Optimise memory view getters

[Liamolucko]

While working on RReverser/serde-wasm-bindgen#34, I noticed that a significant amount of time was spent getting memory views. Most of that time was spent in the getter for WebAssembly.Memory.buffer, so I've changed the memory view getters to check for invalidation without comparing to the wasm memory, by checking if the length is 0 (which indicates the buffer being detached).

wasm-bindgen benchmarks:

Benchmark	Before	After	Change
Access Node.nodeType with final	364/s	383/s	+5.22%
Access Node.nodeType with structural	1,166/s	1,189/s	+1.97%
Call a custom JS class Foo.bar method with final	6,248/s	6,189/s	-0.94%
Call a custom JS class Foo.bar method with structural	6,642/s	6,667/s	+0.38%
Pass ascii_small to/from wasm-bindgen	1,281,297/s	1,823,742/s	+43.04%
Pass ascii_medium to/from wasm-bindgen	1,231,065/s	1,723,498/s	+40.00%
Pass ascii_number to/from wasm-bindgen	1,161,427/s	1,591,248/s	+37.01%
Pass ascii_date to/from wasm-bindgen	918,950/s	1,138,144/s	+23.85%
Pass ascii_url to/from wasm-bindgen	899,162/s	1,111,610/s	+23.63%
Pass ascii_link to/from wasm-bindgen	864,511/s	1,105,580/s	+27.89%
Pass unicode to/from wasm-bindgen	351,071/s	390,481/s	+11.23%

serde-wasm-bindgen benchmarks (not including RReverser/serde-wasm-bindgen#34):

Benchmark	Before	After	Change
Parse canada	44.01 ms/iter	34.52 ms/iter	-21.56%
Parse citm_catalog	10.32 ms/iter	8.78 ms/iter	-14.92%
Parse twitter	8.99 ms/iter	8.23 ms/iter	-8.45%
Serialize canada	11.74 ms/iter	11.59 ms/iter	-1.28%
Serialize citm_catalog	7.51 ms/iter	7.51 ms/iter	-0.00%
Serialize twitter	14.54 ms/iter	14.07 ms/iter	-3.23%

[RReverser]

Great work!

[alexcrichton]

Nice wins, and thanks!