Remove some rerun-if-env-changed lines from build.rs.

[sunfishcode]

Now that we appear to have a known-working fix for #526, we can investigate relaxing the fix. If #526 reappears, we now have a known-working state we can quickly revert to.

Reports in #562 and #575 are that having rerun-if-env-changed=RUSTC and rerun-if-env-changed=RUSTC_WRAPPER are triggering spurious rebuilds in some situations, so remove these. In theory, cargo should already be aware of these environment variables. In practice, there were reports of build failures which had the appearance of using an older version of rustc with build.rs output from a newer version of rustc, and it wasn't clear what was happening. But, it's possible that #563 fixes the issue properly. So let's try removing these rerun-if-env-changed lines, and see if the problem resurfaces.

And, rerun-if-env-changed=TARGET is explicitly documented as being unnecessary here.

[OvermindDL1]

I can reliably test both states, when it is too strict and when it is not strict enough. If you give me a branch or checkout to test then I can change my deps to that and test it out?

[sunfishcode]

@OvermindDL1 I'd much appreciate it. Would you be able to test the "sunfishcode/build-scripts" branch, which is what's in this PR? https://github.com/bytecodealliance/rustix/tree/sunfishcode/build-scripts

[OvermindDL1]

Patched it in via:

[patch.crates-io]
rustix = { git = "https://github.com/bytecodealliance/rustix.git", branch = "sunfishcode/build-scripts" }

Running cargo clean, then cargo build in cli, then cargo build in intellij, success, then cargo build in cli again, success, then cargo clean, then cargo build in intellij, then cargo build in cli, success, then cargo build in intellij, success.

I'm not experiencing either the old or the new issue with it, it's looking good for me at least. :-)

[OvermindDL1]

Also went through my normal workflow (cargo-watch test in terminal on one screen, compiling and running in intellij on the other, while editing code of course), it's much faster now!

[OvermindDL1]

And a quick test with rust-analyzer via emacs and I can't reproduce it there either.

[OvermindDL1]

Note, I tested this with latest stable rust, 1.68.2, I cannot comment about older versions. If you want me to test with a specific older version then I can install it and test?

[sunfishcode]

Thanks for testing this! In #526 the problem was reported in Rust 1.65, so if you could install and test Rust 1.65, that'd be a useful data point.

[OvermindDL1]

Sorry for the delay, just got back, installing 1.65 now, setting up intellij to 1.65, performing the same tests, same tests pass in intllij, testing emacs/rust-analyzer, same tests pass again, cargo watch test, also still showing good, confirmed cargo version is 1.65 reporting in all locations.

Anything else you want tested while I have all this still prepared?

[sunfishcode]

This is really great, thanks! I can't think of anything else that urgently needs testing here. It seems very likely that #563 fixed the root issue here. So I'll plan to merge this PR soon and do a point release, and then keep an eye out to see if any problems surface.