This changes the library's default connection over to use the adapter
for `Net::HTTP::Persistent`, which is a connection pooling library for
Ruby.
In the long run, I think we should probably just drop Faraday ... the
amount of value it's getting us is extremely tenuous and its API is
difficult to work with. I hate to do it at this point though because
technically people could be writing custom middleware for it.
Overrides `#eql?` (hash equality) and `#hash` so that Stripe objects can
be used more easily as Hash keys and that certain other frameworks that
rely on these methods will have an easier time (e.g. RSpec's `change`,
see #687).
I think this might be a little controversial if we weren't already
overriding the `#==` implementation, but because we are, I think it
makes sense to extent it to these two methods as well.
Changes all arrays from classic Rack encoding:
``` sh
arr[]=...&arr[]=...&arr[]=...
```
To integer-indexed encoding:
``` sh
arr[0]=...&arr[1]=...&arr[2]=...
```
We think that this should be tractable now that we've fully converted
all endpoints over to the new AbstractAPIMethod infrastructure on the
backend (although we should do a little more testing to make sure that
all endpoints still work).
As part of the conversion, we also remove any places that we were "spot
encoding" to get required integer-indexed syntax. This should now all be
built in.
Remi pointed out in #666 that we basically just have to keep adding more
more onto the `Max` exception for both these rules every time we add a
new API resource.
Here I suggest that we modify the check on method length in two ways:
1. Permanently disable the cop on `Util.object_classes`. This is just
going to keep growing until we change are approach to it.
2. Choose a more reasonable maximum of 50 lines for elsewhere (IMO, the
default of 10 is just too short). Most of our methods already come in
below this, but there's a couple outliers like `#execute_request` in
`StripeClient`. If we knock over some of those, we could lower this
number again, but I suspect that we'd probably want somewhere closer
to 30 (instead of 10) event then.
I also disable the check on module length completely. I'm not convinced
this is a very good heuristic for code quality.
