Tony Arcieri

This crate generally uses the `*Ref` naming convention for references, leaving `*String` for the owned types, ala `std::string::String`. `StrSlice` is used all over the place in an `AsRef` capacity, so...

I really think the types which represent encodings which aren't subsets of ASCII/UTF-8 should be distinct from each other. The `CHOICE` types that operate over them can map to enums,...

`Cow` abstracts over the owned and borrowed forms, so the encodings which are subsets of ASCII/UTF-8 can be represented using a simple reference conversion, and UTF-16 transcoded. --- Separately I'm...

I opened a proposal for generic string support in #1061. For `BMPString` to work with a type which is `String` internally we'd have to transcode UCS-2(?) to UTF-8 when decoding...

I've added `BmpString` which is the most pressing. Let's track the others in #1061.

This is an extremely immaturely phrased issue. Please act like an adult when reporting issues to this project. Knock off the caps lock. > My .exe file is too large...

I'll also note that on Linux/Mac, the binaries produced by the test cases in this repo are around 1MB-1.5MB, even with debug symbols. This includes particularly complicated cases like `x509-cert`....

We can look into improving `NestedReader` in the next breaking release to reduce the number of combinations that are monomorphized, however your original reported issue was it bloating your binary...

Regarding this suggestion: > i'd suggest using always a wrapped NestedReader with a stack: > > ```rust > stack: Vec > ``` The core functionality of `der` (including `read_nested`/`NestedReader`) supports...

I think the best way to address this would be to get rid of `NestedReader` and make handling nested length boundaries an internal concern of each of the reader types,...