Sorry, I just realized that I unintentionally introduced a change in #251.
Previously if the position of any of the values in the loop was invalid, we would set mb_error. But now we are overwriting mb_error every iteration of the loop, so if the first position is invalid and the second position is valid, mb_error will be ERR_NONE in the end.
Sorry, I just realized that I unintentionally introduced a change in #251.
Previously if the position of any of the values in the loop was invalid, we would set
mb_error. But now we are overwritingmb_errorevery iteration of the loop, so if the first position is invalid and the second position is valid,mb_errorwill beERR_NONEin the end.This fixes my mistake, sorry about that.