Regenerate email bindings with per-mono generics

[logan-gatlin]

Summary

• Enable experimental per-monomorphization generics for both generated Email binding copies
• Regenerate Email APIs with JsStringLike parameters and JsString getter variants
• Advance ts-gen to include the typed-array dictionary-helper ABI fix

Stack

• Depends on Add flagship bindings with per-monomorphization generics #1053, which adds the Flagship per-mono bindings

Validation

• cargo test -p worker --target wasm32-unknown-unknown
• cargo check -p worker-sys -p worker --target wasm32-unknown-unknown
• Email and Flagship Vitest suites pass; full suite passed 150 tests with one unrelated R2 timeout that passed on targeted rerun

[codspeed-hq]

Merging this PR will not alter performance

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

✅ 2 untouched benchmarks

---

_{Comparing generics-per-mono-everything (d3ac101) with main (78dc0cd)¹}

Footnotes

1. No successful run was found on flagship-per-mono-generics (8be4ab2) during the generation of this report, so main (78dc0cd) was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩

[logan-gatlin]

Oh I thought ts-gen was used for way more than just email and flagship, I should probably just fold this into the first PR

[guybedford]

Verified locally: chomp build:types is a clean no-op on this commit (deterministic, and the ts-gen bump leaves flagship.rs unchanged), send_email.spec.ts + flagship.spec.ts pass, and the glue collapses (&str, &str, String) into a single new EmailMessage(...) import as expected.

Requesting changes for two things:

1. worker-sys/src/types/email.rs is a dead file — it is never declared (worker-sys/src/types.rs has no mod email;), it was left behind by #996. Rather than adding a second ts-gen invocation to keep regenerating it, delete the file and drop the second command.

2. This is the first per-mono regeneration of an already-released binding, and it breaks &String callers. JsStringLike is sealed to String, &str, JsString, &JsString; with a concrete &str parameter &String deref-coerced, through a generic bound it does not:

error[E0277]: the trait bound `&std::string::String: JsStringLike` is not satisfied
  --> EmailMessage::new(s, s, s)   // s: &String

The example diffs in this PR (new(&to, &from, &raw) -> new(to, from, raw)) are exactly that break, and worker::EmailMessage::new, SendEmailBuilderBuilder::text(&body) etc. are public API. This needs a wasm-bindgen PR adding IntoWasmAbi / OptionIntoWasmAbi / JsStringLike for &String (delegating to &str) before we regenerate existing bindings with per-mono generics, otherwise every binding we convert will hit the same regression. This PR should then land on top of that submodule bump with the example changes reverted.

[guybedford]

worker-sys/src/types/email.rs isn't wired into worker-sys at all (no mod email; in worker-sys/src/types.rs) — it's an orphan from #996. Delete the file instead of adding this second invocation.

[guybedford]

This change is required because &String no longer satisfies the generic bound — which is a source break for downstream users of the released EmailMessage::new. See the review summary: this needs JsStringLike for &String upstream in wasm-bindgen first, after which this diff (and the send-email / test ones) can be reverted.

[guybedford]

Nit (ts-gen): type__js_string has a double underscore from type_ + _js_string. The keyword-escape underscore should be stripped before suffixing so this becomes type_js_string.