If you think you found a solution, think twice.
If the implementation is taking too long (too many changes in different functions and classes to fix a single bug) there may be a better solution, it's never too late to reverse the changes and start again, it's not a shame, in the worst case you will reimplement the same solution, but better, in the best you'll find an easier and better one.

Don't run, even if there's a deadline.
It's much worse having to deal with negative feedbacks later.