dishmaker

Need to benchmark this and rebase

I don't have an example. This just seems to be reasonable, where PKCS#15 nests objects recursively ([example](https://github.com/monai/node-passport/blob/master/lib/pkcs15/cryptographic_information_framework.asn1#L234-L242)) Actually, this newtype can be in a separate crate.

> but Reader has an impl cardinality of 2 (SliceReader and PemReader) > What you're describing doesn't sound like you actually want to leverage an associated error type? Why would...

Okay, I was too focused on a single solution and need to think that through.

`der::Error` trait with optional functions handling (field|struct|choice) names and (sequence|set) index would be great.

28% slower `SubjectPublicKeyInfoRef` due to indefinite flag in Length Commit b288a6186531203c97ecda5791adb1f70f086993 (after) vs parent cce5ea4a20790c11cdddbb81be8090796eb91166 (before) Generated using `cargo bench -p spki -F alloc` #1947 | |before [ns/iter]|after [ns/iter]| |----------------|----------------|---------------|...

Current master 7674d0c6619bb2dab08fee3a531f1664daddbb44 is also 24% slower than cce5ea4a20790c11cdddbb81be8090796eb91166

Here is a comparison of https://github.com/RustCrypto/formats/pull/1726 500cd652e66dd8191776ccf9720bd75a2bb5b716 vs parent 0b3a0b2e00741e54e81ae0b742cd5fcacd640d2e Looks like indefinite `u32::MAX` would be about 1% slower (similar to `Length::MAX`) | |After / no limit [ns/iter]|Before / with...

I removed the dependency on `tynm`

Yes, the `Tag` alongside `&'a [u8]` is a problem. One parallel solution is: ```rust pub struct AnyCow, } ``` but that wouldn't be `alloc`-gated on allocating variant.