You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I understand that this issue is mostly related to vscode-languageserver but since pyright is the end product I'm trying to use I'm posting the issue here.
The behavior of --socket argument is wired. First it accepts port number, not path to a socket.
Second this port number is a port where client listens . On start this port sends a couple of messages then it just listens to initialize request. If there is nobody listening on that port, the server exists. I see no reason on this behaviour. These messages are not needed to the client, it can send initialise request without them. Additionally, it makes the only order that client and server must be started - first client, bind to port, and then provide this port to the server.
I see no reason, why this order can't be reversed - first run the server, make it listen on some port, then tell the client which port to connect to. And It gives the client an ability to reconnect to the server without restarting it. Isn't it much simpler way than making the client responsible to behave like a server?
The text was updated successfully, but these errors were encountered:
I understand that this issue is mostly related to
vscode-languageserver
but since pyright is the end product I'm trying to use I'm posting the issue here.The behavior of
--socket
argument is wired. First it accepts port number, not path to a socket.Second this port number is a port where client listens . On start this port sends a couple of messages then it just listens to initialize request. If there is nobody listening on that port, the server exists. I see no reason on this behaviour. These messages are not needed to the client, it can send initialise request without them. Additionally, it makes the only order that client and server must be started - first client, bind to port, and then provide this port to the server.
I see no reason, why this order can't be reversed - first run the server, make it listen on some port, then tell the client which port to connect to. And It gives the client an ability to reconnect to the server without restarting it. Isn't it much simpler way than making the client responsible to behave like a server?
The text was updated successfully, but these errors were encountered: