I think that there is up only because it is the weaker version of hover engines.
Forward is the weaker version of main and backward is a weaker version of retro.
I think that the idea was to have the weaker version of all "big" engines.
Yes, that was the idea.
Even so, I also think that having other RCS options could be usefull.
Don't you think that it could become annoying to have to circle through so many engine types?
Anyway, Orbiter is still hanging on my end with BTC 2.6 with that scenario
I posted in the TransX thread.
I'm mentioning it here because,
Orbiter 2010 - Testing BTC 2.6 - YouTube
I hope you didn't feel personally attacked. You, dgat, Ripley, aldarion, and others in the TransX dev thread have been extremely helpful, as well as others who do report bugs. I'll get this hang bug solved. I promise. It just seems that some on-line coordination is needed for this task, but so far I wasn't able to provide it. It all depends on a given week's workload.
To be honest I got a bit mad after, you could say, having to fix somebody else's awful mistake in BTC, that you’ve pointed out in that thread. Of course you had no chance of knowing whether it was my fault or not, and I’ve discovered it in the same time as you, which got me thinking whether nobody else had this problem before in 9 years. However, it seems that our use case (debugging with new TransX) was specific enough to cause this bug, unlike normal user handling.
What added to my frustration, was that on Sunday I wanted just to fly around in Orbiter, using my new addon modifications and I got annoyed even more by Orbiter’s general instability, like crashing when leaving Earth’s SOI, or when entering Mars’ SOI. Hence I decided to write this small, maybe too emotional-laden, survey. I guess that the correct approach is to do exactly what Cras said: grow up to this fact, and save the sim’s state regularly. You’ll probably agree though that you cannot show Orbiter to somebody else (like on a presentation), when it frequently crashes, no matter if you’re a good pilot or not.
This "laziness" argument got me thinking a bit more. Let’s think what happens when somebody reports a CTD. The person is instructed to perform a time consuming test that consists of unloading modules and responding with the minimal setup which still reproduces this bug. So he spends his precious time on finding bugs in something that, as a whole, is somebody else’s intellectual property. No wonder that some people refuse to do it.
On a technical level, the described time-consuming test is required because of the fact, that C++ doesn’t throw any Null Pointer Exceptions, which could be caught by an upper scope, like in Orbiter’s core, and print an error on screen, showing the source of the problem, without crashing the sim. But because of the lack of such feature (and for other reasons), the Fathers of C++ advice not to use pointers at all, only references. A question arises - what should be done in exchange for returning a NULL in the event that a called vessel isn’t available (like GetVessel("GL-01"), which may not be present in the simulation)? The answer is - use a combination of a boolean function and a function that returns a reference on success,
and throws an exception on failure, which can be caught by Orbiter’s core. For example:
PHP:
string name = "GL-01";
if (oapiIsVesselExist(name.c_str()))
{
Vessel & vessel = oapiGetVessel(name.c_str());
// it's legal to use vessel
} else {
// Normally this event wouldn’t be handled
Vessel & vessel = oapiGetVessel(name.c_str()); // this would throw an exception
}
Even without modifications of the Orbiter API, like in the above example, some simpler testing framework could be achieved, if only Orbiter was built with Free compilers, like MinGW, instead of MS-VC. Both MinGW and a full version of MS-VC allow for attaching their debuggers to an external process (Orbiter.exe), which allows you to find out which at least which DLL has caused the crash. The problem is, that you can’t debug code with MinGW which is not compiled by it, and a full version of MS-VC, that allows it, costs 600€. Who wants to waste so much money?
