[cairo] Getting away with dummy final move_to

[Vladimir Vukicevic]

On Feb 16, 2008, at 11:36 PM, Behdad Esfahbod wrote:

> Adrian pointed out that all paths emitted to PS/PDF have a final
> unneeded move_to.  That's because _cairo_path_fixed_close_path()  
> always
> does a move_to too.  I thought instead of fixing this in each backend,
> we should fix it in some higher level, but am not sure where/how.   
> Carl,
> any ideas?

I'd love to see this fixed as well -- it's affecting the Quartz  
surface somewhat as well.  I had a hack at one point while I was  
trying to debug another issue that basically recorded the last moveto  
and only executed it if it was followed by another op; maybe  
fixed_interpret could take a flag saying whether final movetos should  
be ignored, and have it just skip?  (Or just always ignore them...  
only time they'd matter is if you're somehow building up a single path  
from two different ones.)

     - Vlad