[PATCH 0 of 1 V2] test-pathencode: actually compare our current pathencoding implementations

[Adrian Buehlmann]

Reason for sending a new version:

Fixed spelling error in subject line and amended the change message.

The diff itself is unchanged.

General note:

I think it would make sense to push this patch before we resume to
further examine Bryan's pending high risk C patches.

Enabling test-pathencode.t as it is would allow to uncover potential
unwanted differences in the *existing* dual implementations (C and Python)
of the path encoding functions. We get a chance to catch potential bugs
on each testrun for free. The buildbot would even log it.