A naive approach to supporting U+00A0 nonbreakable space in parseurl

Author: jkbzh

Changelog

@@ -5,8 +5,14 @@ Version Changes for Hypermail
 HYPERMAIL VERSION 2.4.1:
 ============================
 
+2020-08-27 Jose Kahan
+	* src/string.c
+	extended parseurl() in a naive approach for supporting U+00A0
+	nonbreakable space inside sscanf. This function should eventually be
+	rewritten to use libpcre
+
 2020-06-19 Jose Kahan
-	* domains.c
+	* src/domains.c
 	valid_root_domains() was validating a domain name against the
 	historical DNS domains such as .org, .com. DNS has evolved and allows a
 	bigger variety of domains than those this function verified. That part

src/string.c

@@ -1941,10 +1941,19 @@ char *parseurl(char *input, char *charset)
 		&& ((istelprotocol && (*inputp == '+' || isdigit(*inputp)))
 		    || (!istelprotocol && !ispunct(*inputp)))) {
 
-                if (set_iso2022jp)
-		    urlscan = sscanf(inputp, "%255[^] \033)<>\"\'\n[\t\\]", urlbuff);
-                else
-		    urlscan = sscanf(inputp, "%255[^] )<>\"\'\n[\t\\]", urlbuff);
+                if (set_iso2022jp) {
+                    if (charset && !strncasecmp(charset, "UTF-8", 5)) {
+                        urlscan = sscanf(inputp, "%255[^] \u00a0\033)<>\"\'\n[\t\\]", urlbuff);
+                    } else {
+                        urlscan = sscanf(inputp, "%255[^] \033)<>\"\'\n[\t\\]", urlbuff);
+                    }
+                } else {
+                    if (charset && !strncasecmp(charset, "UTF-8", 5)) {
+                        urlscan = sscanf(inputp, "%255[^] \u00a0)<>\"\'\n[\t\\]", urlbuff);
+                    } else {
+                        urlscan = sscanf(inputp, "%255[^] )<>\"\'\n[\t\\]", urlbuff);
+                    }
+                }
             }
 
 	    if (urlscan == 1) {