And that's an understatement. The approach also won't work if a filename contains a tab, single quote, double quote, or backslash. There is also the remote possibility that a file name of a single underscore could be problematic.

WARNING: Pedantic piconit (of which you are probably aware) ahead.

That will still have problems with leading/trailing spaces and backslashes. Misinterpretation of such filenames manifests as an inability to find a file that exists or as reading the wrong file.

To remedy those shortcomings:

Code:

while IFS= read -r filename

Regards,
Alister

---------- Post updated at 12:50 PM ---------- Previous update was at 12:45 PM ----------

Alternatively, the task can be accomplished with multiple -exec's (which properly handles every allowable filename character):

Code:

find /etc -type f -exec printf '### START OF FILE ###\n%s\n' {} \; -exec cat {} \; -exec printf '### END OF FILE ###\n' \;

However, I would suggest following Scrutinizer's advice. The while-read loop approach is both robust and readable.

Regards,
Alister