The main change is to the connect() method which now checks the urlOrChannel variable to determine which arguments to pass to WebChannel.open. The url case remains largely the same, we simply check that it is of type string. If a WebSocket or RTCDataChannel is passed, we use the object prototypes constructor name to determine the type of the instance variable provided is one which we expect and then forward it on to WebChannel.open, along with the name so we dont have to do the same work in WebChannel. We use Object.getPrototypeOf because accessing the __proto__ directly is not recommended.
The name WebChannel was chosen to represent the opaque datachannel used, either WebSocket or RTCDataChannel. The open method was refactored to take an object of uri and protocols or webChannel and channelType. We introduce a channelStates enum and helper method that provides compatible states across WebSocket and RTCDataChannel. Call sites for logging are refactored to list the channelType (WebSocket or RTCDataChannel) to be more descriptive.
As a rule, instead of hard-coding a behavior on specific platforms we
should do dynamic detection.
This commit moves away from always hiding scrollbars on Android and iOS
and instead detects the rendered width of scrollbars in the browser.
Internet Explorer seems to flag images as loaded prematurely, which
can result in rendering bugs. We can detect this by looking at the
dimensions though.
When showing a new status popup we want to set a timer for how long to
show it. In cases where we show many statuses in a fast succession we
need to remove any running timeouts when showing a new one.
There are exceptions when new statuses won't be shown, and thats if a
more severe status is already showing, i.e and error or a warning.
Warnings can still have timeouts. There was a bug that occured when we
tried to show a normal status while a warning was showing. The bug
caused the warning status timeout to be removed even if the normal
status was never shown. We should only remove running timeouts if we're
actually going to show a new status.
There has been a lot of renaming and restructuring in babel, so we need
to modify our code to handle the latest version. We also need to adjust
the way we build our babel worker as babel itself no longer runs in older
browsers such as Internet Explorer.
This control flow is difficult enough as it is to follow. Move the
handling of the untransformed files to a separate block to make it
slightly easier to understand.
Caps Lock on iOS only trigged key release or key press events.
When it's clicked it would only send keydown, and next time
it would only send keyup and so on. It should send both a key press
and a key release.
Also added the unit tests for macOS since those were missing.
Co-Authored-By: Alex Tanskanen <aleta@cendio.se>
Ryan Castner