[Spice-devel] [PATCH spice-protocol] build-sys: simplify autogen

[Marc-André Lureau]

Hi

On Fri, Dec 5, 2014 at 6:24 PM, Christophe Fergeau <cfergeau at redhat.com> wrote:
> On Fri, Dec 05, 2014 at 06:03:42PM +0100, Marc-André Lureau wrote:
>> On Fri, Dec 5, 2014 at 5:43 PM, Christophe Fergeau <cfergeau at redhat.com> wrote:
>> > What is so bad with having a commit delayed for a few hours while it's
>> > waiting for reviews?
>>
>> It is mainly the difference between asking someone to take an action
>> or not. And asking someone take time often longer than necessary.
>
> I don't understand why it's bad for the project, or why it makes things
> worse.

It makes the general process more complicated and slower, it prevents
taking responsability. It make us more rely on each other instead of
steping up and doing the right thing without asking each other what
and how to do it. It prioritizes a bit our work. We don't necessarily
need each other for all the thing we do.


> Actually I don't and often feel uncomfortable making changes without
> anyone looking at them.

s/uncomfortable/comfortable I presume. The problem is this is not just
libgovirt and your work but a lot of more important pieces. There are
a lot of projects we rely on without mandatory review, and it's not
that bad. Imho they are more healthy that way.

>> Similarly, I would like you trust contributors to do the right thing
>> for Spice, without mandatory code review.
>
> I'd prefer if we did not frame this in term of trusting people or not,
> this is not what this is about. We are all humans, we all make mistakes,
> and we all have different skills. Patch reviews are just a way to tap
> into other people skillsets in order to improve the project overall, and
> to try to avoid these human mistakes as much as possible.

Isn't it trust when you give resonsabiity to people? If you remove
responsability, and forcefully double check everything they do, you
remove trust.

>> There has been no big issues so far, only you complaining for no valid
>> technical reasons. Let's stick to that before enforcing rules.
>
> Well, I did complain, Jeremy did too. Coming getting pushed "because they
> are trivial" is something newish and mostly done by you. I don't think the
> obvious conclusion to this discussion is to just do whatever you

May be the other raison Jeremy step up is because of this noisy and
endless discussion.

So let's bother more people and ask them what they think, good chances
are that opinions will be divided.

> decided, quite the contrary. Let's stop trying to be smart in deciding
> what is trivial or not, and let's keep sending all patches to the ML,
> this is not a big constraint, and will result in less frictions.

Where did you see that all patches have to be mandatorily reviewed in
Spice? We always had a trivial push rule, and you always complained
about it. That's all I know.

I invite more people to do more patches in general, and if mandatory
review is what slowing them down, I would happily repeat that we don't
have mandatory review in the Spice project and we trust your call for
making the changes without review. If you do something bad or
controversial, some (at least me), will notice it and it will be
discussed.

-- 
Marc-André Lureau