moandor-y

> This is a rather big change in the behavior and impacts a lot of use-cases/tests. > > I understand it's desired to move in this direction but I feel...

> I think we need to consider what is the problem are we trying to solve here with this PR and what is its impact and what's the impact of...

> This behavior is intentional. When forming an Operational Dataset, the Active Timestamp must be different to ensure that one Operational Dataset takes precedence. If this is intentional, I think...

@taoyinzhou I think you can set the second BR's full dataset to the exact same one in the first BR. If you set only the master key, it will only...

> The Extended PAN ID is intended to provide a unique identifier for a Thread network, so it should be randomly generated. > > When joining a node to an...

> Agree that the `GenerateLocal` behavior should be removed. > > @Moandor-y Will you take on this fix, which may involve fixing many test scripts, including OTNS tests? I agree....

@jwhui @abtink what do you think?

> Be careful that it might cause OpenThread THCI or other scripts to fail. @Moandor-y Yes. At least all scripts using regex to match IPv6 addresses need to be changed...