Yeah, it’s just a side project for now, but I’m hoping to make it more solid over time. As for C, I wanted to challenge myself and step away from what I usually do — try something a bit different.

It records escape sequences verbatim during capture, then handles them intelligently during replay.

Recording phase: All terminal output including ANSI escape codes, color sequences, and cursor movements are captured exactly as they appear - no processing or stripping occurs.

Replay phase: - Decodes various escape formats (\u001b, \033, \x1b) back to actual escape characters - Filters out problematic terminal query sequences that could cause artifacts - Preserves visual escape sequences (colors, cursor positioning) for faithful reproduction

So yes, escape sequences are recorded verbatim, but the replayer ntelligently processes them to recreate the original terminal experience while avoiding terminal corruption.

That sounds like a good strategy! I've dabbled in writing task runner, and relaying logs with preserved colors and formatting without messing up the terminal and interleaving messages from different tasks is a huge hurdle.

I wish there was a standard for telling processes "keep the colored and formatted output, but assume it will be read line by line"... It's possible to just let processes write into pty's and then parse the output, but then you pretty much have to implement an entire nested terminal emulator :-(

By the way, debba: execvp()/execlp() and _PATH_BSHELL out of <paths.h> as the fallback are the ways to go, here.