It is hard to know what might pop up in a text, esp. if you consider some of the Native American languages, African languages, etc. As type now supports more and more diacriticals, more and more authors start using them.

Don't know if this helps, but one thing I do with the i-accents is to give then a bit larger setwidth (sidebearings). And not all the same, use whatever works best with individual characters.

And sometimes a slightly shorter macron, a slightly condensed caron, circumflex, dieresis, etc. You can put them up alongside a vowel with unmodified accents in the metrics window, and if a difference isn't objectionable when you view it at at 120 points, it will likely pass muster,

Otherwise, yes, you need to kern. Or do what so many others do, let the typesetter do it.