Thanks for the ping, I had indeed missed that the previous PR did not only remove set_string_function (good) but also introduced an alternative. I don't have a strong opinion, but it does seem to make sense to allow one to override both repr and str. @seberg - are you suggesting that the user-defined function should not just take the array, but also the repr argument? My sense is that it is easier to just provide two arguments, i.e., add add override_str as well (or perhaps in analogy with the old functions, call it str_function and repr_function).

Yeah, that was a thought, just to add fewer arguments. But I don't mind the approach of having two kwargs, either (in which case, there is also no rush).
I thought they might often come together, but if not, then passing repr=True/False might just complicate things anyway (because making the one you don't want to override do the default thing is a bit annoying).

Ping @jni @GenevieveBuckley, I noticed the trouble you had in napari/napari#6776, this should help you get back the shorter repr you had before.

Thanks Mateusz.

I went ahead and merged this because the implementation is already in and the discussion seems to be mostly about alternatives.