[Modpack] Plantlife [rolling release] [plantlife_modpack]

[VanessaE]

Well buildable_to just makes it easier is all. Besides, I expect that most people probably just throw the extra plants in the trash anyway. :-/

Any comments on the speed of the new code? To me, it feels just orders of magnitude faster, but I'm really curious if anyone can do some objective comparisons between this version and any copy from Aug 11 or older.

[Sokomine]

Well buildable_to just makes it easier is all. Besides, I expect that most people probably just throw the extra plants in the trash anyway. :-/

That's forbidden :) Unless, of course, it's a proper compost heap.

[VanessaE]

HAH, a compost heap! This in the same game where people can't even be bothered to re-plant their saplings. You're funny. :-)

[VanessaE]

I've just pushed a small update that fixes a memory leak that Zeno` caught. Block caches weren't being released after being used. I don't know the exact figures (since this depends on Lua's internals) but I would guess from some quick googling that each mapblock would take around 500 kB to 1 MB in the block cache, so the leak will stack up fast. :-)

I recommend all users and admins update right away.

[VanessaE]

Well after sapier's tests with the profiling/tracing code he's working on, it appears that that big update (as used under Dreambuilder anyway) is more like only about twice as fast as before. The previous 10-15x figures were in error. Bummer :-(

So not nearly as fast as I thought, but at least it's markedly faster than before.

Still, more improvements will follow as I find little things I can tweak.

[VanessaE]

More performance tweaks went in today. Hopefully they result in another measurable improvement.

EDIT: And another set of tweaks went in. Probably won't result in significant improvements, but they'll likely show up in a profiler, and they are at least better coding style: made all mods within the pack pass their function calls to plantslib:register_generate_plant() as actual functions instead of strings where possible (only the strings that go to the four legacy mapgen hooks remain now).

[Minetestforfun]

Thank you for your work VanessaE, i update my plantlife mod on my server :)
i hope it will help a bit the little laptop of my girlfriend, for getting more FPS :p

[VanessaE]

Hope it helps :-)

[VanessaE]

And the last set of performance tweaks are up on git (filter out undefined spawn and surface nodes, and also the mod now prints final counts after everything has been registered).

I don't think there's much more I can do to speed it up further. Enjoy anyway! :-)

[Tom]

Thank you Vanessa!

[VanessaE]

You're quite welcome. Another minor tweak just went in, removed change that "fixed" a non-existent memory leak (forgot to account for scoping, so Lua would have freed the memory anyway).

[Inocudom]

https://github.com/VanessaE/mushroom
You do have this mod in your Dreambuilder sub-game, but it might get more use if you add it to the plantlife modpack. It might go well with the custom dirts of undergrowth too and have the potential to add life to the underground caverns.

[VanessaE]

I've just pushed another update to this modpack - fixed a few bugs in ferns and young trees, and more importantly, rewrote the "legacy" mapgen code in plants_lib a bit to use the same "action loop" method that the standard code uses.

Should result in another small speed boost.

[VanessaE]

Inocudom: Actually that's a good idea, and so it has been done.

In plantlife git is a slightly rewritten version of that mod, with Undergrowth/woodsoils support and slightly ligher-weight ABMs than before (they still function the same, they just do their job a little better).

This update has been merged into Dreambuilder also.

Unlike the original mod, the source code of which is that repository you linked to, this one continues to lack those "giant" mushrooms (I frankly think they just don't work well for this modpack or my game).

Also, the recipe for the spore identifier has been changed. It was just a single steel ingot. Now, it's:

- T -
S M S

(T = Torch, M = Mese crystal fragment, S = steel ingot)

Also: I've phased out the old junglegrass-based pie recipes. We have the farming mod in minetest_game, there's no reason not to use its flour where appropriate.

[VanessaE]

Update: added to flowers_plus, a potted cactus and a potted bonsai tree. The latter can be crafted from a flowerpot and a regular sapling. The former...take a guess. ;-) Contributed by jp.

[Minetestforfun]

Update: added to flowers_plus, a potted cactus and a potted bonsai tree. The latter can be crafted from a flowerpot and a regular sapling. The former...take a guess. ;-) Contributed by jp.

Hi,

Thank you again for your good work VanessaE, i've an error when i add your last commit :
http://hpics.li/4f78328

please, can you fix it ?

[Inocudom]

Update: added to flowers_plus, a potted cactus and a potted bonsai tree. The latter can be crafted from a flowerpot and a regular sapling. The former...take a guess. ;-) Contributed by jp.

Hi,

Thank you again for your good work VanessaE, i've an error when i add your last commit :
http://hpics.li/4f78328

please, can you fix it ?

I got that error too.

[VanessaE]

Fixed in git. Sorry about that. "bonzai!" ~= bonsai :-P

[VanessaE]

Update: Two things:

First and foremost, a big thanks goes out to Zeno` for all the hard work he's done over the past days on rewriting the Ferns mod in this modpack. It has been significantly refactored and improved and should be faster than before, and much cleaner code as well. In addition, most ferns now randomly rotate as they're spawned.

This random rotation is also a new feature in the mapgen routines in plants_lib (used to only apply to the spawning ABMs), and can be used with any node that uses param2. Rotation of the plantlike draw type is a feature in the Minetest engine which will only work in the newest development builds (servers and clients that are too old will just see the standard 45-degree rotation).

Second, plants_lib can now accept a table of node names in its mapgen registration calls, from which it will pick randomly at spawn time (this is in addition to accepting a single node, a function, or an L-systems tree definition). Those ferns rewrites use this new feature.

Other mods in this pack will be adapted soon to follow these features.

[balthazariv]

Hello Vanessa,

I have this error :

Code: Select all

17:13:26: ERROR[main]: ========== ERROR FROM LUA ===========
17:13:26: ERROR[main]: Failed to load and run script from 
17:13:26: ERROR[main]: L:\Minetest\minetest - Copie\bin\..\mods\plantlife_modpack\trunks\init.lua:
17:13:26: ERROR[main]: ...opie\bin\..\mods\plantlife_modpack\trunks/generating.lua:6: bad argument #1 to 'pairs' (table expected, got nil)
17:13:26: ERROR[main]: stack traceback:
17:13:26: ERROR[main]: 	[C]: in function 'pairs'
17:13:26: ERROR[main]: 	...opie\bin\..\mods\plantlife_modpack\trunks/generating.lua:6: in function 'clone_node'
17:13:26: ERROR[main]: 	...opie\bin\..\mods\plantlife_modpack\trunks/generating.lua:233: in main chunk
17:13:26: ERROR[main]: 	[C]: in function 'dofile'
17:13:26: ERROR[main]: 	...st - Copie\bin\..\mods\plantlife_modpack\trunks\init.lua:11: in main chunk
17:13:26: ERROR[main]: ======= END OF ERROR FROM LUA ========

[VanessaE]

This error means that the name of an undefined node is being passed to the clone_node() function call. Could be a typo, or a mod that is being called upon that isn't actually installed.

Since I can't reproduce that error here, I need you to make a small code tweak.

Using notepad or some other text editor, open plantlife_modpack/trunks/generating.lua, go to lines 3 and 4 of that file. Please add print(dump(name)) and reproduce the error again. I.e. make the first few lines of that function look like this:

Code: Select all

local function clone_node(name)
	print(dump(name)) 
	node2={}
	node=minetest.registered_nodes[name]
	...
	...

Tell me what node name it prints just before the crash message, or just copy&paste the whole log section for that run (not the entire debug.txt of course), put it on a pastebin, and I'll find it in the log output.

[balthazariv]

I've had print(dump(name)) function and my partial debug.txt is :

Code: Select all

18:04:56: ERROR[main]: ========== ERROR FROM LUA ===========
18:04:56: ERROR[main]: Failed to load and run script from 
18:04:56: ERROR[main]: L:\Minetest\minetest - Copie\bin\..\mods\plantlife_modpack\trunks\init.lua:
18:04:56: ERROR[main]: ...opie\bin\..\mods\plantlife_modpack\trunks/generating.lua:7: bad argument #1 to 'pairs' (table expected, got nil)
18:04:56: ERROR[main]: stack traceback:
18:04:56: ERROR[main]: 	[C]: in function 'pairs'
18:04:56: ERROR[main]: 	...opie\bin\..\mods\plantlife_modpack\trunks/generating.lua:7: in function 'clone_node'
18:04:56: ERROR[main]: 	...opie\bin\..\mods\plantlife_modpack\trunks/generating.lua:234: in main chunk
18:04:56: ERROR[main]: 	[C]: in function 'dofile'
18:04:56: ERROR[main]: 	...st - Copie\bin\..\mods\plantlife_modpack\trunks\init.lua:11: in main chunk
18:04:56: ERROR[main]: ======= END OF ERROR FROM LUA ========
18:04:56: ERROR[main]: Server: Failed to load and run L:\Minetest\minetest - Copie\bin\..\mods\plantlife_modpack\trunks\init.lua
18:04:56: ERROR[main]: ModError: ModError: Failed to load and run L:\Minetest\minetest - Copie\bin\..\mods\plantlife_modpack\trunks\init.lua

[VanessaE]

I need the part of the log preceding this error, where the errant node name is printed.

[balthazariv]

Sorry, i didn't understand.

Code: Select all

"default:tree"
"default:jungletree"
"trees:tree_conifer"
"trees:tree_mangrove"

[VanessaE]

Ok, remove your change and try updating plantlife. See if the change I just pushed fixes it.