impl `MulAssign` (and friends) for `VecN<T>` and `MatNxMf`

The [WGSL spec](https://www.w3.org/TR/2026/CRD-WGSL-20260825/#compound-assignment) allows for "compound assignments", this lets you do things like:

```wgsl
var v = vec2f(0.,0.);
v *= 2.;
var m = mat2x2f(v,v);
m *= 2.;
```

However this is currently not supported in wgsl-rs:

```
error[E0368]: binary assignment operation `*=` cannot be applied to type `wgsl_rs::std::Vec2`
--> crates/app/src/my_shader.rs:11:9
|
11 | v *= 2.;
| -^^^^^^
| |
| cannot use `*=` on type `wgsl_rs::std::Vec2`
|
note: `wgsl_rs::std::Vec2` does not implement `std::ops::MulAssign<{float}>`
--> vendor/wgsl-rs/crates/wgsl-rs/src/std/vector.rs:17:1

error[E0368]: binary assignment operation `*=` cannot be applied to type `Mat2x2f`
--> crates/app/src/my_shader.rs:13:9
|
13 | m *= 2.;
| -^^^^^^
| |
| cannot use `*=` on type `Mat2x2f`
|
note: `Mat2x2f` does not implement `std::ops::MulAssign<{float}>
```

Start with vendor/wgsl-rs/crates/wgsl-rs/src/std/vector.rs and locate the corresponding MatNxMf definitions. Use crates/app/src/my_shader.rs as the reproduction entry point, then verify that the requested compound assignments compile for the vector and matrix types covered by the issue.

Written by the indexing model from the issue text.