[p5.js 2.0+ Bug Report]: nfc() truncates instead of rounding and drops the documented trailing zero padding

Most appropriate sub-area of p5.js?

Utilities

p5.js version

2.x main (4b096e2)

Actual vs expected behavior

nfc()'s own reference examples document rounding and zero padding, but doNfc in src/utilities/utility_functions.js compares the requested decimal count against a substring that still contains the leading ., so it truncates where the docs round and skips the padding branch by one:

nfc(12345.67, 1); // actual '12,345.6'   documented '12,345.7'
nfc(12345.67, 3); // actual '12,345.67'  documented '12,345.670'

nf() already rounds via toFixed, so the two formatters also disagree with each other on the same input. A rounding carry also needs comma re-grouping, e.g. nfc(999.96, 1) should give '1,000.0'.

Steps to reproduce

Outputs above are from executed runs against current main via the unit test harness (npx vitest run test/unit/utilities/utility_functions.js with added cases).

Note

I have a fix ready (route through toFixed like nf, keep the comma grouping, preserve the existing nfc(32000, '3') === '32,000.000' behavior) with unit tests covering rounding, padding, the carry re-group and negatives. Verified the four new tests fail on current main and pass with the change, 23 existing tests unaffected. Filing for approval first per the contributing guide; will open the PR once approved.

Start in src/utilities/utility_functions.js at doNfc, then run test/unit/utilities/utility_functions.js with the reported cases. Confirm that nfc() matches the documented rounding and trailing-zero behavior, preserves string decimal input, handles carry regrouping and negatives, and leaves the existing tests passing.