[cairo] Dash pattern approximation

[Andrea Canciani]

Attached are an explanation and an improvement of the dash pattern
approximation code and a change to use it in gstate, for alla
backends.
The changes in 26e9f149063b9e1fdb54fc54fccbefdf04a68190 affecting
cairo-path-stroke.c can be reverted if 0003 is applied, without
causing regressions in the test suite.
Sanooj, based on your request I tried to explain the
ROUND_MINSQ_APPROXIMATION and why in 0002 those computations are
needed.
Adrianj, in 0003 I rewrote the patch I previously asked you to test by
removing the dynamic allocation (both the style and the dashes are in
the stack, now).
Are code style and explanations ok?
If anybody has objections and/or suggestions about the code, please speak up.
Thank you in advance
Andrea Canciani
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-Round-caps-coverage-extimate-explanation.patch
Type: application/octet-stream
Size: 2014 bytes
Desc: not available
Url : http://lists.cairographics.org/archives/cairo/attachments/20100117/11078de2/attachment.obj 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0002-Improve-dash-pattern-approximation.patch
Type: application/octet-stream
Size: 3275 bytes
Desc: not available
Url : http://lists.cairographics.org/archives/cairo/attachments/20100117/11078de2/attachment-0001.obj 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0003-Approximate-dash-patterns-in-gstate.patch
Type: application/octet-stream
Size: 1830 bytes
Desc: not available
Url : http://lists.cairographics.org/archives/cairo/attachments/20100117/11078de2/attachment-0002.obj