[0.4.14] Spring release, announcing feature freeze

[rubenwardy]

Wait, what? Minetest is banned from Google Play? Are you sure? What for? If you would have said it was for iOS, then I would have immediately believed you (“Ewwww! Free software. We don't want that.”). But Google Play? No way!

The App Store not Google Play. So iOS

[MineYoshi]

Whilst you can technically run minetest on android, you are not legally allowed to put it on the app store. Please do not encourage devs that put ad ridden versions on illegally

I am not encouraging anybody, i am justly showing the fact that Minetest can run in "iOS"...
Sorry if someone was confused...

In my opinion is better have an Android device, the amount of money for an "iPhone" is ridiculous, also there are Android devices with a lot better hardware, a lot cheaper...

If it can be built, we do not necessary have to put it on the app-store, it can just be put on the MT downloads page.

This would probably require those who want to install it to jailbreak their iDevice, since iOS does not allow downloading random apps from the Web, unlike Android.

[Andrey01]

Will be new update 0.4.15?For example, new mobs(for the first time default)

[LNJ]

Will be new update 0.4.15?For example, new mobs(for the first time default)

Here are all changes in MTG since the 0.4.14 was released: https://github.com/minetest/minetest_ga ... 4...master

But no mobs. (It's probably the best way to add a mob api in C++, for speed)

[orwell]

#clientsidescripting

[100best]

Whilst you can technically run minetest on android, you are not legally allowed to put it on the app store. Please do not encourage devs that put ad ridden versions on illegally

This might be interesting Apple's iOS takes a bite out of your freedom and this Android and Users' Freedom.

[Andrey01]

Will new version "0.4.15" in future?For example, new blocks, items and mobs!

[Worldblender]

Will new version "0.4.15" in future?For example, new blocks, items and mobs!

There are new blocks and items coming for sure, along with biome changes, but mobs are not going to land in this version, unfortunately.

[GreenDimond]

Will new version "0.4.15" in future?For example, new blocks, items and mobs!

There are new blocks and items coming for sure, along with biome changes, but mobs are not going to land in this version, unfortunately.

I don't think mobs are gonna arrive any time soon. I don't think they're even in the plans :P

[Spaghetti Developer]

Will new version "0.4.15" in future?For example, new blocks, items and mobs!

There are new blocks and items coming for sure, along with biome changes, but mobs are not going to land in this version, unfortunately.

I don't think mobs are gonna arrive any time soon. I don't think they're even in the plans :P

nothing mobs, will never be added, resign. I agree with the choice of developers! in the future make an API for the management of the mobs for modding, and this useful! but include mobs in Minetest_Game not a right choice for me. what I prefer to add new doors, color glass, and new decorative blocks, because Besides be useful for the palace, they are great added to show the potential of Minetest engine. fix bugs, improve access nodes, better organize the menu never, and optimizing the game engine, these are the important things! for everything else there are the mods!

[GreenDimond]

nothing mobs, will never be added, resign. I agree with the choice of developers! in the future make an API for the management of the mobs for modding, and this useful! but include mobs in Minetest_Game not a right choice for me. what I prefer to add new doors, color glass, and new decorative blocks, because Besides be useful for the palace, they are great added to show the potential of Minetest engine. fix bugs, improve access nodes, better organize the menu never, and optimizing the game engine, these are the important things! for everything else there are the mods!

I agree.

[orwell]

Why don't we have a separate forum for minetest-game?

[paramat]

We do have plans for optional simple mobs in MTGame.

[Alcyone]

We do have plans for optional simple mobs in MTGame.

Could be interesting :)

In fact, it appears that they are some mobs mod, but almost all the mobs are dumb and limited, that's really the main problem. I don't even feel danger to go out during night whatever naughty mobs there are outside and don't really feel a very interesting system to play with the neutral/kind mobs due to the fact that they are limited.

I don't know what is really the cause of that : relevant mob API missing ?

What is you opinion on that :) ?

[sofar]

I don't know what is really the cause of that : relevant mob API missing ?

I've spent a lot of time looking at the existing mob "APIs" but they're very minimal. None of them create a truly neutral API that allows mobs to use the same framework without constantly overriding default functions with more specialized functions, which is an unmaintainable situation. This is the reason that there are so many "mob APIs", but in effect they're not APIs at all - they're just mobs disguised as APIs. Any new monster means you'll have to copy most of the code and make major code adjustments if you want your monster to behave differently.

There's also no easy way for monsters to do scenario-based activities (e.g. find a specific node, dig it, find a different node, place the picked up node there) or change behavior (e.g. give up fleeing and return to an original position) on the fly, without again custom programming very long routines.

This is the reason I wrote entity_ai, since it solves exactly this problem:

- remove code from the API itself
- allow code to be replaced for any functionality (drivers, e.g. "flee", "eat", "roam")
- allow entities to behave differently in states (state driven behavior)
- allow easy creation of conditions that makes entities go from one state to another (get hit -> flee)

This allows us to create monsters without code, and code without monsters:
- code to find food nodes
- code to nicely walk from A to B
- code to hit a player
- conditions to handle getting hit by a player, and then code to run away frmo that player
- conditions to handle getting burnt by lava, and run away from it
- conditions to handle spotting a nearby player, and then attacking it
- code to find and target other entities (e.g. sheep herding themselves)
- code to find and target victoms (for attacking players)
- code to find and target villains (to flee from)
(etc)

So that way we can make new monsters by creating a state model and defining conditions, for example, a simple sheep would look somewhat like this:

- roam state: driver: randomly walk around for X seconds
condition: get hit: change to flee state
condition: near food node: change to eat state

- eat state: driver: consume a node
state end: return to roam state

- flee state: driver: run away from whatever hit it
condition: fleeing too long: change to roam state

As you can see, we can define the sheep's behavior in a very subtle and precise way and make it change intelligently (the sheep won't stop to eat if it's fleeing) without any custom code blocks.

All of this is supported by entity_ai already, it just needs maturity, testing and some more driver code to handle various new drivers (aggressive is missing, abilities are somewhat limited and need more structure), but nobody else appears interested in solving the reason why other mob apis are failing to gain traction.

And ultimately that's why the currently most used mob apis will fail to ever make it into minetest_game.

[Alcyone]

Thank you very much sofar for all your explanations :D !

I think I understand better the matter, it was really missing an unified work to create a good, maintainable and overall improvable API.

I'm a bit surprise that, seeing all the mob mods which exists, that nobody among these mod developers is interested by the matter if you already tried to explain them what you are trying to create, even if I imagine how difficult it could be to design an API which is able to deal with so much different behaviours (like for an interesting use of a subtle breeding system for sheep for example) to let the possibility to create sophisticated mobs with it.

Sadly I'm not a developer yet :/ I can't help you even if I would like to (but if there is something a non-developer could do, tell me)

[taikedz]

I've been hosting a couple servers with a fair few mobs on them, all based on mobs_redo as many currently are.

The main problem I have been seeing is the necessity of some to fork copies of mobs_redo where they cannot easily override / extend the functionality without re-writing some of the custom code (I am thinking most notably of goblins, esmobs and nssm, which all fork mobs_redo in some capacity and create some headaches in compatibility and maintenance...)

I am thinking this is probably more a problem of modularization of the mobs_api code - a lot of it is hard coded all together, hidden away internally, rather than left in the open to extend, append-to, or replace.

What I did want to say is - a lot of the mobs I come across at the moment are based on mobs_redo (or maybe that's just some sort of confirmation bias I have)

What workflow is going to have to happen once a new, "better" API comes along? With more and more mobs using mobs_redo, a standard is being created...

I would more likely focus on refactoring mobs_redo so that it becomes more like a "true API" as you would say, as opposed to a rigid mob engine that needs forking for truly original mobs...

Providing a base table definition of a mobs is a nice way to have handled mobs, it just lacks a whole number of "register_*" calls to add/append and control custom every aspect of the mobs from outside the API's code so to speak...

[sofar]

Providing a base table definition of a mobs is a nice way to have handled mobs, it just lacks a whole number of "register_*" calls to add/append and control custom every aspect of the mobs from outside the API's code so to speak...

my approach solves this by moving the code to `drivers` and `factors`, which are then attached to monsters in the state table. These driver codes can be inserted by any lua mod, and they literally work in the form of `entity_ai.register_driver(name, driverdef) which is all too familiar.

That allows one to provide code bits that handles a state driver that makes the monster spawn a baby monster, and two factors that automatically make the monster proceed through the baby making process:

factor: has been fed breeding food -> switch to "find_mate" driver
factor: has found suitable mate -> switch to "breed" driver
breed: produce offspring. -> switch back to normal driver

Now we have 3 simple drivers and 2 factors that can be re-used by other monsters without writing a single line of code.

[GreenDimond]

my approach solves this by moving the code to `drivers` and `factors`, which are then attached to monsters in the state table. These driver codes can be inserted by any lua mod, and they literally work in the form of `entity_ai.register_driver(name, driverdef) which is all too familiar.

That allows one to provide code bits that handles a state driver that makes the monster spawn a baby monster, and two factors that automatically make the monster proceed through the baby making process:

factor: has been fed breeding food -> switch to "find_mate" driver
factor: has found suitable mate -> switch to "breed" driver
breed: produce offspring. -> switch back to normal driver

Now we have 3 simple drivers and 2 factors that can be re-used by other monsters without writing a single line of code.

Amazing. Makes sense too.
When will mobs be added? Even a simple mob like a pig or something? Any hint whatsoever on when they are planned?

[sofar]

Amazing. Makes sense too.
When will mobs be added? Even a simple mob like a pig or something? Any hint whatsoever on when they are planned?

There already is a simple sheep and a stone monster. The sheep behaves pretty good already (flees when hit, and is really hard to catch again) but doesn't drop anything on death yet.

[GreenDimond]

Amazing. Makes sense too.
When will mobs be added? Even a simple mob like a pig or something? Any hint whatsoever on when they are planned?

There already is a simple sheep and a stone monster. The sheep behaves pretty good already (flees when hit, and is really hard to catch again) but doesn't drop anything on death yet.

Wait, there is? How have I never seen them? :o

[pandaro]

This is the reason I wrote entity_ai

I cant find this mod, where is repo or link?

[Warnerfam]

One word: "Nice"

Translations are still missing, please suspend the release until translations are done.

About the minimap, this is great
Just curious, how does the minimap gets the colour of texture? Client-side or server-side?

For shaders, I can not try it. :

[OldCoder]

0.4.14 was a few years ago :-)

The new group has released Final Minetest 5. Try it and see if it works for you:

https://minetest.org/

https://minetest.org/minefiles/