Issue #538 - Adding bash utility

[jasonrahm]

Updated file:
f5/bigip/tm/util/init.py

Added files:
f5/bigip/tm/util/Bash.py
f5/bigip/tm/util/test/test_bash.py
test/functional/tm/util/test_bash.py

[jasonrahm]

@caphrim007 and @zancas this is ready for review.

Structurally I believe it is sound, but I wonder if we should allow any string in utilCmdArgs that doesn't start with "-c ..." since any other use of bash from REST would be useless. Thoughts appreciated.

[jasonrahm]

@pjbreaux & @zancas, since you're in the neighborhood today, and thoughts on my previous comment?

[pjbreaux]

What would happen if I sent in FakeiControl.exec_cmd('run', utilsCmdArgs='env')?

[jasonrahm]

likely bad things.

From: pjbreaux notifications@github.com
Reply-To: F5Networks/f5-common-python reply@reply.github.com
Date: Wednesday, September 14, 2016 at 5:03 PM
To: F5Networks/f5-common-python f5-common-python@noreply.github.com
Cc: Jason Rahm J.Rahm@F5.com, Author author@noreply.github.com
Subject: Re: [F5Networks/f5-common-python] Issue #538 - Adding bash utility (#690)

What would happen if I sent in FakeiControl.exec_cmd('run', utilsCmdArgs='env')?

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHubhttps://github.com//pull/690#issuecomment-247170183, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AA0SS_eTwoQ6EpeauCZVLwM3dORvUhidks5qqG8WgaJpZM4J8OXS.

[jasonrahm]

The actual answer is:

"/bin/env: /bin/env: cannot execute binary file "

So I'm guessing I should back off from testing for /bin/bash and just test for starting with /bin/ to catch any other unacceptable arguments passed to bash?

[jasonrahm]

@jgruber requested the functionality and really only needs support for bash -c.....

But if there are other important use cases we need to allow, I want to be thorough.

[pjbreaux]

I'm for adding the restriction if it will help protect users and notify them of the requirement. And when we get a user that needs to do something without -c, then we can handle it then.

[jasonrahm]

Ok, I'll rework the code and resubmit tomorrow. Thanks for the feedback.

[jasonrahm]

looks like I'm short a test case in coveralls @caphrim007, I'll investigate.

[jasonrahm]

ok, @caphrim007, ready for review.

[caphrim007]

lgtm

[caphrim007]

"no" instead of "not". We will fix this in a follow-up PR

[caphrim007]

"be". Again, we'll fix it in a follow-up PR

[jasonrahm]

thanks, I'll fix those with next command in util.