lynx-dev Patch for termination of named entity in CGI query string

[Steve Harvey]

Hello,

  A major job search web site doesn't work with lynx because of
the following problem.

  An anchor tag contains a CGI query string with multiple parameters
with names beginning with pi_ .  Lynx sees the preceeding ampersand
and interprets the '&pi' as though it were an entity reference,
with the result that the string '&pi' is replaced by the UTF-8
encoding for π .

I.e.  <a href="http://www.domain.com/some.cgi?pi_x=1&pi_y=0";>link</a>
turns into a request for
http://www.domain.com/some.cgi?pi_x=1%CF%80_y=0

  I don't know why the parser does not insist upon the terminating
semicolon being present, presumably for compatibility with non-w3c
compliant web pages.

  The following corrects this behaviour for this special case.

--- src/LYCharUtils.c.orig      Tue May 28 15:41:42 2002
+++ src/LYCharUtils.c   Mon Aug 12 15:41:19 2002
@@ -1885,7 +1885,7 @@
            **  if &yz string happened to be a recognized entity name. - LP
            */
           if ( ((code = HTMLGetEntityUCValue(name)) > 0) &&
-               !((cpe == '=') && (stype == st_URL)) ) {
+               !(((cpe == '=') || (cpe == '_')) && (stype == st_URL)) ) {
                state = S_check_uni;
                break;
            }
  
-- 
Steve Harvey

; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to address@hidden