Summary: | mirrored glyph is rendered at incorrect position | ||
---|---|---|---|
Product: | Ghostscript | Reporter: | Werner Lemberg <wl> |
Component: | PS Interpreter | Assignee: | Igor Melichev <igor.melichev> |
Status: | NOTIFIED FIXED | ||
Severity: | major | ||
Priority: | P2 | ||
Version: | 0.00 | ||
Hardware: | PC | ||
OS: | Linux | ||
URL: | http://freetype.org/gs-bugs/gsbug.ps | ||
Customer: | Word Size: | --- | |
Attachments: |
gsbug.ps
gsbug-1.ps gsbug-1.ps.CPSI.tif.zip |
Description
Werner Lemberg
2004-08-11 15:51:16 UTC
Reproduced with CVS HEAD. Created attachment 998 [details]
gsbug.ps
A local copy of the test file.
Adobe CPSI rasterizes this document differently than GS and differently then AdobeDistiller+ASdobeReader. IMO the CPSI raster is correct. Will attach a simplified test and its raster from CPSI. Created attachment 999 [details]
gsbug-1.ps
A simplified test.
Created attachment 1000 [details]
gsbug-1.ps.CPSI.tif.zip
A raster of gsbug-1.ps made with CPSI.
The "hand" glyph has Metrics [sbx, wx]=[-904, -939] and the native sbx = 35. I believe that CPSI renders it correctly : mirrorred and then shifted forth in 904 design units. Since 904 = 935-35, both occurances must have same projection to the X axis, and we do observe so with CPSI. Others render it wrongly : Distiller misses sbx, and pdfwrite applies sbx with a wrong sign (i.e. in the opposite direction). A fix to pdfwrite: HEAD : gdevpdte.c Revision 1.65 GS_8_1X : gdevpdte.c Revision 1.33.2.15 A fix for raster devices : HEAD : gstype1.c Revision 1.32 gstype2.c Revision 1.36 zchar1.c Revision 1.41 GS_8_1X : gstype1.c Revision 1.22.2.2 gstype2.c Revision 1.25.2.3 zchar1.c Revision 1.28.2.4 Now I can list references to related patches, which correspond to revisions listed above : HEAD : http://www.ghostscript.com/pipermail/gs-cvs/2004-October/004976.html http://www.ghostscript.com/pipermail/gs-cvs/2004-October/004978.html GS_8_1X : http://www.ghostscript.com/pipermail/gs-cvs/2004-October/004977.html http://www.ghostscript.com/pipermail/gs-cvs/2004-October/004979.html |