Release of 0.18

[abdullah]

I am not all that sure release in December will happen. Due to the fact a lot of people do other things (holidays) through second half of the December. But if there is a real desire to do a release this year. We should i guess still have the time to do it before that. And should i guess rather target the end of November. And to do most of the release related work through November.

December is also 1st December, which is almost end November

If to have one release in December we have to aim for 1st December or end November, then we should aim at it there. Then, if it does not happen we have some days for contingencies and still release in December before Christmas.

[NormandC]

So please @yorik, @sliptonic, @bernd try to get ready for py3!

@ does not ping a user yet, and sliptonic has not participated in this topic, so if you want to make sure he is notified, you should quote him with his user_id:

Code: Select all

[quote=sliptonic user_id=708]ping[/quote]

ping

[ezzieyguywuf]

For a release with py3 I think we need to check and fix at least 4 wb's:

- draft
- arch
- fem
- path

I can't do this, because I am not very familiar with these wb's... So please @yorik, @sliptonic, @bernd try to get ready for py3! Py3-ignorance must finally come to an end.

@looo, can you elaborate on specifically what is lacking in these work benches that is preventing the py3 update? I don't mind taking a look if everyone else is good with that.

[sgrogan]

On one side, I think that it makes no sense at all is not to release for mac.

On the other side, Debian plays a major role in the Linux universe, and we run a high risk of being late:

kkremitzki wrote:
I think the first thing to mention is where we fall in release cycles currently. The Ubuntu 18.10 Debian import freeze is in a few days, so we'll have to handle packaging for that version by PPA. However, the soft-freeze for Debian 10 is upcoming (February 12, 2019), and since this will be the basis for things like the next wave of Raspberry Pi's and other long-term devices like school computers, I think we should try to aim for an early, before end-of-year release.

So I think it is important to have a before end-of-year release, at the latest very very beginning of January. With the proposed delay we have no margin for contingencies.

I'm torn here. If I understand correctly FreeCAD supports OSX fine, Homebrew doesn't support FreeCAD's dependencies.
FreeCAD supports QT5, PY3, pyside2 etc.

I'm just sensitive to "FreeCAD is supported on OSX, you just need to compile some stuff yourself"
https://forum.freecadweb.org/viewtopic. ... 94#p263665
https://forum.freecadweb.org/viewtopic.php?f=10&t=13341

[wmayer]

I think it is time for a new evaluation on the progress of these topics. Are they going to be ready soon? and relating to this: a) should we stick with our initial idea of including them in v0.18, or b) should be delay them to v0.19? I really hope for the former.

The main idea of making this release was to having finished Qt5 and Py3 because most Linux distributions will drop Qt4 and Py2 soon. Thus it's for me the topic with highest priority.

For py3, it is already stable enough in the sense that it passes all the unit tests. There are still some encryption kinks here and there, though. And many external workbenches are still not ported.

On my virtual Ubuntu 18.04 still many unit tests fail. The main problem is that somehow the shiboken2 package for Py3 is broken because it lacks of a correct cmake file (there is only one for Py2 but not Py3) and the shiboken2 Python module is missing too. I installed the packages from Launchpad created by Kurt.

[kkremitzki]

I installed the packages from Launchpad created by Kurt.

Those packages are in quite an old state and not super representative of how things are now, let me do some rebuilding and see where we are.

[looo]

@ does not ping a user yet, and sliptonic has not participated in this topic, so if you want to make sure he is notified, you should quote him with his user_id:

Code: Select all

[quote=sliptonic user_id=708]ping[/quote]

ping

yes, I know. I don't want to be to persistent (german: aufdringlich) as I leaned that I can't push anyone to anything... But thanks anyway.

@looo, can you elaborate on specifically what is lacking in these work benches that is preventing the py3 update? I don't mind taking a look if everyone else is good with that.

fem: trying different tutorials. Check if everything work the same way as with python2
path: check if pybind11-bindings for libarea works the same as boost-bindings. So doing some tutorials should also help to see what's not working currently.
arch, draft: I guess there are still some unicode problems hidden in the code. So also here we need testing...

[Jee-Bee]

For FEM see https://forum.freecadweb.org/viewtopic.php?f=18&t=30954

[looo]

For FEM see https://forum.freecadweb.org/viewtopic.php?f=18&t=30954

Should bei solved with https://github.com/FreeCAD/FreeCAD/pull ... e704e6cbf0

[abdullah]

I'm torn here. If I understand correctly FreeCAD supports OSX fine, Homebrew doesn't support FreeCAD's dependencies.
FreeCAD supports QT5, PY3, pyside2 etc.

I'm just sensitive to "FreeCAD is supported on OSX, you just need to compile some stuff yourself"
https://forum.freecadweb.org/viewtopic. ... 94#p263665
https://forum.freecadweb.org/viewtopic.php?f=10&t=13341

I am quite ignorant about macos builds.

Correct me if I am wrong, but if I understand it correctly the problem is packaging. Then no code changes are required, but making "Homebrew" support FreeCAD's dependencies (if the sentence makes sense, as I do not know what Homebrew is).

If such is the case, the work to generate the macos package is to be aligned with the packaging work. For the case of Debian 10, the packaging would be starting in January 2019, to be in-time for mid-February. This seems to be "more or less" aligned with Peter's availability. So generally speaking we could expect to have packages for the release in February 2019.

The main idea of making this release was to having finished Qt5 and Py3 because most Linux distributions will drop Qt4 and Py2 soon. Thus it's for me the topic with highest priority.

Definitely. Wrong choice of words from my side.

We would only made it into Debian 10 with Qt5 support, as Qt4 will not be supported. So even if we would love to catch the Debian 10 train, we could not do it if we do not have Qt5 support. The next FC release must have Qt5.

Debian 10 will support py2, as will ubuntu until 20.04 (not included). So py3 is way much desirable to have in a release. I would love to have it. However, it appears not to be a showstopper for Debian 10 (I am aware that debian is not ALL linux distributions). Maybe I am wrong.

Point releases might be consider if we want to catch the train, but miss some non-essential support. However, from what I read here, it appears that it is feasible to have all, py3, qt5 and in time for Debian 10.

I am very bad at python (or better said I am not at python at all), but if help is needed to be in time, just point me in the right direction. I will try to catch up.