3 ms·
They're forking a new process for each request. Passing the request data to that process on stdin when you create it is much simpler than spawning a process, wa
by autarch 8y ago
They're forking a new process for each request. Passing the request data to that process on stdin when you create it is much simpler than spawning a process, waiting for it to be ready, and then communicating with it via http.
- apendleton 8y ago> They're forking a new process for each request. Hm, I don't think they are. I think they one persistent process managing the state of each document, and many requests might interact with that document, via stdin/stdout channels on that document's process. Makes the case for the rest of what you said all the stronger, though.