buffer: add isLatin1

[jasnell]

Implements a fast check to determine if a string is Latin1 a byte string Latin1.

Can be many times faster than the equivalent regex check.

input	isByteString	regex	loop
short, one-byte	232M	122M	55M
short, two-byte	207M	130M	56M
short, invalid	106M	48M	53M
long, one-byte	221M	93M	230k
long, two-byte	6.7M	913k	186k
long, invalid	75M	914k	234k

(the larger perf gap there is with concatenated strings)

[nodejs-github-bot]

Review requested:

• @nodejs/performance

[aduh95]

I don't think isLatin1 is a good name, \u0100 can definitely be a 2-byte latin1 string

[jasnell]

Yeah, I'm going to rename to isByteString ... it matches the webidl usage better.

[nodejs-github-bot]

CI: https://ci.nodejs.org/job/node-test-pull-request/77917/ 💚

[codecov]

Codecov Report

❌ Patch coverage is 85.00000% with 3 lines in your changes missing coverage. Please review.
✅ Project coverage is 90.36%. Comparing base (cd908df) to head (67b095c).
⚠️ Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
src/node_buffer.cc	76.92%	0 Missing and 3 partials ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #66298      +/-   ##
==========================================
- Coverage   90.37%   90.36%   -0.02%     
==========================================
  Files         792      792              
  Lines      275324   275344      +20     
  Branches    52764    52760       -4     
==========================================
- Hits       248828   248805      -23     
- Misses      16918    16973      +55     
+ Partials     9578     9566      -12     

Files with missing lines	Coverage Δ
lib/buffer.js	99.74% <100.00%> (+<0.01%)	⬆️
src/node_buffer.cc	70.56% <76.92%> (+0.06%)	⬆️

... and 35 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[addaleax]

isLatin1String would be much better naming because it describes what the method actually tests for. The WebIDL ByteString type describes a legacy type that stems from the lack of commitment to a specific encoding in early HTTP versions other than vague ASCII backwards compatibility, and it's not actually used widely outside of HTTP-specific APIs for that (very good) reason.

I don't think isLatin1 is a good name, \u0100 can definitely be a 2-byte latin1 string

This is not a statement that makes sense, \u0100 cannot be represented in Latin-1, and there are, by definition, no 2-byte Latin-1 strings to begin with -- maybe you're referring to the idea that the byte sequence 01 00 can be Latin-1, but this method takes a string as an argument, i.e. a character sequence, not a byte sequence

[jasnell]

The challenge with isLatin1 is that there are two interpretations for latin1 (ISO/IEC 8859-1 or Windows-1252 and we're not really verifying against either.... may it should just be isOneByteString(...)?

[addaleax]

may it should just be isOneByteString(...)?

Without knowledge of the V8 OneByteString/TwoByteString terminology, I'd probably assume that this means one-byte when encoded as UTF-8 (i.e. ASCII).

The challenge with isLatin1 is that there are two interpretations for latin1 (ISO/IEC 8859-1 or Windows-1252 and we're not really verifying against either

That's a fair point, although in the context of Node.js APIs, latin1 already has the meaning that this PR uses, and that consistency combined with being relatively "close" to the actual mechanics of this method still makes me think it's the best choice for naming (that isn't excessively verbose).

[jasnell]

Fair... ok... renaming it back to isLatin1 :-)

And if @aduh95 disagrees we'll have you both play rock/paper/scissors at nodeconf to decide.

[panva]

can you add the missing typings here? rather than doing it in a follow up #66328, stringLengthUtf8 is missing too.

[jasnell]

Typings added. If you can give it another stamp I'd appreciate it.