feat(cli): Replace bundling with eszip in deno compile

[williamtetlow]

#13430

Replace bundling with eszip in deno compile.

Also adds support for dynamic imports in standalone executables #8655

[CLAassistant]

All committers have signed the CLA.

[bartlomieju]

@littledivy please take a look

[williamtetlow]

@littledivy please take a look

@littledivy @bartlomieju this isn't ready for review still very much WIP, haven't started impl. yet.

I was making draft PR but it auto added you as reviewer.

I will tag you once it's ready for review

[williamtetlow]

@littledivy @bartlomieju this is ready for review

[bartlomieju]

@williamtetlow looks good! Before landing we should ensure that stack traces behave correctly. Can you add an integration tests that compiles following script?

// a.ts
import "./b.ts";

// b.ts
console.log("hello");
throw new Error("boom!");

[bartlomieju]

This seems like unrelated change. I suggest to revert

[williamtetlow]

@bartlomieju this was done as both extract_standalone and standalone::run have been converted to async functions.

I wrapped the calls to these and also the call to the async fn get_subcommand in an async block so we call run_basic only once.

[littledivy]

JSON import assertions no longer work.

~/g/deno (williamtetlow/replace-bundling-with-eszip-for-deno-compile)> target/release/deno compile -A ./a.ts
Check file:///Users/divy/gh/deno/a.ts
Compile file:///Users/divy/gh/deno/a.ts
Emit a
~/g/deno (williamtetlow/replace-bundling-with-eszip-for-deno-compile)> ./a
error: Expected a "JSON" module but loaded a "JavaScript" module.

[williamtetlow]

I think this change will allow us to support dynamic imports in standalone executables too f2b55e8

[littledivy]

LGTM! Thanks @williamtetlow

@bartlomieju PTAL

[bartlomieju]

I'm still not happy about changes to cli/main.rs but I can live with them and refactor them later.

LGTM, nice work @williamtetlow

[vectronic]

@littledivy @williamtetlow

I'm curious as to what changed in this PR that meant you had to remove this test:

0ff2ce0

I assume you originally had this working due to this comment:

#13563 (comment)