* Add serialize and deserialize features with examples

* Fix formatting

* Make some types (de)serialize transparently

* Fix more formatting

* Remove new examples and add (de)serialization support to the convert example

* Remove extraneous file

* Rename param file

* Initial backend implementation

* Refractored glsl450 backend to have a common module between further glsl backends
Implemented more missing functionality

* Error handling

* Implemented most of the suggestions

* Addresed all comments
Fixed some bugs

* Made code style consistent

- Fix clippy issues

Focus on handling global in/out vars

Add mat2x3, mat2x4, mat3x2, mat3x4, mat4x2, and mat4x3 keywords

- Initial work for handling expressions that also generates statements

* [glsl-new] Add initial handling of stmts and exprs

* [glsl-new] use &'static str for NotImplemented msg

* [glsl-new] Don't lazy create function context

* [msl] Cast single-component scalars

* [msl] Allow `BinaryOperator::And`

* [glsl-new] handle gl_Position builtin

- Also fix width of vec4 type

* [glsl-new] add shader stage as arg to program ctor

* Fix TokenMetadata for #version

* Add glsl-new #version parsing

* Fix clippy issues

* glsl-new only core profile, allow 440-460

* Cleanup glsl-new profile matching

Adds constant evaluation for addition, and uses it for globals and array
specifiers.

* Add inital ast for glsl-new

* Fix clippy for glsl-new

Deref instead of cloning

Previously the lexer would accept any alphanumeric character as part of an
identifier. The spec only allowes identifiers to consist of ASCII alphanumeric
characters though.

* Add multi line comment support to glsl-new

* Fix glsl-new TokenMetadata for multi line comments

Fix all the usage of Bytes type to really mean bytes.
Have logic in the SPIR-V front-end to detect the comparison properties of samplers and images.

- Includes test

Until now the WGSL parser would interpret a character index as a byte index.
This could lead to a panic on invalid input strings like "\"\u{2}ПЀ\u{0}\"",
because it would use that index to slice a string without ensuring the slicing
happens on a character boundary.

One possible fix would have been to call `str::find` instead of `position`,
however by relying on `splitn` instead of slicing a str manually it is
easier to convince ourselves that this code can no longer panic.

Fixes https://github.com/gfx-rs/naga/issues/90

- Require 4.1 as 4.0 now gives compile errors

* Add initial pomelo glsl work

* Fix ref to glsl_old

* Fix idents with numbers in lexer

* Use glsl_new for .vert in convert example

Controlled by PREFER_GLSL_NEW=1 env var

Also add simple test files

* Start making parser look like spec

* Port 'old' lex.rs to glsl_new

* Apply fixes after rebasing

* Fix clippy issue

* Fix PR comments

- Make into optional feature glsl-new

- Minor code style improvements