Re: calc and calcout records in EPICS 3.14.6

[Andrew Johnson <anj@aps.anl.gov>]

Ronald L. Sluiter wrote:
> I reported a problem with the Tornado 2.2.1 supplied compiler
> via the Mantis Bug Reporting mechanism; Bug ID #82.
>
> The jest of the report is that the epicsThreadSleepQuantum() function
> returns erroneous values when -O3 optimization is used.

We're turning down the optimization level of the 68K C compiler in Base 
R3.14.7 to -O0 as a result of this bug.  The C++ compiler is not affected. 
 I guess this will also fix the calc record problem.

For those who don't know, the Tornado 2.x C compiler is actually pretty 
ancient, being based on gcc 2.7.2 (1996).  Their C++ compiler has a 
somewhat more recent origin, although it's still old (EGCS, 1997).  WRS 
are rightly assuming that their customers need object code compatibility 
for existing object-only libraries, which any significant gcc upgrade 
would break.  The best way to find your compiler version is to run the 
relevent compiler with a '-v' switch.

- Andrew
--
Dear God, I didn't think orange went with purple until I saw
the sunset you made last night.  That was really cool. - Caro