GitLab

A link to the solutions is provided at no extra charge.

A node having no TPM info is not an error (right now anyway) so don't
complain about it, since it confuses Mike.

Also, do a little Old Fart Normal Form code formatting enforcement.

Plug one hole in the dike.  Don't blindly atoi() a DB return value.
One down, a zillion to go.  Fixed a couple of nits while in there.

Added TBCOREDIR define for the place where tmcd can go to die.
It is the directory where tmcd should run and is intended to insure that
it can always drop core.

Up past 3:30AM this morning.  Despite a constant stream of caffinated products,
I am incapable of anything more demanding than cut and paste...

These variables were previously constants hardwired into create_image:

images/create/maxwait:
    Max time (minutes) to allow for saving an image
images/create/idlewait:
    Max time (minutes) to allow between periods of progress
    (image file getting larger) when saving an image (should be <= maxwait)
images/create/maxsize:
    Max size (GB) of a created image

Max image size should really be in DB as a per-testbed (per-project? per-user?
per-something!) value, and not hardwired here.  In the meantime, we set this
big and let disk quotas do the dirty work of limiting size.

Making crossdomain more permissive like this allows us to host the flash client on www.emulab.net which has a CA-certified certificate.

the toplevel makefile. Add a quiet mode. Minor cleanup. Also make
it possible to run the from object directory before the initial
install.

fragment scripts, much like we do with the DB updates. The intent
is to simplify testbed update for remote sites.