![Logo](https://imgur.com/7NdnoFl.png")
MaterialFX is an open source Java library which provides material design components for JavaFX
Explore the wiki »
Download Latest Demo
·
Report Bug
·
Request Feature
- Important Notes!
- About the Project and History of JavaFX
- About the Logo
- Some GIFs
- Getting Started
- Documentation
- Changelog
- Roadmap
- Theming System
- Contributing
- License
- Contact
- Donation
- Supporters
Please, before using this library and submitting an issue complaining that controls are not styled and bugged check how the styling system has changed since version 11.14.0
JavaFX is a software platform intended to replace Swing in creating and delivering rich client applications that operate consistently across diverse platforms. With the release of JDK 11 in 2018, Oracle has made JavaFX part of the OpenJDK under the OpenJFX project in order to increase the pace of its development.
Key features:
- FXML and SceneBuilder, A designer can code in FXML or use JavaFX Scene Builder to interactively design the graphical user interface (GUI). Scene Builder generates FXML markup that can be ported to an IDE where a developer can add the business logic.
- Built-in UI controls and CSS, JavaFX provides all the major UI controls required to develop a full-featured application. Components can be skinned with standard Web technologies such as CSS.
- Self-contained application deployment model. Self-contained application packages have all the application resources and a private copy of the Java and JavaFX runtimes. They are distributed as native installable packages and provide the same installation and launch experience as native applications for that operating system. JavaFX is a software platform for creating and delivering desktop applications, as well as rich Internet applications (RIAs) that can run across a wide variety of devices.
Over the years the way of creating GUIs has often changed and JavaFX default appearance is still pretty much the same. That's where this project comes in. The aim of my project is to bring components which follow as much as possible the Google's material design guidelines to JavaFX. The second purpose is to provide a successor to the already available JFoenix library, which is a bit old and has a lot of issues.
In recent months the project has evolved a lot, to the point that it is no longer a simple substitute.
To date MaterialFX offers not only restyled controls, but also: new and unique controls such as the Stepper,
controls completely redone from scratch such as ComboBoxes or TableViews (and many others),
and many utilities for JavaFX and Java (NodeUtils, ColorUtils, StringUtils ...).
MaterialFX v11.13.0 brought a lot of fixes and new features, but it also brought a new logo, something that is more
meaningful for me and that somewhat represents the new version.
The new logo is a Phoenix, the immortal bird from Greek mythology, associated to regeneration/rebirth.
When a Phoenix dies it obtains new life by raising from its ashes.
MaterialFX v11.13.0 fixed many critical bugs and broken features, I like to think that it is reborn from
the previous version, so I thought a new logo would have been a good idea.
Imgur Link: Gallery
Buttons
Check Boxes, Radio Buttons and Toggles
Combo Boxes
Dialogs
Fields
Lists
Notifications
Pickers
Progress
Scroll Panes
Sliders
Stepper
Tables
In this section you can learn what do you need to use my library in your project or see a preview/demo which I'm planning to release as runtime images here on github.
To build MaterialFX, execute the following command:
gradlew build
To run the main demo, execute the following command:
gradlew run
NOTE: MaterialFX requires Java 11 and above.
repositories {
mavenCentral()
}
dependencies {
implementation 'io.github.palexdev:materialfx:11.17.0'
}<dependency>
<groupId>io.github.palexdev</groupId>
<artifactId>materialfx</artifactId>
<version>11.17.0</version>
</dependency>You can read MaterialFX's documentation at javadoc.io
See the CHANGELOG file for a list of changes per version.
See the Open Issues for a list of proposed features (and known issues)
.
See the ROADMAP for a list of implemented and upcoming
features.
Since MaterialFX 11.14.0 the way controls are styles through CSS has drastically changed. Before telling you about the new Theming System, and about its pros and cons, let's talk a bit on the history of this project, the causes that brought to this drastic change.
When I started developing MaterialFX I was a complete noob, I knew nothing about JavaFX. But I really wanted to use it
and to make it look good. Competitors had broken libraries that made usage difficult for the end user, I didn't like
them at all.
And so the journey begun, MaterialFX is born. Like any newbies, what do you do when you know nothing but want to
learn?
You check others work and try to copy them but still make the changes you want to implement.
This lead me to create controls that made use of the infamous getUserAgentStylesheet() method. For those of you that
do not know about it, a developer of custom controls is supposed to override that method to provide a CSS stylesheet to
define the author's intended style for the custom control.
Sounds great right, just the thing I need... Well, I'd say that if only it worked properly. This system has been the
root cause of CSS issues right from the start of the project, with little I could do to fix it properly.
(Little secret that almost no one know: I actually sent a PR on the JavaFX repo to improve the system and make it
dynamic, guess what, it's still there lol)
The two most annoying issues caused by this system are:
- The little buggers didn't think of nested custom controls. For example, if a custom control(parent) has a skin that
uses other custom controls(children), the user agent of the parent will be completely ignored by the children,
the result is a bunch of children that cannot be styled in any way unless you create a custom skin yourself.
A fix I implemented for this in the past, was to override inline the
getUserAgentStylesheet()method of each children node to use the one of the parent, and even this drastic solution was working half the time (didn't work in some user cases) - For some reason, sometimes stylesheets provided by the user were half or completely ignored leading to half/not styled(as intended) custom controls. This was the most annoying issue, as the causes would vary from case to case, not always there was an easy/feasible solution, a nightmare, really
End of the rant
How can I fix it? I asked myself many many times.
Recently I've been working on a rewrite of MaterialFX, this new
version will have controls based on the new Material Design 3 Guidelines, will implement modular themes thanks to the
usage of SASS and a Theming API that will let user change themes, implement new ones, very easily.
So the idea is to backport at least the concept on the main branch at least until the rewrite is done.
The Theme API
Previous System (changed because of performance issues)
An interface called `Theme` allows users to define custom themes entries. It defines the bare minimum for a theme, its path and a way to load it. There are two implementations of this interface:Themes: this enumerator defines the default themes of MaterialFX, there is theDEFAULTtheme that includes the stylesheets of all MaterialFX controls, as well as dialogs, popups, menus, etc...Stylesheets: this enumerator defines all the stylesheets of every single control, allowing the user to not use a theme (for whatever reason) and instead choose which component he wants to style
MFXThemeManager is a utility class that will help the user add/set themes and stylesheets (which implement Theme) on
nodes or scenes.
Pros
- The biggest pro is to have a more reliable styling system. With this users shouldn't hava any issue anymore while styling MaterialFX controls with their custom stylesheets. Of course, I consider the system experimental, I don't expect to not have even a single report about CSS bugs, but they should be way less and much easier to fix
- Another pro is to have less code duplication as now I don't need to override the infamous
getUserAgentStylesheet()anymore anywhere - This change should have also impacted on memory usage in a good way as now controls do not store the "url" to their stylesheet anymore
Cons
- One con is that now themes must be managed by the user. Since controls are not styled by default, the user must
use the aforementioned manager or enumerators to load/add the themes on the App.
The preferred way to do so would be to add the themes/stylesheets on the root scene, like this:public class App extends Application { @Override public void start(Stage stage) { ... Scene scene = ...; MFXThemeManager.addOn(scene, Themes.DEFAULT, Themes.LEGACY); } }
- Another con that derives from the above example are dialogs/popups or any separate stage/scene.
Since you are applying the themes on the primary stage's scene, it means that all other scenes will be un-styled.
You have to add the Themes on every separate scene.
To simplify things, MaterialFX automatically applies the Themes on its dialogs and popups, but since now they
are added to the
getStylesheets()list it's easy to remove them and define your own The last con I can think of is SceneBuilder. As of now there is no support for it, I have some ideas on how to style controls inside of it though. The issue is that even if I figure out a way,I doubt the system will be flexible enough.What I mean is, I can probably set the default themes on the SceneBuilder' scene,but it's very unlikely there will be a way to choose which themes/stylesheets will be applied.
Since version 11.15.0, MaterialFX controls are capable of detecting if they are being used in SceneBuilder and can automatically style themselves. From my little testings, it seems that this doesn't break the styling system in any way, I was able to style a button by adding a custom stylesheet on itself or on its parent. There's also an emergency system to completely shut down the SceneBuilder integration, more info here: Themable
New system
The best way to style a JavaFX app is to set its User-Agent Stylesheet by calling `Application.setUserAgentStylesheet(...)`. The issue with this is that JavaFX's default theme is overridden, and so, any other non-custom control will not be styled. This is good if you are going to use only custom controls, but obviously bad if you are going to use JavaFX's controls or any other control from other libraries that rely on the JavaFX's default theme. There is no easy way around this. There is a proposal to enhance the system on the JavaFX page, but it won't come anytime soon. So, I came up with a pretty decent workaround. If I can't set multiple User-Agents then I will build a single one. So, let's see how the API works.An interface called Theme, allows users to define custom theme entries. It specifies the bare minimum properties a
theme must have: its path and a way to load it. There are two implementations of this interface:
JavaFXThemes: this enumerator defines the JavaFX's default themes. Since JavaFX 8 the default one isMODENA. The funny thing about all this crap is that, JavaFX's User-Agents are actually split in multiple CSS files, there is a main file and then others that are added according to certain system's properties the app is running on. So, they are allowed to use multiple files, but we are not....thanks for nothing I guess.MaterialFXStylesheets: this enumerator defines all the stylesheets for each MaterialFX control. However, to make things work as intended, all of them have been merged into a single CSS file calledDefaultTheme.css. The same thing has been done for legacy controls too, the aio CSS file is calledLegacyControls.css
Now, the missing core part. The class responsible for building a single User-Agent stylesheet is: UserAgentBuilder.
There are three main aspects of the system you should know, and I'm going to explain them after giving you a short code
example:
UserAgentBuilder.builder()
.themes(JavaFXThemes.MODENA) // Optional if you don't need JavaFX's default theme, still recommended though
.themes(MaterialFXStylesheets.forAssemble(true)) // Adds the MaterialFX's default theme. The boolean argument is to include legacy controls
.setDeploy(true) // Whether to deploy each theme's assets on a temporary dir on the disk
.setResolveAssets(true) // Whether to try resolving @import statements and resources urls
.build() // Assembles all the added themes into a single CSSFragment (very powerful class check its documentation)
.setGlobal(); // Finally, sets the produced stylesheet as the global User-Agent stylesheet- First and foremost, the build has to know which themes/stylesheets you want to combine, add them through the
themes(...)method. - Themes are allowed to have assets. You see, when they are combined into one single stylesheets, resources are likely
to fail loading. So, fonts, images and such will not be available anymore. To overcome this issue, there's only one way,
deploying all the needed resources on the disk. Assets should be packed in a
.zipfile, and you should be very careful at the structure inside it. When extracting the files the structure will be honored. - Deploying the assets in most cases is not enough. Let's say a CSS file has this statement
@import ../fonts/Font.css. You know how the API works and packed the font into a zip file, the structure is as followsfonts/Font.css. The file is going to be extracted in a directory on the disk, let's sayosTempDir/myassets/fonts/Font.css. (You can change the root dir's name by overriding the Theme'sdeployName()method) Now, as you may guess, that import statement needs to be changed so that it points to the resource on the disk. This is exactly whatsetResolveAssets(true)attempts to do. The result will be something like this:@import osTempDir/myassets/fonts/Font.css
Pros
- The biggest pro is to have a more reliable styling system. With this, users shouldn't have any issue anymore while styling MaterialFX controls with their custom stylesheets. Of course, I consider the system experimental, I don't expect to not have even a single report about CSS bugs, but they should be way less and much easier to fix. Since the theme is set as the app's global User-Agent, it will be applied on all the Stages/Scenes of the app, which is great
- Another pro is to have less code duplication as now I don't need to override the infamous getUserAgentStylesheet() anymore anywhere
- This change should have also impacted on memory usage in a good way as now controls do not store the "url" to their stylesheet anymore
- Potentially, you could now create a single theme, that is applied consistently everywhere in your apps, with as much stylesheets as you want
Cons
- The main con is that now theming must be managed by the user. Since controls are not styled by default anymore, the user
must use the
UserAgentBuilderto create the theme and set it as the application User-Agent (you can check the code snippet above to see how to do it). - The system is fragile and naive. CSS files need to be well written and formatted, and even in that case, there may still be issues while parsing or once the aio User-Agent is set.
- The generated theme won't be able to access "local" resources anymore. For this reason, themes now have to deploy assets if needed, which surely introduces some overhead. Not only that, for the previous point, resolving "local" assets to "absolute" assets may fail and can be hard to set up.
- SceneBuilder integration is tricky and error/bug prone. MaterialFX controls are capable of detecting if they are being used in SceneBuilder and can automatically style themselves. From my little testings, it seems that this doesn't break the styling system in any way. I was able to style a button by adding a custom stylesheet on itself or on its parent. There's also an emergency system to completely shut down the SceneBuilder integration, more info here: Themable
Contributions are what make the open source community such an amazing place to learn, inspire, and create. Any contributions you make are greatly appreciated.
- Fork the Project
- Create your Feature Branch (
git checkout -b feature/AmazingFeature) - Commit your Changes (
git commit -m 'Add some AmazingFeature') - Push to the Branch (
git push origin feature/AmazingFeature) - Open a Pull Request
Distributed under the GNU LGPLv3 License. See LICENSE for more information.
Alex - [email protected]
Discussions
Project Link: https://github.com/palexdev/MaterialFX
It's been more than a year since I started developing MaterialFX. Implementing cool looking, fully functional controls,
introducing new components and features as well as providing many utilities for JavaFX and Java is really hard,
especially considering that developing for JavaFX also means to deal with its closeness, its bugs, its annoying
design decisions. Many times I've honestly been on the verge of giving up because sometimes it's really too much
stress to handle.
But, today MaterialFX is a great library, supported by many people and I'm proud of it.
If you are using MaterialFX in your projects and feel like it, I recently
activated GitHub Sponsors so
you can easily donate/sponsor.
(If you want your github page to be linked here and you didn't specify your username in the donation, feel free to contact me by email and tell me. Also contact me if for some some reason you don't want to be listed here)
- Alaa Abu Zidan
- Alex Hawk
- Aloento
- brr53
- Eugen Gubin
- Mauro de Wit
- Mohammad Chaudhry (thank you very much for the huge donation, YOU are the legend)
- Jtpatato21
- Sourabh Bhat
- stefanofornari (thank you very much for the big donation!)
- Ultraviolet-Ninja
- Yahia Rehab
- Yiding He
- Your name can be here by supporting me at this link, GitHub Sponsors
Thank you very very much to all supporters, to all people who contribute to the project, to all people that thanked me, you really made my day
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent