Update the example to showcase some new features

[aldanor]

@mulimoen Here's the updated example (I'll add it to the README as well) which showcases several new features (anything else? need to keep it as short as possible though)

Closes #156.

Interestingly enough, it currently fails (!) unless you disable blosc compression (in which case it runs fine), see below for examples.

This is definitely a bug somewhere and needs investigating.

  left: `[[Pixel { xy: (0, 0), color: B }, Pixel { xy: (0, 0), color: B }],
 [Pixel { xy: (0, 0), color: B }, Pixel { xy: (0, 0), color: B }]], shape=[2, 2], strides=[2, 1], layout=Cc (0x5), const ndim=2`,
 right: `[[Pixel { xy: (3, 4), color: G }, Pixel { xy: (4, 5), color: R }],
 [Pixel { xy: (5, 6), color: B }, Pixel { xy: (6, 7), color: G }]], shape=[2, 2], strides=[2, 1], layout=Cc (0x5), const ndim=2`', examples/simple.rs:42:9

And if you replace read_slice with read_2d:

  left: `[[Pixel { xy: (618879123714, -6917529027641081712), color: B }, Pixel { xy: (0, 0), color: B }],
 [Pixel { xy: (0, 0), color: B }, Pixel { xy: (0, 0), color: B }],
 [Pixel { xy: (0, 0), color: B }, Pixel { xy: (0, 0), color: B }]], shape=[3, 2], strides=[2, 1], layout=Cc (0x5), const ndim=2`,
 right: `[[Pixel { xy: (3, 4), color: G }, Pixel { xy: (4, 5), color: R }],
 [Pixel { xy: (5, 6), color: B }, Pixel { xy: (6, 7), color: G }]], shape=[2, 2], strides=[2, 1], layout=Cc (0x5), const ndim=2`', examples/simple.rs:42:9

Am I missing something obvious again? Hmm... 🤔

[aldanor]

Also, how to only run this example on CI with blosc enabled so it doesn't fail to compile? 🤔

[aldanor]

Few more observations: LZF filters work fine. SZip filters couldn't make work due to 'unknown library error', perhaps the dataset is too small or something.

[mulimoen]

It seems blosc can not compress the data on some algorithms and returns 0. This should cause hdf5 to fall back to no compression, but data written is garbled

[mulimoen]

Found and fixed the compression error, was a too eager comparison which overwrote the data buffer. For running in CI I inserted the feature flags, this might not be the easiest to read, but should compile on all platforms and still illustrate the principle

[mulimoen]

Build failure shows btime does not have to equal ctime for LocationInfo

[aldanor]

(While resolving an "add blosc_*() methods" todo, I couldn't help but refactor some impl_builder_stuff!() mess, it will help in the future iteration to work on consistent documentation etc)

Relaxed location info timestamp tests, let's see if it's green now.

I think the next release we should focus 99% on tests and documentation, it kind of got to the point where it's more important than adding/improving new features, at least in the state it's in now.