1. 26 May, 2005 8 commits
  2. 25 May, 2005 6 commits
  3. 24 May, 2005 14 commits
  4. 23 May, 2005 3 commits
    • Timothy Stack's avatar
    • Leigh B. Stoller's avatar
      Add a set_time_limit(0) call. I'm thinking this is why the webcams stop · ca5b55f1
      Leigh B. Stoller authored
      spewing after a while.
      ca5b55f1
    • Timothy Stack's avatar
      · a1000b60
      Timothy Stack authored
      Some RMCD debugging and improvements.
      
      	* robots/emc/GNUmakefile.in, robots/rmcd/GNUmakefile.in: Fix some
      	of the dependencies.
      
      	* robots/emc/emcd.c: Bump default speed for events to 0.2 m/s.
      	Send an event when a dynamic obstacle is updated.
      
      	* robots/emc/emcd.h: Send position updates three times a second.
      
      	* robots/mtp/listNode.h, robots/mtp/listNode.c: Add a function to
      	prepend a list onto another.
      
      	* robots/mtp/mtp.x, robots/mtp/mtp.h, robots/mtp/mtp.c: Add
      	support for sending updates concerning dynamically detected
      	obstacles.
      
      	* robots/rmcd/masterController.h, robots/rmcd/masterController.c:
      	When a robot is stationary, add it as an obstacle.  Play with the
      	pause time a bit.
      
      	* robots/rmcd/obstacles.h, robots/rmcd/obstacles.c: Add robots as
      	a third type of obstacle.  Do not expand static obstacles when the
      	robot detects something with its sensors.  Add some more checks to
      	make sure the obstacle list is sane.  When
      	creating/growing/removing dynamic obstacles, send a notification
      	to emcd so it can be passed on to the applet.  More comments.
      
      	* robots/rmcd/pathPlanning.h, robots/rmcd/pathPlanning.c: When
      	monkeying with the obstacle list, we need to make sure the
      	obstacles are put back in the right order (dynamic before static).
      	More comments.
      
      	* robots/rmcd/pilotConnection.c: Add some more stuff to the
      	SIGINFO print out.
      
      	* robots/rmcd/rmcd.c: Initialize the emcd mtp handle for sending
      	updates for obstacles.
      
      	* robots/mtp/mtp_dump.c, robots/vmcd/visionTrack.c: Remove macros
      	that are now defined elsewhere.
      a1000b60
  5. 21 May, 2005 1 commit
  6. 20 May, 2005 8 commits
    • Russ Fish's avatar
      Add camera checker daemon. · 392a65b2
      Russ Fish authored
      392a65b2
    • Leigh B. Stoller's avatar
      Oops, flip back to user after the fork. · 57501b7d
      Leigh B. Stoller authored
      In case I didn't mention it; this script is setuid so it can kill
      itself later on the swapout path. It drops privs right away.
      57501b7d
    • David G Andersen's avatar
    • Leigh B. Stoller's avatar
      Checkpoint some robot changes. · 5f67fe09
      Leigh B. Stoller authored
      * New robot event listener:
      
          * It is intended to be started and stopped from the experiment
            swapin path instead of as a global daemon. It takes the pid/eid
            of the experiment, and will deal with events only for those
            nodes that are allocated to the experiment. We have some long
            range plans of time sharing the robot lab, so I figured we might
            as get a little bit of a start on that.
      
          * Once it fires up, it subscribes to the usual assortment of
            events, just like the loclistener does.
      
          * It then binds a socket on which to listen for connections from
            the web server.
      
          * Then it loops, looking for events and for connections from the
            web server. Connections from the web server are for forwarding
            the event stream in real time to whatever applets are currently
            viewing the robot lab.
      
          * As each event comes in, it is parsed, entered into the DB (nodes
            and location_info table), and fired out (in a textual form) to
            all the applet listeners. The web interface just acts as pipe in
            this case for the data.
      
          * The event stream is also duplicated to a file in the experiment
            directory (the same stuff that is piped to the applet), named by
            the current resource record ID. I hope to use this stream to
            playback the motion in the applet (coupled with webcam images
            once I figure out how to sync them).
      
      * tbswap: Start and stop the new listener.
      
      * Robotrack: I changed the interface for how we actually communicate
        the event data. Much more reasonable then it was (a comma separated
        string of numbers!).
      
      * new database fields in the experiments table to hold the process ID
        of the listener and the port it is listening on. The port is not
        used yet, as the robot lab is still not shared. Will need to revist
        this later. Currently uses a fixed port number.
      
      * www/robotrack/robopipe.php3: Killed most of the old code and replace
        with simple socket connect to the listener.
      5f67fe09
    • Leigh B. Stoller's avatar
      Change to stats gathering code. Generate a new experiment_resources · c5b79546
      Leigh B. Stoller authored
      record for each swapin (previously, it was just at swapmod). The
      reason for this is that as the testbed gets more fragmented in terms
      of hardware, it is less and less likely that consecutive swapins of
      the same experiment will use the same number of physical resources.
      
      We end up with some duplication of data inside the table, but no big
      deal. I suspect we will revisit this per experiment state as the
      workbench stuff proceeds.
      c5b79546
    • Leigh B. Stoller's avatar
      Build new jar file for Tim. · 29759ea3
      Leigh B. Stoller authored
      29759ea3
    • Timothy Stack's avatar
    • Leigh B. Stoller's avatar
      Cheesy hack to avoid confusing people with output. The stats scripts · 58919d32
      Leigh B. Stoller authored
      need output formatting work, but its too low priority to find the time.
      58919d32