Summary: @public
Since we added packager-managed assets -- internally we still think of asset dependency as a single "module". In reality there are multiple files that represent this module. This becomes important with the `getDependencies` API which is used by Buck to inform it on what to rebuild. Since `getDependencies` deals with modules, and is more of an internal API, I've introduced a new one and would go on to deprecate this.
Reviewed By: @frantic
Differential Revision: D2487207
Summary: @public
Have a top-level debug namespace: `ReactNativePackager`
And add a couple of debugs in the transformer. This is ground work for adding a verbose option.
Reviewed By: @DmitrySoshnikov
Differential Revision: D2489960
Summary: @public
We swallow errors like it's nobody's business :(
Having an error handler that `reject`s after the promise has been resolved is a no-op. In node, if there is no `error` event handler then the error would throw.
So after we start listening and we want to resolve the promise, we remove the error listener so that we make sure errors actually throw.
Finally, I made the `uncaughtError` handler log `error.stack` so we can get an idea of what's going on.
Reviewed By: @martinbigio
Differential Revision: D2468472
Summary: @public
The server dies after 30 seconds if it has no jobs on it's queue. The problem is that the jobs counter gets decreased before returning the bytes to the client. As a consequence, it's possible that the server dies while it's returning the bytes to the client, or just after it finished returning the bytes to the client.
To avoid both issues lets move the counter decrease a few lines below and bump the timer to make sure we have time to fully write the bytes on the socket and let the client close the connection before the server dies.
Reviewed By: @vjeux
Differential Revision: D2445264
Summary:
1. When the server starts up, it only gives itself 30 second to live before receiving any connections/jobs
2. There is a startup cost with starting the server and handshaking
3. The server dies before the client has a chance to connect to it
Solution:
1. While the server should die pretty fast after it's done it's work, we should have a longer timeout for starting it
2. I also added accompanying server logs with client connection errors
Summary:
Saw an issue with a build because of an ENONT error: https://fb.facebook.com/groups/716936458354972/permalink/923628747685741/
My hypothesis:
1. We issue a ping to the socket (in SocketInterface/index.js) a decides if the available socket is alive
2. We see that it's alive but by the time we actually connect to it the server would've died
Solution:
1. The server shouldn't die as long as there are clients connected to it (currently it only stay alive as long as there are jobs)
2. The "ping" should only disconnect once the client is connected
3. Finally, have a better error message than ENOENT
Summary:
Fix failing test that matches the exact error string to match using `contains`.
I was under the impression that jest tests were running in CI -- turns out not yet.
Summary:
A few potential races to fix:
1. Multiple clients maybe racing to delete a zombie socket
2. Servers who should die because other servers are already listening are taking the socket with them (move `process.on('exit'` code to after the server is listening
3. Servers which are redundant should immediatly die
Summary:
Buck (our build system) currently starts multiple packager instances for each target and may build multiple targets in parallel. This means we're paying startup costs and are duplicating the work. This enables us to start one instance of the packager and connect to it via socket to do all the work that needs to be done.
The way this is structured:
1. SocketServer: A server that listens on a socket path that is generated based on the server options
2. SocketClient: Interfaces with the server and exposes the operations that we support as methods
3. SocketInterface: Integration point and responsible for forking off the server
