Regression: Click on a submit button triggers an exception

[linkdotnet]

We have a regression in version 1.17 where "clicking" a submit button doesn't trigger the onsubmit event of the form itself.
Now technically the button does not have a onclick event but we allowed that prior to version 1.17. So it is open for discussion if we are now technical correct (even though in a browser it would work differently).

Reproduction

Following component:

<form id="my-form" @onsubmit="() => FormSubmitted = true">
    <input id="inside-form-input" type="submit" value="Submit" />
</form>

@code {
    public bool FormSubmitted;
}

Test:

[Fact(DisplayName = "Should submit a form on submit button")]
public void Test308()
{
	var cut = RenderComponent<SubmitFormButton>();

	cut.Find("#inside-form-input").Click();

	cut.Instance.FormSubmitted.ShouldBeTrue();
}

But leads to:

Bunit.MissingEventHandlerException: The element does not have an event handler for the event 'onclick', nor any other events.

Bunit.MissingEventHandlerException
The element does not have an event handler for the event 'onclick', nor any other events.
   at Bunit.TriggerEventDispatchExtensions.TriggerBubblingEventAsync(ITestRenderer renderer, IElement element, String eventName, EventArgs eventArgs) in /Users/stgi/repos/bUnit/src/bunit.web/EventDispatchExtensions/TriggerEventDispatchExtensions.cs:line 125
   at Bunit.TriggerEventDispatchExtensions.TriggerEventsAsync(ITestRenderer renderer, IElement element, String eventName, EventArgs eventArgs) in /Users/stgi/repos/bUnit/src/bunit.web/EventDispatchExtensions/TriggerEventDispatchExtensions.cs:line 90
   at Microsoft.AspNetCore.Components.Rendering.RendererSynchronizationContext.<>c.<<InvokeAsync>b__9_0>d.MoveNext()
--- End of stack trace from previous location ---
   at Bunit.TriggerEventDispatchExtensions.ThrowIfFailedSynchronously(Task result) in /Users/stgi/repos/bUnit/src/bunit.web/EventDispatchExtensions/TriggerEventDispatchExtensions.cs:line 199
   at Bunit.TriggerEventDispatchExtensions.TriggerEventAsync(IElement element, String eventName, EventArgs eventArgs) in /Users/stgi/repos/bUnit/src/bunit.web/EventDispatchExtensions/TriggerEventDispatchExtensions.cs:line 77
   at Bunit.MouseEventDispatchExtensions.ClickAsync(IElement element, MouseEventArgs eventArgs) in /Users/stgi/repos/bUnit/src/bunit.web/EventDispatchExtensions/MouseEventDispatchExtensions.cs:line 373
   at Bunit.MouseEventDispatchExtensions.Click(IElement element, Int64 detail, Double screenX, Double screenY, Double clientX, Double clientY, Double pageX, Double pageY, Double offsetX, Double offsetY, Int64 button, Int64 buttons, Boolean ctrlKey, Boolean shiftKey, Boolean altKey, Boolean metaKey, String type) in /Users/stgi/repos/bUnit/src/bunit.web/EventDispatchExtensions/MouseEventDispatchExtensions.cs:line 289
   at Bunit.GeneralEventDispatchExtensionsTest.Test308() in /Users/stgi/repos/bUnit/tests/bunit.web.tests/EventDispatchExtensions/GeneralEventDispatchExtensionsTest.cs:line 286
   at System.RuntimeMethodHandle.InvokeMethod(Object target, Void** arguments, Signature sig, Boolean isConstructor)
   at System.Reflection.MethodInvoker.Invoke(Object obj, IntPtr* args, BindingFlags invokeAttr)

[egil]

Clicking a submit button inside a form should submit the form if there is not a "prevent default" flag on the form, right? That was my impression of the fix in 1.17.

The test you have here should pass as far as I can tell, so yeah, this should be fixed.

[egil]

Looking at the code, I do not think it is a regression. It probably never worked like that before.

[linkdotnet]

Looking at the code, I do not think it is a regression. It probably never worked like that before.

It did work before - I used that code in one of my projects, which after the update to 1.17 started failing ;)
I also checked the code and reverted to 1.16 with thegiven example and it also passes

[egil]

Weird. Just checked out "#e361552 - build(deps): Bump System.Text.Json from 7.0.1 to 7.0.2" the commit right before at "#c094d30 - Honor preventdefault for events so they don't bubble (#982)", and the test and component from this issue passes on my computer.

[linkdotnet]

Weird. Just checked out "#e361552 - build(deps): Bump System.Text.Json from 7.0.1 to 7.0.2" the commit right before at "#c094d30 - Honor preventdefault for events so they don't bubble (#982)", and the test and component from this issue passes on my computer.

Exactly ;) It should - therefore I do think it is a regression.

[egil]

Ahh of course, nevermind, I'm tired :)