more advanced repl backends, rules-based perhaps?

Conditions of interest would apply to, I think, 3 classes of rules:

A) Line matching a regexp

B) Line has not been transmitted to any session yet

C) Other lines

For each class, the user should be able to define a minimum number of seconds/minutes/hours, or number-of-subsequent lines for which that line should be replicated. Lines matching more than one rule class would be replicated for the largest amount of any of its matching rules.

Under such a scenario, I could say that all lines about me or jabber are replicated for a week, lines not transmitted to an active session are held for up to 3 days, and all conversations/channels get a minimum of 100 lines held, so when I connect I always get some context for where things had been before I connected.

Just some ideas: [[BR]]
- limit replicated lines per channel to a global number of n lines (#71) [[BR]]
- set a limit n1 for channel #a, #b and #d, but don't set it for #c and #e, and set limit n2 for #f [[BR]]
- replicate all lines that match the current nick [[BR]]
- replicate all lines said by some other nick [[BR]]
- set matching to nick!user@host mask [[BR]]
- matching could be either a wildcard string or a regexp [[BR]]

I propose a setting where you define an x amount of lines. Then ctrlproxy would replicate x-lines around highlights and also mark that the replication was started/ended.[[BR]]

Point of this is that often the line, in which someone highlighted you carries very little information and the actual char is around that one specific highlight-line.
[[BR]]
In addition, there should be a setting for maximum lines to replicate to prevent flood and/or other problems.