I have a command that runs forever and periodically output stuff until it gets killed by something else (similar to tail -f), and I want to make it so that whenever there's new output another command gets executed.
Caveats:
tail, it just behaves in a similar mannerHere is an example of using dd to trigger a shell command whenever something is read from stdin, whether terminated by newline or not:
{
printf not-nl-terminated
sleep 1
printf '%s\n' nl-terminated
sleep 1
echo
sleep 1
printf 'binary\0data'
} |
while : ; do
input=$(dd bs=1M count=1 2>/dev/null | tr '\n\0' __)
[ "$input" ] || break
printf 'input of size %d\n' "${#input}"
done
will give
input of size 17
input of size 14
input of size 1
input of size 11
This snippet from the standard spec may help understand dd's behavior when bs= is used explicily:
If the
bs=expr operand is specified and no conversions other thansync,noerror, ornotruncare requested, the data returned from each input block shall be written as a separate output block; if the read returns less than a full block and thesyncconversion is not specified, the resulting output block shall be the same size as the input block. If thebs=expr operand is not specified, or a conversion other thansync,noerror, ornotruncis requested, the input shall be processed and collected into full-sized output blocks until the end of the input is reached.
input=$(dd bs=1G count=1 2>/dev/null) is going to cause the trailing newline to be eaten.
input=$(...; printf x); input=${input%x}