Allow store and retrieve for workspace/symbols request

For a modest project with 40k symbols (mainly system SDKs I like to browse easily) this translates to 10MB of JSON text. Even though 99% of the files don't change often, or ever I am required to send this text for every request back to the client who must then deserialize it so it can be displayed/sorted in the editor (I am aware of query string but that's another issue with potential solutions).

I would like to propose a syntax which allows me to send this data once to the client (which can then store it in client-side memory) and then request it to be retrieved later from server.

The model follows:

• Associate a unique ID with a symbol (perhaps an additional field in SymbolInformation which invalidates the need for other fields)
• Send all symbols in workspace to client upon the initial workspace/symbols request (or when files change)
• On subsequent requests send the ID to the client which can then use that as key to look up the data which was sent previously.

Is this a feasible approach? Currently as it's designed sending 10MB of text for each request is untenable for performance and abusive on system resources.

Start by reviewing the workspace/symbols request and the SymbolInformation model described in the issue. The issue proposes a protocol-level storage and retrieval design, so done would require an agreed feasible approach and corresponding specification changes; no implementation file or test is identified.