I do agree completely with Andy, that first step should be a branch
master_py3.
So we can all work together on this branch till it is usable, not stable.
Then merge this branch to master and set default python to 3.

I do know that we will for sure brake some machine configurations, but
is master not exactly for this?
I do not know if it possible to give the user a big warning if he try to
install master, telling him
"This release may brake your machine configuration, as it will use
python3, so if you go on be sure all your configs and files are python3
ready!"

If the user still goes on, at least he know what risk he goes.

And when master get a stable release, all should be good.

I as maintainer og gmoccapy do know how much people may suffer under
this. (I have about 400 contacts world wide with people using gmoccapy).
That is exactly the reason, why I do think about letting die gmoccapy
and name the new gtk3 GUI then "MINOS" (Machine Interface NOrbert
Schechner). I tried to just convert gmoccapy to py3/gtk3, most is
working, but I needed to use a lot of work around and some steps are
still to go.

Even destroying thinks, I do not see any other solution for this.

Norbert

Am 12.05.21 um 19:15 schrieb andy pugh:
On Wed, 12 May 2021 at 14:43, Chris Morley <[email protected]>
wrote:

I just think should have a better plan then 'sometime someone will break
master for python3'

We may not need to break master for as-installed LinuxCNC. We will almost
certainly end up breaking existing user configs which rely on LinuxCNC Py2
/ Gtk2 infrastructure.

I think that step 1 is for Rene to push his work as a new, public, branch
to the LinuxCNC github. I suggest calling it master-py3.

We will then attempt to get that into an equivalent state to the main
master branch, and will merge any changes in master to that branch.

At some point the changes in master-py3 will be merged back in to master.

This was what happened with joints-axes. Though I would hope that the
process does not take 10 years this time.




_______________________________________________
Emc-developers mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-developers

Reply via email to