Can you explain the issue with having a hard-coded key for the local replica environment? Is there some issue with the local dfx replicas all having the same key for development?
Seems like a simple fix to just hard-code that if using dfx. I understand that there might be other deployments of the IC, such as test networks or who knows other production or private networks later on. But I think there should be an option to hard-code it for this case.
Perhaps the agent can just use the hard-coded keys basing it off of the URL (isnāt that how it knows to use the hard-coded production key?), if itās a URL that looks local (localhost or 127.0.01 etc) then it can use the hard-coded development key. And only outside of those two situations (which currently would be very rare), a custom key can be fetched.
URL is only a thing in a browser context though. Also, when custom domains come into play, it becomes even more complicated.
the host passed to the httpAgent would be the best existing feature to hook into, but at the moment thatās private, so the only time weād know whether to use it for fetchRootKey would be to check it in the createActor args, so that might be a safe enough enhancement
Hi. I have a question (or suggestion) in the local environment if several Actors (Canisters) are randomly generating identifiers. Is it possible to somehow hard-link them in the dfx.json file?
Weirdly, we seem to have released 0.11.2 recently, but it doesnāt seem to have the feature yet. The feature will be in 0.12.0, and you can try it out today by installing 0.12.0-beta.1