We still have the --http option as a fallback for now. Once we're confident the
--no-http version works, we'll drop the --http support.

Also, create the FLX in a temp directory and then delete the temp directory
when we're done. Finally, pull the Linux artifacts from the cloud storage
bucket that the buildbot is uploading to.

add a persistent daemon/server mode to sky_tools

wrap up first daemon implementation

...by requiring a new version of analyzer.

Also, when the subprocess is terminated unexpectedly, report the actual
problem in more detail.

abstract some OS operations

When the engine dies unexpectedly during test execution, we have to
terminate any tests running in that engine. Previously, they would just
hang. For some reason that I was never able to satisfactorily explain,
the WebSocket doesn't die in a way I can detect in this case. So
instead, we hand in a future that we only complete when we detect the
server subprocess ends.

use stack_trace to create more readable exception traces

The previous attempt to port the 'test' framework to the new framework
wasn't super-successful. This does a better job, hopefully.

skip a test on windows

This flag builds a local FLX file and pushes that to the device instead of
using an HTTP server.

increase test timeout

Stop requiring customers to use a dev SDK

Modify run_mojo command to take path to devtools.

Generalize path handling in flutter init

Made changes as discussed.
R=@jamesr

In case we run into trouble, this at least gives us an escape hatch.

This will make "-o ." work as expected to initialize a project in the
current directory

Wait for sky server to start before starting device

Instead of just waiting for the sky server process to start
before we start the activity on the device, this causes us to
wait for the sky server to actually start listening on its port

Fixes #141

Depend on exact version of flx package so I can make breaking changes.