Fix PartialOrd implementation for prune candidate

[suchapalaver]

In the previous code, the line Some(Ordering::reverse(self.0.cmp(&other.0))) was incorrect because it would return the opposite of what you intended, potentially leading to incorrect ordering of elements.

Signed-off-by: Joseph Livesey jlivesey@gmail.com

[vaporos]

Can you elaborate in the commit message on the reasoning on why this is correct vs. the previous code?

[suchapalaver]

Can you elaborate in the commit message on the reasoning on why this is correct vs. the previous code?

@vaporos, thanks. I've updated the commit message to read:

Fix 'PruneCandidate' 'PartialOrd' impl to return result of 'self.cmp(other)', not its reverse

[peterschwarz]

Can you elaborate in the commit message on the reasoning on why this is correct vs. the previous code?

@vaporos, thanks. I've updated the commit message to read:

Fix 'PruneCandidate' 'PartialOrd' impl to return result of 'self.cmp(other)', not its reverse

I would still think of that comment as the “what” and not the “why”

[suchapalaver]

Can you elaborate in the commit message on the reasoning on why this is correct vs. the previous code?

@vaporos, thanks. I've updated the commit message to read:

Fix 'PruneCandidate' 'PartialOrd' impl to return result of 'self.cmp(other)', not its reverse

I would still think of that comment as the “what” and not the “why”

Any suggestions? I think for people working in Rust it's clear why the "what" is correcting something wrong. You have a custom implementation of Ord. So you don't want your implementation of PartialOrd to reverse that.

[suchapalaver]

This, btw，fixes the build issues on PR 2458.