(cherry picked from commit 940c2a0c)

Detect location of boost headers, since they moved in some recent
version.
(cherry picked from commit b7f6ae13)

Have I mentioned that I hate it when library developers move their
headers around? Change the location of the property_list.hpp header
to a macro and add a new -DNEWER_BOOT define that swtiches this to
the newer location.

Not automatically detected yet; should be checked in configure.
(cherry picked from commit 2403b264)

Minor units conversion problem here. IO::Poll() takes seconds as it argument,
not milliseconds as we were doing (by multiplying the arg by 1000 before
calling).

Unfortunately, this is not the Big One (memory corruption) that we have been
chasing for so long. Sigh...
(cherry picked from commit 1ee85494)

Introduced when I 'fixed' warnings earlier. Found by Mike

can create slices, and be notified of expiring slices and slivers. The
problem is the PL credentials do not inclide an email address, so we
have to get it from the user's registry. This required a bunch of goo
top find the proper place to do a resolve, since PL credentials do not
include a URL either.

push to stable.

null, so need to check that before using it.

interface for the underlying physical node.

This header contains the reference name (eg. branch name or tag name)
that was modified by the commits in the mail.

... so that we can specify a different path to git if necessary

'archivemail' works exactly like 'alwaysmail', except that any
addresses it contains get separate copies of the mail. This is
intended to keep cleaner 'To' lines when sending mail to
archive-type lists.

This script now generates its own Message-Id header, so that
if more than one piece of mail is generated, they can easily
be detected by duplicate-suppression features of mailers.

Make sure "tmcd imagekey" returns info for all images.
Make Linux rc.frisbee handle it.

Also: change to elabinelab setup to not load secure image info
into the elabinelab DB (i.e., so we don't propogate the secret keys).

'git config' lets you add mutliple values for the same key - support
this for the alwaysmail config variable.

Also add support for adding in a Reply-To header, so that discussions
can be directed to mailing lists, etc.

we still look at the local policy since we use the slice credential
expiration.

even built.

Did this a long time ago, but was afraid to check it in since Leigh said
"do NOT mess with tmcd until further notice!" I'm hoping he forgot about
that...

nodes.

equivalence in the presence of bridge nodes.

Use that to spit out proper topomap so the clients can do distributed
route calculation.

Add some code that is not enabled (and not working) for linktest in
the presence of bridge nodes. This is is really a pain and I have not
figured it out yet.

bunch of work for bridge nodes.

CM) and max_slice_lifetime (for the SA) so that slice credentials and
slivers can have max different expirations. This is so a CM can can
limit easily limit slivers so they do not hang out for months doing
nothing.

Add call to ipassign_wrapper to fill in IPs since we are not running
the parser, and the stuff that is in the CM is bogus. Actually need to
remove that. Also need to do this for new bridge support.

Check new users for a valid email, so that we can send warnings from
the expire daemon when slivers are going to expires (see above).
Anyone from an Emulab will have an email address in their certificate,
but PLC users do not. Try to contact the PLC registry to resolve the
user, and fill in the email. If we cannot get the email, we fail.
This is not tested yet, cause PLC does not trust the CM certificate,
and this is probably the case for PLE and MAX, but this will force
things to get fixed.

CM) and max_slice_lifetime (for the SA) so that slice credentials and
slivers can have max different expirations. This is so a CM can can
limit easily limit slivers so they do not hang out for months doing
nothing.

of time that slivers are allowed to hang out without being renewed.

breaks and since we always run the distributed route calculator on the
nodes, just let it do it (module fixes to gentopofile).

check at BBN's Emulab. No clues yet.