Update mdrepo.py

[mrvanes]

This fix enables "bare" aliases to be succesfully retrievable via /alias requests. The main reason turns out to be an error in inlince doc in builtins.py.

[leifj]

Can you help me figure out why the test fails? I'm not ruling out the possibility that the test is broken or just failed randomly but I also wan't to make sure this fix doesn't break anything.

[mrvanes]

I'm not a travis expert, but I can try.
The test PR seems ok. But here I see things like this:

root: ERROR: Generating RSA private key, 2048 bit long modulus
...........+++
...........................................................+++
unable to write 'random state'
e is 65537 (0x10001)

I just don't understand why a pull request with harmless comment does not trip on this as well?

[leifj]

Thats actually not an error but stdout from a successful test. Look for "FAIL"

[mrvanes]

The

- select as /ndn

in test_alias_ndn PyFFDTest fails, as is to be expected. Back to the drawing board to see how we can discriminate between true and false positives for the strip("/").

Or, correct the inline documentation in builtins.py?

[coveralls]

Coverage remained the same at 85.501% when pulling eecaf95 on mrvanes:patch-1 into 1d032fe on leifj:master.

[coveralls]

Coverage remained the same at 85.501% when pulling eecaf95 on mrvanes:patch-1 into 1d032fe on leifj:master.

[mrvanes]

If you agree with the replacement of the strip("/") to store.py this PR looks fine now.

[leifj]

Reviewing old PRs (sorry about the delay) - can you help me unconflict this PR and also perhaps provide a testcase that demonstrates the breakage?

[coveralls]

Changes Unknown when pulling 6e0d25d on mrvanes:patch-1 into ** on leifj:master**.

[mrvanes]

It's clear from the docs now that the alias is actually a path and needs the leading /