Add -S option to pvget/pvmonitor to stringify byte arrays

[JJL772]

This emulates the behavior of the caget -S option. Byte arrays are displayed as strings.

Requires slac-epics/pvAccessCPP#2

[mskoufis]

Some comments so I understand this better.

• For this check below:

for (shared_vector<const int8_t>::iterator it = byteArray.begin(); it != byteArray.end(); ++it) {
                            if (!*it)
                                break;
                            strm<<static_cast<char>(*it);

What happens if we encounter a legitimate 0x00 byte within the binary data? Does it exit prematurely?

• Also, are all byte arrays NUL terminated? If not, then this loop may keep on reading passed the array byte content or no?
• Do we need to consider the case of pvUByte for the byte array or only pvByte? Seems that we convert to characters for pvByte only. Was this the intent?
• Do we expect to have any special characters in the byte array that shouldn't need to be stringified or may cause printing issues?
• Do we need to keep the floating point precision if we just print characters?
  strm<<std::setprecision(6);

[JJL772]

Also, are all byte arrays NUL terminated? If not, then this loop may keep on reading passed the array byte content or no?

Since it is using an iterator and checking against byteArray.end() it shouldn't overrun.

Do we need to consider the case of pvUByte for the byte array or only pvByte? Seems that we convert to characters for pvByte only. Was this the intent?

This is a good point. Initially I was going to have it apply to pvUByte as well, but looking at the caget code, it only stringifies signed byte arrays on request.

Do we expect to have any special characters in the byte array that shouldn't need to be stringified or may cause printing issues?

This is a good point; perhaps I should escape unprintable chars? Then they'd be printed as \x12 (for example). I'm not sure what the caget code does, and I'd like to match the behavior with that somewhat.

Also worth noting that non-ASCII chars encoded with UTF-8 (or ANY chars encoded with UTF-16/32) will be printed incorrectly. I dont think we have to worry much about encoding though,

Do we need to keep the floating point precision if we just print characters?
strm<<std::setprecision(6);

Good point, I will apply this change

[JJL772]

Seems like I can just use epicsStrnEscapedFromRaw, which is what caget uses.

[JJL772]

@mskoufis this is how it behaves now with unprintable chars:

$ pvget -S mystring
mystring 2026-05-28 16:44:48.472  *@,\x16\n\x02
$ caget -S mystring
mystring *@,\x16\n\x02