client software to widearea nodes. Most of these changes were to
reduce the embarrassment factor. At some point we need a proper
autoconf and such, but for now there is a makefile in the src dir for
creating the distribution.

I've tested it on a local linux node and mostly on a freebsd node, but
I've moved things around and so updating the RON nodes will require
some hand intervention by me at some point.

nodes.
Add sanity check in doaccounts to make sure that project is approved
(specific to wide remote nodes).

solution to account management problem on remote nodes.

	num_pcplab int(11) default '0',
	num_ron int(11) default '0',
	pcremote_ok set('pcplab','pcron') default NULL,

driven by new field in the project table called pcremote_ok, which is
the set of remote node types that a project is allowed to have
accounts on. When a node of that type checks in, match its type
against the project table, and return accounts for everyone in all of
the projects with that type listed in pcremote_ok.

Kill the NOSHAREDEXPTS ifdefs in doaccounts since that stuff is dead
for now.

boss setup docs.

Add mode_transitions table, and add data that gets filled into state_transitions and mode_transitions.

Add a .htaccess for /tutorial/ like we have for / . One benifit is now if they request emulab.net/tutorial/ it sends them to tutorial.php3 as would be expected.

Remove sharks from this test so we can keep it around even though we no longer support sharks the old way.

traffic between subnets

Chris and Dave.  Somebody please test this; I don't know how.

(if I am wrong on this, let me know, guys).

Add Austin to current people.

Fix broken flest tests: remove shark tests, since we are giving up support for them until we make proper heirarchical nodes, and change the numbers we compare the tbcmd results against to be correct.

removing all of the ports in it.

top level ones.

Changed behavior:
If "-s <seed>" is not specified, a timeofday-dependent random
seed will be used.

In any case, when in verbose mode, the random seed used will be
printed out at the beginning of the run.

record is killed and it does not matter.

the sleep time a bit, so that we don't end up with multiple processes
continually sending requests at the same times.

condition that was probably caused by unexpected parallelism.

One semantic change is that we used to block SIGTERM (but not SIGINT, which
also would quit) during read/write operations.  I don't do that anymore,
so if you kill capture, the log could lose some console output.