Comment 101 for bug 179988

Ulf, I understand what you're saying. I am also conscious of the release date. I've had someone express interest in testing the GTK patch on the GTK bug tracker, but I don't know whether that will result in getting our hands on a windows build, and I agree that this will be untested. About these patches breaking non windows versions, isn't it just a matter of carefully making sure that there's always an #ifdef WIN32 before any subroutines added? I can help there. (Yes I am picking up just enough of a sliver of C++ to become at least vaguely useful in this limited fashion).

I think we (well not so much me) have made great strides, and I don't want my hesitance at adopting an 'acceptable output' patch to be confused with being ungrateful. I appreciate that, if nothing else, the many comments on this bug show the huge effort by many people. I am astounded by the great teamwork, in spite of some people like me making inflammatory and somewhat unjustified comments in the heat of the moment (sorry everyone!).

I don't care anywhere near as much about the bitmaps or the blurs to get acceptable output, but the white boxes still bother me as they fundamentally interfere with almost every image. It's a question of what is 'acceptable output'.

I agree with Adrian that the best solution is if we can get a patched GTK that works, which the test seemed to show. But a patched GTK it might break other things. I have also found that the test code crashes on more complex images - I haven't worked hard at narrowing that down, but the test code could be exposing a GTK or Cairo problem (???). Is it better for now for us to revert back to the 0.45.1 printing system for 0.46.0, and work on the GTK common print dialog for 0.46.1?

Currently building with the newest patch now, plus the other patches mentioned too.