Significant bug is discovered in an accepted answer, peer review fails

Question

Put an accepted answer into production code for a commercial website, which was then subjected to pre-production testing. A significant, easily-reproducible bug was found, tracked down and fixed. The fix was edited into the accepted solution, which was rejected during peer review. It seems unlikely that the peer review involved any testing of the validity of the edits. The reason given for rejection appears to have been purely pro-forma.

What now?

This is the q/a:
https://stackoverflow.com/posts/9220653/revisions

Answer 1

Post your own complete solution pointing out the bug in the accepted answer, then post a comment on the accepted answer referring back to your corrected answer. If the correction is valid, then it should either be edited into the accepted answer, or the new answer should be accepted instead.

Answer 2

The fix was edited into the accepted solution, which was rejected during peer review. It seems unlikely that the peer review involved any testing of the validity of the edits. The reason given for rejection appears to have been purely pro-forma.

I saw, but did not take any action on, your second suggested edit. My opinion was "well, this might well all be true, but I don't think much of the tone. I'll let this be someone else's problem^[1]".

As to what did eventually happen to your suggested edits, as the faq says,

What is reputation?

Reputation is a rough measurement of how much the community trusts you; it is earned by convincing your peers that you know what you’re talking about

When you come along with three days of membership and no more than the starting amount of rep, it's not particularly likely that peer review of your suggested edits will necessarily involve rigorous testing of the change; and (as might be learned from experience of suggesting edits), nearly all rejections of suggested edits will be with reasons that are "purely pro-forma", because the volume of suggested edits is such that it's impractical to do much more.

None of the answers here are offered with any warranty or guarantee as to fitness; therefore it's not really surprising that no major fire alarm was instituted when mistake was reported to have been discovered in one. There's no SLA here, or response time guarantees, or anything like that: just a bunch of volunteers who, like any group of humans, sometimes make mistakes.

---

^{[1] I'm allowed to think like this, because I'm a volunteer}

Answer 3

I always reject edits that attempt to fix technical errors in the answer. If the answer includes code that doesn't check for something, and I see a suggested edit that adds the check, I reject the edit as

This edit is incorrect or an attempt to reply to or comment on the existing post.

Comment on the answer if you have the rep for it, or add your own answer if you don't. (Your own answer should not say just "the answer by Bob has a horrible flaw"; it should present the correct answer with the flaw fixed.) Edit-reviewers can't tell if you're right that it might leak memory or delete a good file or whatever the horrible problem is, so editing the accepted answer is not the right mechanism for fixing the Internet in this case.