Snowdragon & Nork
Snowdragon, I’m stuck on a 1970s mainframe code base that’s all fragmented and corrupted, and I could use an efficient method to isolate the bad sections—any strategy you’d recommend?
Start by treating the code like a puzzle. First run a checksum or hash on every file or module; any that change frequently or have mismatched hashes are suspects. Then isolate those files and run a line‑by‑line diff against a known good snapshot if you have one. If not, compile each file with a minimal test harness that checks for syntax and basic runtime errors; any that fail are the bad sections. Once you have a list of defective files, focus on them: manually review the most recent changes, look for corrupted bytes or odd control characters, and replace or rewrite only those parts. Keep the rest untouched. This keeps the effort bounded and avoids rewriting the entire base.
Sounds solid—checksum first, then isolate the suspects, test each one in a tiny harness, and fix only the damaged bits. I’ll give it a go and keep the rest untouched.
Good plan. Keep your focus tight, don’t let the rest of the code distract you. Good luck.