Michael Stefaniuc <mstefani@redhat.com> writes:

>> The pending state is feedback. It means that the patch is not clearly
> yes, but the worst possible feedback.
>
> New people assume you or the area maintainer need to still make up their
> mind on the patch but that's not the case, it is a done deal.

Not necessarily. Patches are automatically marked pending as soon as I
look at them, until they get a more permanent resolution. There can be
many different reasons why they don't get one.

> Iff one knows you and knows to read in between the lines then the above
> can be reworded as:
>   "Current implementation rejected; idea might have some merit."
>
> That's what "Pending" means most of the times.

Sometimes, yes.

> Now for the other meanings of "Pending":
> - "Waiting for feedback from the main developer in that area."
>   Don't remember to have ever seen that. Those patches stay normally in
> "New" as you don't look at them before the area developer ACKs.

No, I usually do look at them. Sometimes I go back and mark them as New
again, but not always.

> - "preferably in the form of a test case"
>   That should be "Needs tests".

It would be, in cases where tests are clearly the right way. There are
cases where it's not clear whether a test is feasible, and using "Needs
tests" may lead the submitter down the wrong path.

> - With the two (minor) other meanings of "Pending" handled by other
> patch states we can rename "Pending" to "Convince me" or "Needs work".
> That would make it more obvious what is meant.

I can certainly rename "pending" to "convince me" if it helps, but I
think it's important to have some sort of open-ended resolution to leave
room for developer's judgement. There are many cases where I'm genuinely
undecided about the right resolution, and I feel it's preferable to show
the undecision and leave it to the developer to try to address one way
or the other, even if that's of course harder than being told exactly
what to do.

--
Alexandre Julliard
julliard@winehq.org