XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
由于宏展开的机制限制，（至少在最新的 Rust 中）
不可能做到把一例宏的展开结果作为有效信息提供给另一例宏。
这为宏的模块化工作施加了难度。
使用递归并传递回调 (callbacks) 是条出路。
作为演示，上例两处宏调用的展开过程如下：
recognize_tree! { expand_to_larch ! ( ) }
println! { "I don't know; some kind of birch maybe?" }
call_with_larch! { recognize_tree }
recognize_tree! { larch }
println! { "#1, the Larch." }
可以反复匹配 tt 来将任意参数转发给回调：
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
如果需要的话，当然还可以在参数中增加额外的标记 (tokens) 。