cg_llvm: Perform target-machine command-line quoting as bytes, not characters

[Zalathar]

(Slightly modified version of the original experiment at #146804.)

This works because the only characters we care about escaping are ASCII characters, so forwarding non-ASCII UTF-8 bytes as-is still produces correct results.

A perf run at #146804 (comment) indicates that this has a big impact on several versions of the large-workspace benchmark.

---

This PR doesn't address the underlying problem, which is that we're doing this quoting far too many times overall. Really we shouldn't be doing it more than once per process, but in practice we do it multiple times per CGU, which is bad news for incremental builds.

But these changes are basically a free win, so I'm happy to merge them anyway.

[rustbot]

r? @petrochenkov

rustbot has assigned @petrochenkov.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[Zalathar]

I did make some tweaks since the original experiment (pre-computing a buffer size and re-validating the output), so let's do another perf run.

@bors try @rust-timer queue

[rust-bors]

☀️ Try build successful (CI)
Build commit: 341189b (341189b5f9cbc286c17fbfc1309105ccf8051986, parent: ae12bc21d8ec76bbb753d4da168e9b08e1b09ebf)

[rust-timer]

Finished benchmarking commit (341189b): comparison URL.

Overall result: ✅ improvements - no action needed

Benchmarking this pull request means it may be perf-sensitive – we'll automatically label it not fit for rolling up. You can override this, but we strongly advise not to, due to possible changes in compiler perf.

@bors rollup=never
@rustbot label: -S-waiting-on-perf -perf-regression

Instruction count

Our most reliable metric. Used to determine the overall result above. However, even this metric can be noisy.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	0.3%	[0.3%, 0.3%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-1.5%	[-8.8%, -0.2%]	10
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

Results (primary -1.2%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	2.0%	[2.0%, 2.0%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-2.3%	[-3.7%, -1.1%]	3
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-1.2%	[-3.7%, 2.0%]	4

Cycles

This benchmark run did not return any relevant results for this metric.

Binary size

Results (secondary -0.0%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.0%	[-0.0%, -0.0%]	1
All ❌✅ (primary)	-	-	0

Bootstrap: 470.741s -> 469.606s (-0.24%)
Artifact size: 387.87 MiB -> 387.89 MiB (0.00%)