WebDriver now only has an IPC connection to WebContent. WebDriver still
launches the browser, but now when the session ends, we simply send a
SIGTERM signal to the browser.
There are a couple changes here from the existing Get All Cookies
implementation.
1. Previously, WebDriver actually returned *all* cookies in the cookie
jar. The spec dictates that we only return cookies that match the
document's URL. Specifically, it calls out that we must run just the
first step of RFC 6265 section 5.4 to perform domain matching.
This change adds a special mode to our implementation of that section
to skip the remaining steps.
2. We now fill in the SameSite cookie attribute when serializing the
cookie to JSON (this was a trival FIXME that didn't get picked up
when SameSite was implemented).
Requests to maximize and minimize Browser windows will be coming from
the WebContent process rather than the WebDriver process. Add hooks to
propagate these requests back up to the Browser.
Requests to restore, resize, and reposition Browser windows will be
coming from the WebContent process rather than the WebDriver process.
Add hooks to propagate these requests back up to the Browser.
The spec notes "The specification does not guarantee that the resulting
window size will exactly match that which was requested", so these new
methods return the actual new size/position.
I had originally thought to just leave these and remove them all at once
at the end of the WebContent migration. But it is kind of confusing to
have them around, so this removes the endpoints that have already been
ported.
First, this moves the WebDriver socket to the /tmp/websocket/ directory,
as WebDriver now creates multiple sockets per session. Those sockets are
now created with Core::LocalServer rather than manually setting up the
listening sockets (this was an existing FIXME which resolved some issues
I was hitting with creating a second listening socket).
WebDriver passes both socket paths to Browser via command line. Browser
continues to connect itself via one socket path, then forwards the other
socket path to the WebContent process created by the OOPWV. WebContent
then connects to WebDriver over this path.
WebContent will temporarily set the navigator.webdriver flag to true
after connecting to WebDriver. This will soon be moved to its own IPC to
be sent by WebDriver.
From https://www.w3.org/TR/webdriver/#security:
It is also suggested that user agents make an effort to visually
distinguish a user agent session that is under control of WebDriver from
those used for normal browsing sessions. This can be done through a
browser chrome element such as a “door hanger”, colorful decoration of
the OS window, or some widget element that is prevalent in the window so
that it easy to identify automation windows.
This doesn't follow the spec to a tee. Our OutOfProcessWebView already
has a bitmap that can be used as the window screenshot. Therefore, we
can bypass the steps that assume we need to access the window's frame
buffer in-flight.
We also don't create an HTMLCanvasElement. We would need a Document in
the WebDriver process to do so. Instead, we can still run the encoding
steps exactly as-is using the screenshot bitmap.
Otherwise, we end up propagating those dependencies into targets that
link against that library, which creates unnecessary link-time
dependencies.
Also included are changes to readd now missing dependencies to tools
that actually need them.
