Use doc_auto_cfg

[tcharding]

If we use #![cfg_attr(docsrs, feature(doc_auto_cfg))] instead of #![cfg_attr(docsrs, feature(doc_cfg))] we no longer need to manually mark types with #[cfg_attr(docsrs, doc(cfg(feature = "std")))].

Sweeeeeet.

Props to pezcore for the lesson :)

[Kixunil]

I'm not really sure, if we do this we will need to check if any odd cfgs (like rust_v_x) get rendered correctly. IIRC there were some issues with this. If this renders correctly today then I guess I'm fine with this.

[tcharding]

From my investigation just now it looks like this PR actually improves the situation.

Build docs on master

impl<'a> From<&'a Script> for Arc<Script>
Available on target_has_atomic="ptr" only.

Build docs on this branch

impl<'a> From<&'a Script> for Arc<Script>
Available on non-rust_v_1_60 or target_has_atomic="ptr only.

(Did not check other instances so there may be regressions elsewhere. I only checked this one and Bound (in script/borrowed.rs), and found that we forgot to mark the delegate_index! call with the correct attribute.)

[tcharding]

If this renders correctly today then I guess I'm fine with this.

I'm surprised you are not more enthusiastic for this change, 100 lines of red in the diff coupled with less maintenance and never having to remember to add the additional attribute - seems like a big win to me.

[tcharding]

Note to self, if this merges we need to change #1127 to remove the attribute from CONTRIBUTING.md.

[apoelstra]

Nice! I also suspected that it would improve the situation somehow, because surely we couldn't have gotten it all correct doing things by hand ... am glad to be proven correct.

[apoelstra]

ACK a189942

[Kixunil]

Cool, you've convinced me.

[Kixunil]

ACK a189942