Can you add the reasoning you hinted at in the PR description also to the commit that adds the feature to QEMU? It would be nice to be able to judge why this was done just from the commit logs without yet another indirection.

The reason would likely be

So we don't need to patch the vendored libslirp.

Not sure why you left that out.

I think use shared libslirp instead of vendored is accurate and sufficient as a commit message.

I think use shared libslirp instead of vendored is accurate and sufficient as a commit message.

But it doesn't contain the why we did it. Stating what has been done can ofte be extracted from the changes but not the reason for that change.

Let me propose the following:

qemu: use shared libslirp


In the past we had to patch the bundled libslurp multiple times (e.g. #76065 &
#79050) while we have & had a proper package for libslurp. With this change we
can apply patches for libslurp once in a straight forward fashion and those
will also apply to QEMU.

I'm a bit disappointed to see you close the PR and delete the branch over what looks like a really good commit message suggestion...

And I'm disappointed that a commit message that does explain the change is being bikeshedded.

Heaps of stuff gets merged with commit messages vastly worse than this.

If what I had isn't good enough I've got other things to do rather than argue.

Okay, thanks.

Just as reference: I am not trying to pick on you but I am trying to make sure everything I handle follows a good style. We even have that documented in the CONTRIBUTING.md

WFM macos 10.14, LGT ... oh.

I was being serious when I said "I've got other things to do rather than argue" but I see that this is now being discussed on discourse so I will respond. I'm not going to open an account on discourse.

I am not trying to pick on you but I am trying to make sure everything I handle follows a good style.

If you want people to follow a style, find a way for it to be enforced consistently tree-wide. Otherwise it isn't anything except unsustainable pedantry.

We even have that documented in the CONTRIBUTING.md

As we don't even enforce format the commit messages in the following way from CONTRIBUTING I'm actually kind of annoyed by writing good commit messages being used as some sort of justification here.

During a review that I did today a contributor decided to close their PR

s/contributor/committer/

Can we pick this up, please? If we had merged this we'd not be affected by CVE-2020-29129.

My bad, libslirp is affected, not QEMU itself. But I still believe this is a good idea.

I'll reopen if the request to change the commit message is withdrawn.

I'll not withdraw that request. It should clearly state the motivation for this change.

I'm losing the will to live here.