Thomas Reinke <[EMAIL PROTECTED]> writes:

> Does raise an interesting question re
> a bug introduced by the patch - is https://HOST the the same host
> when it is referred to as http://HOST?

Good question.  After first, second, and third thought, I'd say yes.
But then I remembered that Wget, rightfully, treats different ports as
different sites.  So what we have is:

URL1                  URL2                   SAME-HOST?
http://server.com     http://server.com:80   yes
http://server.com:81  http://server.com:82   no
http://server.com     http://server.com:81   no
http://server.com:81  http://server.com:81   yes

I.e. both host and (possibly implied) port need be equal to produce
equality.

Since https uses different implied ports, the equality table might be
more convoluted.

URL1                     URL2                    SAME-HOST?
http://server.com        https://server.com      yes
http://server.com:100    https://server.com:100  yes
http://server.com:100    https://server.com:101  no (of course)
http://server.com:80     https://server.com:443  ??? (yes if consistent with
                                                     1st entry)
                                                     (no if consistent
                                                     with non-https table)
http://server.com:80     https://server.com:80   ??? (the latter is a
                                                     non-standard
                                                     port, so perhaps
                                                     it shouldn't be
                                                     "equal" to the
                                                     standard one.)
http://server.com:100    https://server.com:100  yes (probably allow this)

The whole thing is probably quite unimportant in practice, but I have
a feeling that if it's possible to get it wrong, that I'll get it
wrong.  :-)

Reply via email to