Comparison to stripe-rs

[seanpianka]

Hi @arlyon, thanks for making this fork!

Besides the async support, could you describe how else async-stripe differs from stripe-rs? Is there support for other API features that are currently missing from the codegen in the stripe-rs project?

Namely, support for:

• idempotent requests, connection pooling, retries
• 2020 API
• Checkout,
• webhooks (also, there's a security vulnerability in the upstream crate)

I'd be willing to help port over any missing functionality that's in stripe-rs and not available in this project yet. Hopefully, Stripe is start supporting an official Rust crate soon...

[arlyon]

Hey Sean!

I have a post here in the discussions that lists changes. #79

While I originally was planning on upstreaming (which is why I was pushing to get us maintainer access), this diverged fairly significantly in the interim and I don't feel comfortable essentially 'overwriting' the other project without support (or interest) from the original maintainers. That said, if they are on board I am very happy to merge all these changes back in. Though I will say, as time went on I became more and more opinionated on this fork.

To answer your bullets:

• no work has been done on idempotency, though I'd like to see it included at some point. it's just a question of finding the time
• the api is regenerated weekly and should be mostly up to date (though there is an issue with the openapi definitions from the last 2 months or so that causes builds to fail; there are some types that aren't generated correctly)
• I added some codegen changes that I believe fixed the checkout api (see this issue CheckoutSession::create ? #24)
• no work has been done on webhooks

[seanpianka]

Thanks for the quick response!

I'd bet the upstream maintainers would be open to upstreaming each of the items in #79. They're all major improvements compared to what stripe-rs has on master or (especially) crates.io currently. I'm pushing for this effort to upstream solely because (I estimate that) it has more active users presently, and all this good work would get into the hands of a greater number of people faster.

My estimation might be off of course (and the recent downloads on crates.io do tell a different story), but I predict there's a decent group of users currently pulling it via a git dependency. The larger the community, the better? 🤔

[arlyon]

(converted this into a discussion) If we can get the green light from the core maintainers then lets go for it: the sooner the better really!