Closed martinthomson closed 1 year ago
@chris-wood do you have any opinion about this relative to #45?
My take is that if callers never have to inspect KeyConfigs, then I would go with the former option, since this particular change forces callers to deal with a KeyConfig when otherwise they could just deal with encoded configs (byte strings).
This builds on https://github.com/martinthomson/ohttp/pull/45 and requires that the client provide a key configuration object, rather than an encoding of one.
Closes https://github.com/martinthomson/ohttp/issues/44. Closes https://github.com/martinthomson/ohttp/pull/45.