Normalize discovered URIs before comparing them

[cjwatson]

python-openid2's urinorm is stricter than that in the old Python-2-only
version: in particular, it always normalizes the reserved character '+'
to '%2B'. The discovery process normalizes URIs, so
GenericConsumer._verifyDiscoverySingle could end up comparing normalized
and unnormalized identifiers.

In particular, this broke with the Ubuntu single-sign-on system
(login.ubuntu.com / login.launchpad.net) which includes a '+' character
in the path part of its identifiers.

It makes sense to normalize both expected and discovered identifiers
before comparing them, so do that according to the rules in OpenID
Authentication 2.0 section 7.2.

[cjwatson]

#42 fixes the CI failures here. I plan to rebase this once that PR lands.

[cjwatson]

Rebased now, so CI should be happy shortly.

[codecov-commenter]

Codecov Report

Merging #41 into master will increase coverage by 0.00%.
The diff coverage is 92.75%.

@@           Coverage Diff           @@
##           master      #41   +/-   ##
=======================================
  Coverage   89.48%   89.49%           
=======================================
  Files          93       93           
  Lines       12015    12058   +43     
  Branches     1084     1087    +3     
=======================================
+ Hits        10752    10791   +39     
- Misses       1094     1096    +2     
- Partials      169      171    +2     

Impacted Files	Coverage Δ
openid/consumer/consumer.py	95.00% <82.14%> (-0.48%)	⬇️
openid/test/test_verifydisco.py	98.97% <100.00%> (+0.13%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 4b0cbbf...9b980fd. Read the comment docs.

[cjwatson]

@ziima @tpazderka Would there be any chance of getting a review of this, please? I have I think three other projects so far that are blocked on this as part of their Python 3 upgrades.

[ziima]

Sorry about the delay and thanks for poking us :-)

I believe the cause is in the urinorm itself, so I prepared a fix only for that in #43. I'd appreciate if you could check that it really helps with this issue.

[cjwatson]

@ziima #43 seems to fix my problem just as well, so I'm happy with that. Thanks!

[ziima]

Closing, since #43 fixed the issue.