Minutes from the 02/02/05 LCLS SLC IOC meetings.

(1) New Meeting Room - Next week we'll start meeting in B280, 9:30.
    Working on a room - maybe POM 280C, Rm 114.

(2) cstrAsync and async utilities:
    Discussed some requirements and implementation issues.  RonM will have
    a requirements review on Feb 16 or 17.  He has a "hello world" cstrAsync
    ready which Diane will include in the main distribution.  

    We asked that create his first utility in slcAsync.c called 
    slcAsyncSleep that is called by cstrAsync:
    (1) Check if the "stop" flag for this task is set.  If set, immediately 
    exit with the proper status.  cstrAsync should then exit with the proper 
    cleanup (reset "active" flag).  
    (2) Call epicsThreadSleep with a 1 second timeout.  Go back to (1).
    Later he will add the async logic (for determining when it is time to
    do something) into this utility.  At least now cstrAsync will stay up 
    until commanded to "stop" by slcExec.

    For testing, he will add either calc or soft ao records to the slc ioc
    test EPICS database (records names = CSTR:$(micro):1:CPU and 
    CSTR:$(micro):1:RMX to hold CPU and memory usage).

(3) Database Service:
    Debbie is finished with the database creation (with the iocsh dbdumphash
    working) and has moved on to the database utilities.  Spec has been 
    updated.  She'll do these tasks in this order:
    * db utilities except dbupdate and hi/lo logic
    * iocsh routines to test db utilities
    * dbupdate and hi/lo logic
    * add/correct cmlog messages
    * add any missing diagnostics
    Done by April 1.

    Debbie mentions that we may want to set the size of the memory chunks
    used to create the dictionary depending on how large the SLC database
    is on that IOC.  She was thinking of an environment variable - I think
    a iocsh routine to set a global or file-scope variable might be better 
    (as is done to set sizes for other EPICS "services").  This routine
    would be called before the slc tasks are started or restarted.  

(4) Message Service and General Utilities:
    Diane is finished testing the message service.  Specs have been 
    updated.  Last week a bug was introduced that killed the proxy.  
    KenU is fixing the proxy with Diane's and Debbie's help.  Diane 
    will now do these tasks in this order:
    * get shared development area up-to-date (Ron's and Debbie's code)
    * make sure what is there so far builds for RTEMS and linux
    * finish data conversion utilities
    * implement code for the CTL socket needed for Debbie's db update logic
    * handle the SLC_NOTIFY function code (for SCP restart)
    * add/correct cmlog messages
    * add any missing diagnostics
    Done by April 1.

    Diane wants to test sending large buffers from the IOC to tha Alpha.
    We wonder if this can be done with the TEST_ECHO_MWORD function code.
    Diane will find out where in the SCP the TEST_ECHO_MWORD is passed
    down.

(4) IOC Shell Routines:
    All IOC shell routines will be registered in one file under exec
    (ie, slcRegistrar.c).  That way, IOCs will only need to add one registrar
    line in their *Include.dbd file:
    registrar(slcRegistrar)
    To keep all the slc ioc routines in one place when "help" is typed from
    iocsh, all iocsh routines will start with "slc" with the next character
    capitalized (ie, "slcDbedit" or "slcDBedit").

(5) CVS, Code Review, Development:
    No CVS yet until after the cmlog messages are cleaned up.  Before CVS,
    Diane will add in Debbie's name in the review line for her files and 
    vice-versa.  I will review Ron's code.  Diane asked how best to import:
    Clean up the directories.  No garbage or O.* directories.  Then:
    cd to the working directory right above configure
    setenv CVSROOT /afs/slac/package/epics/slaconly/cvs
    cvs import -d site/src/slc SLC SLC-1-0
    cd /afs/slac/package/epics/R3.14.6/
    cvs checkout site/src/slc
    cd site/src/slc
    gmake
    More instructions for the LCLS area later.

    For development, slcsun1 is XL02 and slcs6 is XL01.  Diane is using ddd
    for debugging (works on solaris and linux) and she finds it very useful.
    The keyboard doesn't work right when using XWin32 on a PC so she uses
    slcsun1 directly instead of a PC xterm.  It would be nice if a simple
    method to reserve XL01 and XL02 on unix was available so developers don't
    stomp on each other.  Another test micro needed?
 
(6) iocCmlog:
    I found some problems which I cleaned up.  I need to test and then will
    CVS into site/src/iocCmlog.  This week.