It happens regularly that the code in the post is broken (or otherwise not ready for review), but the problem is too complex to explain or demonstrate in the limited space afforded by a comment, and needs to be posted as an answer. Here's an example from earlier today, here's another from a week ago, here's a third with a detailed explanation of how to discover the bug, and here are some more: two1 more2 3.
Current procedure
What we (Code Review regulars) would like to happen in these cases is:
We vote to put the question with the broken (or otherwise unreviewable) code on hold.
The OP fixes the code.
The OP posts a new question with the improved code.
The reason we would prefer this is that it leaves a record of what the problem was — sometimes the technique needed to discover a bug is interesting, but even if it's just a reminder not to post code without testing it, that's still valuable.
Is the procedure too onerous on the OP? It doesn't seem difficult to me to copy the text of the question that was closed, and paste it into a new question. But I guess it might not seem so easy for a beginner.
Alternative procedure
But if we are agreed that is too hard, then what's the alternative? Perhaps:
We vote to put the question on hold.
The OP fixes the code.
The OP updates the question.
We vote to reopen the question.
The reviewer who explained the original problem deletes the answer in which they explained it (since this answer is now unrelated to the question). Or someone else flags it for deletion by a moderator if the reviewer does not delete their answer in a timely fashion.
This is more complex than the first procedure, it takes substantially longer (because of the delay at step 4), it erases the record of original problem and its discovery, and denies credit to the person who discovered and described the problem. I don't think it's an improvement at all.