I've seen chunks of code in lots of scripts that look something like this:
What does this do? It looks like some kind of comparison, but it dumps everything in /dev/null. I don't understand.

-elyk

1> redirect stdout to /dev/null
2> redirect stdin to /dev/null


> is not acting as the greater than operator in this case but as a redirection operator.
for example

will write the string "rubbish" to temp.txt

you might want to read the man page for bash

just a small correction - 2> is for stderr, not stdin (stdin is 0).

Doesn't 2> redirect stderr? If so, this redirects stdout and stderr to /dev/null, so shouldn't this always give you nothing?

oh my.... i think sleep is my only cure.

Yes, adding this to the end of a command basically hides all output. This is usually done in scripts to prevent too much junk printing up on the console when it's not necessary.

yeah, you squeezed in your reply just before mine.

Just to add some more info to this thread...

This sort of output redirection is typically used (but not limited to) by distro init scripts, so instead of seeing the output of all the startup commands you see the much more friendly:
etc...

can be written more concisely

Yes, all output from the command in the if-statement gets discarded this way. But the output is not what "if" uses to decide to execute some other commands or not. It is the exit-code from the program that matters for the if-statement. More precise: the exit-code of the last command inside the if-statement.

The exit-code is a number from 0 to 127 each program returns to its parent process (often the shell) when it ends. By convention exit-code 0 (zero) is returned when the program ran succesfully, and some other exit-code (non-zero) when some error ocurred, or when the program did not succeed in some other way.

You can see the exit-code of the last command executed in bash with "echo $?". Example:
As you see, apart from the output of ls it returns 1 when there was an error, so ls didn't succeed, and 1 when ls ran fine.

When you want a script to do something if a file or directory exists, you normally do:Here, "test" is a program that checks if the file or directory exist, and return an exit-code of 0 or 1 depending on whether the file /tmp/some_file exists or not. Note that the "test" program doesn't do any output. It is the exit-code that matters to the "if" statement.

Also note that when you do this instead:You should think of "[" being a program that returns an exit-code ( "[" actually is a bash-builtin command, as is "test" by the way).

Another way to check whether a file (or directory) exists in a bash-script could be:
The output of the "ls" command is directed to /dev/null to prevent it from appearing as output from the script. It's the exit-code what is needed by the "if" in the script.