Change SymbolName::name to a &str.

[nnethercote]

This eliminates a bunch of Symbol::intern() and Symbol::as_str()
calls, which is good, because they require locking the interner.

Note that the unsafety in from_cycle_error() is identical to the
unsafety on other adjacent impls.

r? @eddyb

[nnethercote]

This gives some very small perf wins. At least they're mostly on the real-world benchmarks. Here is a subset of results:

clap-rs-debug
        avg: -0.4%      min: -0.4%      max: -0.4%
webrender-debug
        avg: -0.4%      min: -0.4%      max: -0.4%
cranelift-codegen-debug
        avg: -0.4%      min: -0.4%      max: -0.4%
piston-image-debug
        avg: -0.3%      min: -0.3%      max: -0.3%
webrender-wrench-debug
        avg: -0.3%      min: -0.3%      max: -0.3%
hyper-2-debug
        avg: -0.3%      min: -0.3%      max: -0.3%
tokio-webpush-simple-debug
        avg: -0.2%      min: -0.2%      max: -0.2%
deep-vector-debug
        avg: -0.2%      min: -0.2%      max: -0.2%
cargo-debug
        avg: -0.2%      min: -0.2%      max: -0.2%
ripgrep-debug
        avg: -0.2%      min: -0.2%      max: -0.2%
regression-31157-debug
        avg: -0.2%      min: -0.2%      max: -0.2%
regex-debug
        avg: -0.2%      min: -0.2%      max: -0.2%
syn-debug
        avg: -0.1%      min: -0.1%      max: -0.1%
encoding-debug
        avg: -0.1%      min: -0.1%      max: -0.1%
issue-46449-debug
        avg: -0.1%      min: -0.1%      max: -0.1%
inflate-debug
        avg: -0.1%      min: -0.1%      max: -0.1% 
futures-debug
        avg: -0.1%      min: -0.1%      max: -0.1%
html5ever-debug
        avg: -0.1%      min: -0.1%      max: -0.1%

[nnethercote]

@bors try @rust-timer queue

[rust-timer]

Awaiting bors try build completion

[bors]

⌛ Trying commit 61e2a3949b118fd04592e73c8ca4ef4b805076a2 with merge 28b4962ee286e2682945dd2a9b3dec058d518ec7...

[bors]

☀️ Try build successful - checks-actions, checks-azure
Build commit: 28b4962ee286e2682945dd2a9b3dec058d518ec7 (28b4962ee286e2682945dd2a9b3dec058d518ec7)

[rust-timer]

Queued 28b4962ee286e2682945dd2a9b3dec058d518ec7 with parent e59b08e, future comparison URL.

[rust-timer]

Finished benchmarking try commit (28b4962ee286e2682945dd2a9b3dec058d518ec7): comparison url.

[nnethercote]

I only measured check-full and debug-full runs locally, and the best improvement was 0.4%. But this change helps incremental runs more, and when all runs are done (as on CI) we see improvements of up to 2.5%.

@bors rollup=never

[eddyb]

Why can't you write ty::SymbolName<'tcx>? That's why Value has a 'tcx parameter.

Oh this is a specialization issue. IMO it should be noted as such rather then left as a vague FIXME (cc @matthewjasper).

[nnethercote]

What exactly would you like me to write in the comment?

[eddyb]

I hope @matthewjasper can come up with a better comment. Otherwise it's probably fine as-is.
Although I kind of doubt this from_cycle_error implementation is needed at all.

[bors]

☔ The latest upstream changes (presumably #74175) made this pull request unmergeable. Please resolve the merge conflicts.

[eddyb]

@bors r+

[bors]

📌 Commit a1b8540 has been approved by eddyb

[bors]

⌛ Testing commit a1b8540 with merge d9e8d62...

[bors]

☀️ Test successful - checks-actions, checks-azure
Approved by: eddyb
Pushing d9e8d62 to master...

[nnethercote]

This was a perf win upon landing, as expected.