[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: [pygame] pygamezero moving to github discussion thread

Indeed. Me too. However, SDL uses the zlib license. But I guess that doesn't mean we should automatically do that too. It might make sharing code with them easier.

On Sat, Mar 18, 2017 at 3:17 AM, Daniel Foerster <pydsigner@xxxxxxxxx> wrote:
I'd prefer use of the MIT or Apache 2.0 licenses to the less common zlib, for what it's worth.

On Thu, Mar 16, 2017 at 7:52 AM, René Dudfield <renesd@xxxxxxxxx> wrote:
I'm not sure if we should move it right now. I think so. I guess it'd be a bit of work updating the build tools. But already many of them work off the github mirror(s). The main work is filling in the member name translations (but I think I could do that in an hour or so).

Yeah, let's all shed a tear for the monoculture. But I don't see bitbucket/Atlassian, or gitlab going away anytime soon. But if it makes things a tiny bit easier for contributors, I'm ok with that. I've wasted time explaining hg, and bitbucket so many times to people already.

I think making it easier for people to publish commercial games will be good for everyone using pygame. Currently Apple prevents people doing that with LGPL. Otherwise I don't really care. I'm pretty sure most people would be happy with a license like what SDL moved to (zlib). But haven't explicitly asked everyone if they are apposed to zlib. Also, I don't know if there is a foundation that would be willing to take it. Perhaps PSF, Raspberrypi, or Kivy foundations are all options to ask. Just re-licensing, and keeping the copyrights as is, is also an option.

Why not a ctypes/runtime binded way? The main issue with ctypes, is again Apple, but also emscripten/asmjs/webassembly.

Why pygame using SDL2 patches by Lennard? Good question, I'm not sure which is the better way myself. This would allow us to gradually improve things, and reuse much of our existing code. To introduce new APIs where needed, and keep the old ones where they are. Doing big changes can be risky. pygame_sdl2 will require quite a bit of backwards compatibility work... who will do that work? Was the only reason for pygame_sdl2 to avoid the LGPL? Does it still make sense with a zlib pygame? What other benefits does pygame_sdl2 provide?

I'd really be interested to know from Tom and Lennard what they think. It doesn't sound like Tom is interested in working so much on backwards compatibility. But also, Lennard mentioned quite some time ago that pygame_sdl2 could be a way forward.

The main benefit of a foundation (if we don't need to run it!) is that donations, grants and such are much easier to deal with. Whilst dealing with GSoc and other grants things was something I was happy to be involved with it was a lot of admin, for something which can be more easily shared in a larger organisation (like for example the PSF).

For now, my own main pygame priorities are:
  • Making the website a bit more usable.
  • Trying to bring some of the pygame projects together, on github on the website (maybe with pygame 1.9x).
  • Coming to a consensus on way forward with everyone regarding licensing, and what to base future pygame releases on (probably pygame_sdl2, zlib, or Lennards SDL2 patches).
  • Also working to make distribution of peoples games better. Both for people who want to just show off their work, and people who want to try and publish games.
  • Fix a few remaining packaging issues, and install issues with pygame 1.9x (like the macos thing)
  • Modernise the pygame code with regards to type hints and docs. (again, so people and newbies especially get better feedback in their editors).
  • ... make it work better on raspberrypi.
Finally I'd like to highlight some of cool bits in pygame 1.9.3. There's some pretty major pieces in there, that I don't know so many people know about.

I think it would be pretty good if pygame zero, pygame2, and pygame 1.9 were on github in the same organization. Also if we decided on a way forward (with pygame_sdl2 or however) for pygame2. It feels like there's a bit of momentum now with pygame, and getting the projects together will only increase the excitement further. But more importantly it will let us share our limited resources in a much better way.


On Thu, Mar 16, 2017 at 11:09 AM, Thomas Kluyver <takowl@xxxxxxxxx> wrote:
On 16 March 2017 at 10:01, René Dudfield <renesd@xxxxxxxxx> wrote:
Yeah, with the python community moved to github, it's really past time to move.

Do you also want to move the pygame repo itself onto Github rather than using it as a mirror? Or is pygame in maintenance-only mode, and it's less important to attract new contributors?

I agree that it's easiest to attract contributors if your project is on Github, but at the same time, I'm a bit sad that we seem to be heading for a total Github monoculture.