UNIX for Dummies Questions & Answers

Hello,

I am editing a script which has a .sh extension but is listed as #! /bin/ksh that has output files where process ID's are somehow getting tacked onto the end of the filename. The programmer codes these scripts as "/tmp/email_file$$" saying that he put the $$ in there to be able to handle the process ID that unix seems to want to give it. I have never see this or had problems with this in the past in my other job - is there a way to output these files so that the process ID isn't produced at the end of the file? Or do I have to put $$ at the end to accomodate it?

Thanks,

What the original code was trying to do - create unique filenames using a poor man's shell script method. It's a common thing to do when the job ran yesterday or runs concurrently with the same filenames, and there is a chance that your script could minsterpret the presence of those files to mean something bad. Like not send email when it was supposed to.

If it offends you, use the mv command at the end of the script:

Code:

mv /tmp/email_file$$ /tmp/email_file

to clean up.

Be aware of concurrency issues before you do this.

You can also have the script cleanup (assuming it always runs under the same username)
after itself, so you don't have to.

Code:

find /tmp -name 'email_file*' -mtime +3 -exec rm -f {} \;