建议做表达式缓存

[killme2008]

建议做表达式缓存

[killme2008]

启用 cache 模式

sum += (Long) AviatorEvaluator.execute("a+b*c", env, true);

1万次只要 110 毫秒左右。

[killme2008]

你怎么测试的。。怎么可能

    public static void main(String[] args) {
        long start = System.currentTimeMillis();
        Map<String, Object> env = new HashMap<String, Object>();
        env.put("a", 100);
        env.put("b", 101);
        env.put("c", 102);
        long sum = 0;
        for (int i = 0; i < 10000; i++)
            sum += (Long) AviatorEvaluator.execute("a+b*c", env);
        long end = System.currentTimeMillis();
        System.out.println((end - start) + "ms.");
        System.out.println(sum);
    }

1 万次调用 2.5 秒左右，平均 2.5 毫秒。

[lfkdsk]

Hi @killme2008 ~
1 million 110 ms efficiency does perform well, but I noticed that the default situation is using Eval mode, the general Eval model on the efficiency of this short expression should have a greater room for improvement.
I tried to implement an expression engine that supports compilation and interpretation of execution. Trying to use meta-cyclic explanations in the explanation to directly evaluate the performance of Ast's current performance.

I think aviator can also try to optimize the AST / eval mode to make it better.
Thank you.