+ '/
+ \[\[
+ :? # ignore optional leading colon
+ ([^\]|]+) # 1. link target; page names cannot include ] or |
+ (?:\|
+ # 2. a pipe-separated substring; only the last is captured
+ # Stop matching at | and ]] without relying on backtracking.
+ ((?:]?[^\]|])*+)
+ )*
+ \]\]
+ ([^[]*) # 3. link trail (the text up until the next link)
+ /x',