Junio C Hamano [off-list ref] writes:
Johannes Schindelin [off-list ref] writes:
quoted
On Tue, 25 Sep 2007, Glenn Rempe wrote:
quoted
+if (($smtp_server =~ /:\d+/) && (defined $smtp_server_port)) {
Not that I want to be a PITA, but this breaks down with IPv6, right?
Right. Do we care about symbolic "server.addre.ss:smtp"
notation as well, I wonder?
Well, does it break?
BTW, I do not think we care about ":smtp"; it was a
tongue-in-cheek comment.
On Sep 25, 2007, at 4:25 PM, Junio C Hamano wrote:
Junio C Hamano [off-list ref] writes:
quoted
Johannes Schindelin [off-list ref] writes:
quoted
On Tue, 25 Sep 2007, Glenn Rempe wrote:
quoted
+if (($smtp_server =~ /:\d+/) && (defined $smtp_server_port)) {
Not that I want to be a PITA, but this breaks down with IPv6, right?
Right. Do we care about symbolic "server.addre.ss:smtp"
notation as well, I wonder?
Well, does it break?
BTW, I do not think we care about ":smtp"; it was a
tongue-in-cheek comment.
Unfortunately, I know little about IPv6 and whether this breaks IPv6
addressing or not. So I'll leave that question for others. Does the
unpatched code work with IPv6? Does anyone currently use it that way?
Junio, are you suggesting that I should remove the host:port form
support entirely and leave only the --smtp-server port option as
valid? I kind of liked that this new method allows both forms of
specifying port as equal citizens. :-) If you are suggesting that
it be removed, I think we would have to reject the host:port form
smtp-server addresses so we don't break when both --smtp-
server=host:port and -smtp-ssl are provided. (which brings us back to
the ipv6 question). No?