Comment 98 for bug 179988

Thanks Ulf for the latest patch, and Adrian for many suggestions and clarifications.

> That's the test rectangle I used to find the scaling bug
Well at least we're on our way to drawing crop marks for 0.47 now! ;^)

But seriously, this latest patch compiles OK for me, but seems to take us back a step .

Can I ask more about the previous patch?
- It seems to call cairo_win32_printing_surface_create() from ui/dialog/print.cpp to print clean vectors with no white blocks. Is that true?
- Was the win-print test code just a workaround that partly bypasses GTK to narrow down whether using the cairo_win32_printing_surface_create() call fixes the issues? That's how it seemed to me.
- Is the latest patch just waiting for the GTK stuff to be sorted, and then you reckon we'll get clean vectors with no white/black blocks?
- If so, what is stopping us just using that code temporarily, even if it's a kludge until the GTK stuff gets properly sorted? I'm guessing this bypasses other stuff, but how important is that? If it's only the look of the dialog, but it's still using Cairo to print nice vectors behind the scenes, does that matter for 0.46.0, and could we fix it properly for 0.46.1?