The comment added in commit 8a0bc97e ("[egui_glium] Fix paste") seems to
assume that `winit` "should have translated" common "paste" keyboard
combos to a `Cut`/`Copy`/`Paste` "KeyCode", but completely glossed over
the fact that this `KeyCode` (now also `NamedKey`) maps to a special key
dedicated to this purpose found on some keyboards and OSes. Make sure
that this key is still handled in addition to the combo that is
detected.
---
Note that this PR does not compile as it is (and I have hence not tested
this nor even ran into this limitation), just noticed this inconsistency
while failing to understand a code comment. We'd have to decide if the
variants should be added to `egui::Key` or if these helper functions
need to take `winit` keys (`ScanCode` or `NamedKey`) directly?
I should have an old keyboard with a physical paste key in a drawer
somewhere. And on Android there are special copy/paste events that can
be sent by a virtual keyboard or the debug shell, so that this can be
properly tested before it is merged.
(Except that the current `clipboard` implementation is not supported on
Android)
<!--
Please read the "Making a PR" section of
[`CONTRIBUTING.md`](https://github.com/emilk/egui/blob/master/CONTRIBUTING.md)
before opening a Pull Request!
* Keep your PR:s small and focused.
* If applicable, add a screenshot or gif.
* If it is a non-trivial addition, consider adding a demo for it to
`egui_demo_lib`, or a new example.
* Do NOT open PR:s from your `master` branch, as that makes it hard for
maintainers to add commits to your PR.
* Remember to run `cargo fmt` and `cargo cranky`.
* Open the PR as a draft until you have self-reviewed it and run
`./scripts/check.sh`.
* When you have addressed a PR comment, mark it as resolved.
Please be patient! I will review your PR, but my time is limited!
-->
Since accesskit_winit version 0.15.0, it is necessary to call
`Adapter::process_events` to let AccessKit fire up window focus events
on Unix and macOS. Furthermore this has always been needed on Unix (X11
only) to let AccessKit know about the window bounds, which is needed to
perform hit-testing on nodes.
<!--
Please read the "Making a PR" section of
[`CONTRIBUTING.md`](https://github.com/emilk/egui/blob/master/CONTRIBUTING.md)
before opening a Pull Request!
* Keep your PR:s small and focused.
* If applicable, add a screenshot or gif.
* If it is a non-trivial addition, consider adding a demo for it to
`egui_demo_lib`, or a new example.
* Do NOT open PR:s from your `master` branch, as that makes it hard for
maintainers to add commits to your PR.
* Remember to run `cargo fmt` and `cargo cranky`.
* Open the PR as a draft until you have self-reviewed it and run
`./scripts/check.sh`.
* When you have addressed a PR comment, mark it as resolved.
Please be patient! I will review your PR, but my time is limited!
-->
Closes <https://github.com/emilk/egui/issues/3677>.
This can naturally be obtained by starting the drawing a bit "later",
but providing support for this directly will simplify interleaved
drawing of dashed lines.
An even more general approach would be to allow passing an even-length
vector of lengths and gaps, in addition to offset. Primarily thinking if
it is better to go even further if a new dashed-line-method is
introduced. The second commit introduce this (I can naturally squash
these!).
This PR fixes an issue where a bad pivot stored in app state would
override the pivot actually set from code. The root code of this issue
if the `Window`'s `State` containing `pivot` in the first place, as
`pivot` is not part of a state to be tracked. It makes the `State`
structure more ergonomic though, so this PR leaves the `pivot` field but
always overrides it with the value provided by user code.
#### Repro of the original issue
This issue can be reproduced using the `re_ui_example` app from the
Rerun repo, at this commit:
fb5add0047.
By using this `app.ron` file, the bug will appear:
```
{
"egui": "(
areas:{
((0)):(
areas:{
(13430889033718688666):(pivot_pos:(x:565.5,y:328.0),pivot:((Min,Min)),size:(x:364.0,y:75.5),interactable:true),
},
)
}
)",
}
```
The modal is entered based on it's top-left corner even though the code
actually specifies a center-center pivot:
<img width="1312" alt="image"
src="https://github.com/emilk/egui/assets/49431240/6df1a8bb-d1ea-4e15-866c-5f14138dda0e">
With this PR, the centring is correct even with the "poisoned"
`app.ron`:
<img width="1312" alt="image"
src="https://github.com/emilk/egui/assets/49431240/7f50b2d4-9245-48f3-91de-c747bbc8c40e">
This introduces a special `Color32::PLACEHOLDER` which, during text
painting, will be replaced with `TextShape::fallback_color`.
The fallback color is mandatory to set in all text painting. Usually
this comes from the current visual style.
This lets users color only parts of a `WidgetText` (using e.g. a
`LayoutJob` or a `Galley`), where the uncolored parts (using
`Color32::PLACEHOLDER`) will be replaced by a default widget color (e.g.
blue for a hyperlink).
For instance, you can color the `⚠️`-emoji red in a piece of text red
and leave the rest of the text uncolored. The color of the rest of the
text will then depend on wether or not you put that text in a label, a
button, or a hyperlink.
Overall this simplifies a lot of complexity in the code but comes with a
few breaking changes:
* `TextShape::new`, `Shape::galley`, and `Painter::galley` now take a
fallback color by argument
* `Shape::galley_with_color` has been deprecated (use `Shape::galley`
instead)
* `Painter::galley_with_color` has been deprecated (use
`Painter::galley` instead)
* `WidgetTextGalley` is gone (use `Arc<Galley>` instead)
* `WidgetTextJob` is gone (use `LayoutJob` instead)
* `RichText::into_text_job` has been replaced with
`RichText::into_layout_job`
* `WidgetText::into_text_job` has been replaced with
`WidgetText::into_layout_job`
Please take a look @Bromeon and @EmbersArc!
You have both contributed substantially to `egui_plot` (more than I
have).
This is me inviting you to co-maintain the `egui_plot` crate.
If you accept, you'll get a notification when someone opens a PR
touching `egui_plot`, and you can review and merge it as you see fit.
You can also merge your own egui_plot PRs without waiting on me to
review them.
Update `resvg` from v0.28 to v0.37.
Remove related, unnecessary entries from `deny.toml`.
⚠ In example `images` ferris is scaled differently, but I guess that now
it scales in expected way (takes all available space; before this PR it
takes up to space that, was available at first render- it does not
upscale).
This PR is minimal adaptation to new `resvg` api and small related
simplification, however it should be considered to update loaders
(currently if svg image initially was small and was scaled up it will be
blurred, see https://github.com/emilk/egui/issues/3501). As svg image
now scales over render size, problem will be more often seen now.
(currently `SvgLoader` theoretically should rerender for different sizes
(but I guess it will result in memory leak in that case), but refreshing
is stopped earlier in `DefaultTextureLoader`).
I have initial version of loaders update, that will fix issue with svg
scaling (and also enable e.g. reloading image if file has been changed),
I will submit these changes in separate PR once this one is merged.
Closes <https://github.com/emilk/egui/issues/3652>.
These PR recently cleaned up the code around auto-bounds, but introduced
an involuntary change whereby auto-bounds would not be enabled by
default. All plots would default to being not properly centred as a
result.
- #3587
- #3586
This PR changes the default back to enabled. It also deprecates
`auto_bounds_x()` and `auto_bounds_y()`, which could only enable
auto-bounds (which is not very useful as auto-bounds were, and now are
again, enabled by default). A new `auto_bounds()` API can now be sued to
disable auto-bounds if needed.
Fixes#3712
Fixes https://github.com/rerun-io/rerun/issues/4503
<!--
Please read the "Making a PR" section of
[`CONTRIBUTING.md`](https://github.com/emilk/egui/blob/master/CONTRIBUTING.md)
before opening a Pull Request!
* Keep your PR:s small and focused.
* If applicable, add a screenshot or gif.
* If it is a non-trivial addition, consider adding a demo for it to
`egui_demo_lib`, or a new example.
* Do NOT open PR:s from your `master` branch, as that makes it hard for
maintainers to add commits to your PR.
* Remember to run `cargo fmt` and `cargo cranky`.
* Open the PR as a draft until you have self-reviewed it and run
`./scripts/check.sh`.
* When you have addressed a PR comment, mark it as resolved.
Please be patient! I will review your PR, but my time is limited!
-->
Small spelling error - changed `Addintional` to `Additional`
I broke this in https://github.com/emilk/egui/pull/3665
For some reason, when using multiple viewports, the first "clear"
doesn't take.
I don't have time to investigate further, so I am adding a hack here
which:
* For single-viewport apps, the clear is done before `App::update` so
user can paint there
* For multi-viewport apps, the clear is done after `App::update`, so
that it works
This means painting in `App::update` won't work with multi-viewports.
Ideally, users should use paint callbacks instead of painting directly
in `App::update` anyways.
<!--
Please read the "Making a PR" section of
[`CONTRIBUTING.md`](https://github.com/emilk/egui/blob/master/CONTRIBUTING.md)
before opening a Pull Request!
* Keep your PR:s small and focused.
* If applicable, add a screenshot or gif.
* If it is a non-trivial addition, consider adding a demo for it to
`egui_demo_lib`, or a new example.
* Do NOT open PR:s from your `master` branch, as that makes it hard for
maintainers to add commits to your PR.
* Remember to run `cargo fmt` and `cargo cranky`.
* Open the PR as a draft until you have self-reviewed it and run
`./scripts/check.sh`.
* When you have addressed a PR comment, mark it as resolved.
Please be patient! I will review your PR, but my time is limited!
-->
This is a very simple change so you don't have to do this:
https://github.com/zeozeozeo/egui-keybind/blob/master/src/bind.rs#L149-L172
Useful when debugging, e.g. if the default choice is one of many, and
perhaps the other ones would be better.
**EDIT**: Oh, `enumerate_adapters` doesn't work on web… will fix.
It required some ugly code in `egui-winit` in order to fix this without
breaking sem-ver (I want this to land in a 0.24.1 patch release). I'll
clean up in time for 0.25.
There is still a font rendering bug when using immediate viewports
across multiple viewports, but that's harder to fix.
We were using [`tts`](https://github.com/ndarilek/tts-rs) for the
web-only screen reader. This was overkill, to say the least. It is now
replaced with ten lines of `web-sys` calls.
This is a bug that can occur on Windows, previously it was fixed by only
showing the window after the first frame was rendered, but the bug
appeared again with egui 0.24.0. This commit fixes this bug by making
the window invisible on startup.
## Before
## After
<!--
Please read the "Making a PR" section of
[`CONTRIBUTING.md`](https://github.com/emilk/egui/blob/master/CONTRIBUTING.md)
before opening a Pull Request!
* Keep your PR:s small and focused.
* If applicable, add a screenshot or gif.
* If it is a non-trivial addition, consider adding a demo for it to
`egui_demo_lib`, or a new example.
* Do NOT open PR:s from your `master` branch, as that makes it hard for
maintainers to add commits to your PR.
* Remember to run `cargo fmt` and `cargo cranky`.
* Open the PR as a draft until you have self-reviewed it and run
`./scripts/check.sh`.
* When you have addressed a PR comment, mark it as resolved.
Please be patient! I will review your PR, but my time is limited!
-->
Closes https://github.com/emilk/egui/issues/3625.
This appears to have been 'consumed' by the viewport changes at one
point (by updating `viewport_info`), but this is no longer the case. We
should avoid marking this as 'consumed', so that applications know to
continue processing this event after passing it to egui.
<!--
Please read the "Making a PR" section of
[`CONTRIBUTING.md`](https://github.com/emilk/egui/blob/master/CONTRIBUTING.md)
before opening a Pull Request!
* Keep your PR:s small and focused.
* If applicable, add a screenshot or gif.
* If it is a non-trivial addition, consider adding a demo for it to
`egui_demo_lib`, or a new example.
* Do NOT open PR:s from your `master` branch, as that makes it hard for
maintainers to add commits to your PR.
* Remember to run `cargo fmt` and `cargo cranky`.
* Open the PR as a draft until you have self-reviewed it and run
`./scripts/check.sh`.
* When you have addressed a PR comment, mark it as resolved.
Please be patient! I will review your PR, but my time is limited!
-->
* Follow-up to https://github.com/emilk/egui/pull/3621 and
https://github.com/emilk/egui/pull/3513
To work around a Safari limitation, we run the app logic in the event
handler of copy, cut, and mouse up and down.
Previously the output of that frame was discarded, but in this PR it is
now saved to be used in the next requestAnimationFrame.
The result is noticeable more distinct clicks on buttons (one more frame
of highlight)
Bonus: also fix auto-save of a sleeping web app
Marijn Suijten
Arnold Loubriat
Emil Ernerfeldt
Oscar Gustafsson
Wybe Westra
Antoine Beyeler
PingPongun
crispy-strawberry
Fredrik Fornwall
zeozeozeo
Brian Janssen
Sebastian Urban
Vaqxai
Aaron Hill
libertyteeth