Even though it was well known before that ssmtp has working but just
not enabled support for IPv6, it appears that somehow no one bothered
noting this fact in the bug log; Magnus Holmgren did it just a few
days ago.
I'm guilty myself, too: I have ssmtp with --enable-inet6 on most
boxes for several months, yet I didn't notice that no one mentioned
this in the BTS.
Anyway, on #debian-devel today:
<vorlon> KiloByte: is there a bug report filed for that? Has someone
verified that it doesn't go goofy in IPv4 environments?
<KiloByte> #349457
<KiloByte> Let me do a basic sanity test on the second question.
<vorlon> if it's actually non-broken, then it falls under the NMU
policy for release team-sanctioned release goals
Here are the results of my tests:
* IPv4/6 kernel, no working v6 route
* by name: ok
* by numeric v4 IP: ok
* IPv4-only kernel
* by name: ok
* by numeric v4 IP: ok
* IPv4/6 kernel, working v6 route
* by name: ok
* by numeric v4 IP: ok
* by numeric v6 IP: fails: send-mail: Cannot open 2002:5033
So, IPv4 environments work as well as they did before, IPv6 ones work
perfectly unless you specify literal numeric addresses. Neither
2002:5033:a761:3::7 nor [2002:5033:a761:3::7] works -- however, this
is an upstream bug.
Anyway, the set of working cases with --enable-inet6 is a strict
superset of cases working without that switch; thus I see no reason
not to enable it.
Cheers and schtuff,
--
1KB // Microsoft corollary to Hanlon's razor:
// Never attribute to stupidity what can be
// adequately explained by malice.
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
