August 14, 2007, 12:15 PM — There are numerous ways to redirect standard error in a Perl script. You can redirect the output from one particular command, you can combine standard error with standard out so that the two are handled together or you can send all standard error to the bit bucket. None of these techniques depends on what the person running the script does. Instead, they can all be set up in your Perl scripts. Let's take a look at how this works.
First, you can redirect standard error from any particular system command in the normal way. In the command shown below, for example, we are throwing away any errors that the nmap command might be generating. This allows us to control the output that the user sees.
@results=`nmap -p $range -P0 -sT $hostname 2> /dev/null`; |
Perl also allows you to refer to standard error with the file handle "STDERR". If you want to generate messages that are specifically sent to standard error, you can set them up like this:
print STDERR "This message goes to standard error\n"; |
If you want to allow your users to redirect the expected output from a script to a file, you might want to send alerts and other messages detailing the script's progress to standard error to avoid having them redirected as well and the user never seeing them.
You can also combine standard error with standard out using a line like this:
open STDERR, '>&STDOUT'; |
Reminiscent of the 2>&1 syntax of many shells, this line instructs perl to send standard error to the same place it sends standard out.
To send standard error to the bit bucket starting at some particular point in your script, you might use this simple straightforward command:
open STDERR, '>/dev/null'; |
Any commands following this one in your script can be coded without regard for whether they might generate errors that would normally be sent to standard error. All output will be sent to /dev/null. If you want to reverse this effect later on in your script, you can close standard error much as would close any output file in a perl program:
close STDERR; |
Now let's see how all this works by examining a script that uses these various ways of redirecting output and looking at the text the script generates.
Here's the script:
#!/usr/bin/perl -w print "This goes to standard out\n"; print STDERR "This goes to standard error\n"; print STDOUT "This goes to standard out\n"; # let's start sending error output to /dev/null open STDERR, '>/dev/null'; # errors generated by trying to open a non-existent file will not appear open MISSING, "< nosuchfile"; # not let's combine standard error with standard out open STDERR, '>&STDOUT'; print "This goes to standard out\n"; print STDERR "This goes to standard out\n"; print STDOUT "This goes to standard out\n"; open MISSING, "< nosuchfile" or die "Can't open nosuchfile: $!"; |
As you can see, this script is displaying a number of messages and predicting where they will go. Writing to STDOUT is, of course, the same as using the print statement without a file handle, but I've included a print statement with and without the handle just to make this obvious.
The first three print statements will print to standard out (first and third) and standard error (second). Once we associated STDERR with /dev/null, the error that would be generated by the "open MISSING" line is not printed at all. When STDERR and STDOUT are then combined, all three of the following print lines along with the error generated by the "open MISSING" line are sent to standard out. If you run the script like this, most of your output will be sent to the file named "one".
% testout 1>one 2>two |
File "one" will contain:
This goes to standard out Can't open nosuchfile: No such file or directory at ./testout line 20. This goes to standard out This goes to standard out This goes to standard out This goes to standard out |
File "two" will contain:
This goes to standard error |
