sewenew

> When a redis instance is going down, I need the socket_timeout small enough to return the error immediately. First of all, how can you tell if a redis instance...

@wingunder > I actually have a use-case for this request: > .... > In order not to use a Redis instance per thread In this case, you CAN have a...

@wingunder > Dead Redis nodes can be determined independently, by subscribing to specific keys of a Sentinel server. Yes, you can be notified by Sentinel, but only when the node...

Hi @wingunder Thanks for your suggestion! I've updated the doc, and you can take a look: [commit ](https://github.com/sewenew/redis-plus-plus/commit/cba5f853c4fb0bccc957f7a481869cafd72ce03b). Regards

> One more benefit of setting socket_timeout per request is that application can control what command should be blocked until server replied, or should we have some timeout for it....

What do you mean by reusing Redis object out of scope? If an object, no matter it’s Redis or other c++ object, is out of scope, you can not use...

> Since redis-plus-plus has gone down the return-code route of error handling NO. In fact, redis-plus-plus uses exception, instead of return code, to report error. Some `Redis`' methods return bool....

@wingunder Thanks for the suggestion! I did some rough research on this `PUSH` feature. It needs Redis 6.0 and higher version of hiredis support. Also if it's used with synchronous...

Hi @mellis13 First of all, it's not a good idea to get/set large message from/to Redis. It's bad for Redis performance. Since we need to convert a C char-array to...

@BratSinot Good try! However, we cannot return `std::string_view` from a redis command, such as `Redis::get`. Because `std::string_view` doesn't own the underlying memory, and once the command returns, the underlying memory...