Re: R3.12 Environment Parameter system

[Andrew Johnson <anj@ast.cam.ac.uk>]

Encouraged by Steve Lewis' agreement that "things could be neater", and as I now 
have the Beta9 release, I have implemented an improved system for Environment 
parameter defaults.  It works like this:

A new configuration file, config/CONFIG_ENV contains *all* the default (site- 
specific) values.  A new script bldEnvData scans the base/include/envDefs.h file 
for all of the ENV_PARAM declarations, and combines these with the defaults to 
create a header file envData.h in the libCom target directories.  envData.h is 
#included by the envDefs.h file when ENV_PRIVATE_DATA is defined, and contains 
all of the ENV_PARAM *definitions* with their defaults.  The whole process is 
automated through the libCom Makefile.* so that changes to envDefs.h or 
CONFIG_ENV will result in a rebuild of the envData.h file.

As a result, I have deleted base/src/libCom/epicsEnvParams.h and removed the 
routine epicsSetEnvParams from libCom/envSubr.c and db/iocInit.c where it is 
called.  This leaves a question - we don't really need INITHOOKafterSetParams 
any more. Should it be removed, or left in?

Disadvantages: the vxWorks routine envShow now only displays those parameters 
you set within the resource.def file or have changed by hand; before it also 
gave the settings from epicsSetEnvParams().  However, epicsPrtEnvParams still 
works as before.  
CONFIG_ENV is interpreted using the Borne Shell /bin/sh so the settings are not 
quite as free-format as they are with Makefile variables; you can't use spaces 
inside the EPICS_ENV_PARAMETER=VALUE pair.

Advantages: Provided they were compiled with the same settings, the host tools 
and IOC will always have the same defaults.  To change a site-wide default, you 
only need to edit config/CONFIG_ENV and rebuild.  
It would be trivial to create epicsPrtEnvParams at the same time as envData.h to 
guarantee it would know about all the declared variables, but I haven't done 
that yet.

Any comments?  Have I been wasting my time?

- Andrew

          ,_}\_.        Andrew Johnson, Technology Division
         /      \       Royal Greenwich Observatory
 ~~~~~~~| ,--/;. )~~~~  Madingley Road, Cambridge, UK
    .,  |/  (/  U       Phone: +44 (0)1223 374823
  .`., /_\ . .          Email: Andrew.Johnson@ast.cam.ac.uk
     wwWWWww'           WWW: http://www.ast.cam.ac.uk/~anj/