Miscellaneous usability improvements

[jaseg]

Here's three small changes I found useful. Tell me what you think.

First, using @functools.wraps, docstrings and function signatures can be preserved when using @autocommit. This is useful in interactive usage. This change should be backwards-compatible to at least python 2.7.

Second, I found Table.chains not very useful as a plain list, since I'd only iterate over it to get a particular chain anyway. A dict makes more sense to me here. This is obviously breaking the API, so I don't expect you to merge this right away.

Third, I wrote some nice __repr__s for the most common ip4tables classes. This also is really useful in interactive usage, or even when just printing debug output. Be aware that as-is the formatting code pretty-prints IP adresses using the ipaddress module found in python 3.3 upwards.

Thank you for the nice module, jaseg

[jaseg]

I'm not sure what's up with the Travis tests here. All unit tests pass on my machine, and the error it throws looks like it might be unrelated to my changes?

[ldx]

I think the tests fail due to the new dependency. Can you try applying this diff:

diff --git a/setup.py b/setup.py
index 145f0c3..f40a9d9 100644
--- a/setup.py
+++ b/setup.py
@@ -23,6 +23,7 @@ setup(
     ext_modules=[Extension("libxtwrapper",
                            ["libxtwrapper/wrapper.c"])],
     test_suite="tests",
+    install_requires=["ipaddress"],
     classifiers=[
         "Development Status :: 5 - Production/Stable",
         "Environment :: Console",

[coveralls]

Coverage increased (+0.3%) to 59.606% when pulling fdf0a0b on jaseg:master into 583f939 on ldx:master.

[jaseg]

That was it. I have added a version check, since ipaddress is a built-in module starting in python 3.3. I think it is better to not have that dependency on these versions.

[ldx]

Looks good, thanks!

Can you add a few test cases for testing the new __repr__ methods?

[jaseg]

Here you go.

Two observations:

• Travis does not actually run the unittests with the python 3.4 it installed. Instead they're run with python2.7 as evidenced here: https://travis-ci.org/ldx/python-iptables/jobs/335130444#L549
• There is an upstream bug in the ipaddress module, as the failure we can observe there does not happen on the built-in version of python 3.5.4. I'll see what we can do there. Maybe I'll just remove the fancy formatting.

[jaseg]

Turns out, ipaddress is actually behaving to its spec, that spec is just very dumb. I'll just rip out that part of the repr code.

[ldx]

@jaseg thanks, please let me know once it's ready for review. 👍

[jaseg]

There. In my enthusiasm I changed quite a lot, have a look whether that works for you.

[ldx]

LGTM, and the tests are passing. The only thing remaining is to update doc/examples.rst - can you look into that as well? Thanks!

[Troyhy]

Is this going to be merged? I've been waiting for a while :)

[ldx]

@Troyhy if you (or someone else) can add a few lines of documentation to doc/ and resolve the merge conflicts, I'm happy to merge it :)

[Troyhy]

Ok, I'm using your library in one project. I will try to resolve it as part of the project.

[ldx]

Looks stale, closing. Feel free to reopen if this is still relevant.