A good method of determining if bugs like this are an Inkscape or cairo bugs is with the cairo-trace utility.
http://www.cairographics.org/FAQ/#profiling
Using a recent cairo from git it is simply a case of running:
cairo-trace inkscape --export-ps=file.ps file.svg
The export should be done from the command line to avoid tracing all the cairo calls that draw the gtk widgets.
Then you will get a file like inkscape.23639.trace with a log of all the cairo functions calls and their arguments.
To find out the cause of unnecessary rasterization, check the log with the list of PS supported features at bottom of this page: http://www.cairographics.org/documentation/using_the_postscript_surface/
It should be as easy as grepping for cairo_mask, set_operator, paint_with_alpha, and rgba (and check the alpha).
A good method of determining if bugs like this are an Inkscape or cairo bugs is with the cairo-trace utility.
http:// www.cairographi cs.org/ FAQ/#profiling
Using a recent cairo from git it is simply a case of running:
cairo-trace inkscape --export-ps=file.ps file.svg
The export should be done from the command line to avoid tracing all the cairo calls that draw the gtk widgets.
Then you will get a file like inkscape. 23639.trace with a log of all the cairo functions calls and their arguments.
To find out the cause of unnecessary rasterization, check the log with the list of PS supported features at bottom of this page: www.cairographi cs.org/ documentation/ using_the_ postscript_ surface/
http://
It should be as easy as grepping for cairo_mask, set_operator, paint_with_alpha, and rgba (and check the alpha).