- Set the url resolver to escape illegal chars instead of stripping.

4 files changed, 21 insertions, 12 deletions

diff --git a/ChangeLog b/ChangeLog
index b5d5d935..886c20aa 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -89,6 +89,8 @@ dillo-fltk2
    Patch: Vincent Thomasset
 +- Fixed void to int conversions for 64bit-arch.
    Patch: Jorge Arellano Cid, higuita
++- Set the url resolver to escape illegal chars instead of stripping.
+   Patch: Jorge Arellano Cid, Jeremy Henty
 +- Added a strndup() replacement in dw2
    Patch: Alexander Becher, Johannes Hofmann, Jorge Arellano Cid
 +- Fixed calcHashValue() to only return non-negative numbers (was SEGFAULT).
diff --git a/src/dialog.cc b/src/dialog.cc
index 08316d7d..9b53d5c6 100644
--- a/src/dialog.cc
+++ b/src/dialog.cc
@@ -103,7 +103,7 @@ char *a_Dialog_open_file(const char *msg,
  */
 void a_Dialog_text_window(const char *txt, const char *title)
 {
-   int wh = 500, ww = 480, bh = 30;
+   int wh = 600, ww = 650, bh = 30;
    int lines, line_num_width;
    TextBuffer *text_buf = new TextBuffer();
    text_buf->text(txt);
diff --git a/src/html.cc b/src/html.cc
index 64e72433..ca9b9ef4 100644
--- a/src/html.cc
+++ b/src/html.cc
@@ -468,15 +468,16 @@ static DilloUrl *Html_url_new(DilloHtml *html,
       const char *suffix = (n_ic) > 1 ? "s" : "";
       n_ic_spc = URL_ILLEGAL_CHARS_SPC(url);
       if (n_ic == n_ic_spc) {
-         MSG_HTML("URL has %d illegal character%s [%d space%s]\n",
+         MSG_HTML("URL has %d illegal character%s (%d space%s)\n",
                    n_ic, suffix, n_ic_spc, suffix);
       } else if (n_ic_spc == 0) {
-         MSG_HTML("URL has %d illegal character%s [%d in (00-1F or 7F)]\n",
+         MSG_HTML("URL has %d illegal character%s (%d in {00-1F, 7F} range)\n",
                    n_ic, suffix, n_ic);
       } else {
-         MSG_HTML("URL has %d illegal character%s "
-                  "[%d space%s and %d in (00-1F or 7F)]\n",
-                  n_ic, suffix, n_ic_spc, n_ic_spc ? "s" : "", n_ic-n_ic_spc);
+         MSG_HTML("URL has %d illegal character%s: "
+                  "%d space%s, and %d in {00-1F, 7F} range\n",
+                  n_ic, suffix,
+                  n_ic_spc, n_ic_spc > 1 ? "s" : "", n_ic-n_ic_spc);
       }
    }
    return url;
diff --git a/src/url.c b/src/url.c
index c21a458a..2a81e74b 100644
--- a/src/url.c
+++ b/src/url.c
@@ -53,6 +53,8 @@
 //#define DEBUG_LEVEL 2
 #include "debug.h"
 
+static const char *HEX = "0123456789ABCDEF";
+
 
 /*
  * Return the url as a string.
@@ -358,12 +360,17 @@ DilloUrl* a_Url_new(const char *url_str, const char *base_url,
       n_ic += (*p != ' ' && *p > 0x1F && *p != 0x7F) ? 0 : 1;
    }
    if (n_ic) {
-      /* Strip illegal characters (they could also be encoded).
-       * There's no standard for illegal chars; we chose to strip. */
-      p = str1 = dNew(char, strlen(url_str));  /* Yes, enough memory! */
+      /* Encode illegal characters (they could also be stripped).
+       * There's no standard for illegal chars; we chose to encode. */
+      p = str1 = dNew(char, strlen(url_str) + 2*n_ic + 1);
       for (i = 0; url_str[i]; ++i)
          if (url_str[i] > 0x1F && url_str[i] != 0x7F && url_str[i] != ' ')
             *p++ = url_str[i];
+         else  {
+           *p++ = '%';
+           *p++ = HEX[(url_str[i] >> 4) & 15];
+           *p++ = HEX[url_str[i] & 15];
+         }
       *p = 0;
       urlstr = str1;
    }
@@ -572,7 +579,6 @@ char *a_Url_decode_hex_str(const char *str)
 char *a_Url_encode_hex_str(const char *str)
 {
    static const char *verbatim = "-_.*";
-   static const char *hex = "0123456789ABCDEF";
    char *newstr, *c;
 
    if (!str)
@@ -595,8 +601,8 @@ char *a_Url_encode_hex_str(const char *str)
          *c++ = 'A';
       } else {
          *c++ = '%';
-         *c++ = hex[(*str >> 4) & 15];
-         *c++ = hex[*str & 15];
+         *c++ = HEX[(*str >> 4) & 15];
+         *c++ = HEX[*str & 15];
       }
    *c = 0;