Making __import__ get the dynamically added methods

Question

Hey,

At my work place there is a script (kind of automation system) that loads and runs our application tests from an XML file.

In the middle of the process the script calls __import__(testModule) which loads the module from its file.

The problem starts when I tried adding a feature by dynamically adding functions to the testModule at runtime.

As expected, the __import__ gets the old version of the module which doesn't have the methods I just added at runtime.

Is it possible to make the __import__ calls import the newer version of the class (which includes the methods I added)?

Please note that I prefer keeping the automation system untouched (even when it would help solving the problem faster).

Thanks

Tal.

Answer 1

reload(testmodule)

might work.

Answer 2

You need to be aware that reloading a module won't magically replace old instances. Even if you do reload, only new objects will use the new code!

The only way to replace code during runtime is to wrap everything in a proxy object! You can sometimes do this, ie for specific, self-contained modules, but in most cases it's simply not a reasonable approach.

Quick demonstration:

>>> import asd
>>> asd.s
'old'
>>> t = asd.s
>>> reload(asd) # I edited asd.py before
<module 'asd' from 'asd.py'>
>>> asd.s # new module content
'new'
>>> t # but this is still old!
'old'

Most applications that looks like it reloads code actually just restart!