[libsoup] soup-uri: %-encode non-ASCII characters when parsing URIs

[Dan Winship <danw src gnome org>]

commit 226f6382496941a4baea6c1c0d22c8d3aed634a1
Author: Dan Winship <danw gnome org>
Date:   Sat Nov 12 09:27:33 2011 -0500

    soup-uri: %-encode non-ASCII characters when parsing URIs
    
    https://bugzilla.gnome.org/show_bug.cgi?id=662806

 libsoup/soup-uri.c  |   16 ++++++++--------
 tests/uri-parsing.c |    5 ++++-
 2 files changed, 12 insertions(+), 9 deletions(-)
---
diff --git a/libsoup/soup-uri.c b/libsoup/soup-uri.c
index 1aa801b..9170feb 100644
--- a/libsoup/soup-uri.c
+++ b/libsoup/soup-uri.c
@@ -673,7 +673,7 @@ uri_normalized_copy (const char *part, int length,
 				*d++ = *s;
 			}
 		} else {
-			if (*s == ' ')
+			if (!g_ascii_isgraph (*s))
 				need_fixup = TRUE;
 			*d++ = *s;
 		}
@@ -681,16 +681,16 @@ uri_normalized_copy (const char *part, int length,
 
 	if (need_fixup) {
 		GString *fixed;
-		char *sp, *p;
 
 		fixed = g_string_new (NULL);
-		p = normalized;
-		while ((sp = strchr (p, ' '))) {
-			g_string_append_len (fixed, p, sp - p);
-			g_string_append (fixed, "%20");
-			p = sp + 1;
+		s = (guchar *)normalized;
+		while (*s) {
+			if (g_ascii_isgraph (*s))
+				g_string_append_c (fixed, *s);
+			else
+				g_string_append_printf (fixed, "%%%02X", (int)*s);
+			s++;
 		}
-		g_string_append (fixed, p);
 		g_free (normalized);
 		normalized = g_string_free (fixed, FALSE);
 	}
diff --git a/tests/uri-parsing.c b/tests/uri-parsing.c
index e8568a9..285b41e 100644
--- a/tests/uri-parsing.c
+++ b/tests/uri-parsing.c
@@ -76,7 +76,10 @@ static struct {
 	{ "http://host/path%%%";, "http://host/path%%%"; },
 	{ "http://host/path%/x/";, "http://host/path%/x/"; },
 	{ "http://host/path%0x/";, "http://host/path%0x/"; },
-	{ "http://host/path%ax";, "http://host/path%ax"; }
+	{ "http://host/path%ax";, "http://host/path%ax"; },
+
+	/* Bug 662806; %-encode non-ASCII characters */
+	{ "http://host/p\xc3\xa4th/";, "http://host/p%C3%A4th/"; }
 };
 static int num_abs_tests = G_N_ELEMENTS(abs_tests);