Vega Strike Code

Bugfixing highend.fp:
- wrappers and macros for easier gamma control
- degamma/regamma fixes:
* degamma-ing of texture input must happen right at sampling, almost any operation will be nonlinear
* regamma-ing output must happen as absolute last step (even after cloaking and alpha application)
- GAR fix: eye vector isn't always normalized, and GAR needs it normalized.

Implemented a system of folders under /techniques. The work is neither complete nor correct nor yet in use. I'm committing it for feedback and peer review.
Essentially, the numbered folders contain shaders and techniques for a spectrum of hardware capabilities. As you can see from the opaque_onetex.technique I implemented as example, the technique has the same in all the folders, --or most, anyways. The user setting for shader in vssetup determines the folder; the xmesh determines the name of the technique. There are also 3 folders for STAGE, meaning the render stage: FAR = planets and stars; MEDIUM = ships and stations and asteroids; NEAR is the cockpit. Shaders and techniques that are common across stages are in the numbered folders under /techniques. Stage-specific shaders and techniques are in the number folders under the appropriate stage folders. Thus, fireglass is in the MEDIUM stage folder, under folder 6_ps4.0, as it applies only to ships and stations. An atmosphere shader would be under the FAR tree. Note that there can only be one shader for all blended applications under each stage, for performance reasons. Thus "fireglass", which currently only renders glass, but will soon also apply to engine exhausts, etc.

Uber technique, that's chuck_starchaser's cinemut pretty much

Fixed the "fixed" shader types to use environment mapping properly

merging trank -r12669 to safemode branch

No need to disable linear filtering on non-MIPped textures. Only clamp mip level.

Added logging output for successful technique compilation.
Helps to know what techniques are being used.

Use dynamic libboost_python when --with-boost=system.
My distro (and many others it seems from the reported troubles in the forum) don't ship with a libboost_python-st (static version).
I checked the repository and it's not even there.

Please, everybody that used --with-boost=system re-check your build to see if it works for you.
If not, we may need to add a --with-boost=system-static.

Add 2.6 to the list of supported versions
(yes, it works)

Changing the default optimization level from 3 back to 2, as -O3 makes vegaserver segfault.