Concerning the internal hardcoding of servers, a few avenues are available:
Consider all caching short-term and push our hard-coded servers to upstream sources
The above, but pubkey cache is preserved as is, or in one of the manners specified below
Introduce a submodule repository for known servers and pubkeys, updated by less frequent CI using Shodan and other scrapers
Same as above, except results are simply polled by script instead of using a submodule, which requires pushes to update
Leave as is
Decisions should consider the following:
How much work is adding new hard-coded Communities?
How much work is adding new pubkeys to ensure security?
How much censorship liability does hard-coding give the project?
How much is it in the project's interest to be a pure aggregator with no hard-coded baggage?
Concerning the internal hardcoding of servers, a few avenues are available:
- Consider all caching short-term and push our hard-coded servers to upstream sources
- The above, but pubkey cache is preserved as is, or in one of the manners specified below
- Introduce a submodule repository for known servers and pubkeys, updated by less frequent CI using Shodan and other scrapers
- Same as above, except results are simply polled by script instead of using a submodule, which requires pushes to update
- Leave as is
Decisions should consider the following:
- How much work is adding new hard-coded Communities?
- How much work is adding new pubkeys to ensure security?
- How much censorship liability does hard-coding give the project?
- How much is it in the project's interest to be a pure aggregator with no hard-coded baggage?
Concerning the internal hardcoding of servers, a few avenues are available:
Decisions should consider the following:
Related issue on PySOGS: https://github.com/oxen-io/session-pysogs/issues/185