refactor: remove return value from v8 throw

[ValoChet]

When throwing a v8 exception, there is no need to set the return value as v8 ignores it.

This PR removes args.GetReturnValue().Set() from throwing paths.
And uses NewFromUtf8Literal() instead of NewFromUtf8().ToLocalChecked() for static error messages.

It also updates Callback and NativeString to throw directly inside the constructor instead of in isInvalid().
And removes the args argument from isInvalid(), making it a simple getter.

This PR only makes the code cleaner, with no performance changes.

[github-actions]

Benchmark: 9fbe959 against 6aa8a01

Scenario	Base req/s	Head req/s	Head / base	Rounds	Noise	Busy	CPU per request
http/hello-world	116408	117460	1.008x (+0.8%)	1.01 to 1.01	±2.7%	100% / 100%	8.58 / 8.51 us (0.992x)
http/headers	104064	103860	0.998x (-0.2%)	0.99 to 1.01	±2.5%	100% / 100%	9.60 / 9.62 us (1.002x)
http/json-post 🏆	86440	88335	1.024x (+2.4%)	1.01 to 1.03	±1.7%	100% / 100%	11.55 / 11.31 us (0.977x)
http/cached	128637	128864	1.002x (+0.2%)	1.00 to 1.01	±0.7%	100% / 100%	7.77 / 7.75 us (0.998x)
ws/echo-20b	118858	118227	0.996x (-0.4%)	0.99 to 1.00	±1.8%	100% / 100%	8.41 / 8.45 us (1.004x)
ws/echo-4kb	104449	104625	1.002x (+0.2%)	1.00 to 1.01	±1.0%	100% / 100%	9.56 / 9.55 us (0.998x)

4 rounds of 3s per scenario, each round loading base, head and a second process of each one, from a copy of its build, after the other, in an order that changes every round. "Head / base" is the median of the per-round ratios of req/s and "Rounds" their range. "Noise" is how far base/base and head/head, the same code on both sides, got from 1 in this same run: that is what the machine did, so a row is marked only when the median moved further than that, at least 2%, and every round moved the same way: 👀 slower, 🏆 faster. "Busy" is the cpu time of the server process over the round and "CPU per request" that time divided by the requests it answered, with the head/base ratio judged against its own same-code band and bold when it cleared it. When busy is well under 100% the load generator set the pace, the req/s are its and the cpu per request is the column to read. Only the ratios are comparable across runs, the absolute figures depend on the runner.

Node v26.12.0, AMD EPYC 7763 64-Core Processor, 4 cores (server on cpu 0, load on 2,3), http_load_test and load_test with 100 connections.

[ValoChet]

I see some other areas we could refactor to clean up the codebase. Let me know if you're open to that.

[uNetworkingAB]

It doesn't simplify anything. What simplifies is that you have (apart from your description) also moved the place of exception throwing from within isInvalid, to instead inside the constructor. That's not due to skipping args.GetReturnValue().Set(, that's due to you changing the interface.

Your PR should really read: "I changed the interface of all Callbacks and NativeStrings to throw inside the constructor instead of in the getter isInvalid".

None of that is related to V8 ThrowException. I don't like this disconnect between what you say you do, and what you actually do. Your PR should be a simple 1-to-1 removal of the args.GetReturnValue().Set( if your PR description is like that. But you also changed a bunch of other things like NewFromUtf8Literal. I'm not saying this broke anything, but you don't really describe your PRs thruthfully.

Also, I don't like it because it makes no difference. Not setting a return value in the throw path has negligible/irrelevant perf. impact since it already is the very slow path. And It's charming to have legacy V8 usage where that is still backwards compatible.

[ValoChet]

You are right, my PR message was not very descriptive of the actual changes.
I am not very good at writing descriptions, I have updated it.

This PR does not give any performance gain (it may help the compiler).
It only simplify/modernise the code base, but it is still important.
Feel free to close it if you prefer to keep the current implementation.

[uNetworkingAB]

but it is still important.

It's not.