summary |
shortlog | log |
commit |
commitdiff |
tree
first ⋅ prev ⋅ next
Mark Wooding [Wed, 17 Oct 2012 23:11:56 +0000 (00:11 +0100)]
Dispatch to methods handling address-family specifics.
This is a fairly invasive change to the program. Rather than carry
around an address family type and switch on it throughout, carry around
a pointer to a vtable containing methods. Methods needed by the system-
specific machinery are properly integrated.
Mark Wooding [Wed, 17 Oct 2012 23:09:00 +0000 (00:09 +0100)]
Makefile.am, configure.ac, ident.c -> linux.c: System specifics.
Rename `ident.c' to indicate that this is where the system-specific
machinery lives, and provide build system support for other platforms
should they be contributed.
Mark Wooding [Wed, 17 Oct 2012 04:40:57 +0000 (05:40 +0100)]
ident.c, yaid.c, yaid.h: Make `get_default_gw' be internal to `ident.c'.
Don't really need the diagnostics any more.
Mark Wooding [Wed, 17 Oct 2012 04:39:41 +0000 (05:39 +0100)]
yaid.c: Don't fail if either IPv4 or IPv6 is unavailable.
Do continue to fail if the ports can't be bound, or if neither protocol
is available.
Mark Wooding [Wed, 17 Oct 2012 04:42:09 +0000 (05:42 +0100)]
policy.c: Implement IPv6 matching in match_addrpat.
Missing piece. Whoops.
Mark Wooding [Sun, 14 Oct 2012 23:21:21 +0000 (00:21 +0100)]
Apparently working version, but still ugly.
It needs a lot of commentary, a proper command-line interface, logging,
pidfiles, and so on.
Also, there's a bunch of inconsistency surrounding IPv4/IPv6 handling:
the core uses a vtable, while most of the rest of the code uses
switches. This ought to be sorted out before a proper release.
Mark Wooding [Sat, 29 Sep 2012 14:37:27 +0000 (15:37 +0100)]
Initial ugly non-portable core of an ident daemon.