Moving Cyclops Level Builder out of the viewport editor

[blackears]

Cyclops Level Builder started out with the design idea that you could drag-create objects directly in the viewport editor. As the project has become more complex, though, I keep running into more and more issues trying to keep Cyclops integrated.

The way the Godot editor is designed prevents me from responding to input or drawing graphics the way I would like to. And a lot of users are having installation issues with the 1.0.2 release due to how Godot's autoload feature works. I'm thinking of breaking Cyclops into a separate window for block editing, or maybe making it into a stand alone program, which should hopefully resolve these issues.. Let me know what you think.

[jkulawik]

I think turning it into a new tab (along 2D, 3D and Script at the top of the editor window) would be fine. I think that turning it into an external program removes the biggest advantage Cylcops has over Qodot+Trenchbroom, which is that it doesn't need an asset import pipeline.

I imagine that you would need to turn data into a resource that would be edited in that new tab (editor) and Cyclops nodes would hold those resources. One disadvantage this whole thing has is that you might have to essentially reimplement the whole 3D editor, but I haven't done anything like that and can't tell how much is reusable. Speaking of reusability, is it not possible to reuse Godot's built-in gizmos?

[mrezai]

I think same as jkulawik, the biggest advantage of Cylcops over Trenchbroom and Qodot combination is its editor integrated nature.
My suggestion is to list your problems and discuss about them in https://chat.godotengine.org maybe there are some proposal, API or workaround for them.
For graphics part of problems maybe this proposal and this pull request could be helpful.

[blackears]

A separate main window within the Godot editor I think makes the most sense. I'll need to see how feasible it is.

[jkulawik]

As to the Autoload issues, I believe that #121 should resolve them. I might implement it myself but would like some input on it first because I suspect there is more to it than meets the eye.

[blackears]

I liked your suggestion - I've not had time to look into it yet, though. Hopefully it will make the installation easier. I hope to get to it before next week.

[jkulawik]

You know what, I also had another idea. Perhaps you can just make this a Blender plugin? A large portion of the tooling is already there. Collision can be added by Godot's glTF importer. All it really lacks is the automatic UV handling, perhaps some snap options and a faster way of adding new blocks instead of Shift+A > Mesh > Cube.

[blackears]

I've already had a go at it in Blender. But I found Blender even harder to work with than Godot. Blender's Python API is really unpleasant to work with and missing a lot of features. You're also working in a text editor and having to reload all the time. It's UI controls are an after thought - you can't even use them when a tool is engaged. I'd much rather work in Godot. I have made a suite of UV tools for Blender, though: https://blendermarket.com/products/uv-tools

[Roosader]

If you made this standalone, I'd LOVE to see this on iOS/Android. I had asked you on Ko-Fi back in like May about making some touch friendly changes like making the points use screenspace (which you did which is pretty cool), but I think even just straight up porting this tool to those platforms and just recommending users to use a stylus without making really any significant changes (to limit the amount of work) would still be sufficient.

[blackears]

I'm focusing on stability and some core improvements at the moment. My main priority is desktop, but I know a lot of people would like this to work for tablet too. Maybe I can add a special mode for tablet, but that is likely going to be a while off.

[Roosader]

I'm not sure if it would even really be needed for you to add a special mode for tablets. Maybe just an as-is release with minimal changes like larger UI targets. Even released with no changes at all, I think it'd be super usable. If you'd be planning on keeping this open source, I'd even look into making some slight tweaks myself for that purpose and then sharing it as a fork or something. Either way, I really appreciate the work you've done on this and I think it's quite impressive :).