-20 09:54:22 dockes Exp $  (C) 2004 J.F.Dockes */
+/* @(#$Id: pathut.h,v 1.8 2006-03-29 11:18:15 dockes Exp $  (C) 2004 J.F.Dockes */
 
 #include <string>
 #include <list>
@@ -40,4 +40,8 @@ extern std::string path_canon(const std::string &s);
 /// Use glob(3) to return a list of file names matching pattern inside dir
 extern std::list<std::string> path_dirglob(const std::string &dir, 
 					   const std::string pattern);
+/// Encode according to rfc 1738
+extern std::string url_encode(const std::string url, 
+			      std::string::size_type offs);
+
 #endif /* _PATHUT_H_INCLUDED_ */

src/utils/transcode.cpp

@@ -1,5 +1,5 @@
 #ifndef lint
-static char	rcsid[] = "@(#$Id: transcode.cpp,v 1.6 2006-01-23 13:32:28 dockes Exp $ (C) 2004 J.F.Dockes";
+static char	rcsid[] = "@(#$Id: transcode.cpp,v 1.7 2006-03-29 11:18:15 dockes Exp $ (C) 2004 J.F.Dockes";
 #endif
 /*
  *   This program is free software; you can redistribute it and/or modify
@@ -40,14 +40,15 @@ using std::string;
 #endif
 
 bool transcode(const string &in, string &out, const string &icode,
-	       const string &ocode)
+	       const string &ocode, int *ecnt)
 {
     iconv_t ic;
     bool ret = false;
     const int OBSIZ = 8192;
     char obuf[OBSIZ], *op;
     bool icopen = false;
-
+    if (ecnt)
+	*ecnt = 0;
     out.erase();
     size_t isiz = in.length();
     out.reserve(isiz);
@@ -79,6 +80,8 @@ bool transcode(const string &in, string &out, const string &icode,
 			 ip - in.c_str(), out.length() + OBSIZ - osiz));
 		out.append(obuf, OBSIZ - osiz);
 		out += "?";
+		if (ecnt)
+		    (*ecnt)++;
 		ip++;isiz--;
 		continue;
 	    }

src/utils/transcode.h

@@ -16,7 +16,7 @@
  */
 #ifndef _TRANSCODE_H_INCLUDED_
 #define _TRANSCODE_H_INCLUDED_
-/* @(#$Id: transcode.h,v 1.3 2006-01-30 11:15:28 dockes Exp $  (C) 2004 J.F.Dockes */
+/* @(#$Id: transcode.h,v 1.4 2006-03-29 11:18:15 dockes Exp $  (C) 2004 J.F.Dockes */
 /** 
  * A very minimal c++ized interface to iconv
  */
@@ -24,6 +24,7 @@
 
 extern bool transcode(const std::string &in, std::string &out, 
 		      const std::string &icode,
-		      const std::string &ocode);
+		      const std::string &ocode, 
+		      int *ecnt = 0);
 
 #endif /* _TRANSCODE_H_INCLUDED_ */