https://github.com/trekhleb/javascript-algorithms/blob/master/src/algorithms/sets/cartesian-product/cartesianProduct.js#L8

I think it makes more sense to return an empty array

// invalid input return null or throw an error
if (!setA || !setB) {
    return null;
}
if (!setA.length || !setB.length) {
    return []; 
}
// or do nothing since the for loop will not trigger anyway

集合 S 包含从1到 n 的整数。不幸的是，因为数据错误，导致集合里面某一个元素复制了成了集合里面的另外一个元素的值，导致集合丢失了一个整数并且有一个元素重复。

给定一个数组 nums 代表了集合 S 发生错误后的结果。你的任务是首先寻找到重复出现的整数，再找到丢失的整数，将它们以数组的形式返回。

示例 1:

输入: nums = [1,2,2,4]
输出: [2,3]
注意:

给定数组的长度范围是 [2, 10000]。
给定的数组是无序的。

来源：力扣（LeetCode）
链接：https://leetcode-cn.com/problems/set-mismatch

Describe the bug

To Reproduce
Steps to reproduce the behavior:

1. Go to 'Edit cells -> Transforms…'
2. Enter 'true.type()' in the Expre

The documentation of the boltons.timeutils.daterange function only mention date objects as input, but I discovered that it works with datetime objects as well. Explicitly stating so would be great. Thanks!

首先，十分欢迎你来给 OI WIki 开 issue，在提交之前，请花时间阅读一下这个模板的内容，谢谢合作！

• 希望添加的是什么？

git 使用指南

• 英文叫什么？

git-guidebook

• 有什么参考资料？

https://git-scm.com/docs
https://git-scm.com/book/zh/v2

issue 标题请写为 'add ' + 要添加的内容

然后可以在 htc 里 link 一下。（我刚开始写 Wiki 的时候就不知道怎么做到“一个 pr 只干一件事”）

I'd suggest a better differentiation between Best and Breadth First Search in the Graphs section.

Acronyms BFS and DFS are used all over the place, but "breadth" and "depth" aren't mentioned anywhere, making it hard for a newcomer to understand what is being referred to.

steps to reproduce

Write a loop, from 1 to 80000, each time add a random int to the max heap.

In theory it takes very little time(NlogN, N=80000, <1sec ), but the program does take a long time.

I'v also tested the BinaryHeap in https://github.com/SolutionsDesign/Algorithmia, it performs well, so it is probably due to the bad algorithm.

Hi, I'm not sure if more platforms are interesting to you, but I've made a slightly more complete implementation of padding a struct to the cache line size here:

https://github.com/tinco/cache_line_size

I got the information on the cache line sizes from the Go compiler repository, so I'm pretty sure these are correct for all platforms Go runs on, which is a whole bunch.

It's not easy to understand what the code generators under jctools-build are supposed to do. I think one or two phases documenting their goal would be useful.

I am proposing to document these classes:

• JavaParsingAtomicArrayQueueGenerator.java
• JavaParsingAtomicLinkedQueueGenerator.java
• JavaParsingAtomicQueueGenerator.java

Yes, I know I could examine all the headers available, but it'd be nice if each documentation page listed which header(s) you need to include to use it.

The documentation gives some hints that the default hashtable and hashset implementation are assuming C strings as keys. However, the "official" documentation use some defines that have been removed in the meantime, i.e. in 9b4c3cb. As of now using the hashing data structures with anything but strings as keys is rather a PITA. There is al