perf: warm expensive init code at snapshot time

[mmastrac]

WIP -- slightly different approach to similar changes in #22386

Note that this doesn't use a warmup script -- we are actually just doing more work at snapshot time.

[littledivy]

I like this approach more. bootstrap code ∝ startup time

[bartlomieju]

Everything looks okay to me, but this completely breaks cargo build --features __runtime_js_sources.

When I try to run a file:

console.log(Deno);

I get this:

$ cargo build --features __runtime_js_sources
$ ./target/debug/deno run foo.js

============================================================
Deno has panicked. This is a bug in Deno. Please report this
at https://github.com/denoland/deno/issues/new.
If you can reliably reproduce this panic, include the
reproduction steps and re-run with the RUST_BACKTRACE=1 env
var set and include the backtrace in your report.

Platform: macos aarch64
Version: 1.41.2
Args: ["./target/debug/deno", "run", "foo.js"]

thread 'main' panicked at /Users/ib/.cargo/registry/src/index.crates.io-6f17d22bba15001f/deno_core-0.269.0/runtime/jsruntime.rs:603:9:
Failed to initialize a JsRuntime: TypeError: Cannot read properties of undefined (reading 'length')
    at Object.internals.__bootstrapNodeProcess (node:process:888:30)
    at initialize (ext:deno_node/02_init.js:30:15)
    at ext:runtime_main/js/99_main.js:917:1
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

Since we're working on a lot of Node compat APIs at the moment, we should preserve __runtime_js_sources working.

[mmastrac]

@bartlomieju Confirmed that __runtime_js_sources continues to work with this patch.

[littledivy]

Awesome! Excited to see its impact

invalid