Capturing Dialog output for further processing

There are several ways you can get output from Dialog or know what action a user took or any data that was entered.

Return Codes

The most basic dialog will always have at least one button which when clicked, will exit with a return code of 0.

The optional buttons, initiated via --button2, --infobutton (or derivatives), when clicked will give return code 2 or 3 depending on which one was clicked.

The following return codes are sent after each action. These can be ignored or used in a calling script for further action

Action	Optional	Hotkey	Return Code
button1	No	Enter	0
button2	Yes	Esc	2
infobutton	Yes		3
<timer>	Yes		4
command file quit:	Yes		5
--quitkey	Yes	cmd+q	10
Key Auth Failed	Yes		30
Image Resource Not Found			201
File Not Found			202

To process the return code on your script simply capture the result from $? in bash or via your chosen scripting language.

Example, processing return codes in zsh:

#!/bin/zsh
dialog ...<options_here>
case $? in
  0)
  echo "Pressed OK"
  # Button 1 processing here
  ;;
  2)
  echo "Pressed Cancel Button (button 2)"
  # Button 2 processing here
  ;;
  3)
  echo "Pressed Info Button (button 3)"
  # Button 3 processing here
  ;;
  4)
  echo "Timer Expired"
  # Timer ran out code here
  ;;
  5)
  echo "quit: command used"
  # post quit: command code here
  ;;
  10)
  echo "User quit with cmd+q"
  # User quit code here
  ;;
  30)
  echo "Key Authorisation Failed"
  # Key auth failure code here
  ;;
  201)
  echo "Image resource not found"
  ;;
  202)
  echo "Image for icon not found"
  ;;
  *)
  echo "Something else happened"
  ;;
esac

stdout

In addition to return codes, output will be sent to sdtout if one or more --textfield or --selectvalues are presented. The default output is non-formatted other than to split the label and output with a : character.

Given the following dialog and user input:

dialog --textfield "text 1" --textfield "text 2" --selecttitle "make selection" --selectvalues "one,two,three,four"

The output would look as follows:

SelectedOption : three
SelectedIndex : 2
text 1 : one
text 2 : two

SelectedOption is the text of the drop down option that was selected SelectedIndex is the index of the option that was selected (indexes start at 0 so index 2 means the third option as listed was selected)

Text fields are presented using the label passed in to dialog as the label in the output, so for the above example:

text 1 displays the text that the user input in the text field labelled "text 1" and the same goes for text 2

To split these up you could use awk to process each line of output you were interested in, e.g.

/usr/local/bin/dialog <...> | grep "SelectedOption" | awk -F " : " '{print $NF}'

For more advanced processing you can pass in the --json parameter which will then format the output as json for processing using a json parser (e.g. in python).

json output for the above example would appear as follows:

{
"SelectedOption" : "three",
"SelectedIndex" : 2,
"text 1" : "one",
"text 2" : "two"
}

Additional examples

capturing the output and return code using python:

import subprocess

result = subprocess.Popen("/usr/local/bin/dialog <...>")
text = result.communicate()[0]   # contents of stdout
return_code = result.returncode  # return code