[sldev] Re: 1.14.1.2 protocol changelog

[Joshua Bell]

Argent Stonecutter wrote:
>
> Are the comments accurate, because they imply the actual resolution 
> has not been improved and U8 is plenty big enough to hold 0..1 in 
> steps of 0.01.
The comments were incorrect before, and the developer didn't update 
them. There's a lot of inconsistent conversion from U8/U16 to F32, so 
it's hard to... um... quantize the quanta.

Quoting the developer:

        to minimize trouble, i followed the existing adhoc techniques,
        but upped the precision.

        specifically: path begin, path end, profile begin, profile end,
        and hollow have all be raised to U16 (and the conversion
        mechanisms take advantage of that added precision.)  not all of
        this precision has been exposed to the gui (nor would we want
        to, i think.)  but the visible precision is now maintained (it
        wasn't in the past) and any precision set via LSL calls is
        maintained.

        also - the texture scaling parameters (repeat) have been raised
        from U16 to F32.