Searches by IPv6 address need to include [], and partial searches need to start with [. Example: "[2001:858:2:2:aabb:0:563b:1526]". That's not as intuitive as it could be. We should consider making the [] part optional for IPv6 searches. Not sure if that's a good idea, but I can't come up with major arguments against it.
This enhancement request originates from #10128 (moved) and has low priority.
To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information
Child items
0
Show closed items
No child items are currently assigned. Use child items to break down this issue into smaller parts.
Linked items
0
Link issues together to show that they're related.
Learn more.
I still didn't come up with reasons against making this change, so I wrote a patch.
The only catch is that we're changing the semantics of the current search parameter, because a search for, say "2001", might now return even more relays: matching hex fingerprint starts, matching base64 fingerprint starts, matching nickname parts, and matching IPv6 address starts. Admitted, it's a tiny change, but it's a change. That's why we should do the major-protocol-update dance where we announce a change 1 month in advance, set next_major_version_scheduled, and merge 1 month later. But before we start that dance, let's get the patch ready for merge.
So, the reason for raising the major protocol number is that a search for 0815 could now return all relays having ipv6 addresses starting with the search string?
Clients supporting 3.1 shouldn't break, b/c of this change. In a way I would count loosening the restriction on searching for ipv6 addresses to adding a new fields. Thus, this could be in 3.2-1.1.0?
Merge ready, but I would say this change doesn't warrant protocol version 4.0.
But not raising the protocol version to 4.0 feels wrong. We are changing the most important parameter by returning potentially more results than previously. That won't break clients, but it might confuse users. In a way, protocol changes leading to 2.0 and to 3.0 were comparable to this one: in both cases we returned potentially more results than before. Those are major protocol changes, because clients either need to adapt or decide they don't care. That's different from minor protocol version updates which clients can ignore if they wish without even looking.
I'll keep this as merge_ready until we tag the next release, in which case I'll commit a placeholder for version 4.0 and set next_major_version_scheduled to 1 month later.
I just rebased this branch to current master and prepared the release that is scheduled for next Tuesday. The new branch is task-16552-3. Feel free to review that branch once more before I merge it next Tuesday, but I'll also ask you to review the release before it goes out. Thanks!