Different "time" API ?

[Amerio.Stephane]

Hello,

Could you precisely detail the differences with these different time API. Exactly how do they increase from example, and what would be the differences between them:

• #include <UnigineTimer.h> getSeconds()
• #include <UnigineGame.h> getTime()
• #include <plugins/UnigineIG.h> getInterpolationTime()

Which one would be closer to the "wall clock"? Does it have millisecond precision?

[cash-metall]

Hello! 

Engine::getTime - time of application. It returns seconds as a float, so precision after the decimal point is not guaranteed. It deliberately does not account for any spikes and freezes in the application. You can use it e.g. for UI animation.

Game::getTime - time if game. This method is the least accurate. based on Engine::time it also ignore spikes and freezes. Suitable for creating games. Moreover you can change speed of this timer using Game::setScale and Game::setEnabled. That is, after a game pause, the game state will remain where it was before the pause, without drifting far away, and animations will continue from the same position, etc.

Timer::getSeconds method is the most accurate and closest to wall-clock time. It returns seconds as a double precision, allowing operation with microseconds/nanoseconds. It does not account any spikes and freezes (uses QueryPerformanceCounter for win; gettimeofday for unix inside). Can be used for any application independent time calculations.

Syncker::getTime() method returns seconds with double precision. Based on Timer, it also ignores spikes but may be shifted forward or backward in case of synchronization disruption between master-slave. Should be used for critical animations and logic synchronized through syncker mechanisms.

IG::Manager::getInterpolationTime() method can also be shifted forward or backward in case of synchronization loss between host-ig. Between host-frames, it uses Syncker::getTime, so it can also be shifted in case of master-slave synchronization disruption. This is used only for interpolating motion of entities obtained with timestamps from the host. 

[Amerio.Stephane]

Thanks a lot! All this would be a great addition to the doc as a summary for future reference