UNIX for Advanced & Expert Users

I'm more familiar with bash/ksh that csh/tcsh. With that said, I recently ran across a problem with tcsh.

Our system admin recently installed purify on our solaris 8 system. In order to use purify you have to execute a Rational script in order to setup the paths and some environment variables correctly. Since we only have a single license for purify, our admin setup the license for a single "common" user account that the developers could use. This common user account happens to be defined to use tcsh as its default login shell.

Our system admin went on to add a 'source' command to the .tcshrc login file for this common user so the Rational script would automatically get executed upon logging in.

At this point everything seems to work just fine. However, it was later discovered that commands from a remote host to the user of the host upon which purify was installed no longer work. Commands like ssh, scp, and rsync.

After a bunch of experimenting, I was able to track this problem down to some output statements (in particular, 'echo') within the Rational script. An example would be... let's say you execute an scp command on a remote host to copy some files from the user on the system for which purify was setup. You actually see the output from the echo commands within the Rational script, but then scp will finish without performing the copy.

Instead of modifying the third party Rational script, I decided to mess around with the .tcshrc login file. Since I'm not that familiar with the tcsh shell, the first thing I tried was to simply redirect the output of the sourced Rational script a multitude of different ways, none of which worked. Here's an example of one attempt.

source /Rational/purifyplus_setup.csh > /dev/null

Eventually I stumbled across a solution by checking the TERM environment variable prior to executing the Rational script. The script will only execute if TERM exists.

This brings me to my questions...

1. Why would output statements cause remote commands to fail?

2. Why does redirecting the output not work? In fact, redirecting the output did absolutely nothing. You still see the output.

3. Is there a better solution than checking the TERM environment variable? I don't know if doing such a check will cause some other login or remote login senario to fail.

I can't tell without seeing the script. Did you try redirecting stderr, too? or just stdout? Depends what you're trying to do, which is not clear at this point.

Which script, the Rational script or the .tcshrc login script? I'm afraid I cannot get the scripts off of the system they're on as it's a classified system. However, if it's the Rational script you're interested in, here's a real simple test you can try which I've done to duplicate this problem in a very simple form.

1. Create a script, say abc, with the following...

echo hello
set xyz=aaa

2. At the bottom of your .tcshrc login file enter the following...

source abc

3. Now login and make sure you see the output from your echo, and then check your xyz environment variable to make sure it's set. If everything works up to this point, proceed with the next step.

4. Get on a remote host, or you can actually try this from the same host if you wish, but as a different user. Execute the following command.

$ scp -pr user@hostname:filename ~/.

...where user is the user you set up your abc script above and modified your .tcshrc file. hostname is the name of the host for the same user. filename can be any file you want to copy from that users account. When you execute this command you should observe that the echo from your script is still being output, but no file will get copied. Why???

5. Now add some redirects to your sourced abc script within your .tcshrc file. Try this for example...

source abc > /dev/null

6. Perform step 4 again. You should observe the output from the echo command still being shown. Why???

This should duplicate the kind of problem I'm experiencing.

I believe so. Like I said, I'm not that familiar with tcsh. According to the man pages for tcsh it says... "The shell cannot presently redirect diagnostic output without also redirecting standard output...", so I tried this...

source /Rational/purifyplus_setup.csh >& /dev/null

This also did nothing. Output was still visable. There was a bunch of other syntaxes I tried as well, but I can't remember all of them now.

We're simply trying to get the Rational script to execute upon login so the paths are automatically setup correctly (which already works), and at the same time get remote commands to successfully execute (which doesn't work, unless you do what I've already mentioned in my original post, upon which I still have my question of concern in #3).

Does anyone have any answers or ideas regarding my questions in the original post?