jgraham

Require pointerMove and wheel coordinates to be integers

2

Prevously we allowed these properties to be undefined, and then later tried to do computations on the values. That doesn't work; they either need to have an integer default value...

Associate known elements with a window rather than browsing context

6

This makes the lifecycle of this element list clearer; instead of having to keep element references alive as long as the browsing context, you only need to keep it alive...

It's currently unclear what happens in actions when pointer property values are omitted. I think the implied behaviour is that if a value (say `pressure`) is omitted, we use the...

Add some security checks when handling a HTTP connection.

6

Check the Host and Origin headers for the incoming connection to verify the connection is allowed. The language is intended to allow the specific behaviour to be largely implementation defined,...

Add the Element Origin concept

1

This is really designed for BiDi, allowing actions to specify the origin relative to an element in the form: { type: "element" sharedId: } rather than sending in a WebDriver-style...

WebDriver key actions don't cover the case of IME inputs, in which multiple composition events end up inserting one or more characters that don't directly map to the keys pressed....

Execute script tries to set a data property on errors

6

e.g. https://w3c.github.io/webdriver/#execute-script step 8 says: > Upon rejection of promise with reason r, let result be a JSON clone of r, and return error with error code javascript error and...

Refactor session handling for better integration with BiDi

2

This is a large change, but it's almost entirely refactoring. The main aim here is to split the session concept so that there's a base "session" which holds the session...

[actions] Handling of "display:none" element as origin

5

This should probably return "move target out of bounds" or something, but at the moment the spec algorithm doesn't account for the possibility of an element without bounding client rects.

We currently have some tests that assume that a keyDown action that depresses the shift key, followed by a keyDown action with value "a" will result in text input "A",...