Experimental Physics and
Industrial Control System

"Kasemir, Kay" <[email protected]> · Mon, 20 Sep 2010 09:01:23 -0400

Hi:

On 9/18/10 24:57 , "Szalata, Zenon M." <[email protected]> wrote:
> ... I know, I could get off my butt and start
> learning all there is to learn about eclipse and java and become my own
> integrator and stop complaining....
In general, people who use EPICS need somebody at the local site capable of
downloading EPICS, compiling it, setting the site-specific preferences, ...
This applies to EPICS base, Striptool, EDM, ....
CSS is no different.
We've been making it easier by providing free access to the sources via
SourceForge; BNL is spearheading an effort to standardize the headless
build; DESY is working on unit tests that are no longer hardcoded to depend
on for example a specific IOC or RDB setup. However, some local butt
movement might always be required.

> In spite of your elaborate argument, quite convincing at that, I am not
> convinced that there is no problem....
There is a problem. Xihui and I have spent several days and emails to
analyze and explain it. On the other hand, we concluded that right now for
us it's not on the top of the things-to-fix list.

There is no blatant memory leak in the BOY Intensity (Image) display.
There is a less than optimal use of memory.
It works for us, and you can also make it work for you by lowering the
update rate or image size.

If we had nothing else to do, we could fix it now by supporting more data
types in the CSS utility.pv layer, which for now generalizes all integer
type arrays into Long arrays, instead of having EPICS-specific
specializations for signed & unsigned 8, 16, 32, 64 bit types, i.e. one
instead of 2*4=8 types.
At the upcoming EPICS meeting, Gabrielle will present the current state of
his PV access library. It is designed to efficiently handle all data types
of every operating system from now until near eternity, and we will adapt
our CSS code to use it instead of wasting time on intermediate fixes that we
don't desperately need right now.

Thanks,
Kay

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System