Hi Phil, Hi Igor, On Mon, 2009-05-04 at 21:57 +0200, Mark Wielaard wrote: > On Mon, 2009-05-04 at 12:34 -0700, Phil Race wrote: > > Igor did the freetype work and in he commented in email on 23rd July 2007 : > > > > > - Changed required freetype version from 2.3.4 to 2.3.0 > > > (it did compile with 2.2.1 too for me but Font2DTest was showing > > garbage > > > and i had no chance to investigate why this happens) > > > > So I don't think we want to move it down without knowing what that was > > about. > > It may have been platform-specific. Igor - can you remember two years back ? > > Interesting, to make sure everything was fine I did also check with that > Font2DTest (and some others) and the fact that everything looked just > fine was the reason to push it. > > > I also recall there was a nasty bug where having fonts with embedded > > bitmaps on your system caused an infinite loop inside freetype. > > I think we needed to update the freetype lib because of that, but > > I can't remember the details to be sure. > > The freetype 2.2.1 package on RHEL and CentOS 5 do have some fixes > applied to them. And the java-1.6.0-openjdk package has been deployed on > those platforms for some time. So I assume if there were any real bugs > they will have been backported at least on those platforms. > > Ideally of course we would have a autoconf test for the features that > the backend relies on. The problem with the hard version check is that > it prevents compilation for everybody, even when they really just want > to compile against an earlier (patched/fixed) version that might just be > what is available for the platform. > > Let me know if I can help with coming up with a feature test for any > issues.
For now to not forget this patch it has been filed at: https://bugs.openjdk.java.net/show_bug.cgi?id=100061 Please do let me know if you find out anything about your earlier concerns and what I can do to help move this forward. Thanks, Mark
