Update to rules_js 1.6.4 to pick up fs patches performance fixes

[gregmagolan]

//:thing_test                                                            PASSED in 5.1s

to

//:thing_test                                                            PASSED in 1.6s

Also updates to rules_ts rc5 and Bazel 6.0.0rc1

[gregmagolan]

Some more metrics here,

Pure jest outside of Bazel 1.0s (with the .ts files manually pre-transpiled to js):

$ time pnpm jest
Browserslist: caniuse-lite is outdated. Please run:
  npx browserslist@latest --update-db
  Why you should do it regularly: https:/browserslist/browserslist#browsers-data-updating
  console.log
    [ 2, 3, 4 ]

      at log (thing.js:10:11)

 PASS  ./thing.spec.js
  ✓ a (16 ms)

Test Suites: 1 passed, 1 total
Tests:       1 passed, 1 total
Snapshots:   0 total
Time:        0.281 s, estimated 1 s
Ran all test suites.

real	0m1.064s

With rules_js v.1.6.3 and fs-patches disabled: 1.5s

jest_test(
    name = "thing_test",
    config = ":jest_config",
    data = [
        ":thing_test_src",
    ],
    patch_node_fs = False,
)

With rules_js v.1.6.3 and fs-patches and no sandboxing: 1.3s

jest_test(
    name = "thing_test",
    config = ":jest_config",
    data = [
        ":thing_test_src",
    ],
    patch_node_fs = False,
    tags = ["local"],
)

[gregmagolan]

@jbedard So looks another 1s of overhead due to the fs patches which we need to workout. I figure there will be some overhead with these patches but we should minimize it as much as possible.

[gregmagolan]

@jbedard We're down to 1.6s now with Bazel 6.0.0rc1 and rules_js v1.6.4 with your last per fix included in that release.

//:thing_test                                                            PASSED in 1.6s

Still a bit slower than no patches which I measured at 1.2s

//:thing_test                                                            PASSED in 1.2s

With patches on and sandbox disabled I measured 1.4.s

//:thing_test                                                            PASSED in 1.4s

With patches off and no-sandbox I measured 1.1s

//:thing_test                                                            PASSED in 1.1s