feat: support resume for mainframe ZIP downloads

[asimtahir-karakus-kd]

Large mainframe ZIP downloads had no resume capability — any interruption required a full re-download from scratch.

Changes

lib/commands/serve/streempublic.js

• Partial file detection — checks fs.existsSync / fs.statSync on the destination before starting; captures existingBytes
• Range request — sends Range: bytes=N- header when existingBytes > 0; uses plain http.get otherwise
• Append mode — opens write stream with flags: 'a' when resuming so existing data is preserved
• HTTP 206 — confirms server support; seeds downloadedBytes with resumeFrom so progress bar reflects true overall completion from the first chunk
• HTTP 200 fallback — if the server ignores the Range header, warns and restarts a clean download after deleting the partial file
• HTTP 416 — treats Range Not Satisfiable as "already complete" and resolves immediately
• Resume message — prints Resuming MainFrame Xpublic Zip download from byte N... when continuing an interrupted download
• fs.unlink callbacks — all unlink calls now include error-handling callbacks for consistency

// Partial file detected → sends Range header, appends to file
const makeRequest = (callback) => resumeFrom > 0
    ? http.get(url, { headers: { 'Range': `bytes=${resumeFrom}-` } }, callback)
    : http.get(url, callback);

// Progress bar accounts for already-downloaded bytes
const totalBytes = isResuming ? resumeFrom + responseBytes : responseBytes;
let downloadedBytes = isResuming ? resumeFrom : 0;