Fix valid_link() to avoid scheme repetition

valid_link("mailto:x@y") returns "mailto:mailto:x@y". This patch fixes it,
and simplifies the code path.

Signed-off-by: Alberto Bertogli <albertito@blitiri.com.ar>

diff --git a/blitiri.cgi b/blitiri.cgi
index bac048b9b237294313aea2ca279baebce0978d47..0c4847583a96ab5a9180655c631357a9bc8b86f0 100755 (executable)
--- a/blitiri.cgi
+++ b/blitiri.cgi
@@ -524,20 +524,28 @@ def validate_rst(rst, secure = True):
 
 def valid_link(link):
        import re
 
 def valid_link(link):
        import re

-       mail_re = r"^[^ \t\n\r@<>()]+@[a-z0-9][a-z0-9\.\-_]*\.[a-z]+$"

        scheme_re = r'^[a-zA-Z]+:'
        scheme_re = r'^[a-zA-Z]+:'

+       mail_re = r"^[^ \t\n\r@<>()]+@[a-z0-9][a-z0-9\.\-_]*\.[a-z]+$"

        url_re = r'^(?:[a-z0-9\-]+|[a-z0-9][a-z0-9\-\.\_]*\.[a-z]+)' \
                        r'(?::[0-9]+)?(?:/.*)?$'
        url_re = r'^(?:[a-z0-9\-]+|[a-z0-9][a-z0-9\-\.\_]*\.[a-z]+)' \
                        r'(?::[0-9]+)?(?:/.*)?$'

-       scheme = ''
-       rest = link
+

        if re.match(scheme_re, link, re.I):
                scheme, rest = link.split(':', 1)
        if re.match(scheme_re, link, re.I):
                scheme, rest = link.split(':', 1)

-       if (not scheme or scheme == 'mailto') and re.match(mail_re, rest, re.I):
+               # if we have an scheme and a rest, assume the link is valid
+               # and return it as-is; otherwise (having just the scheme) is
+               # invalid
+               if rest:
+                       return link
+               return None
+
+       # at this point, we don't have a scheme; we will try to recognize some
+       # common addresses (mail and http at the moment) and complete them to
+       # form a valid link, if we fail we will just claim it's invalid
+       if re.match(mail_re, link, re.I):

                return 'mailto:' + link
                return 'mailto:' + link

-       if not scheme and re.match(url_re, rest, re.I):
-               return 'http://' + rest
-       if scheme:
-               return link
+       elif re.match(url_re, link, re.I):
+               return 'http://' + link
+

        return None
 
 def sanitize(obj):
        return None
 
 def sanitize(obj):