Re: vxWorks variable device support

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

William,

On Fri, 12 Apr 96 13:55:41 HST William Lupton <wlupton@keck.hawaii.edu> wrote:
>  There is vxWorks global variable device support for ai, ao, longin,
>longout, stringin and stringout records. The vxWorks global variable is
...
>  I propose extending device support to make use of the INP and OUT
>fields and to use an INSTIO string to hold the variable name (36
>characters are available) plus to support a leading "*" character to
>force an extra dereference. If the field is not INSTIO, existing
>behavior will occur, so the change is upwards-compatible.

Good idea - we had an application where the existing name derivation was useless 
so in the end we had to design it a different way, but it would have been nicer to 
use vxWorks variables with INSTIO and a name unrelated to the record name.

>  One could also support trailing "[n]" or "+n" syntax to reference the
>nth element of an array or at a byte offset of n. Is this worth it?

IMHO the "a[n]" array syntax is worth it; "a+n" may be confusing because some 
people might think you were providing C pointer arithmetic rather than a byte 
offset.  What about "a-n", "a[n]+m", "a[n][m]" etc?  I'm not suggesting you 
implement these, you obviously need to draw a sensible line.

- Andrew

  O o ._    Andrew Johnson, Royal Greenwich Observatory
     __\\_      Madingley Road, Cambridge, CB3 0EZ
  __|ooooo|__   Tel +44 (0)1223 374823  Fax 374700
~~\_________/~~	WWW http://www.ast.cam.ac.uk/~anj/