The most obvious edge-case was mentioned by other siblings.
My "edge case" is probably not really even an edge-case though.
SPOILER ALERT ---
My implementation stopped looking when it found a number. So "one2one" or "1twone" and so on, would find the numbers "1,2" and not "1,2,1". There were no examples where a number appeared multiple times in a line AND this affected the first-last pair. So e.g. abc1ninexyz841 would result in 14 in my case but should've been 11.
Again: it's rather implied and quite obvious if you interpret the description as human, but I worked at it from TDD, and the example missed this situation and the actual input had only a relative few of them, so debugging was hard.
My "edge case" is probably not really even an edge-case though.
SPOILER ALERT ---
My implementation stopped looking when it found a number. So "one2one" or "1twone" and so on, would find the numbers "1,2" and not "1,2,1". There were no examples where a number appeared multiple times in a line AND this affected the first-last pair. So e.g. abc1ninexyz841 would result in 14 in my case but should've been 11. Again: it's rather implied and quite obvious if you interpret the description as human, but I worked at it from TDD, and the example missed this situation and the actual input had only a relative few of them, so debugging was hard.