Yes. I was well aware of that PR. Despite making an honest effort, it was never close to the finish line and still had major issues to be resolved - not to mention backwards compatibility as well as worse performance & memory consumption on 32-bit devices. The core devs decided to shut it down for a reason (partially the reasons I outlined above, I assume). Proller made two PRs, both with massive diffs, which would've occupied practically most if not all available dev time had they been reviewed, conflicted with plenty of other PRs and were not sufficient to implement the feature to begin with. The devs would have had to put all other PRs aside to review & merge risky changes where a final stage might never come.debiankaios wrote: ↑Thu Sep 29, 2022 13:32Really? https://github.com/proller/minetest/tree/minetest32
"Never" is also an option - those "many more" features may take priority, especially considering the value / effort ratio :)
Is Minetest engine development painfully slow? Yes. sfan5 has recently compared Minetest with a "government organization" due to 100loc taking a whopping year to be merged. See Zughy's work for how to speed it up through prioritization. Our main bottleneck probably is coredev approval. I hope it's getting better with the latest addition of TurkeyMcMac to the team.debiankaios wrote: ↑Thu Sep 29, 2022 13:32I don't know what we do wrong. Mindustry is also so much fun like minetest and here are stats of minetest and Mindustry:
That are only 3 of many stats which show that other repos do anything other and get them goals. I know last time that a update(I think it was 5.5.0) comes 5 months to late and some features which should been added weren't added. I think if we do anything other there wouldn't be so much problems.
- minetest has 625 cotributors, Mindustry has 482 contributors
In minetest are 994 of 6551 Issues closed and in Mindustry are 3283/3283 Isues closed
minetest-devlopment begann at end of 2010(12 years ago) and Anukens(maintainer of Mindustry) made his first commit in Mindustry at end of 2017(5 years ago)
Yet this is an absolute apples-to-oranges comparison.
First of all, Minetest being around longer is not a plus point as in "there was more time to work on it" - we have a decade-old legacy codebase. Mindustry is only half a decade old; they probably have a much cleaner codebase & stack (yay JVM) and didn't make mistakes like depending on abandonware like Irrlicht.
Second: It is apparent from your stats that Mindustry is more popular and sees more activity (consider e.g. contributors / year or issues / year).
Third: Issues don't scale linearly. At first, issues are probably low-hanging fruits such as bugfixes. Later on, only nontrivial feature request issues or hard-to-fix bugs remain. You got your numbers for Minetest swapped: The larger number of almost 6.5k is the number of issues closed, which is pretty impressive. "3283/3283 [issues] closed" for Mindustry also should be suspect to you: It means issues - especially nontrivial feature requests - are rigorously closed (on the plus side, it probably means bugs are worked on - unless they were sweeping them under the rug). You are advocating for a feature request - for which there are good reason to dismiss it, and where core devs have dismissed the PRs already - to not be dismissed here. Think about it.
Fourth: Counting contributors is insufficient as it says nothing about how invested contributors are. I would assume that Mindustry contributors may be more invested.
Fifth: The scope of Mindustry and Minetest is way different.
Sixth: Let me add another stat: Number of open PRs - the lower the better. Zughy has been triaging hard to keep it low for Minetest, managing to have just ~80. Mindustry has twice as many at around 160. So clearly Mindustry is again getting more contributions or managing PRs worse.
TL;DR: Your comparison with Mindustry is not constructive at all. Involve yourself in engine dev before you start criticizing it.