Dirkjan Ochtman

So as I understand it: * The default currently is to use reqwest with native-tls (OpenSSL on Linux, native stacks on macOS and Windows) * If `RUSTUP_USE_CURL` is set, we...

> > At least for some of the issues, there's the suggestion that native-tls might actually be at fault, such that reqwest + rustls might actually do better than the...

> > Yes, I think it would make sense to flip the default for the reqwest backend to rustls, but in that case I feel like we should keep the...

> IRL this can happen for multiple reasons, e.g. the permission is not correct, the binary is corrupted, etc. It would probably be helpful if we can propagate/show the exact...

Adding rustup to thanks seems nice, but I worry that it might be quite a bit of work, since from the description it looks like it might currently be tied...

I'm fine with holding this back until after 1.27.1 is released, though I also think the risk is pretty limited.

> I'm fine in principle with this. there is a similar structure in the fork of `home` that we use - https://github.com/rbtcollins/home/blob/master/src/lib.rs#L41 - so you may want to tackle that,...

> > Personally, I've grown to substantially dislike the style where the parent module root lives in a separate directory from the child modules, since it keeps them farther away...

I've scaled this back a little for now -- it no longer moves the code around and just gets rid of the trait indirection in favor of an enum.

> I think we are somewhat on this journey already, but would like to make sure we don't build larger aggregates than we actually want - and just doing the...