On Tue, Oct 27, 2026 at 08:25:14PM +0100, Omar Polo wrote:
> maybe wouldn't hurt.
> 
> in any case, none of this was meant to be a blocker for the diff =)
> 

I remember another reason for wanting the diff the way it's written.

The original loop was actually a bit tough to write correctly. If you look
at its history, I had to fix a bug in the fringe case where the file is
exactly a multiple of blocks.

Having all the new code be directly under n != blocksize makes it more
self-contained and less likely to interfere with correct code.