Because our recent development work, I was put into a position where I have to monitoring a few of our application's output logs, specifically I am looking for a certain error code(s). Because my limited scripting skill right now the only thing I can do is to tail -f the log file and just keep looking for update. I also use PuTTY. I am hoping someone can make a few suggestions on how I can accomplish either one of the following, if possible:
It's weird that this not was answered completely taking into consideration that @bjarneh presented almost final solution.
So for the sake of other people coming here from search engines - here it is.
To achieve colored text in the terminal it's necessary to use ANSI color codes.
Like in this example:
echo -e "\e[$3{COLOR_CODE};1mMy colored text\e[0m"
Where ${COLOR_CODE} can be a number from 1 to 6, but base is:
Other colors can be achieved by adding those components together - like 3 - yellow, 5 - magenta, 6 - cyan. Whole sequence - for example \e[31;1m - is an ANSI terminal color code. Similarly \e[0m is an ANSI terminal code, but this one "resets" all text attributes, and stop coloring text.
How to achieve colors with sed? There needs to be an "escape" sequence present. echo -e replaces \e for specific binary code. To do this in sed \x1b is needed.
So - to color any text from stdout of any application use following snippet:
COLOR_CODE=1 #red
tail -f somefile.log | sed s/keyword/\x1b[3${COLOR_CODE};1m\0\x1b[0m/g"
You can also chain coloring, and use more complex regex:
WARNING_COLOR=3 #yellow
ERROR_COLOR=1 #red
tail -f somefile.log | sed -e "s/warning/\x1b[3${WARNING_COLOR};1m\0\x1b[0m/Ig" \
-e "s/\(error\|fault\)/\x1b[3${ERROR_COLOR};1m\1\x1b[0m/Ig"
Obviously there are more ANSI color sequences, but use this just as an example.