Hi all, the bash reference manual states in chapter 3.5.7: "The shell scans the results of parameter expansion, command substitution, and arithmetic expansion that did not occur within double quotes for word splitting.". I thought I had understood this, but then came across a problem.
Let's look at a very simple script: echo 'line1' > file echo 'line2' >> file echo $(cat file) When you run this script, you get the following output: line1 line2 This is like expected: The output of "cat file" contains at least one embedded newline; since the command substitution does not take place within double quotes, the newline is replaced by a space when the result of the expansion is word-splitted. The script echo 'line1' > file echo 'line2' >> file echo "$(cat file)" gives the output line1 line2 which is also the expected one because now the word splitting does not occur after the command substitution because of the double quotes;thus, the newline is conserved and the output is two lines. So far, so good. The problem comes up if I am working with variables that store the result of the command substitution. The script echo 'line1' > file echo 'line2' >> file RESULT=$(cat file) echo "$RESULT" gives the following output: line1 line2 I don't understand why: the command substitution $(cat file) is not within double quotes, so word splitting should be applied to the result of the substitution and any embedded newlines should be removed. Thus, $RESULT should contain text without newlines after line three in the above script is executed. But obviously, the original output of $(cat file), including embedded newlines, has made it into $RESULT. Otherwise, the output would be just one line and not two. So why does this happen? I would be grateful if someone could explain when the word splitting is exactly taking place. Thank you very much, Peter
