hunterhug

my docker deploy ``` version: '3' services: mynsqlookupd: container_name: "mynsqlookupd" image: nsqio/nsq:v1.2.1 command: /nsqlookupd --http-address=0.0.0.0:5161 --tcp-address=0.0.0.0:5160 restart: always ports: - "5160:5160" - "5161:5161" mynsqd: container_name: "mynsqd" image: nsqio/nsq:v1.2.1 command: /nsqd...

> I think this change in behavior was caused by #1159 > > Previously, the "memoryMsgChan" would exist, but be unbuffered, but this allowed messages to instantly pass through the...

> > if my messages over 100, is it will store in disk too? > > Not deferred messages, I think: both the topic and the channels have a mem-queue,...

> Is it possible to use singleton NSQ Producer? I'm running an API with NSQ Producer and it seems like i have to open and close connection everytime i hit...

etcd use `github.com/spf13/pflag`, but it forks from`github.com/ogier/pflag`. Which is the main version? funny~

> 作者你好，在step内部使用的好像不是插入排序，更像是冒泡排序，代码对比如下，请指教： > > ``` > package main > > import "fmt" > > // 增量序列折半的希尔排序 > > func ShellSort(list []int) { > > // 数组长度 > > n :=...

> 多次分组进行插入排序，使得最后对整个列表排序时，列表已经是大致有序的，那么最后一次插入排序就会时间复杂度就趋向线性的。 > > ``` > func ShellSort(list []int) { > n := len(list) > > // 增量每次减半 > for step := n / 2; step >= 1; step /=...

> > @hunterhug > > > 多次分组进行插入排序，使得最后对整个列表排序时，列表已经是大致有序的，那么最后一次插入排序就会时间复杂度就趋向线性的。 > > > ``` > > > func ShellSort(list []int) { > > > n := len(list) > > > > > >...

https://github.com/hunterhug/gomap.git

Good