Re: [AD] test 'projection'

[SiegeLord <slabode@xxxxxxxxxx>]

On 01/07/2014 06:51 AM, Peter Wang wrote:
> Maybe drawing a rectangle instead of a circle would have been more
> reliable.  Well, it's up to you.

The test needs one of the curved primitives which compute the number of 
segments used based on the scaling factor transformations used (the 
original bug involved them not taking into account the projection matrix 
and thus using too few segments in some cases), I suppose one of the 
other ones could work better, but there's no guarantee.

-SL