Ensure send() method uses drain() to handle writes correctly

[allrob23]

This PR updates the send() method to ensure it properly queues data when a write operation fails and uses drain() to handle writes efficiently.

Changes

• Added await self._output.drain() after writing data to ensure the write operation is properly flushed.
• This prevents potential issues where data might not be sent immediately, aligning with the expected asyncio usage pattern.

I opened this PR because the asyncio documentation recommends using drain() when writing to a stream. This ensures that writes are properly handled without blocking or data loss, making it a best practice for managing asynchronous I/O.

[mxschmitt]

Looks like all the bots are failing.

[allrob23]

I updated the change

[mxschmitt]

I don't think thats how async methods supposed to be called:

/Users/runner/work/playwright-python/playwright-python/playwright/_impl/_transport.py:179: RuntimeWarning: coroutine 'StreamWriter.drain' was never awaited
self._output.drain()

I'll close it for now since it doesn't seem to fix a real word issue users run into. Feel free to revisit if you have a specific repro which is broken.