Colouring stderr in terminal

[afora]

Hi guys,

I wanted to colour red all stderr to quickly differentiate between different messages in one of my LM setup scripts. I found this one online which sort of works fine but not quite:

Code: Select all

# Colour stderr red for bash terminal
export COLOR_RED="$(tput setaf 1)"
export COLOR_RESET="$(tput sgr0)"
exec 9>&2
exec 8> >(
    perl -e '$|=1; while(sysread STDIN,$a,9999) {print
"$ENV{COLOR_RED}$a$ENV{COLOR_RESET}"}'
)
function undirect(){ exec 2>&9; }
function redirect(){ exec 2>&8; }
trap "redirect;" DEBUG
PROMPT_COMMAND='undirect;'

The problem with it is that it colours red also some irrelevant messages (like if I do $read "Press Enter to continue"), and also the most important it prevents me to run $sudo su as terminal would simply hang. I have to literally comment out the code in .bashrc and restart terminal before being able to elevate myself to root.

I have no idea what this piece of code does frankly. Could anyone suggest either how to fix it or may be there's an easier way to accomplish this.

Many thanks!

[xenopeek]

A quick test shows that read outputs the prompt on stderr, hence it gets colored with your setup:

Code: Select all

$ read -p "Press enter to continue" 1>/dev/null
Press enter to continue
# ^ text was shown so stdout isn't used for outputting the prompt

$ read -p "Press enter to continue" 2>/dev/null

# ^ no text was shown so stderr IS used for outputting the prompt, hence it is red with your setup

A workaround for that specific case is to redirect stderr for the command to stdout:
read -p "Press enter to continue" 2>&1
With this it won't color the prompt.

I only ran your commands in a shell, didn't put them in my .bashrc. With that it also hangs on sudo -i for me but also if I simply run bash to start a new shell. Does it also hang for you if you simply run bash?

Where did you find this code. Perhaps the discussion on that page provides more clues.

[afora]

Fantastic, thank you for resolving the very first question. I should have thought about it myself!

As to the second point - I got it from a now defunct forum threat. I do not think I understand what the code is doing, so I thought I would ask it here. It's a really great solution when one is running a whole bunch of scripts in an automated fashion to distinguish between failures and normal execution.

Perhaps any suggestions from anybody who's seasoned here?

Many thanks!