ireneista@irenes.space ("Irenes (many)") wrote:
okay. so first and most obvious, there are error messages from the hex transform's internals, interspersed with the actual hex dump. we had made the call (documented in comments) that this particular failure mode should really run to completion even if it's going to produce invalid output, because it's a lot easier to diagnose this complex a system when you can see evidence that it mostly hangs together