[Bug 2267] Host matching uses modified hostname as well as original

[bugzilla-daemon at mindrot.org]

https://bugzilla.mindrot.org/show_bug.cgi?id=2267

--- Comment #5 from Richard Birkett <openssh at richard.birkett.com> ---
Hi Damien,

The diff doesn't include the changes to readconf.h, so all of the
defines are missing... can you upload that diff, please?  (Don't worry,
I know all about working-while-asleep!)

I've double-checked the previous patch, though, and Host is definitely
behaving like "Match host" (ie. taking HostName into account), even
with canonicalisation off.  I'll check the new patch as soon as
readconf.h is there.

I'd like to have a go at a proof of concept of my idea, since I still
have concerns about the double-pass parsing and its scope for
surprises.  Obviously it needs to be (a) easy to understand, and
unfortunately (b) backward compatible with the behaviour documented in
6.6 (though luckily the 6.6 docs are a little ambiguous!).

I know 6.7 is already in its final testing phase, but is there any
chance either of our solutions might still get into that?  Or are we
now targetting 6.8 (or 6.7.1)?

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.