tar_writer.py: address duplicate dir warning regression

[Illedran]

It looks like #850 removed the check below, resulting in warning being printed on duplicate directories in addition to files:
https://github.com/bazelbuild/rules_pkg/pull/850/files#diff-9a83a2a408d04cf1937eb6d1eac00438774072bfe46a19082c7f8729d3ac4a27L133-L135

Fixes #899

[google-cla]

Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

View this failed invocation of the CLA check for more information.

For the most up to date status, view the checks section at the bottom of the pull request.

[cgrindel]

@aiuto Some CI tests are failing. However, the failures do not appear to be related to this change. Do we know if main is green? Are there periodic tests that run on main?

[Illedran]

Hold off for a second on this one - I've just tested it locally in my project to build tars for image layers to be used w/ rules_oci and it results in the tarball being unable to be docker load-ed:

1b7379da5 add: Loading layer [=>                                                 ]  32.77kB/837.1kB
duplicates of file paths not supported

I suspect that this is an issue specifically in the consumer/docker since duplicate directories should be valid - normal GNU tar handles archives produced from this PR just fine. While this does fix the warning, I'll take some time to give it another look

EDIT: switched to just suppress the warning for dirs which was my original intention. The remaining test failures are actually unrelated now I think