Iced: Hello World application panics "OutOfMemory" resizing the window

[TheLonelyAdventurer]

Hello, I was looking for a cross-platform gui crate and found this project.
I really like the idea of using Elm's architecture.

Version

cargo rust --version: rustc 1.46.0 (04488afe3 2020-08-24)
iced: 0.1.1
other dependencies: none

Bug description

I copied the "Hello World" example from https://docs.rs/iced/0.1.1/iced/trait.Application.html and, out of frustration, I shaked the mouse cursor for about 7 to 12 seconds (I am able to reproduce it, but varies) while resizing the application window. It panicked and crashed.

Backtrace in debug run (RUST_BACKTRACE=full)

debug.log

Behaviour with --release (RUST_BACKTRACE=full)

The issue is present as well. Produces this backtrace.
release.log

Steps to reproduce the error

Run cargo run or cargo run --release and try to resize the window shaking the mouse cursor all around the screen for about 10 seconds. I also experienced issue #546 .

Machine and OS details

• Windows 10 Home
• Screen resolution: 4K UHD (3840x2160@60hz, 250% zoom)
• RAM: 16GB
• CPU: Intel Core i7-9750H (Coffee Lake-H SRF6U)

Also, I read wgpu-native in the log, so I thought it was related:

• GPU: Intel UHD Graphics 630 (Coffee Lake-H GT2) / NVIDIA RTX 2070 Max-Q (TU106M)

I'm sorry if this isn't really a use case.

If you need anything more, I'll be pleased to help.

[hecrj]

Thanks for the detailed report!

master has changed considerably since the latest release. Does this happen with any of the current examples?

[TheLonelyAdventurer]

Thanks for the detailed report!

master has changed considerably since the latest release. Does this happen with any of the current examples?

You're welcome.

I cloned the repo with git clone and ran cargo build.
Then, I ran cargo run --package clock and tried to reproduce the error, and yes, it panicked with the same behaviour.

Actually, it happens with a variation. It looks like it takes less to panic if I try to resize it from the bottom-right corner.

Also, issue #546 still occurs.

I also noticed a 500ms to 1second lag for the clock to render the first time.

I'm attaching the log here.

debug.log

[hecrj]

Thanks for testing!

Do the examples in the wgpu-rs repository have the same issue?

[Kaiden42]

Could be related to #518

[TheLonelyAdventurer]

Thanks for testing!

Do the examples in the wgpu-rs repository have the same issue?

Just cloned wgpu v0.6.0 and tried to reproduce the error.

Yeah, it does occur. I tried with the hello-window example.

I'm attaching the log here, but I think I should open an issue to wgpu if there isn't already.

debug.log

Notice: the windows took about 1 second to render and the issue occurred immediately and not just after a few seconds.

Also, #546 occurs as well.

[TheLonelyAdventurer]

Could be related to #518

I just checked the attached video to that one issue. I think I didn't check the opened issues list very well.

It behaves exactly the same, but I think we could consider this to be wgpu's fault.

[TheLonelyAdventurer]

@hecrj @Kaiden42 There is in fact an open issue https://github.com/gfx-rs/wgpu-rs/issues/571 on their repo.

I'm going to report it there as well.