fs: use sized reads for large files in readFileUtf8 - #65328

fs.readFileSync(path, 'utf8') gets +19…42 % on 4 MiB files and +60…231 % on 256 KiB files, with small files untouched, by
reading large files into one right-sized buffer instead of 8 KiB at a time.

fs/readfile-utf8-fastpath.js source='path' content='ascii'      size=262144    ***   231.14 %  ±2.91%
fs/readfile-utf8-fastpath.js source='path' content='utf8_mixed' size=262144    ***    60.16 %  ±0.93%
fs/readfile-utf8-fastpath.js source='path' content='ascii'      size=4194304   ***    41.86 %  ±5.85%
fs/readfile-utf8-fastpath.js source='path' content='utf8_mixed' size=4194304   ***    19.06 %  ±2.92%
fs/readfile-utf8-fastpath.js source='fd'   *                    size=4194304   ***   +22…27 %
fs/readfile-utf8-fastpath.js *             *                    size≤16384            0…+2 %
fs/readFileSync.js (1 KiB)                                                              ~0 %  n.s.

(Linux x64, 30 runs. A 300 KB file goes from 39 syscalls to 5.)

ReadFileUtf8 - behind readFileSync(…, 'utf8') and therefore every CommonJS module load - read in 8 KiB read()s,
appending each to a std::string. Files that fit the first 8 KiB read behave exactly as before. Once a read fills the stack
buffer, the rest goes straight into one heap buffer: a 64 KiB step first (so files up to 64 KiB take the same reads as today
and no fstat), then sized from fstat() if that fills too, with geometric growth as the fallback. The size is only an
allocation hint - reading continues until read() returns 0, so procfs/sysfs and concurrently changing files behave as
before and StringBytes::Encode() sees exactly the bytes read. Allocation failure throws ERR_MEMORY_ALLOCATION_FAILED.

Tests: test-fs-readfilesync-utf8-sizes.js (new): equals readFileSync(path).toString('utf8') around every internal
boundary (0 … 8 MiB+5) with multi-byte characters straddling chunk edges; by fd from start / at EOF / mid-file; binary input;
/proc/self/maps and friends; EISDIR. fs, module and require suites pass.

Disclosure: the code, test, measurements and this description were written by Claude Code, directed and reviewed by @codebytere.