Fix external state in benchmarking (edit.rs)

[kanishkarj]

This is in reference to the discussion in PR #1152. This PR fixes the external state of the iterations in the benchmarks.

[cmyr]

maybe it still makes sense to do some loop inside the iter? we can just make it some fixed length.

[cmyr]

For many of these functions we still may want to be doing a loop in the iteration, we just want to be doing it the same each time. This is the right general idea though!

[kanishkarj]

Hmm, So you mean to say all the functions use the same offset update ? for eg

offset += 1

would be the offset for all the benchmarks?

[cmyr]

I would use whatever they were using before, but inside an explicit loop, and having offset declared fresh on each b.iter(|| {. So we want to be doing the exact same thing inside each b.iter. (We might also want to clone over the rope in b.iter, so we're using the same underlying rope each time.)

[cmyr]

@kanishkarj this is very close to mergeable if you want to update again!