Re: [mosquitto-dev] 答复: A non-centralize Mosquitto cluster design.

[Tatsuzo Osawa <tatsuzo.osawa@xxxxxxxxx>]

Hi Jianhui,

I'm looking forward to hearing your result.

Please remember the steps I recommended as followings:
1) Evaluate outcomes of your creation roughly and prove the values for
proceeding.
2) Make plan to divide the whole of your creation into some minimized PRs.
3) Create the 1st PR. It is recommended less than 1000 lines.
4) Post the PR on develop branch, get review by others, merged by committers.
4) Repeat 3)-4) with remained PRs.

You need to improve the quality at step3.
So far you are at 1) , and should focus to demonstrate your outcomes
(etc. performance improvement).

See also https://github.com/eclipse/mosquitto/blob/master/CONTRIBUTING.md .

I'd like to talk with you on this dev-ML until the step 2).
After the step 3), we should move to
https://github.com/eclipse/mosquitto/pulls .

Regarding MQTTv5, see:
http://docs.oasis-open.org/mqtt/mqtt/v5.0/csprd02/mqtt-v5.0-csprd02.html
https://github.com/eclipse/mosquitto/pull/519 (It is my PR and seems
to be a good sample for you as the "1st PR".)

Regards,
Tatsuzo

2017-12-08 1:18 GMT+09:00 zhan jianhui <hui6075@xxxxxxxxxxx>:
> Hi Osawa,
>
> No problem, but I need some time to make sure every line of code goes well
> such as the link list add & remove,  memory allocate & free, to prevent
> potential crash and memory leak since I wanna make it as a product. Then I
> will do the benchmark. Could you share some benchmark tools? I have use
> Tsung but seems it is not flexible to simulate my IoT scenario.
>
> Btw, could you share some material about iot.mosquitto and MQTTv5? I got
> nothing on the internet.
>
> BRs,
> Jianhui