Adds support for "app info" (a mechanism that allows a plugin's author
to identify that plugin) in Ruby. This is already supported in PHP and
we're adding it elsewhere.
Previously, the value of whatever accessor was missing was left out of
the call to build it. This had the effect of skipping over the
lazily-built predicate method when the missing accessor is for a
boolean.
Of course, I realize this all is a bit contrived as most of the time
folks aren't assigning these values manually to stripe objects. However,
in my testing it caught me by surprised that the behavior is asymmetric
depending on how and when values are assigned.
As such I believe this is a less surprising implementation.
The Transfer object used to represent all movements of funds in Stripe. It
split in three resources:
- Transfer: this describes the movement of funds between Stripe accounts
and is specific to Stripe Connect.
- Payout: this describes the movement of funds from a Stripe account to a
bank account, debit card or any future payout method.
- RecipientTransfer: this describes the movement of funds from a Stripe
account to a Recipient's card or Bank Account. This is here for legacy
reasons and can only be accessed from an expanded BalanceTransaction.
This change is behind an API version so old API versions would still use
the Transfer object for everything while new API version would see the
split.
This applies beyond the new object as some properties/methods are removed
from Transfer and other properties are renamed on other objects.
The Faraday::Response object does not respond to #code. The correct
message is #status. It seems this case was previously unhandled by the
test suite as there was no case for the server responding "success" with
an invalid JSON body.
As described in #506, file uploads were broken on the way over to
Faraday and unfortunately I didn't catch it because the file upload
creation test wasn't using a matcher that was strict enough to really
catch anything.
Here we add the multipart middleware to our Faraday connection, add a
compatibility layer to `FileUpload` so that we can support `File` like
the rest-client based version always did (Faraday generally expects an
`UploadIO` object), and then tighten up our tests so that we'd be able
to catch future regressions.
Fixes#506.
Follows the path established in 2d75c8f by porting the rest of
stripe-ruby's tests over to OpenAPI. There are a few other changes here
where I've removed some tests that are duplicated or don't make much
sense, or reorganized how we test certain things, but this commit is
largely the same migration operation applied in bulk a few dozen test
suites.
Colocates the helper methods for looking up a uname by renaming them to
have the same prefix as the base method (i.e. `get_uname`).
Also adds an additional rescue in case we try to run an executable on a
system but it wasn't founded (this should never happen).
Also adds some tests to make sure that each method gets at least a very
basic amount of exercise in the test suite.
Right now we have every error class in a separate file which doesn't do
anyone a lot of good, especially given that most of them are just stubs
that derive directly from `StripeError`.
This patch pulls them all into one file and gives them some
documentation. It follows #487 and #488 as minor refactoring work.
Alphabetizes the list of named model requires in stripe.rb. This makes
it a little easier to spot things, and makes the correct spot to insert
a new model fully deterministic.
This continues along the path of #487 in introducing minor cleanups.
This one is unfortunately a lot of churn, but it's incredibly
frustrating how difficult it is to find methods in this file.
Here we alphabetize the methods, but do nothing else. Alphabetization
only occurs within visibility blocks, so visibility is not affected at
all.
As described in #481, adding a protected field like `legal_entity` as
part of an update API operation can cause some issues like a custom
encoding scheme not being considered and special handling around empty
values being ignored.
As a an easy fix for this, let's disallow access to protected fields in
the same way that we disallow them from being set directly on an
instance of a given model.
Helps address (but is not a complete fix for) #481.
