Compiz is Dead - Beryl Lives Again? Enter - Northfield/Norwood

[Pekka Paalanen]

On Mon, 25 Mar 2013 12:51:07 -0600
Scott Moreau <oreaus at gmail.com> wrote:

> Hey Casey,
> 
> On Mon, Mar 25, 2013 at 12:46 PM, Casey Dahlin <cdahlin at redhat.com> wrote:
> > On Mon, Mar 25, 2013 at 12:41:59PM -0600, Scott Moreau wrote:
> >> "The key point to understand is, that this is not a new protocol in its
> >> own right. It *is* the wayland protocol, with a few minor additions
> >> that make it possible to do new interesting things."
> >
> > Then don't fork the library.
> 
> I did not want to call it a fork since I don't feel that it is.
> However, I have now changed the name and announced it as a fork at
> Kristian's request.

You seem to be unable to distinguish Wayland and Weston.

Fork Weston all you want. Not Wayland.

> > "A few minor additions" can exist outside the codebase. Hell, all of the shell
> > protocol exists outside the wayland codebase. All of THE DRM CODE exists
> > outside of the wayland codebase.
> >
> > If what you say is true about not forking the protocol then there is NO reason
> > to fork libwayland.
> >
> > --CJD
> 
> Yes, there is no reason to fork libwayland. And I don't feel this is a
> true fork, just a temporary rename to avoid the confusion it might
> otherwise cause, remaining under the 'wayland' name. Wayland has been
> in my github repo since I've uploaded it there. The only difference
> now is, the name has been changed and an official fork announcement
> has been made by request.

I totally agree with Casey here. Why do you need to rename it, or
point people to your fork of libwayland, if you are not modifying
it?

There is no need for you to fork or rename libwayland, or advertise
your own repo for libwayland. It is hard to not break the protocol
or ABI - the very thing you said you would not do. And if you
commit there something, it will stick there for a long time, and
you will have to live with it. So if you can't deal with the review
process, don't go there.

Or are you perhaps planning to break your own protocol additions?

I even moved the sub-surface protocol from libwayland to weston,
because it is easier to develop that way.

If you want to change wl_shell_surface interface, you can start by
copying it under a different name. That only gives you more power,
as you can now also modify and remove already existing protocol
that was not good enough. That is the kind of power you really
want, to not be hindered by existing things.

If you don't want to copy it and rename a load of stuff, you can
add a new interface similar to wl_shell_surface, but parallel to
it. That way you don't even need to care about protocol versioning
at first. You can change the protocol all you want for a long time,
and when you are happy with it, then see how it will integrate into
wl_shell_surface, or should we just deprecate wl_shell_surface
altogether.


You only need the Weston fork, just use libwayland from upstream.
- pq