Ensure command names in search path of run() or spawn() commands do not match folders.

[lcm337]

fix for incorrect directory match

[takluyver]

That makes sense. Could you add a test for this in tests/test_misc.py?

[takluyver]

No need for this check - pexpect only works on Unix-y systems anyway.

[takluyver]

@lcm337 , I'll be happy to merge this if you fix the couple of things I pointed out - the change to command lookup, and the unnecessary Windows test code.

[lcm337]

Will do.
On 13 Feb 2014 19:13, "Thomas Kluyver" notifications@github.com wrote:

@lcm337 https://github.com/lcm337 , I'll be happy to merge this if you
fix the couple of things I pointed out - the change to command lookup, and
the unnecessary Windows test code.

Reply to this email directly or view it on GitHubhttps://github.com//pull/37#issuecomment-35014126
.

[takluyver]

Ping.

[jquast]

weird, wouldn't this never fail?

>>> assert True, 'never fails'
>>>

[jquast]

just got to this to say i like the isdir() check. the test is kindof ok.
The ENV thing is pretty strange, I don't get it. this is to help with testing?

This could be done with mock -- or more simply, overriden.
The child process is an exact duplicate of the parent, environment values and all.

os.environ['PATH'] = os.path.pathsep.join(('/XYZZY', os.environ['PATH']))
if not os.fork():
    time.sleep(0.1)
    print('child: {}'.format(os.environ['PATH']))
else:
    print('parent: {}'.format(os.environ['PATH']))

outputs

parent: /XYZZY:/Users/dingo1/.virtualenvs/pexpect/bin:/usr/local/bin:/usr/local/share/python3:/usr/local/share/python:/Users/dingo1/bin:/usr/local/sbin:/usr/sbin:/usr/bin:/bin:/usr/sbin:/sbin:/usr/local/bin:/opt/X11/bin
child: /XYZZY:/Users/dingo1/.virtualenvs/pexpect/bin:/usr/local/bin:/usr/local/share/python3:/usr/local/share/python:/Users/dingo1/bin:/usr/local/sbin:/usr/sbin:/usr/bin:/bin:/usr/sbin:/sbin:/usr/local/bin:/opt/X11/bin

[jquast]

does sh have the same bug? Interesting.

https://github.com/amoffat/sh/blob/master/sh.py#L193

[jquast]

just submitted a similar fix to sh. I can see @lcm337 timed out regarding your suggestions, will pick it up.

[jquast]

Closed by pull #58