Co-authored-by: Cheme
Co-authored-by: Brian Anderson <andersrb@gmail.com>
Co-authored-by: Bastian Köcher <bkchr@users.noreply.github.com>
Co-authored-by: David <dvdplm@gmail.com>

* Extend the benchmarks

* Optimize encoding/decoding of primitive type vectors

* Adds a benchmark for a complex type

* Optimize `u8` and `i8`

* Implement some more `size_hint` and improve array `encode_to`

* Fix the range...

* Update dependencies

Signed-off-by: koushiro <koushiro.cqx@gmail.com>

* Update

Signed-off-by: koushiro <koushiro.cqx@gmail.com>

* Revert some deps; Fix some warnings

Signed-off-by: koushiro <koushiro.cqx@gmail.com>

* Improve utils of derive

Signed-off-by: koushiro <koushiro.cqx@gmail.com>

* Bump bitvec to 0.15

Signed-off-by: koushiro <koushiro.cqx@gmail.com>

* Improve utils of derive

Signed-off-by: koushiro <koushiro.cqx@gmail.com>

* Rename to `parity-scale-codec` and revert version to 1.0

* Fix compilation after renaming

* Fix benchmarks

* Update src/lib.rs

Co-Authored-By: thiolliere <gui.thiolliere@gmail.com>

* Remove last parity-codec leftovers

Encode/Decode has now a new associated const that tells if the type is u8, that allow some kind of specialization when implementing Encode/Decode for `Vec<u8>`

* Introduce `EncodeAppend` and `CompactLen`

`CompactLen` is implemented by `Compact` to return the encoded length of
a given value.
`EncodeAppend` is a trait that can be implemented by types that support
appending without decoding all previously encoded items. Currently it is
implemented by `Vec<T>` and the benchmark shows a speed-up of factor
800x when appending items.

* Change `EncodeAppend` signature

* Apply suggestions from code review

Co-Authored-By: bkchr <bkchr@users.noreply.github.com>

* Use .extend from ArrayVec

* Bench Vec<u8> as Output

* Add size_hints and introduce WrapperType.

* Remove redundant Compact implementations.

* Add docs and more size_hints.

* Add some more detailed docs for read.

* Update src/codec.rs

Co-Authored-By: tomusdrw <tomusdrw@users.noreply.github.com>

* Address review grumbles.

* Make note a header.

* Bring back optimized impls for empty types, use faster Output to ArrayVec