From: hniksic <devnull@localhost>
Date: Sat, 9 Apr 2005 11:48:31 +0000 (-0700)
Subject: [svn] Don't mark ~ as unsafe, it confuses too many sites.
X-Git-Tag: v1.13~1215
X-Git-Url: http://sjero.net/git/?p=wget;a=commitdiff_plain;h=8b9cabe00460434123bd265a2ab60b0a4f9e048c

[svn] Don't mark ~ as unsafe, it confuses too many sites.
---

diff --git a/src/ChangeLog b/src/ChangeLog
index 43d7494f..9be27d80 100644
--- a/src/ChangeLog
+++ b/src/ChangeLog
@@ -1,3 +1,14 @@
+2005-04-09  Hrvoje Niksic  <hniksic@xemacs.org>
+
+	* url.c: Use "static const" in preference to "const static".
+	Sun's cc warns that "storage class after type is obsolescent".
+
+	* url.c (urlchr_table): Don't mark ~ as unsafe, too many broken
+	web sites are confused when ~ is changed to %7E.  Their servers
+	redirect /%7Efoo/ to /~foo/, which Wget again accesses using %7E,
+	causing further redirections, therefore looping infinitely.  See
+	Debian bug #301624 for an example.
+
 2005-04-09  Hrvoje Niksic  <hniksic@xemacs.org>
 
 	* alloca.c: Include wget.h to be able to use xmalloc.  In addition
diff --git a/src/url.c b/src/url.c
index 82b4bcea..2606c4de 100644
--- a/src/url.c
+++ b/src/url.c
@@ -87,13 +87,14 @@ static int path_simplify PARAMS ((char *));
    changing the meaning of the URL.  For example, you can't decode
    "/foo/%2f/bar" into "/foo///bar" because the number and contents of
    path components is different.  Non-reserved characters can be
-   changed, so "/foo/%78/bar" is safe to change to "/foo/x/bar".  Wget
-   uses the rfc1738 set of reserved characters, plus "$" and ",", as
-   recommended by rfc2396.
-
-   An unsafe characters is the one that should be encoded when URLs
-   are placed in foreign environments.  E.g. space and newline are
-   unsafe in HTTP contexts because HTTP uses them as separator and
+   changed, so "/foo/%78/bar" is safe to change to "/foo/x/bar".  The
+   unsafe characters are loosely based on rfc1738, plus "$" and ",",
+   as recommended by rfc2396, and minus "~", which is very frequently
+   used (and sometimes unrecognized as %7E by broken servers).
+
+   An unsafe character is the one that should be encoded when URLs are
+   placed in foreign environments.  E.g. space and newline are unsafe
+   in HTTP contexts because HTTP uses them as separator and line
    terminator, so they must be encoded to %20 and %0A respectively.
    "*" is unsafe in shell context, etc.
 
@@ -117,7 +118,7 @@ enum {
 #define U  urlchr_unsafe
 #define RU R|U
 
-const static unsigned char urlchr_table[256] =
+static const unsigned char urlchr_table[256] =
 {
   U,  U,  U,  U,   U,  U,  U,  U,   /* NUL SOH STX ETX  EOT ENQ ACK BEL */
   U,  U,  U,  U,   U,  U,  U,  U,   /* BS  HT  LF  VT   FF  CR  SO  SI  */
@@ -134,7 +135,7 @@ const static unsigned char urlchr_table[256] =
   U,  0,  0,  0,   0,  0,  0,  0,   /* `   a   b   c    d   e   f   g   */
   0,  0,  0,  0,   0,  0,  0,  0,   /* h   i   j   k    l   m   n   o   */
   0,  0,  0,  0,   0,  0,  0,  0,   /* p   q   r   s    t   u   v   w   */
-  0,  0,  0,  U,   U,  U,  U,  U,   /* x   y   z   {    |   }   ~   DEL */
+  0,  0,  0,  U,   U,  U,  0,  U,   /* x   y   z   {    |   }   ~   DEL */
 
   U, U, U, U,  U, U, U, U,  U, U, U, U,  U, U, U, U,
   U, U, U, U,  U, U, U, U,  U, U, U, U,  U, U, U, U,
@@ -1269,7 +1270,7 @@ enum {
    translate file name back to URL, this would become important
    crucial.  Right now, it's better to be minimal in escaping.  */
 
-const static unsigned char filechr_table[256] =
+static const unsigned char filechr_table[256] =
 {
 UWC,  C,  C,  C,   C,  C,  C,  C,   /* NUL SOH STX ETX  EOT ENQ ACK BEL */
   C,  C,  C,  C,   C,  C,  C,  C,   /* BS  HT  LF  VT   FF  CR  SO  SI  */