On 23-Jul-2014 17:47, mathog wrote:
Sorry, forgot to post the link to the mangled one. The preceding one is what comes out of the EMF read and is passed to createNewDocFromMem, the next one is what Inkscape then makes of it.
http://saf.bio.caltech.edu/pub/pickup/mangled_clip.svg
Look at, for instance, all the pieces related to "clipEmfPath1". The units for the clipPath are the same in the two files
d="m 993.743,142.499 14.997,0 0,30 -14.997,0 z"
and the units for the "star" are the same in its <path>
d="m 1016.02,168.299 -7.95,-6.975 -8.102,6.825 4.202,-9.675 -9.002,-5.625 10.572,1.05 2.48,-10.35 2.32,10.35 10.58,-0.75 -9.15,5.4 z"
but note that the unmangled one has no associated transform, and the mangled one now includes
transform="matrix(0.28222223,0,0,0.28222223,0,-9.9911637e-7)"
The clip on that path still works in the mangled one because the clip is applied before the coordinates are scaled.
Not so for text. In the unmangled one it is:
transform="matrix(1.000000,-0.000000,0.000000,1.000000,991.492631,146.998908)" x="0.000000" y="2.497559"
which is overlaps with the clip. However the mangled one has:
x="279.82126" y="42.191227"
which does not overlap. In this case the transform has been applied (permanently) before the clip is applied, so it does not work.
Regards,
David Mathog mathog@...1176... Manager, Sequence Analysis Facility, Biology Division, Caltech