Creating this issue preemptively to capture the rationale behind this choice.
Search requests that come directly from the server or from the distributed network don't have an associated IPEndPoint. Only search requests that come directly from a user do.
For this reason it doesn't make sense to pass an IPEndPoint argument to the resolver because it will be null 99.999% of the time, and because including it (even marked nullable) would be a breaking change.
I'm marking this wontfix to hopefully keep myself from investigating this again and coming to the same conclusion. If anyone can make a convincing case for why this is useful, I'll reconsider.
This check should be done client side, and a user endpoint cache should be used to avoid a double-fetch when resolving the user's address to send the results.
Creating this issue preemptively to capture the rationale behind this choice.
Search requests that come directly from the server or from the distributed network don't have an associated
IPEndPoint
. Only search requests that come directly from a user do.For this reason it doesn't make sense to pass an
IPEndPoint
argument to the resolver because it will be null 99.999% of the time, and because including it (even marked nullable) would be a breaking change.I'm marking this
wontfix
to hopefully keep myself from investigating this again and coming to the same conclusion. If anyone can make a convincing case for why this is useful, I'll reconsider.