MOON-CLJ

@SergiusTheBest because we want to use util::FormatTime for output time format like "2018-01-20 22:24:35.496" in our code。

together with struct Time and util::ftime

@mckelvin i also think it's good practice to separate 3rd party's code from the source code，but i don't insist。maybe we will add more deps or add more tools like cppcheck?

@mckelvin this pr contains two points， 1，as suggestion by https://man.openbsd.org/connect.2 ，also see the https://man.openbsd.org/connect.2#EXAMPLES 。 add getsockopt with option SO_ERROR after poll indicate ok to double check after connect。it's not...

@mckelvin this pr add some log in info level。specially, pointer address of "this" is new。pointer address is useful to track reconnecting，host:port or alias is not enough。these logs help us discover...

test still not success ``` chenlijun05@MBA-FGW0YMQ6LX-0335 gomonkey % docker run -it --rm -v /Users/chenlijun05/go:/go --privileged golang bash ``` ``` root@a8e7d44bfc3c:/go# cd src/github.com/agiledragon/gomonkey root@a8e7d44bfc3c:/go/src/github.com/agiledragon/gomonkey# cd test/ root@a8e7d44bfc3c:/go/src/github.com/agiledragon/gomonkey/test# GOARCH=arm64 go test -gcflags=all=-l...

introduce pre-commit to this boilerplate，not make sense to me。

@antirez thx for reply! i understood the design that may failover the same master with same or differ epochs almost at the same time, but as you say this is...

@antirez i got what you said! your worry about delayed failover again and again is not unnecessary, in actually, this happened in our test env after we patch this pr...

hi @antirez i think your desc is a possible scenario. the experience happened several month ago, unfortunately we clear our test data again and again until now. i dig into...