So one of the things our Professor has told us about is using a command, like




And putting it at the end of a script or something. So I have an idea of what this does roughly (takes the stderr output from the script you ran it in and sends it to some_file), but what all will it catch specifically?

For scripts does it make sense to use this instead of try/catch blocks? (If they even exist in korn shell, I don't know yet)

Also, is there a similar functionality for modern high level languages like c++/java? Something that will just take all the errors from stderr and funnel them somewhere?

This post was edited by Eep on Feb 21 2013 01:18pm

programs only write to stderr if they were made to. you most likely never knew it but, cin is stdin, cout is stdout, cerr is stderr. only when you write to cerr will anything be written to it.

now for most standardized linux applications they do write their errors to stderr. take gcc for example. any errors during compiling will be sent to stderr.

as for it itself it does not add any form of error handling it simply pipes all output from stderr to a file (or if you use 2>&1 i believe it is you can pipe all errors to stdout).

the reason why one would want to pipe errors to a file is so that he could view the errors later or potentially parse the logfile for specific errors.

should read that link i posted in your other bash thread. it explains all these perfectly well.

This post was edited by AbDuCt on Feb 21 2013 01:59pm

I made a bash thread?

also thanks for the reply. I see what you mean now. Try/catch still necessary for programs to handle errors.

I guess in a program (rather than a script) you would want to have each try/catch block send errors they find down cerr to a file as well? Seems a little tedious to do it for every block.