How to grep from tee?

Question

I want to check wether the output of my command contains "rerun" (and then rerun) but I also want to display the whole output. I know that I could use one of these:

command | tee >(grep rerun)
command | grep rerun

The first one prints the whole output as expected, but I can't use it as a condition because it always returns 0. The second one only prints the lines that contain rerun, but it returns 1 if there's no match.

My usage example:

while pdflatex paper.tex | grep -E "rerun LaTeX|run Biber"; do
    biber paper
done

The answers provided here also don't help because there grep always returns 0.

Answer 1

Just use

command | tee outfile | grep rerun

or

while pdflatex paper.tex | tee outfile | grep -E "rerun LaTeX|run Biber"; do

You can check the status of the grep command, and you can later look at the file "output".

Answer 2

you could tee to a file and perform the grep on the file. Then you can use the grep exit code (0 when there's a match):

RERUN=1
while [[ $RERUN == 1 ]] ; do
  biber paper
  ! pdflatex paper.tex | tee output.txt && grep -E -q "rerun LaTeX|run Biber" output.txt
  RERUN=$?
done

The ! on the 4th line inverses the exit code of the grep process because grep returns 0 when it finds a match and 1 when no match, see the grep man page:

EXIT STATUS

   Normally the exit status is 0 if a line is selected, 1 if no lines were
   selected, and 2 if an error occurred.  However, if the -q or --quiet or
   --silent is used and a line is selected, the exit status is 0  even  if
   an error occurred.

5th line puts the last exit code ($?) in the RERUN var which is used in the loop condition.

I also added the -q option to grep to not write to stdout

Answer 3

If we stick with displaying stdout right away then the grep output in Process Substitution should be redirected to a file lest you get duplicated lines in the output, so you have to do for example this: command | tee >(grep rerun >/tmp/my.log).

Once you have a file produced, all you have to do is to use [[ -s /tmp/my.log ]] as your rerun condition.