egui
has been an on-and-off weekend project of mine since late 2018. I am grateful to any help I can get, but bear in mind that sometimes I can be slow to respond because I am busy with other things!
/ Emil
You want to contribute to egui, but don't know how? First of all: thank you! I created a special issue just for that: emilk#3742. But make sure you still read this file first :)
You can ask questions, share screenshots and more at GitHub Discussions.
There is an egui
discord at https://discord.gg/vbuv9Xan65.
Issues are for bug reports and feature requests. Issues are not for asking questions (use Discussions or Discord for that).
Always make sure there is not already a similar issue to the one you are creating.
If you are filing a bug, please provide a way to reproduce it.
For small things, just go ahead an open a PR. For bigger things, please file an issue first (or find an existing one) and announce that you plan to work on something. That way we will avoid having several people doing double work, and you might get useful feedback on the issue before you start working.
Browse through ARCHITECTURE.md
to get a sense of how all pieces connects.
You can test your code locally by running ./scripts/check.sh
.
When you have something that works, open a draft PR. You may get some helpful feedback early! When you feel the PR is ready to go, do a self-review of the code, and then open it for review.
Don't worry about having many small commits in the PR - they will be squashed to one commit once merged.
Please keep pull requests small and focused. The smaller it is, the more likely it is to get merged.
Most PR reviews are done by me, Emil, but I very much appreciate any help I can get reviewing PRs!
It is very easy to add complexity to a project, but remember that each line of code added is code that needs to be maintained in perpetuity, so we have a high bar on what get merged!
When reviewing, we look for:
- The PR title and description should be helpful
- Breaking changes are documented in the PR description
- The code should be readable
- The code should have helpful docstrings
- The code should follow the Code Style
Note that each new egui release have some breaking changes, so we don't mind having a few of those in a PR. Of course, we still try to avoid them if we can, and if we can't we try to first deprecate old code using the #[deprecated]
attribute.
See https://docs.rs/egui/latest/egui/#integrating-with-egui for how to write your own egui integration.
If you make an integration for egui
for some engine or renderer, please share it with the world!
Make a PR to add it as a link to README.md
so others can easily find it.
- Build with
scripts/build_demo_web.sh
- Host with
scripts/start_server.sh
- Open http://localhost:8888/index.html
While using an immediate mode gui is simple, implementing one is a lot more tricky. There are many subtle corner-case you need to think through. The egui
source code is a bit messy, partially because it is still evolving.
- Read some code before writing your own
- Leave the code cleaner than how you found it
- Write idiomatic rust
- Follow the Rust API Guidelines
- Add blank lines around all
fn
,struct
,enum
, etc // Comment like this.
and not//like this
- Use
TODO
instead ofFIXME
- Add your github handle to the
TODO
:s you write, e.g:TODO(emilk): clean this up
- Avoid
unsafe
- Avoid
unwrap
and any other code that can cause panics - Use good names for everything
- Add docstrings to types,
struct
fields and allpub fn
- Add some example code (doc-tests)
- Before making a function longer, consider adding a helper function
- If you are only using it in one function, put the
use
statement in that function. This improves locality, making it easier to read and move the code - When importing a
trait
to use it's trait methods, do this:use Trait as _;
. That lets the reader know why you imported it, even though it seems unused - Avoid double negatives
- Flip
if !condition {} else {}
- Sets of things should be lexicographically sorted (e.g. crate dependencies in
Cargo.toml
) - Break the above rules when it makes sense
/// The name of the thing.
pub fn name(&self) -> &str {
&self.name
}
fn foo(&self) {
// TODO(emilk): this can be optimized
}
//gets the name
pub fn get_name(&self) -> &str {
&self.name
}
fn foo(&self) {
//FIXME: this can be optimized
}
The left-top corner of the screen is (0.0, 0.0)
,
with Vec2::X
increasing to the right and Vec2::Y
increasing downwards.
egui
uses logical points as its coordinate system.
Those related to physical pixels by the pixels_per_point
scale factor.
For example, a high-dpi screen can have pixels_per_point = 2.0
,
meaning there are two physical screen pixels for each logical point.
Angles are in radians, and are measured clockwise from the X-axis, which has angle=0.
The code should never panic or crash, which means that any instance of unwrap
or expect
is a potential time-bomb. Even if you structured your code to make them impossible, any reader will have to read the code very carefully to prove to themselves that an unwrap
won't panic. Often you can instead rewrite your code so as to avoid it. The same goes for indexing into a slice (which will panic on out-of-bounds) - it is often preferable to use .get()
.
For instance:
let first = if vec.is_empty() {
return;
} else {
vec[0]
};
can be better written as:
let Some(first) = vec.first() else {
return;
};