Adapt tests to the output of gnumeric 1.12.43

Hm? This was meant to be covered by https://salsa.debian.org/reproducible-builds/diffoscope/commit/e9d3062

that URL 404s for me…?

(and also locally it gives unknown commit e9d3062)

Oh, race condition… 255cbe59

Ah, I see what you meant there. But that wouldn't have been enough, as also the values in the diff headers are different (i.e. @@ -1 +1 @@ vs @@ -1,3 +1,3 @@)

but now, doesn't adding that .strip() make the new test of mine fail?

It does. Reverted it in 344ae4d6.

(The .strip() /when I wrote it/ did cause the test to succeed as planned, but perhaps I didn't think about those diff headers which were not changing for me... the strip surely occurrs before diff is called anyway?)

Oh right. I think I was thinking of me while I was looking at fixing it, I did try to add a diff, but I must have placed it somewhere different (clearly after diff was called…)

Anyway, if you think your fix is better, do feel free to revert my change and add yours!

Done in 47588095. Again, sorry about the mess...

Instead, thank you for solving the issue in a much smaller (!) commit than mine!