* ADD a new option, `COLLECT_TIMESTAMPS` to the config file to configure
   whether or not to run the timestamp collection code, including
   `Stability.addOrUpdateBridgeHistory()`, in `Main.load()`.
 * FIXES #10724

There was a newly-added escaped double quote without a corresponding,
escaped pair, followed by a newline, i.e.: `\"\n` in the Thai
translation which didn't seem to mess up the template parser nor the
browsers I tested it in, but it did add an unnecessary double quote on
the page so I removed it.

Philipp Winter's scramblesuit paper has a section on creating easily
handwritable shared secrets. [0]

To my memory, this excludes using the character `O` (capital o) and the
integer `1`, because they are easily confused for the integer '0' and
the character 'l', respectively. However, Philipp mentioned on IRC that
scramblesuit `password=` fields in the `transport` line are actually
just the set of base32 characters. [1]

 * ADD a line to doc/DESCRIPTORS.md which shows an example
   scramblesuit-enabled bridge which uses a shared secret passed through
   BridgeDB. See #9013.

[0]: http://arxiv.org/abs/1305.3199
[1]: https://tools.ietf.org/html/rfc4648

Otherwise, we'd also have to pass the `transport`, and `addressClass`
parameters to renderAnswer().

 * CHANGE bridgedb.parse.networkstatus.parseRLine() to check the OR
   nickname validity only after parsing (and readding the base64 padding
   to) the OR ID digest.
 * ADD inline comments about why this is done.

 * MOVE logic for parsing filternames from the "unique string" in method
   FilteredBridgeSplitter.addRing() to new method
   FilteredBridgeSplitter.extractFilterNames().
 * ADD a FilteredBridgeSplitter.distributorName attribute which is set
   when Dist.Distributor.setDistributorName() is called.
 * REFACTOR FilteredBridgeSplitter.addRing() to use the new subring names.

This is to see if they are ever actually passed into this function. They
weren't being used anywhere inside BridgeRing.getBridges(), nor anywhere
else.

 * REMOVE the comment and other cruft around it.

 * FIXES #10559
 * FIXES a bunch of PEP8 problems.
 * CHANGE getTransportLine() to use `' '.join()` rather than formatting the
   bridge line inside the return statement. It's faster and safer.

 * ADD a wrapper which catches all unhandled exceptions which occur
   while generating an HTML response. If such an exception occurs, this
   ensures that the error is logged and the default "No bridges
   currently available" response is served to the client (rather than
   display a traceback to the client, which is what BridgeDB did
   previously).
 * FIXES #6127

 * MOVE logic for generating HTML responses out of
   bridgedb.HTTPServer.WebResource.getBridgeRequestAnswer() and into a
   new method, bridgedb.HTTPServer.WebResource.renderAnswer(). This
   allows other methods to generate responses, even if the request was
   malformed.