Implementation Notes

Float16 Conversion

The TRX specification allows streamline positions to use the float16 datatype, which is not native to JavaScript. This code converts float16 values to float32 using a lookup table for performance.

Endianness

The TRX specification requires little-endian byte order. The current code only supports little-endian systems, which covers all modern Android, iOS, macOS, Linux, and Windows devices.

Decompression

• gzip / zip: Uses fflate, which is significantly faster than the pako and jszip alternatives.

• zstd: Uses fzstd for zstd-compressed TRK files.

64-bit Integer Limitation

JavaScript does not natively support 64-bit integers. For uint64 and int64 data in TRX files, only the lower 32 bits are read. A warning is raised if any value exceeds the 32-bit range.