[llvm-dev] RFC: Introducing an LLVM Community Code of Conduct

Renato Golin via llvm-dev llvm-dev at lists.llvm.org
Tue Oct 13 14:32:01 PDT 2015

On 13 October 2015 at 21:59, Joerg Sonnenberger via llvm-dev
<llvm-dev at lists.llvm.org> wrote:
> I think this goes much more to the heart of community interaction than
> the CoC did. Things like seemingly ignored contributions and bug reports
> are IMO far more damaging and off-putting for newcomers.

This is a very good point, and one that agrees with my idea that we're
only doing this because everyone else is doing, too.

> "We welcome contributions, but we also require contributors to work with us."

This is what he was getting at. Final words: "They don't want your
contribution and they don't want you, because you are not helping.
Unless you do, in which case you are like them". Basically, as a
spectrum, we can all use similar rationales, with very different

We had people *demanding* their bugs being fixed because millions of
users rely on their products. The use case for the unfixed bugs are
normally so specialised, that virtually nobody else in the world care.
The come back is usually in the form of "you are the experts, I'm not,
therefore you have to fix the bugs I don't understand about".

That kind of behaviour, in my view, is disrespectful to begin with. It
assumes a the commercial model should apply for everyone. In some
cases, I replied saying it would be unlikely that those bugs would be
fixed, to which the OP probably got upset and left thinking our
community is unhelpful. Should I have acted differently? Should we
strive to fix every possible bug, ever?

Another bug in RedHat and Debian unstable for AArch64 which I can't
reproduce because I don't have those systems installed. Should I
uninstall one of my production systems to install theirs and fix the
problem? Should I ask them for hardware / software so I can reproduce
the problem? Or should I rely on them to fix their own problems? My
choice was the second one, but I can see reason on all three. Can we
blame people that don't care at all? Should we force people to care
more than they are able to?

Open source is past the time of peace and love, and now it's all about
money. If we force too much, companies will leave, fork, re-invent the
wheel. If we force too little, benevolent dictatorships will form, and
then become the norm, even later, when behaviour is forced.

All in all, our community has stayed very sane over the same years
many others went in disarray.

The only issue I see recently is an potential internal power struggle
encouraging fast and incomplete commits, as long as the bots are
green, which IMHO is unhelpful and often destroy other people's work.
The view that more contributions equals more community kudos is the
enforcer of that behaviour and I don't think that's a valuable metric.
I'd rather have fewer good quality and considerate commits than
thousands of half-backed features, but that's probably just me.

Even that, is completely independent of ethnicity, religion, gender,
blah, blah, blah. so the original argument is *still* invalid.


More information about the llvm-dev mailing list