GitLab

More bug fixes for vmcd and friends:

	* GNUmakerules: Set SRCDIR as an environment variable so every
	test doesn't have to be a template.

	* robots/emc/emcd.c: Correctly handle a vision position that is
	dead-on the initial position.  XXX Turn off orientation matching
	for now, it is too strict at the moment so that it favors close
	orientation over close position.

	* robots/mtp/mtp.c: Fix some broken printfs in mtp_print_packet.

	* robots/vmcd/GNUmakefile.in: Add another test.

	* robots/vmcd/test_emcd2.config, robots/vmcd/test_vmcd2.pos,
	robots/vmcd/etst_vmcd2.sh: A test that matches the initial
	positions of two robots.

	* robots/vmcd/test_vmcd.pos: Reduce the distance the robot moved
	between frames.

	* robots/vmcd/vmcd.c: Fix a bugfix and some formatting changes.

	A script for powering on, off, and cycling motes attached to
	stargates

Requires /usr/local/bin/motepower (our program, not yet in CVS) to be
installed on the hosting stargate.

	media "media 100baseTX mediaopt full-duplex";

since they will be in autosense mode by default, which does not play
well with switchports that are fixed to speed/duplex.

meter for the image.

Some bugfixes and an improved test case for vmcd:

	* robots/emc/emcd.c: Add a REQUEST_POSITION handler for the
	emulab_callback.  Fix some simple bugs when processing REQUEST_ID
	packets.

	* robots/vmcd/test_emcd.config: Use just one robot for simple
	test.

	* robots/vmcd/test_vmc-client.pos: A position file for the
	vmc-client test.

	* robots/vmcd/test_vmc-client.sh.in: Test with a position file
	instead of hardcoded numbers.

	* robots/vmcd/test_vmcd.pos: Position file for the vmcd test.

	* robots/vmcd/test_vmcd.sh.in: Make the test a little simpler, but
	add a check to see if vmcd finds the robot initially.

	* robots/vmcd/vmc-client.c: Add a "-f" option that lets you use a
	"position" file that contains the object positions for a series of
	frames.

	* robots/vmcd/vmcd.c: Some bug fixes.

Removed it.

   - oops, now its fixed!
* vmcd/GNUmakefile.in:
   - linked to math lib as well...

   - compilation fixes

   - added the final code necessary to fully handle a vmc client.

Cleanup rmcd a bit:

	* GNUmakerules: Quick fix to the "check" target.

	* robots/rmcd/rmcd.c, robots/rmcd/test_rmcd.sh.in: Cleanup the
	code a bit, add some comments, etc...

Rmcd and garcia stuff:

	* configure, configure.in: Add robot related template files.

	* robots/GNUmakefile.in: Add primotion directory.

	* robots/emc/emcd.c: Debugging printfs, check the status for
	update-position messages from rmc, and add a basic handler for
	emulab clients.

	* robots/emc/test_emcd.sh.in: Update for changes in mtp.

	* robots/mtp/mtp.h, robots/mtp/mtp.c: Changes for the garcia.

	* robots/mtp/mtp_send.c: Fixes so that it will compile under
	linux.

	* robots/primotion/GNUmakefile.in: Makefile for building a fake
	gorobot in the testbed tree.

	* robots/primotion/Makefile: tweaks

	* robots/primotion/gorobot.cc: First draft with sort-of working
	networking code.

	* robots/primotion/test_gorobot.sh.in: Test case for the fake
	gorobot.

	* robots/primotion/dgrobot/GNUmakefile.in: Makefile for building a
	fake grobot class in the testbed tree.

	* robots/primotion/dgrobot/grobot.h: Add #if !defined(GROBOT_SIM)
	conditionals.

	* robots/primotion/dgrobot/grobot_sim.cc: Empty impl of grobot
	class used for testing.

	* robots/rmcd/GNUmakefile.in: Targets for building rmcd and
	running its test case.

	* robots/rmcd/rmcd.c: First draft with sort-of working networking
	code.

	* robots/rmcd/test_emcd.config: emcd configuration for the rmcd
	test case.

	* robots/rmcd/test_rmcd.sh.in: Test case for rmcd.

pid/eid argument.

Minor cleanup:

	* robots/emc/test_emcd.sh.in, robots/mtp/mtp_send.c,
	robots/mtp/mtp_test.c: Bring up-to-date with the latest version of
	mtp stuff.

	* robots/vmcd/vmc-client.c: Add "valid" field to the mezz stub
	struct.

	* robots/vmcd/vmcd.c: Some compile fixes...

* emcd.c:
   - minor changes, mostly comments. this is next to fix up.
* vmc-client.c:
   - minor changes to tell vmcd about the last packet in a frame
processing cycle (via MTP_POSITION_STATUS_CYCLE_COMPLETE being set in the
last update_position packet of a frame processing cycle).
   - maybe other stuff...
* vmcd.c:
   - massive changes to handle tracking of robots from data passed from
vmc-client, and to handle ID association.
* mtp.h:
   - addded MTP_POSITION_STATUS_CYCLE_COMPLETE #define
   - updated structs request_id and update_id to include a `request_id'
field so that we can uniquely track requests (since they're asynchronous).
* mtp.c:
   - revamped code for above mtp.h changes.

Oops, here's the changes to mtp.h to add support for 64bit timestamps.

info from other mtp structs, added support to encode 64bit floats.

Frame out vmcd:

	* GNUmakerules: Fix "check" target so it runs more than one test.

	* robots/emc/emcd.c: Send the vmc config packet out when vmcd
	connects.

	* robots/vmcd/GNUmakefile.in: Add vmcd-related stuff.

	* robots/vmcd/test_emcd.config: Emc config for the test case.

	* robots/vmcd/test_vmcd.sh.in: Simple test case for vmcd.

	* robots/vmcd/vmcd.c: Initial framing of vmcd, just receives and
	dumps packets, actual logic still needs to be put in.

least get the rest of the users.

have to set $UID=$EUID for sshtb as root to work right.

Temporarily treat garcia 'hosts' like stargates.

Add support for mica1 motes.

real ones (doh!).

assume it always has its default OSID (from node_types) loaded and
ready to go (add this as the OS- feature). This so assign will agree
to the allocation (assign_wrapper adds a desire that says it has to be
running the OSID the user has selected, or the default OSID from the
node_types table).

Watch out for problems ...

stargate and running uisp there.

Untested so far.

Unfortunately, tbuisp now needs to be setuid root because it has to
ssh into the stargates as root. So, take some precautions, like taint
checking and dropping root privs before running uisp locally, to
mitigate the risks.

Start on vmc:

	* configure, configure.in: Add vmcd related template files.

	* robots/GNUmakefile.in: Switch order of vmcd/rmcd.

	* robots/emc/GNUmakefile.in: cleanup

	* robots/mtp/GNUmakefile.in: Add mtp_dump tool.

	* robots/mtp/mtp.c: Change mtp_encode_packet to use a passed in
	buffer pointer or allocate a buffer if its NULL, probably gonna be
	a big source of errors...

	* robots/mtp/mtp_dump.c: Another command-line tool that connects
	to a server and dumps mtp packets that are received.  Useful for
	seeing output from the vmc-client.

	* robots/vmcd/GNUmakefile.in: Add vmc-client and test case.

	* robots/vmcd/test_vmc-client.sh.in: Test case for the vmc-client.

	* robots/vmcd/vmc-client.c: First cut of the vmc-client, it reads
	mezzanine output and sends it to any connected clients.

desktops, nodes, etc.

Thanks to Tim for actually writing most of this code.

subnodes - this was only a good idea in the IXP case. Instead, skip
reloading if they do not have their imageable bit set.

'rebootable' flag set to 0.