base line project for "spec-driven" ai dev, included allot of yelling at ai to listen to the specs, trying and probaly failing at catching it deleting unit tests

A D implementation of Unicode Separated Values (USV) parser following the official IETF specification.

Unicode Separated Values (USV) uses Unicode separator characters to structure data:

• Unit Separator: U+001F (␟)
• Record Separator: U+001E (␞)
• Group Separator: U+001D (␝)
• File Separator: U+001C (␜)
• End of Transmission: U+0004 (␄) - supported as stop marker only

This project does NOT use DUB (per NO DUB directive). Build manually with:

dmd usv.d  # Compile the library

import usv;

// Write USV data
int[] data = [1, 2, 3];
writeusv(data, "output.usv");  // Write to file

// Read USV data
int[] readData = readusv!(int[])(data, "output.usv");  // Read from file

• writeusv(T, int N=-1,alias TO=void)(T data, string file): Write data to USV file
• T readusv(T, int N=-1,alias TO=void)(string file): Read USV file to data

• isvoid(alias A): Check if type is void
• dim(T): Determine dimension of a type
• isnotstructure(A): Check if type is not a structure

This project follows these core principles:

• NO DUB: No DUB build system - simple manual compilation
• KISS (Keep It Simple Stupid): Simple, straightforward implementations
• OBEY MY API: Strict adherence to defined interfaces

Based on the official IETF draft specification for Unicode Separated Values.

Run the unit tests with:

dmd -unittest -main -run src/usv.d