问题描述
我有一个关联操作>>.问题在于它的成本线性取决于其左操作数的大小.所以一个表达式由一系列 n 的 >>> 应用组成，比如
I have an associative operation >>. The problem is that its cost linearly depends on the size of its left operand. So an expression formed by a sequence of n applications of >> like
a >> a >> a >> a >> a >> ... >> a
就 n 而言，它具有二次成本，因为默认情况下中缀运算符是 左结合.如何使其右结合，以使这样的表达式的成本在 n 方面保持线性?
it has quadratic cost in terms of n, because by default infix operators are left-associative. How to make it right-associative so that the cost of such an expression is kept linear in terms of n?
推荐答案
我找到了解决方案.Scala 参考 在 6.12.3 中缀操作部分说:
I found a solution. Scala reference says in section 6.12.3 Infix Operations:
运算符的结合性由运算符的最后一个字符决定.运营商以冒号:"结尾的是右结合.所有其他运算符都是左关联的.
因此，将 >> 重命名为 >>: 就足够了.
Therefore it was enough to rename >> to >>:.
我花了一些时间才意识到，虽然 a >>b 脱糖为 a.>(b)，a >>: b 脱糖为 b.>>;:(a).所以我必须将 >>>: 定义为
It took me some time to realize that while a >> b is desugared into a.>>(b), a >>: b is desugared into b.>>:(a). So I had to define >>: as
def >>:(x: T): T = x >> this
这篇关于如何制作右结合中缀运算符?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持！