Experimental Physics and
Industrial Control System

Andrew Johnson <[email protected]> · Fri, 11 Dec 2009 10:08:23 -0600

Hi Simon,

On Thursday 10 December 2009 22:22:26 Simon Hoyle wrote:
>
> I'm attempting to use an asub  to read an array of
> strings from a waveform record with Epics 3.14.11

...

>   void loadETA(aSubRecord * asub)
>   {
>     int etaLen = *(short*)(asub->b);
>     char** etaIn = (char**)(asub->c);
>
>     for (int i = 0; i < etaLen; ++i)
>         printf("\t%s\n", *(etaIn++));
>   }

What's wrong is that with FTC=STRING asub->c is not a char**, it's actually a 
pointer to a char[MAX_STRING_SIZE][asub->noc] i.e. a single flat buffer of 
size NOC*MAX_STRING_SIZE.  The IOC code is very careful to avoid allocating 
and freeing memory buffers during record processing, which is why all our 
string fields are fixed length.

There is a typedef named epicsOldString which is useful in this case; if you 
change your subroutine code to this you'll have more success:

long loadETA(aSubRecord * asub)
{
    int etaLen = *(short*)(asub->b);
    epicsOldString *etaIn = (epicsOldString *)asub->c;
    int i;

    for (i = 0; i < etaLen; ++i)
        printf("\t%s\n", etaIn[i]);

    return 0;
}

tux% caput -a wf "*ignored*" Alas poor Yorik, I knew him Horatio.
Old : wf 10
New : wf 10 Alas poor Yorik, I knew him Horatio.

epics>  Alas
        poor
        Yorik,
        I
        knew
        him
        Horatio.

HTH,

- Andrew
-- 
The best FOSS code is written to be read by other humans -- Harald Welte

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System