ngc92
commented
1 year ago While I agree that we shouldn't focus too much on old compiler compatibility, I also don't think we should update just for the sake of updating. Keeping with the current minimum of gcc-7, we could still switch to C++17, though. GCC-7 doesn't yet support std::filesystem, but that should be fine given that we use physfs anyway.
In terms of our infrastructure, I think we should use the process of releasing 1.1 to ensure we have automated building/testing for all our platforms (i.e., #39).
Personally, I also really want to make some improvements to the actual game simulation:
- Switching to fixed-point numbers (#38), so we can finally remove the old assembly hack, and get reproducible physics across all devices.
- Decouple rendering and game simulation, as in #100
Regarding a league, I think this is a good idea, but really should be coordinated with www.blobby-liga.de. If we have some league option already in the game, that might help get a few more players, as no external sign-in would be required.
Given the age of RakNet, would it make sense to at least handle all new league-related stuff with a newer networking library? I guess we cannot really get rid of RakNet for the actual game-play that easily, but at least for new things that don't require low-latency networking, using a more modern lib would be great.
Getting closer to 1.1 release we can start thinking about what we can do next. My idea would be to extend blobby volley with features that bring real added value to the player and this as fast as possible. Blobby Volley 1.1 has many smaller improvements all over the code and makes sure it works on many old platforms with old compilers. But that also slowed us down on the other side. I would suggest that we move away from these restrictions with the start of development of 1.2.
Proposal:
Let us focus one single main feature: What do you think of developing a league with rank system?
Any other good ideas or wishes?