[Mod] Pipeworks [git] [pipeworks]

[mtmodder148]

Just reporting this bug here too.
https://github.com/VanessaE/pipeworks/issues/78

[CrisZ03]

thank you

[Jesseman1]

Hello, I'm having a problem with this mod; I set up a small pump that goes through my forcefield (technic mod), then connects to a switch before ending with a spout. after finishing the pump and saving the game, I came back later but found that when I turn off the switch, the water empties from the pipe at first, but then refills again. The same thing happens when I turn off the pump itself too. Also, the spout continues to pour water even when I destroy the pipe supplying water to it, and the water source block that appears directly beneath the spout doesnt go away after I destroy the spout. Can someone please help me? (In case it's relevent, the world is in creative mode, and I also play it in both singleplayer and multiplayer.)

[VanessaE]

Please file a suitable bug report on the pipeworks git repository:

https://github.com/VanessaE/pipeworks/issues/new

[Jesseman1]

Please file a suitable bug report on the pipeworks git repository:

https://github.com/VanessaE/pipeworks/issues/new

I don't have a github account...

[mtmodder148]

This is the same bug reported here. https://github.com/VanessaE/pipeworks/issues/78

[VanessaE]

I'll get those bugs sorted out soon. Meanwhile, big appearance upgrade: All nodebox-based pipe-related nodes in this mod have been converted over to mesh nodes (all of them made by hand, starting from Blender's geometric primitives).

No more ugly jagged-edged pipes! Requires Minetest 0.4.11 or newer (otherwise they'll look like oddly-textured cubes).

This should also result in a small FPS improvement. I'm pretty sure the total number of faces/polys is far lower now than it was with nodeboxes, and there should be fewer textures as well.

[mtmodder148]

Yay for meshnodes! Thank you for your hard work. I figured you were working on these bugs so I'm being patient. I know a work around so it's not too much trouble.

[Minetestforfun]

Thank you for your work !

[VanessaE]

Between t4im and kaeza's patches, this mod just got a major speed boost in its sorting and teleport tubes. Enjoy!

[mtmodder148]

I have a question about which method is better?

Code: Select all

after_place_node = pipeworks.after_place,
after_dig_node = pipeworks.after_dig

or

Code: Select all

after_place_node = pipeworks.scan_for_pipe_objects(pos),
after_dig_node =pipeworks.scan_for_pipe_objects(pos)

[VanessaE]

Both are exactly equivalent at this point, but the former is better because it allows other mods to possibly hook into pipeworks' route-on-place/dig functions.

[VanessaE]

Due to a small change in the way pipeworks handles tube autorouting internally, it's introduced an extra requirement to make tubes auto-route to your items when they're placed. This is due to the removal of global register_on_placenode and register_on_dignode callbacks. Instead, per-node callbacks are used. If they aren't already present in your mods that use pipeworks, just add:

Code: Select all

pipeworks?

...with the question mark (hence optional) to your depends.txt, and this line:

Code: Select all

local use_pipeworks = minetest.get_modpath("pipeworks")

...to the top of your init.lua. Then add these two lines:

Code: Select all

after_place_node = use_pipeworks and pipeworks.after_place,
after_dig_node = use_pipeworks and pipeworks.after_dig,

...to each of your nodes that need tube support.

[mtmodder148]

Thanks VanessaE. Oh one more question what's the purpose of the inventory slot in the node breaker?

[VanessaE]

That's for when you want the nodebreaker to dig with some tool other than the built-in mese pick. For example, you could put a tree tap or a shovel in there, and the nodebreaker will dig using that, with all the effects that tool should have on the target.

[mtmodder148]

Wow! You guys are awesome with lua. That's a great feature. Thanks again.

[VanessaE]

I'd like to ask a favor of all singleplayer pipeworks users and server admins who have authority to alter servers' mod settings.

I've been working off and on on a branch in this mod called "new_flow_logic", where water in pipes is tracked and moved around explicitly and pipes change their visible state to reflect the presence of that water, rather than just pretending there's water and changing state if a neighbor does.

Originally I started that branch when Minetest gained finite water support, and lost interest when that support was dropped, but I decided to pick it up again. With much help from est31 and t4im (namely rebasing it and getting the code working again), I've brought it up-to-date and expanded it. Everything that works in master now seems to work in the new branch.

A few bugs in the old logic have been fixed here as well, such as flow sensors, valves, and entry panels picking up crossflow from a not-connected pipe above/below them, and fountains and spigots not shutting off when you dig the pipe feeding them. Also in this branch, valves, sensors and panels will re-orient themselves if a screwdriver gets used on them and they end up in an unusable orientation. The only things I haven't done anything with yet are the tanks and gratings, but they don't do anything in the "master" branch either.

Anyway, please try out the "new_flow_logic" branch and let me know how it behaves with your existing maps and pipelines as well as whatever new stuff you come up with. As the new logic uses metadata to track the water content of the pipes, and ABMs are still used to drive the pumps, spigots/fountains, and pressure averaging, there are several things to keep in mind that differ slightly from the master branch:

1. water propagates a little slower through the pipes, but not by much.
2. since water is stored as sort of a "pressure" value now, a pipeline that seems to only be partially "on" will still drive a spigot or fountain for a while after the pump is turned off.
3. placing too many spigots/fountains sequentially down a pipeline from a pump will cause the water to flow out of the ones closest to the pump first, leaving inadequate pressure to drive the ones that are further down the line.
4. existing pipelines that were showing as "on" before will probably revert to empty initially, and will take a while to "refill" as the pipes' metadata catches up.
5. pipes no longer just arbitrarily shut off: since a pipeline behaves as if it's always capped (as far as water flow is concerned anyway - we don't care about air that should theoretically be trapped in the pipe), a spigot or fountain must be attached to a pipeline in order to "drain" the pressure off.
6. A pump that's turned off will also gradually bleed-down the pressure in the circuit it is attached to.

[mtmodder148]

This branch seems to be very hacky. One problem I encountered was with water flowing into the next pipe, it flashes full and empty, occasionally stopping it's flow to stall for a long time or not flow at all.

[VanessaE]

That probably just means I need to adjust the "shows as full" threshold. Similarly, the spigots/fountains need a similar adjustment.

[mtmodder148]

Okay cool. I love this change to a more realistic pressure value.

[VanessaE]

Some changes just went up in new_flow_logic that should greatly improve those flow propagation issues. Mind you, pipes will still seem to turn on and off from time to time, but it should work better now.

[mtmodder148]

It's a little better. Sometimes pumps don't work and I have to dig it up and replace it a few times. Also after placing pipes some distance from the pumps the "pressure" is too low for it to continue flowing and the same bug as before occurs. They really do turn off; I had a spigot at the end and the spigot kept turning on and off with the pipes. Is there a limit to how far pipes can be placed from a pump?

[VanessaE]

There's no limit in my code, though the engine will limit your distance to whatever the ABM range is (I think 3 mapblocks), but the further downstream a spigot is, the longer it will take for the pipes to build up enough pressure to keep the spigot running. Once it's running continuously, it'll continue to do so until the pressure drops too low again.

[mtmodder148]

Okay, I'm talking like no more than ten nodes away :). I'll test for a longer period of time. How long do you think it should take for let's say 6 pipes to continuously stay running?

[VanessaE]

Maybe 10-15 seconds, assuming you're actually on the new_flow_logic branch ;-) and have the latest code.

(I never actually timed it, just subjectively compared it to the old behavior)

Can you give me a screenshot of the malfunctioning pipeline?