Kavinli

docs: Chaosblade documentation improvement program.

2

## Issue Description The chaosblade documentation has received a lot of flack for lacking a good documentation for a tool that is known for its simplicity and power. 1. https://harness.io/blog/devops/chaos-engineering-tools/...

I'm a newbie and I tried to source compile influxdb on mac m1, but I get the following error: I followed the error and executed `rustup target add aarch64-apple-darwin`, but...

符号从中文改为英文

在`example/counter/run_xxx.sh`和`example/block/run_xxx.sh`中IP变量的获取是执行如下语句`hostname -i | awk '{print $NF}'`。 容器的网络模式默认是bridge，在这种模式下容器启动后会通过DHCP获取一个地址，这就导致`hostname -i`会生成两项，导致在宿主机中使用了容器生成的虚拟网卡，从而导致执行样例失败：   现在的问题是如何优雅的解决这个问题，如果IP的获取不通过shell，而使用C++，这种问题很好解决，我们只需要跑自连接，成功当前环境就是宿主机，否则就是容器环境。但是是否有方法在不改变现有获取配置的方法的情况下解决这个问题。

Replace raxIteratorAddChars with raxStackPop to improve performance

3

It takes a little time to clarify this problem, but it does make our code more efficient. First, before judging lt and gt, we executed `(i != len && it->node->iscompr)`,...

fix: dragonfly compilation failure due to glibc version less than 2.30

1

Signed-off-by: Super-long This is a problem to be solved, but not necessarily the final version. I encountered the following problem when compiling dragonfly： In the man documentation I found this...

feat(server): implement heavykeeper data structure. #257

15

Signed-off-by: Super-long This commit implements a hotspot identification data structure - HeavyKeeper, and here is some description of this implemented version. **heavy_keeper.h**： 1. The basic version of the original paper...

This PR adds the get option to stringfamily. Two modifications. 1. add memcache_reply_builder/redis_reply_builder, replace the existing static_cast. 2. StringFamily::Set returns a different return value to the client in the presence...

Why Expirable-LRU delete Close method？

7

In using Expirable-LRU it was found that if we wanted to dynamically close Expirable-LRU, a deleteExpired goroutinue would remain. and it looks like this can be done by simply deleting...

引入内存池提高内存分配的效率

1

这其实是很现实一点，举个最明显的例子，当使用多个共享栈的时候 ，co_routinue.cpp的save_stack_buffer中会多次对save_buffer进行拷贝，而这会调用大量的malloc/free。这里显然我们可以引入一个内存池（不是对象池），在我仿照libco实现的协程库中我引入了gperftools中的tcmalloc，在协程的创建和5000个以上的协程的频繁切换时均优于libco。当然直接使用tcmalloc会引入一个较重的框架，所以此时实现一个内存池就是一个较为正确的选择。