Thanks Chris!

Chris fixes up the handle_virt script.

Not really needed, so no big deal.

by the web server forks a child to do the actual work of calling tbend
and other stuff. The parent returns right away and the script ends.
When the experiment termination (child) ends, an email message is sent
to the user that issued the termination request. To prevent multiple
clicks, I added a DB field called expt_terminating that is a DATETIME
field.  If the field is set, the script fails and the user is told to
be more patient. I used a DATETIME field mostly for debugging purposes
so we can track and future problems.

the web wrapper script.

(or was it months)?

where script email goes. Defaults to testbed-ops@fast.cs.utah.edu. Change
it for your development tree with: --with-TBOPSEMAIL=stoller@fast.cs.utah.edu

blocks in a single lseek, more powerful than gzip on steriods!

handle_ip: Removed virt_names code.
handle_virt: This sets up the virt_names table.  The vast majority of it
is to handle the ugly cases LAN scenerios.

Note: If we ever want to add broadcast loop detection in, handle_virt
would be a very easy place to do it.

this out of development trees.

Previously on the rare occasion that a run scored an integer score
assign_wrapper would break.

shark.tcl - The file to create them.  Good example for creating custom
labels.

pid/eid before releasing from the reloads/reserve table, since an
admin initiated reload might be pending when a user does his own
load. Don't want to clear the reload state, or even worse, the
reserved state. There is a cross dependency created bewteen tmcd.c
and sched_reload on the pid/eid testbed/reloading. No biggie. I put in
a warning in both places.

the network to copy the image off.

to do when you find out

now in the DNS map. Hopefully people will see this and understand
what it means.