How to test whether a command output is non-empty in POSIX shell?

Question

How to efficiently check whether an arbitrary command produced any output in POSIX shell? (Let's say git ls-files --killed.)

There are three obvious ways, but each seems ugly or evil to me:

1. [ -n "$(git ls-files --killed)" ] - the complete output of the command has to be captured in memory and passed to test (which is at least hopefully a builtin)
2. [ "$(git ls-files --killed | wc -c)" -gt 0 ] - two fork-execs involved
3. TMP=$(tempfile); git ls-files --killed >"$tempfile"; [ -s "$tempfile" ] && ...; rm "$tempfile" - an intermediate temporary file (again capturing all the output, too) is required

Answer 1

You can check the exit status of the command also. Usually, commands if successfuly run, return an exit status of 0.

git ls-files --killed > /dev/null
if [ $? -eq 0 ]

OR if you want to only depend on the output of the command , you can use the "head -1" along with you 1st option since anyway it looks you are not doing any processing with your command output apart from knowing the result.

Answer 2

My preferred solution uses POSIX shell builtin read

if git ls-files --killed | read REPLY; then
  echo "Some output"
else
  echo "No output or git failed"
fi

read attempts to read a line from stdin (and into the variable REPLY), and returns 0 if successful, or a positive exit code if failed. The rest of the pipeline is not read. In bash only, REPLY is the default name if omitted, so read REPLY may be shortened to just read.

A potential downside of this approach is that the exit code of the program in question (git in your example) is not checked.