faketerm.Diff doesn't seem to correctly handle partial cells

[mum4k]

This only has impact on the readability of test outputs.

Partial cells are cells in the buffer that should not contain a rune, because the previous cell contains a full-width unicode rune. Full-width unicode runes that two cells when printed on the terminal.

The partial cell is printed out even though it should be skipped:

               got:
                    │         ⢀⠎
                    │        ⡠⠃
                    │       ⡔⠁
            世^@世^@世^@世^@│     ⢀⠎
                    │    ⡰⠁
                    │  ⢀⠜
                    │ ⢠⠊
                 abc│⡰⠁
                    └───────────
                     0         1

The bug seems to be here, we print out the rune even if it is is partial:

termdash/internal/faketerm/diff.go

Lines 72 to 75 in 07850c7

	} else if r == 0 && !partial {
	r = ' '
	}
	b.WriteRune(r)

And here:

termdash/internal/faketerm/faketerm.go

Lines 130 to 133 in 07850c7

	if r == 0 && !partial {
	r = ' '
	}
	b.WriteRune(r)