mjhjmtu

I want to do something like:

command || log $error_from_last_command

Is there a way to use || and still access stderr like a pipe?

My intention here is to process the error message from command, using log, but only if command fails.
I'm reading through the marked duplicate but I don't see how to apply that to my situation.

If you want to use the output of a program in another that only runs after the first command completed, it's probably easiest to store the output in a file.

errfile=$(mktemp)
if ! somecommand 2> "$errfile" ; then
 log < "$errfile" # or log "$(cat "$errfile")" ?
fi
rm "$errfile"

Piping the output would require the commands to run at the same time, but we only get the exit code when the first command finishes.

log < "$errfile" above would of course direct the error message to stdin of log (like you'd get with a pipe). To get it as a command line argument, use log "$(cat "$errfile")" (one argument), or log $(cat "$errfile") (with word splitting, log sees multiple arguments), or log "$(< "$errfile")" (non-standard, works at least in Bash).

If it would be sufficient for your purposes to know the exit code of the first element of the pipe (or any element of any pipe), you can avail yourself of the bash variable PIPESTATUS, which according to the bash man page is: "An array variable ... containing a list of exit status values from the processes in the most-recently-executed foreground pipeline ...".

I realize that technically you are asking for something a bit different, but you might consider whether using this variable might meet your need in a way you didn't originally anticipate.