[ft] Loading, Hinting and Rendering Without Translation

[Mickey Gabel]

If I understand the docs correctly:

1) Hinting is done during FT_Load_Glyph().
2) The hinting process only cares about current EM size and the 
fractional part of the current translation (assuming the current 
transform matrix is identity).
3) Rendering using FT_Glyph_To_Bitmap() receives a translation vector, 
but all it really cares about is the actual fractional part. Translation 
by integer amounts only affects the "left" and "top" values, and the 
actual rendered bitmap (size and contents) will be the same.

Putting this together, it seems that as long as I'm fine with integer 
translations, I don't ever need to use FT_Set_Transform() and 
FT_Glyph_Transform().

Assuming I am not trying for WYSIWYG, and don't care about 
rotation/shearing/etc, (but I do care about hinting, anti-aliasing and 
visual quality) is there some special reason for me to use FT for 
translation?

Here is my current pipeline that seems to work:

When loading:
        FT_Load_Glyph()
        FT_Get_Glyph()
        use glyph metrics to calculate pen position

When rendering:
        FT_Glyph_To_Bitmap()
        calculate bitmap origin from previous metrics
        copy the bitmap to origin on the destination image

Note that no FT transformation is done anywhere at all, I simply use 
whatever defaults FT2 starts with (I assume (0,0) and identity matrix).

Thanks,
        Mickey