…stdweb to wasm-bindgen

Only a couple of global functions as well as Room objects with some of their
functions are implemented so far

… room object and structure prototypes and positions

…eep this?

… for id and the least significant 8 store the padding length to support variable padding

…me doc updates

…ct id string, and some resolve wrappers

… files

…target environments and fmt

… returns

… observation.

…imits

…urrent versions of bindgen

…erCore

…lean up lib.rs

…4 chars

…d Game.symbols

* refactor(position): converted Position.x and Position.y to using u8 instead of u32

In other places, we use u8 for the in-room (x,y) positions. This isn't reflected in Position, where
instead u32 is used. This fixes that.

BREAKING CHANGE: Position only accepts and returns u8 instead of u32 for the in-room x and y.

* fix(position): fixed typos in casts and serde readable type

* refactor(local_cost_matrix): changed LocalCostMatrix to use a fixed-size array internally

Instead of using Vec<u8>, LocalCostMatrix now holds a [u8; 2500] directly. Also replaced the
Into<Vec<u8>> for LocalCostMatrix with From<LocalCostMatrix> for Vec<u8>, to be idiomatic.

* feat(local_cost_matrix): added indexing implementations for (u8, u8)

LocalCostMatrix's can now be indexed by (u8, u8) pairs.

* feat(local_cost_matrix): added unchecked get and set

* fix(local_cost_matrix): added asserts that x and y are valid coordinates

Before, we just relied on the bounds checking the index that pos_as_idx spits out, but that allows
invalid coordinates like (0, 201) to be passed in. This fixes that. As a result of the bounds
checking from the asserts, we can skip the checking that goes on in []::index, and just use the
unchecked versions instead.

* feat(local_cost_matrix): added indexing impls for Position

* refactor(local_cost_matrix): get and set now use indexing, indexing uses unchecked internally

With the (u8, u8) indexing impl, we can rewrite get and set in terms of them. And in the indexing
impl, because we check that x and y are in bounds, we're guaranteed that the index from pos_as_idx
will be in bounds as well, so we can just use the unchecked accessors instead.

…Is. (#349)

* Cleanup duplicated traits across structures.
* Move to using serde-wasm-bindgen instead of built-in bindgen serialization that uses stringification + JSON parsing.
* Implement pathfinder functions back to parity including options callbacks.
* Implement typed find methods. (Still missing 'Terrain' find.)
* Add strongly typed wrapper containers for 'Array' and 'HashMap' JS objects.
* Add 'as_has_store' back to structure enum type. (Other traits need adding still!)
* Add missing structure properties.
* Remove old stdweb code for functionality that is now implemented.
* Convert RoomName conversion that can fail to TryFrom. Use 'expect' in places it should never fail to unwrap.

* refactor(local_cost_matrix): changed LocalCostMatrix to use a fixed-size array internally

Instead of using Vec<u8>, LocalCostMatrix now holds a [u8; 2500] directly. Also replaced the
Into<Vec<u8>> for LocalCostMatrix with From<LocalCostMatrix> for Vec<u8>, to be idiomatic.

* feat(local_cost_matrix): added indexing implementations for (u8, u8)

LocalCostMatrix's can now be indexed by (u8, u8) pairs.

* feat(local_cost_matrix): added unchecked get and set

* fix(local_cost_matrix): added asserts that x and y are valid coordinates

Before, we just relied on the bounds checking the index that pos_as_idx spits out, but that allows
invalid coordinates like (0, 201) to be passed in. This fixes that. As a result of the bounds
checking from the asserts, we can skip the checking that goes on in []::index, and just use the
unchecked versions instead.

* feat(local_cost_matrix): added indexing impls for Position

* refactor(local_cost_matrix): get and set now use indexing, indexing uses unchecked internally

With the (u8, u8) indexing impl, we can rewrite get and set in terms of them. And in the indexing
impl, because we check that x and y are in bounds, we're guaranteed that the index from pos_as_idx
will be in bounds as well, so we can just use the unchecked accessors instead.

* feat(cost_matrix): added a SparseCostMatrix struct, added LocalCostMatrix::{iter, iter_mut}

SparseCostMatrix is the sparse counterpart to the LocalCostMatrix. Also added methods for generating
iterators over both the sparse and dense versions.

* feat(sparse_cost_matrix): added From<CostMatrix> implementation

* perf(sparse_cost_matrix): refactored assert in From<CostMatrix> to only run once

* refactor(local/cost_matrix.rs): refactored index -> (x,y) into a separate function

* feat(local/cost_matrix.rs): added methods for merging cost matrices

Can now merge the two cost matrix types into each other; forms the basis for possible layered cost
matrix behaviour, builder patterns, etc.

* refactor(cost_matrix_iter): rewrote the iter methods to give &u8 values back instead of u8

Some users might only care about a few values, why deref before it's needed? They can always chain
.copied() after if needed.

BREAKING CHANGE: .iter() returns impl Iterator<Item = ((u8, u8), &u8)> instead of impl Iterator<Item
= ((u8, u8), u8)> for Local and SparseCostMatrix.

* fix(local/cost_matrix.rs): fixed impls that broke due to iter output change

* refactor(local_cost_matrix): removed casts of Position::x and Position::y to u8

Now that the position change is merged, these casts are unnecessary.

* style(local/cost_matrix.rs): removed old commented-out code

* docs(sparse_cost_matrix): added comment explaining why 0 entries are filtered out

* style(local/cost_matrix.rs): removed more std-web code

…ment access. Add structure trait access.

…ndency

If we're going to be using wasm-pack test, we need this.

…olations

RoomCoordinate::unchecked_new is a non-const function, and so can't be used when creating
const/static arrays. So instead, provide helper methods that return vecs or arrays.

…es, re-add move_to_with_options

…types)

…lementation. (#354)

* Work in progress on parity API implementation.

* Lots of bindings fixes. Cost matrix pos fix. Better JsHashMap semantics.

* Fix doc test

Co-authored-by: Shane Madden <shanemadden@users.noreply.github.com>

…nning

* Work in progress on parity API implementation.

* Lots of bindings fixes. Cost matrix pos fix. Better JsHashMap semantics.

* Fix doc test

* Convert packed position to u32 from i32. Add flag + construction site container support. Fix type conversion for room status timestamp.

Co-authored-by: Shane Madden <shanemadden@users.noreply.github.com>

(needed for the changes in the starter I pushed a couple days ago for the starter but didn't commit)

…ed for creeps that are currently spawning) (#367)

Co-authored-by: Shane Madden <shanemadden@users.noreply.github.com>