Yes, I have lots of games.
But do they use OpenGL? You're on Windows, so they might just be DirectX.
I've spent an hour researching your rendering issues. I think they're fundamentally mostly the same (the error messages seem fairly similar) - missing buildings with just an outline rendered, black terrain or the no-show of the water - they're all shader compilation errors, and they're independent of rendering framework (i.e. are observable in ALS and outside) just the symptoms are different. The only genuinely different things are the red buildings and the z-ordering problems.
As for the compilation errors:
* I'm seeing none of them, which isn't overly surprising in itself, NVIDIA drivers are fairly permissive, ATI/AMD not necessarily so, so sometimes bad GLSL slips by me and causes issues. Usually the error message is descriptive enough to fix it - but not here.
* they have to my knowledge never been reported before for FG (some of the symptoms are fairly prominent, for instance no seeing buildings after switching them on would bound to be notices, so would black terrain - so it seems no other ATI/AMD user who is on the devel list has seen them either
* The cause reported by the error message (dual definition of a main function) is definitely not the actual cause - none of the affected bits of code declare two main function. The line reported for the message is the declaration of the main function in all cases, but there has not been a main function before - not clear what happens.
* I've tried to google the error, but it seems to be very rare, and the explanation given is only as above, so no help here
* I can't see any common theme among the affected shaders - it's agriculture for ALS, water for ALS and non-ALS, buildings for ALS and non-ALS so far, possibly more. Agriculture has last been modified Dec. 2014, so before the 3.4 release. water-sine.frag has last been modified Nov. 2013, i.e. before the 3.0 release. There's no change FG-side which could explain why these shaders stopped working for you when they did in 3.4.
* While I might miss some ATI/AMD specific subtlety, Emilian (who made the last changes to the non-ALS variants) is far less likely to do so and I believe he has access to ATI/AMD architectures to test - the fact that his work shows just the same issues makes it further unlikely that this is a combination of simple slips on my side.
Thus, at this point the evidence would suggest some change on your side, leading to some weirdness (otherwise, I can't for the sake of it figure out why code that hasn't changed since 2013 should suddenly start to not compile).
So, I don't know what is going on, and I don't have any leads to investigate the compilation errors further, but I suspect based on th fact that things which haven't changed stopped working that it's not FG-side. A suggestion to proceed by trail and error would be to try to up-/downgrade drivers - there are some reported cases for ATI/AMD architectures where this solved problems. Or maybe someone else can see what I'm missing.