That way, cancelling a context while checking the error is much simpler.
The context data already holds onto the cancel func, so this requires no
internal changes.
This renders the Browser.Shutdown API obsolete, even though it doesn't
do exactly the same. If we want Cancel to do a proper shutdown action
before cancelling a browser context and killing the process, we could do
that change within the cancel logic.
For some reason, this test fails about half the time on Travis, but I
can't get it to fail even after stress-testing hundreds of concurrent
runs.
It might be because Travis is on a much older version of Chrome. We'll
fix that soon, by having chromedp select a specific version of Chrome.
For now, make it more conservative, not assuming that a Location after a
Navigate isn't racy.
Many consecutive calls to Run can be collapsed into a single call. While
at it, make the error handling style more consistent. Overall removes 70
lines of repetitive code.
This can simplify some common use cases, like running a few actions
directly, or running no actions at all. It's also an almost entirely
backwards compatible change, as all Run call sites should continue to
compile and work.
Leave Tasks, as it can still be useful for functions to return complex
sequences of actions as a single Action.
We broke this in the refactor because of a nil pointer dereference, but
we didn't catch that as none of the tests loaded a page with an iframe.
That is, a page with multiple frames.
Add such a test, and fix the bug by creating an almost-empty frame when
we start receiving events about a new frame before it's navigated to.
Using a smaller viewport speeds up both tests, and lets us know what
dimensions to expect in TestCaptureScreenshot.
For TestScreenshot, we can know what dimensions to expect in advance, as
we have the images in testdata.
'go test -run Screenshot' goes from ~0.9s to ~0.5s on my machine.
Finally, don't run ExampleTitle as part of 'go test', as it's slow.
The navigate sleeps can be replaced by appropriate wait actions.
Some other tests don't need any sleeps at all. This might be because
work is done synchronously now; I haven't been able to get test flakes
after hundreds of test runs with flags like -parallel=32 -count=200.
We hadn't noticed a few uncaught exceptions being received from the
browser, because the events were ignored. Start printing them via the
error logger.
The ones we were getting were caused by testAllocate running Navigate
actions when the path argument was empty. Navigating to "testdata/"
causes JS exceptions, as it's not a valid page.
Instead, leave the new target pointing at a blank document.
First, we want all of the functionality in a single package; this means
collapsing whatever is useful into the root chromedp package.
The runner package is being replaced by the Allocator interface, with a
default implementation which starts browser processes.
The client package doesn't really have a place in the new design. The
context, allocator, and browser types will handle the connection with
each browser.
Finally, the new API is context-based, hence the addition of context.go.
The tests have been modified to build and run against the new API.
- Added contrib/grab-headless_shell.sh to grab latest headless_shell
- Changed travis config to use contrib/grab-headless_shell.sh
- Reenabled TestStop since Page.stopLoading is now available with latest
available headless_shell
