perf(codegen): speed up printing StringLiterals

[overlookmotel]

Speed up printing StringLiterals by:

• Iterating over the string byte-by-byte, not char-by-char.
• Not using a peekable iterator.
• Printing string in chunks.
• Using a lookup table for bytes that need escaping.
• Minimizing the amount of code in the hot loop.
• Calculating the quote character to use lazily.
• Faster calculation of quote character when backtick cannot be used as quote.

In common case where string contains no bytes that need escaping, the string will be written to output buffer in one shot, and there's no need for an extra pass through the string's bytes to decide on the quote character to use.

[overlookmotel]

• perf(ast/estree): faster checking if bytes are ASCII #10183
• fix(parser): store lone surrogates in TemplateElementValue as escape sequence #10182
• fix(parser): fix parsing lone surrogates in StringLiterals #10180
• perf(codegen): speed up printing StringLiterals #10046 👈 (View in Graphite)
• test(parser): tests for lone surrogates and lossy escape characters #10175
• main

---

How to use the Graphite Merge Queue

Add either label to this PR to merge it via the merge queue:

• 0-merge - adds this PR to the back of the merge queue
• hotfix - for urgent hot fixes, skip the queue and merge this PR next

You must have a Graphite account in order to use the merge queue. Sign up using this link.

An organization admin has enabled the Graphite Merge Queue in this repository.

Please do not merge from GitHub as this will restart CI on PRs being processed by the merge queue.

This stack of pull requests is managed by Graphite. Learn more about stacking.

[codspeed-hq]

CodSpeed Instrumentation Performance Report

Merging #10046 will create unknown performance changes

_{Comparing 03-26-perf_codegen_speed_up_printing_stringliteral_s (775abac) with main (7672620)}

Summary

🆕 36 new benchmarks

Benchmarks breakdown

	Benchmark	BASE	HEAD	Change
🆕	codegen[checker.ts]	N/A	22.5 ms	N/A
🆕	codegen_sourcemap[checker.ts]	N/A	65.5 ms	N/A
🆕	formatter[antd.js]	N/A	7.6 ms	N/A
🆕	formatter[react.development.js]	N/A	41.4 µs	N/A
🆕	formatter[typescript.js]	N/A	7.6 ms	N/A
🆕	isolated-declarations[vue-id.ts]	N/A	58.3 ms	N/A
🆕	lexer[RadixUIAdoptionSection.jsx]	N/A	21.3 µs	N/A
🆕	lexer[antd.js]	N/A	24.8 ms	N/A
🆕	lexer[cal.com.tsx]	N/A	5.9 ms	N/A
🆕	lexer[checker.ts]	N/A	14.8 ms	N/A
🆕	lexer[pdf.mjs]	N/A	3.9 ms	N/A
🆕	linter[RadixUIAdoptionSection.jsx]	N/A	2.7 ms	N/A
🆕	linter[cal.com.tsx]	N/A	1.2 s	N/A
🆕	linter[checker.ts]	N/A	2.9 s	N/A
🆕	mangler[antd.js]	N/A	16 ms	N/A
🆕	mangler[react.development.js]	N/A	294 µs	N/A
🆕	mangler[typescript.js]	N/A	39.7 ms	N/A
🆕	minifier[antd.js]	N/A	166.2 ms	N/A
🆕	minifier[react.development.js]	N/A	1.9 ms	N/A
🆕	minifier[typescript.js]	N/A	293.1 ms	N/A
...	...	...	...	...

ℹ️ Only the first 20 benchmarks are displayed. Go to the app to view all benchmarks.

[Boshen]

I don't want you to go down the rabbit hole ... but we can probably avoid looping through the string twice by storing the position of the "unknown" values, and change them at the end of the loop.

[Boshen]

Merge activity

• Mar 28, 12:29 AM EDT: The merge label '0-merge' was detected. This PR will be added to the Graphite merge queue once it meets the requirements.
• Mar 28, 9:16 AM EDT: The merge label '0-merge' was detected. This PR will be added to the Graphite merge queue once it meets the requirements.
• Mar 28, 9:18 AM EDT: The merge label '0-merge' was detected. This PR will be added to the Graphite merge queue once it meets the requirements.
• Apr 2, 8:57 AM EDT: Boshen added this pull request to the Graphite merge queue.
• Apr 2, 9:11 AM EDT: Merged by the Graphite merge queue.

[overlookmotel]

As a test, I tried the shortcut of just printing raw - the fastest possible way to print a StringLiteral, as it involves no processing at all. That only gave a 0.15% perf improvement vs this PR.

So basically, I think there's very little gain in trying to optimize it any further. We might get 0.1% perf improvement at best. It was probably a waste of time even going as far as I have in this PR!

[overlookmotel]

This is finally ready to merge.

[overlookmotel]

@Boshen I assume the effect on formatter benchmarks is just noise?