Minetest Forums

I saw the changelog the other day, when is it going to be released?

I'm really excited !! :D

Probably within the next week. No promises, though.

Im excited too :D

sfan5 wrote:Probably within the next week. No promises, though.

Cool, can't wait :)

Here's the 100% accurate release date:
It's done when it's done.

Wuzzy wrote:Here's the 100% accurate release date:
It's done when it's done.

This is, and always will be, the best release date. I'd prefer to wait a few more days than to have a rushed and buggy update.

I love "rushed n buggy" = new challenges & features ; )

TumeniNodes wrote:I love "rushed n buggy" = new challenges & features ; )

I love the “update early, update often” concept. Do not “collect” too many smaller changes/fixes before releasing, release when a change/fix was implemented and properly tested.

This causes overall less issues because there are no giant feature updates at one time but only smaller updates over a period time. Adaption to that smaller updates is easier than adapting to one giant update that changes a lot of features at once.

Also getting rid of “versioned updates” completely is a good idea. Rolling release works wonderfully with “update early, update often” (and absolutely does not mean “bleeding edge” or releasing untested software).

Either you're crazy or trolling. You can't just remove the version number from Minetest!

It would be an absolute nightmare for gamedevs like me because I would be forced to chase after the latest changes and breakages every day … Supporting latest stable is enough, thank you very much!
Also, everyone will have a different version, so that's annoying for debugging too. Also, there aren't even real “versions” anymore, only commit IDs. This is very annoying for debugging as you need to examine the commit log to see the correct order. With real version numbers, the order is obvious. It's not very user-friendly as well, they have no idea whether their version is up-to-date.

Linux package managers will also be royally screwed over this.

Rolling release is a lie. It means that every commit is a "release". This is absurd. Also, I have no idea how deprecation would work in rolling release. And I didn't even start with networking. As soon you realize there's a network, rolling release falls apart. Because clients must be network-compatible, otherwise they won't work!

Rolling release also puts a huge burden on the devs. Since every commit is a release, it means that every commit must be perfect. This is an absurd expectation to have. Development is meant to be “dirty”. Rolling release is especially damning for devs if you ever want to implement a big feature. You're basically forcing devs to do big changes perfectly and squash them into a single commit, because every commit is a release …

I'm pretty sure the core devs also have 1001 reasons why rolling release will NEVER be a thing in Minetest.

Gosh, there is SO much wrong with this idea … Or maybe I just took a bait, who knows?

I do agree that it's a bad idea to accumulate a huge amount of features per release, but that's not an excuse to go the other extreme.

There are three groups of people. One group want big updates, second group wants fast updates and the third group wants bug-free updates. And you can satisfy only two of them at once.

Wuzzy wrote:It would be an absolute nightmare for gamedevs like me because I would be forced to chase after the latest changes and breakages every day

So you say you don’t want to actually care about your code and just want to do it once and then never care about it until the next release in a couple of months?

Wuzzy wrote:Linux package managers will also be royally screwed over this.

There are multiple Linux distributions having a rolling release update strategy. They all work wonderfully.

Wuzzy wrote:Rolling release is a lie.

Can’t confirm from 20+ years of experience with rolling release.

So you say you don’t want to actually care about your code and just want to do it once and then never care about it until the next release in a couple of months?

What's your point? My code is compatible with latest stable, isn't this good enough? Experience shows that chasing after bleeding edge OF MINETEST (!!!) is a nightmare because I depend on code that hasn't stabilized yet. It saves me a lot of stress if I just depend on latest stable. And yes, the latest commit in Minetest totally is the bleeding edge. It would be a terrible idea to depend on it, only to learn that 2 weeks later, my code breaks again …

There are multiple Linux distributions having a rolling release update strategy. They all work wonderfully.

What they are offering is an automatic compiling of the bleeding edge. This is NOT the same as rolling release. It's only rolling release when the developers decide it to be. Which is not the case, currently.
Also, this kind of stuff is out of reach for most users anyway. When I strictly depend on latest stable, I simply reach the biggest audience and have the least trouble. I really hate it when games on Minetest depend on bleeding edge, and don't offer any stable version at all, it puts an unneccessary extra burden on the player.

Explain how rolling release would work out for Minetest in practice. I have a hard time believing you.

Linuxdirk wrote: So you say you don’t want to actually care about your code and just want to do it once and then never care about it until the next release in a couple of months?

That's

• Obviously not what they're doing if you look at their post history.
• Pretty rude of you to say. You basically called Wuzzy, who puts a lot of work in their numerous games and mods lazy.

funny,
if someone mean 5.1.1 is not good enough anymore ==> make in issue, help to fix it and control that it is fixed,
then
make sure no other issues might make you sad = take 5.2.0-dev
otherwise
just wait ... for it, requests like this here will not help anywhere