Ignacio Hagopian

To be clear, on *connected peers*. There might be cases where we know about log addrs, but we aren't connected yet (for some reason). Maybe you're thinking of making sure...

Sounds great

Yes, that shouldn't be fatal since there isn't anything corrupt or similar... just an expected error.

I have the feeling that maybe I left this invariant hold now, but not 100%. Good for checking again, and for future APIs (if they mutate things in thread (or...

@requilence , thanks for your report! Yes, there's a big limitation in the json patcher since it doesn't have any kind of causal ordering. Apart from the scenario you describe...

Currently, it isn't possible, and a bit out-of-scope to what `lotus-devnet` is meant for (allow applications to test Lotus APIs or use in CIs). Tunning the miner's address or private-key...

You don't need to pass an auth token when using this devnet. All API calls should be authorized even if you send an empty token.

For our purposes, this problem isn't a high priority anymore since we worked around relying on `IpfsOnlineMode=true`.

@dtbuchholz, thanks for the details. Testing [that query directly in the gateway](https://testnet.tableland.network/query?s=SELECT%20year||%27.%27||month%20as%20date,avg(eth)%20as%20eth_avg,avg(usd)%20as%20usd_avg,min(eth)%20as%20eth_min,min(usd)%20as%20usd_min%20FROM%20netvrk_trade_137_71%20group%20by%20date%20order%20by%20date%20desc%20limit%201), it gets resolved fast. This is just to confirm that ~1.2s duration isn't related to the query itself,...

@dtbuchholz, an offtopic comment if this wasn't apparent when seeing the result of the query. That table has this schema: ``` CREATE TABLE netvrk_trade_137_71 (id INTEGER PRIMARY KEY, token_id TEXT,...