Jonatan Männchen
Jonatan Männchen
Because i simplified some things that caused problems while refactoring. My first choice was to raise an error, but that breaks a test. Also I want an error if people...
@mathieuprog that would be a possibility. I'm not sure though why we would not allow this in general for other libraries. The complete opposite would work too: * deprecate custom...
@mathieuprog that makes sense We could also forbid cast alltogether and provide a custom ex machina strategy. it seems one should be able to override it. That way we would...
@mathieuprog I'll push a "strict" version with an error being raised in `cast` and I'll see if we can add an ex_machina strategy.
@mathieuprog I've amended the PR. This shows that it is possible in general, but I've copied a lot of code. I think if we like this route in general, we...
@mathieuprog No real reason, I'm not sure which is better.
@mathieuprog Before we talk about details: * How do you like the PRs direction in general? * Would you be willing to coordinate with ex_machina to resolve the issue of...
@mathieuprog The oposite could also be done: Add optional support to ex_machina instead of this way around. That would be easier to implement I think.
@mathieuprog I'm already maintainer of far too many libraries and I'm fine just creating PRs :) What are you saying with this? 1. Should we add the files as they...
@mathieuprog Changed About the files from `ex_machina`: Can we include MIT licensed code in an Apache licensed library?