Fix realtime publisher race condition upon initialization

[saikishor]

Should fix : #24
Closes: #308

@bijoua29 can you test this and see if this helps?

[bijoua29]

Note my comments and see if you agree

[bijoua29]

@saikishor Yes, I will test it today or tomorrow. Thanks for fixing this.

[christophfroehlich]

This looks great.

So we don't need something like

while (!publisher_->is_running()) {
    std::this_thread::sleep_for(std::chrono::milliseconds(10));
}

in the application code, right?

[saikishor]

This looks great.

So we don't need something like

while (!publisher_->is_running()) {
    std::this_thread::sleep_for(std::chrono::milliseconds(10));
}

in the application code, right?

Yess! Exactly

[bijoua29]

@saikishor I ran a test 20 times pre-PR and had a 85% failure rate. I ran another test 20 times post-PR and had 100% success. So I think the PR fixes the problem. I am approving the PR.

[bijoua29]

LGTM