Page 1 of 2

-TESTERS NEEDED- for OpenCS Terrain Editing

Posted: 22 Sep 2019, 17:32
by silentthief
Work on the OpenCS has had a recent breakthrough with unelsson having made land editing now possible. We can now make new land, or modify the original. We need testers to see if this works for you, if there are any issues, and what OS you got it to work on.
unelsson wrote: Land shape editing is ready for testing
Please test it, try to break it, give me feedback. It should be functional, shouldn't crash.
Testing and feedback is most certainly useful, and if you feel like asking for testers, go on!

- how does it work on different platforms? (Mac, win... I've tested myself with Manjaro Linux)
- how does it feel to use?
- any land breaking or other bugs?
- other suggestions?
***The pull for this is found at this link ->

EDIT - land editing should now be merged into masters, so test it and let us know if you have any issues!

Big thanks go out to Unelsson for this pull and the continued work on the CS.

ST *really hyped for land editing*

Re: -TESTERS NEEDED- for OpenCS Terrain Editing

Posted: 22 Sep 2019, 22:42
by Ravenwing
This is great news! If I had a build environment set up I would love to test it. Once it's merged and part of the nightlies I'll certainly be able to test it out and give feedback. Hope some others are able to give some helpful feedback. Great job unelsson!

Re: -TESTERS NEEDED- for OpenCS Terrain Editing

Posted: 23 Sep 2019, 01:27
by AnyOldName3
Remember that if you're on Windows you can use the AppVeyor artefacts produced for a PR or MR.

Re: -TESTERS NEEDED- for OpenCS Terrain Editing

Posted: 23 Sep 2019, 01:31
by Ravenwing
Damn, I've been out of the loop for so long I'd completely forgotten that was an option! Will try tonight!

Re: -TESTERS NEEDED- for OpenCS Terrain Editing

Posted: 20 Oct 2019, 22:19
by Lamoot
It's great to see this developed, being one of the essential editor features. I can't provide feedback testing it until there are builds available, but I can offer some ideas and observations.

1.) From the latest terrain editing video it looks like there is already a solid basic set of brushes available. Perhaps they could be named differently, but other than that, things look very usable. Name suggestions:
* Raise (height raise)
* Lower (height lower)
* Offset (height drag)
* Smooth (smooth)
* Flatten (flatten)

2.) To make terrain editing more streamlined, CS could borrow ideas from 3d sculpting apps (Blender, Zbrush). There you can sculpt regularly with a brush, or by holding CTRL you get the inverse effect of the same brush. In addition, while any brush is active, if you hold SHIFT, the smooth brush is used instead. Not only would this help familiarity, but it's also a very quick way to sculpt, without having to switch the brushes from the menus.

3.) Another quality-of-life improvement would be to see the radius and shape of the brush in the viewport

4.) And the last one, the ability to change the size and strength without going into the menu. In Blender for example, you press a shortcut F after which you move the mouse to increase/decrease the chosen value and confirm with the LMB.

And of course, thank you unelsson for developing this!

Re: -TESTERS NEEDED- for OpenCS Terrain Editing

Posted: 22 Oct 2019, 22:37
by unelsson
Thanks for the feedback! Many ideas are noted, as well as comments on github. I think it's good to have a bit more discussion though, more ideas and planning before I rush into making a lot of changes.

1) naming requires some thought, there are other naming things too at code level too (like.. the name of the editmode in the first place)
2) Holding CTRL or SHIFT is something related to Primary Edit and Secondary Edit, how the OpenMW-CS is planned to function... Currently Secondary Edit is unimplemented for this mode. It was also suggested that a hotkey switch between modes would be implemented, and that's definitely a possibility, but it's subject for another PR..
3) tool outline, seeing the radius and shape, possibly strength, in the viewport is definitely needed, but... another day, another PR. This one should be for all editmodes, e.g. texture editing.. And texture editing should also show what texture you are painting!
4) Very good UI suggestion (the ability to change the size and strength without going into the menu).. another PR!

edit: Compared to the videos, it's also much more difficult to actually break land heights or cell edges right now. There are some evil tricks that may still do that thought.. Another PR! :lol:

Re: -TESTERS NEEDED- for OpenCS Terrain Editing

Posted: 24 Oct 2019, 19:00
by Lamoot
For sure, no need to rush things. It's already a big and welcome step having terrain editing in CS.

1.) Regardless of what name is picked in the end, it would be nice to have simple and clear names for the brushes. It's what I mainly went for with the proposal.
2.) I'm familiar with primary and secondary edit when manipulating objects around the scene, but that is disabled while in terrain edit mode? So CTRL and SHIFT could then be used for the brushes?

Re: -TESTERS NEEDED- for OpenCS Terrain Editing

Posted: 24 Oct 2019, 23:02
by faerietree
Very busy, unelsson! Thank you

@Lamoot: I'm not sure it is a misunderstanding, but I think you and unelsson mean the same? No harm meant.

What I wonder about is whether it could ease our lives to reuse blender? No more worries about the UI. Only maintain our python addons. We'd need cell edges, attach scripts to models and export models with Nif exporter and terrain with a new one - basically the algorithms we have in OpenMW Construction Suite right now?

Couldn't integrating blender as our construction suite be very powerful? People would be used to it like a charm. It would allow us to concentrate on a basic set of functionality. UNIX principle. Manjaro also builds on Arch Linux which makes it very powerful, maintainable and usable.

EDIT: I don't want to undermine current development. I am very happy about it. It is just a proposal because the critics and TODO PRs put forth concentrate on usability and sound like a lot of work to be done while blender is already very efficient in such regards. Why reinvent the wheel? Why not focus on what is essential and specific to OpenMW? Still I respect your work a lot, unelsson! Don't get me wrong

Re: -TESTERS NEEDED- for OpenCS Terrain Editing

Posted: 24 Oct 2019, 23:13
by faerietree
I'm not talking about this without background. I have coded several blender addons already (including a Bill of materials and blueprint generator and render2print. My impression is it is very powerful. Cutting the mesh into cells. But really before being able to give estimates about feasibility I need to study Construction suite code!

Currently I have other code that requires my attention. Once I tame it, I will be of more help. I will very likely even hire a team to help develop the engine if I make it. And I envision a kind of compensation for development efforts of all the volunteers. Work done must be valued. Rewards to the busy and sincere!

As another supporting view I remember words by Ykkrosh, a very capable 3D and hardware coder. If blender had been available back then when they started 0A.D., then they'd used that as a base for their editor. And if this man says something like that then it may be of value to us, because our construction suite is still in a relatively early stage.

Re: -TESTERS NEEDED- for OpenCS Terrain Editing

Posted: 25 Oct 2019, 01:43
by AnyOldName3
It's been mentioned before, but the problem with the idea is that very little of what actually gets changed in the editor has any relation to what Blender can do already. The sculpting tools could be used for terrain (provided we could add extra constraints so they produced something that worked) and object manipulation widgets could be used to move things around the scene, and if we really wanted to push the boat out, we could make mesh editing a native part of the editor to have an excuse to use the rest of Blender's feature set, but all that's a small fraction of what the CS needs to do. Most of it is just loads and loads of database manipulation, and Blender doesn't do that.