Hi David, Werner and others, I just found regression between 2.1.9 and 2.2.0 version of Freetype, reproducible in ftview (It probably also work correctly in 2.1.10, but this version I was not able to compile on Windows using jam and VC 2005). It happens on Dynalab fonts, for example DFHsm5.ttc (length: 3250760 bytes, I will send it to Werner and also to anybody who will request it). This font file contains 3 fonts. Problem is with 3rd one DFGHSMincho-W5. Start ftview -f 719 30 DFHsm5.ttc Press "n" 2 times to see 3rd font index. Now you should see character fullwidth digit three (Unicode: 0xFF13), shape is correct but there is big space before - left bearing is same as in first two fonts, but is it not correct. This glyph has in hmtx table 719. advWid: 547, LSdBear: 30 - this left bearing is correctly returned by Get_HMetrics function in ttgload.c, but it has no impact on actual left bearing of final glyph outline (I change it in debugger and resulting outline was same). Unfortunately don't know how to fix this regression after some code inspection/debugging. It is correctly propagated to pp1 and pp2 phantom points, but then my knowledge ends.
Thanks for outstanding work on Freetype. Best regards, Boris Letocha _______________________________________________ Freetype-devel mailing list [email protected] http://lists.nongnu.org/mailman/listinfo/freetype-devel
