If the loaded shader has uniforms, they will be exposed to the UI, so the users will be able to modify the shader parameters easily. Currently only works with float uniforms.
This new feature allows users to load a .shader file (must be a GLES2 Godot shader) that will modify the image itself. This feature is experimental and possibly buggy, not all shaders are working properly and I'm not sure yet why. As such, this feature may not be included in v0.8 stable.
First option is to affect the selected pixels only. The second it to affect the current cel, or the entire frame (all cels of the frame). Options to affect all frames and all projects will be added next. I also made changes to Canvas.handle_undo() and handle_redo() to make this work.
Once all these options are added successfully in HSVDialog, they will also be added in the rest of the Image effect dialogs.
This will allow us to create more selection tools in the future, that aren't necessarily rectangular (See #129) and even enhance the current rectangle selection tool (See #56)
Current issues spotted so far:
Drawing is slower for large images, and bucket filling is also considerably slower even on a 64x64 image. Optimizations are required.
Also fixed issue with "Show Guides" view menu option and having multiple projects with guides. Only think remaining is to make the Symmetry Guides look different than regular guides. Closes #133.
Two special guides - one horizontal and one vertical - that let you change the axis of symmetry for mirroring. On the next commit(s) I will make them visible only if mirroring is toggled on. Almost solves #133
The default image width, height and fill color are being used only when the program first launches. After that, if the user creates an image with changed settings, these settings are being kept.
* Refactoring tools
* Remove unused code
* Fixed some inferring errors and added translations
* Attempt to fix some Script Errors found in the CI workflow
* Fix bucket crash.
* Fix static type convert.
Co-authored-by: OverloadedOrama <35376950+OverloadedOrama@users.noreply.github.com>
Had to move some palette png code around in order to make it possible to use these image files as palettes. Hopefully loading .pxo files should be next. Addresses #280
This helps in previewing how each spritesheet frame will look when it will be imported in Pixelorama. Note that this does not work well with non-square images at the moment.
More specifically, moved scale, crop, invert, desaturate and generate outline code to DrawingAlgos.gd. I am a little worried of DrawingAlgos.gd getting too big though.
Instead of setting the main arrow as a custom cursor every time the cursor gets inside the viewport, set the viewport's default cursor to be the cross, and change the cross's image once in Global.gd
The "Open" option will handle importing of .pxo files and image files from now on. The "ImportSprites" file dialog node and script are scheduled to be deleted too. Importing spritesheets option will return soon. Next thing I'm going to work on is a preview window that will appear when importing image files.
The structure of the .pxo files is now consisted of a JSON-structured metadata part, where all the data that can be stored as text are, and a binary part, that contain all the actual image data for each cel and project brush.
This makes it easier for users to understand the .pxo structure, easier to add more changes without having to check versions for backwards compatibility, easier to be opened by third-party apps and it allows us to make an "Export JSON metadata" option, that will export just the metadata in JSON format, without the binary image data.
It's backwards compatible and .pxo files from as far as v0.5 are still supported.
A Project class contains project-specific data like name, undo_redo, frames, layers, tags and brushes. These variables have been moved from Global. This is the first step towards multiple tab support, where each tab will be a different Project.
The Canvas is now single node, instead of having multiple canvases for each frame. This should also be a performance optimization, since there are less canvas nodes, so there are less _input() calls. It should also fix a rare Undo/Redo issue with motion drawing and making lines. Could be unstable, needs more testing.
As a side effect, the guides are now the same for all frames, so this should also close #246.
Replaces nested Global.animation_tags arrays. Also replaced array.duplicate(true) with looping through the array and creating a new class for each array element, because duplicate(true) does not create new classes, unfortunately, which was causing issues with undo/redo.
ColorAndToolOptions has now the same code for left and right tool options, with more similar refactoring coming soon to places like Canvas and DrawingAlgos
Instead of having 2 variables for left & right nodes, use an array instead. This will help with better looking code, automation and less repetitive code, as seen in ToolButtons.gd. Move related refactoring will follow.
Moved code from Global.gd and Canvas.gd to DrawingAlgos.gd. Will also move the fill_gaps and draw_brush methods of Canvas.gd next. Maybe even refactor the inside of them a bit to make them easier to read.
* Refactoring Main.gd. Mostly cutting big methods into smaller one.
- Reduced size of _ready method in Main.gd
- Moved code from certain parts of old _ready method into seperate methods
- Fixed the translation bug related to CurrentFrame node in TopMenuContainer scene. The CurrentFrame node wasn't updating the language when I was changing language. I've also changed the translation file for this.
- Fixed Global.palette_option_button.selected related warning. Because of some unknown reasons, git didn't push completed line there.
- Moved code from file_menu_id_pressed and view_menu_id_pressed method in Main.gd to separate methods to make it more readable.
* Removed window_title changes from Main.tscn
Co-authored-by: OverloadedOrama <35376950+OverloadedOrama@users.noreply.github.com>
Removed the tool option signals from Main.gd and put them to ColorAndToolOptions.gd. Instead of having 2 methods for left and right, they are now in one method, using a boolean to differentiate between left and right.
* New remove current palette feature
This commit adds new feature to Pixelorama that will allow user to remove the current selected palette.
* Fixed grammar and added focus_mode = 0 on the remove palette button
Co-authored-by: OverloadedOrama <35376950+OverloadedOrama@users.noreply.github.com>
It used to save the linked cel array which contained the canvas nodes into the .pxo. This, however, does not make sense, because the canvas nodes are being recreated when a .pxo is loaded. Therefore, they are different nodes than they used to be.
Now, an array containing the frame numbers which have cels linked for each frame is being saved and loaded into the pxo.
* Added selective frame tag export for Spritesheets
* Changes LinesCount value and max_value to canvas_size
This does not allow the spritesheet to contain more frames than the tag has, because it limits its size to the number of frames the currently selected tag has.
Co-authored-by: Manolis Papadeas <35376950+OverloadedOrama@users.noreply.github.com>
Removed separators entirely. Also made the AnimationTimeline have a StyleBoxFlat of its own, which we use to expand its margin top to cover the little space left because of the VSplitContainer. We also use a TextureRect as a fake VSplitContainer grabber.
Pixelorama now requires at least Godot 3.1.2 or 3.2.x to work, because TranslationServer.get_loaded_locales() was not included in older versions. If you're using an older version of Godot, please update. Godot 3.2.1 is the currently recommended version, until 3.2.2 is out.
Also updated the list of the translators.
