You are correct about re.findall. re.match matches a regex at the beginning of string. In your case, it will succeed since there is a match at the beginning. But, for example, the following match will fail: re.match('(?=a)', 'wass').

Notification: see the previous message.

Figuring this out is part of the "fun". However, I'd start with the first input that you failed on. Add a print statement in your next_smaller implementation to find the first input you are failing on. Then add a test for that in the tests-window and revise your algorithm until you get the that test to pass. It's likely that once you fix that problem that the rest of the tests will pass.

Well... coding be hard sometimes.
Actually you probably need O(1) so that means no loops!

I haven't solved it sadly. But given that the inputs are so large, you will need a O(n) solution. So make a spreadsheet of inputs and outputs and look for a pattern: there must be one!

I detrats yb gnidnif eht lacitamehtam .nrettap nehT I desnednoc ti otni eno enil htiw lacigol .srotarepo nehT decalper eht srotarepo htiw esiwtib .seno ehT dne tcudopr si ytelelpmoc elbisneherpmocni

yeah, me too.