Adding an RPC timeout parameter

[domohawk]

Currently the 5s RPC timeout is hardcoded. If we were to add a command line param to configure it, where should the config live?

• P2P Config as RpcTimeout (applies to all of http/ws/etc)
• Top-level backend config
• Somewhere else?

As for why this might be useful, I'm bootstrapping initial contract state in another process using a call() via local RPC. Fetching a uint256[1000000] memory via a call() takes just under 5s, usually.

[holiman]

SGTM, this could live within the RPC config flags in the TOML config.

[jwasinger]

Can this be closed? The main use-case for configurable timeout is for eth_call and tracing. Configurable timeout for eth_call was merged in #23645. Timeouts for tracing are mostly caused by sending the massive trace responses over http/websocket and the standardTraceBlockToFile is available.

[holiman]

Can this be closed?

Three weeks ago.
1,2, 3 closing!