And rename IsolateState to ThreadSafeState.

Also make ThreadSafeState directly implement Dispatch. This is simpler.

Reduces generics.

* Moves how snapshots are supplied to the Isolate. Previously they were
  given by Behavior::startup_data() but it was only called once at
  startup. It makes more sense (and simplifies Behavior) to pass it to the
  constructor of Isolate.
* Adds new libdeno type deno_snapshot instead of overloading
  deno_buf.
* Adds new libdeno method to delete snapshot deno_snapshot_delete().
* Renames deno_get_snapshot() to deno_snapshot_new().
* Makes StartupData hold references to snapshots. This was implicit when
  it previously held a deno_buf but is made explicit now. Note that
  include_bytes!() returns a &'static [u8] and we want to avoid
  copying that.

* Refactored the way worker polling is scheduled and errors are handled.
* Share the worker future as a Shared

Using open/read/write

* get_source_code_async
* fetch_module_meta_data_async
* fetch_module_meta_data_and_maybe_compile_async

To better distinguish the deno_core crate from the executable deno,
which will now be called "the cli" internally.

Re-enables arm64 CI test

Removing this reduces tail latency in test/http_bench.ts by half.

This disables a few tests which are broken still:
- tests/error_004_missing_module.test
- tests/error_005_missing_dynamic_import.test
- tests/error_006_import_ext_failure.test
- repl_test test_set_timeout
- repl_test test_async_op
- repl_test test_set_timeout_interlaced
- all of permission_prompt_test

This is in preperation for core integration.

Refactored permissions to be assignable on a per-isolate 
basis, and added a fix for #1858 to op_fetch_module_meta_data.

src/js_errors.rs takes care of source maps and color while
core/js_errors.rs is just the basic struct.

This patch contains a special hack that circumvents the current tokio
seek problem.

tokio `seek` is implemented to take ownership of the original File and
emit a new one in its future, which conflicts with the design of
ResourceTable.

To avoid the problem, the current hack makes the FsFile resource
an Option which we could `take` the value ownership out of it. We then
convert the tokio File into a Rust std File, perform the seek, and then
put it back into the resource.

This might be able to drop this hack after
https://github.com/tokio-rs/tokio/pull/785 lands.