[Teem-users] I hate gage_t

[Gordon K. <gk...@bw...>]

hello,

I think some problems I'm seeing with using gage on tensor  
derivatives have to do with the fact that I'm using float instead of  
double for very small quantities that are being squared and cubed.

So I changed gage_t from float to double, and now everything is  
zero==black.  The cause of this is probably the fact that there is  
something complicated logic in gage to do things differently  
according to whether gage_t is float versus double.  Part of  
implementing this logic involved adding a whole bunch of stupid Ell  
macros that do casting as part of their operation.  Bad scene.

What I plan on doing right now is ripping out gage_t and making all  
its internal representation be double.  Does anyone object?

The reason I think this is okay is that gage_t is never used for  
storage-- Teem never allocates large arrays of type gage_t-- it is  
only used for temporary or intermediate storage.

Gordon