Re: Proposed change in asyn - request for comments

[Bruce Hill <bhill@slac.stanford.edu>]

Interesting!
I can see how that could be quite useful!
As we don't use post_event() for that purpose, I just
assumed it was a general purpose signaling mechanism.
It also works quite nicely for that purpose, as I used it to
implement a history mechanism in our EDT frame grabber.

Regards,
- Bruce

On 6/1/2012 7:55 AM, Andrew Johnson wrote:
> Hi Bruce,
>
> On 2012-05-31 Bruce Hill wrote:
> > My concern was that overloading the TSE field to contain EVR Event codes
> > would be inappropriate for a module like asyn that is used by many sites
> > that don't use EVR's.
> The TSE and TSEL fields were originally /designed/ for use with timing system
> event codes.  The Record Reference Manual describes the TSE field thus:
>
>      This indicates the mechanism to use to get the time stamp. '0' - call
>      get time as before '-1' - call the time stamp driver and use the best
>      source available. '-2' - the device support provides the time stamp
>      from the hardware. Values between 1-255 request the time of the last
>      occurrence of a generalTime event.
>
>    https://wiki-ext.aps.anl.gov/epics/index.php/RRM_3-14_dbCommon
>
> I agree that this mechanism is not currently much use on systems that don't
> use an event timing system, although it would be quite feasible to write a
> plugin for the generalTime event subsystem that provides global event times
> from UDP packets.
>
> - Andrew