As of r3090+8795, ghostxps hangs in the clist playback rendering fts_30xx.xps. xps/obj/gxps -dNOPAUSE -dBATCH -sDEVICE=ppmraw -r300 -sOutputFile="|md5sum" ../tests_private/xps/xpsfts-a4/fts_30xx.xps This is the cause of the recent empty regression reports for ghostxps; the run is hitting the job timeout and being killed, producing no output.
this is a clist problem, so copy in Leo. Test file is here: casper.ghostscript.com:/home/regression/tti/tests_private/xps/xpsfts-a4/fts_30xx.xps It is 30M and it was taking too long to attach to bugzilla. I am not sure why.
Created attachment 4194 [details] fts_30xx_page2.xps The XPS document in question stripped to contain only page 2 (which is where it hangs).
Since the problem is in the clist, assigning to Igor.
Bumping P2 for regression.
The issue is a malformed transformation caused by an XML parsing error. It's causing an image to be scaled by 0 in one dimension. With interpolated images, we get a divide-by-zero error instead. When I fixed the XML parsing error, this file does not hang given the command line. We still get a segfault in the banding code, so leaving the bug open for further investigation.
Tor, Since I've got highjer priority bugs, and sisnce I haven't got a debugging framework for XPS, I can't work on this issue soon. Feel free to investigate the segfault.
Can't reproduce with the test file attached. Do I need the full test ? Please attach it or copy to casper. I can't check out tests_private due to big size. Passing to Tor. Please return back to me with the *full* information how to reproduce the problem.
Created attachment 4253 [details] BuildLog.htm MSVC8 build log. Please note harmful warnings. Command line : nmake -f xps_msvc.mak DEBUG=1 TDEBUG=1 MSVC_VERSION=8 DEVSTUDIO=M: The compiler crashed before I added "MSVC_VERSION=8 DEVSTUDIO=M:" .
Leonardo, it may help if you point out which harmful warnings you mean. All I see is one MSVC #define ifdef-voodoo that's not working (which I'm fixing), and some warnings about harmless signed/unsigned char implicit type conversions. The error moved to page 3 after my first fix, so you will need the original fts_30xx.xps test file. Unfortunately SVN does not allow single file checkouts, but you can use SCP to copy the file from casper.
One cannot check out a single file from svn, but one may 'export' a single file. It's still 28 MB though.
these warnings I believe are harmful : >..\gs\src\gxclrast.c(766) : warning C4293: '<<' : shift count negative or too big, undefined behavior 1>..\gs\src\gxclrast.c(770) : warning C4293: '<<' : shift count negative or too big, undefined behavior 1>..\gs\src\gxcht.c(237) : warning C4293: '>>' : shift count negative or too big, undefined behavior 1>..\gs\src\gxcht.c(302) : warning C4293: '>>' : shift count negative or too big, undefined behavior 1>..\gs\src\gxcht.c(431) : warning C4293: '<<' : shift count negative or too big, undefined behavior 1>..\pl\plchar.c(1466) : warning C4054: 'type cast' : from function pointer 'gs_glyph (*)(gs_font *,gs_char,gs_glyph)' to data pointer 'void *' 1>..\pl\plchar.c(1466) : warning C4153: function/data pointer conversion in expression 1>..\pl\plchar.c(1467) : warning C4054: 'type cast' : from function pointer 'int (*)(gs_show_enum *,gs_state *,gs_font *,gs_char,gs_glyph)' to data pointer 'void *' 1>..\pl\plchar.c(1467) : warning C4153: function/data pointer conversion in expression 1>..\pl\plchar.c(1478) : warning C4054: 'type cast' : from function pointer 'gs_glyph (*)(gs_font *,gs_char,gs_glyph)' to data pointer 'void *' 1>..\pl\plchar.c(1478) : warning C4153: function/data pointer conversion in expression 1>..\pl\plchar.c(1479) : warning C4054: 'type cast' : from function pointer 'int (*)(gs_show_enum *,gs_state *,gs_font *,gs_char,gs_glyph)' to data pointer 'void *' 1>..\pl\plchar.c(1479) : warning C4153: function/data pointer conversion in expression 1>..\pl\plchar.c(1534) : warning C4054: 'type cast' : from function pointer 'gs_glyph (*)(gs_font *,gs_char,gs_glyph)' to data pointer 'void *' 1>..\pl\plchar.c(1534) : warning C4153: function/data pointer conversion in expression 1>..\pl\plchar.c(1535) : warning C4054: 'type cast' : from function pointer 'int (*)(gs_show_enum *,gs_state *,gs_font *,gs_char,gs_glyph)' to data pointer 'void *' 1>..\pl\plchar.c(1535) : warning C4153: function/data pointer conversion in expression 1>..\pl\plchar.c(1466) : warning C4054: 'type cast' : from function pointer 'gs_glyph (*)(gs_font *,gs_char,gs_glyph)' to data pointer 'void *' 1>..\pl\plchar.c(1466) : warning C4153: function/data pointer conversion in expression 1>..\pl\plchar.c(1467) : warning C4054: 'type cast' : from function pointer 'int (*)(gs_show_enum *,gs_state *,gs_font *,gs_char,gs_glyph)' to data pointer 'void *' 1>..\pl\plchar.c(1467) : warning C4153: function/data pointer conversion in expression 1>..\pl\plchar.c(1478) : warning C4054: 'type cast' : from function pointer 'gs_glyph (*)(gs_font *,gs_char,gs_glyph)' to data pointer 'void *' 1>..\pl\plchar.c(1478) : warning C4153: function/data pointer conversion in expression 1>..\pl\plchar.c(1479) : warning C4054: 'type cast' : from function pointer 'int (*)(gs_show_enum *,gs_state *,gs_font *,gs_char,gs_glyph)' to data pointer 'void *' 1>..\pl\plchar.c(1479) : warning C4153: function/data pointer conversion in expression 1>..\pl\plchar.c(1534) : warning C4054: 'type cast' : from function pointer 'gs_glyph (*)(gs_font *,gs_char,gs_glyph)' to data pointer 'void *' 1>..\pl\plchar.c(1534) : warning C4153: function/data pointer conversion in expression 1>..\pl\plchar.c(1535) : warning C4054: 'type cast' : from function pointer 'int (*)(gs_show_enum *,gs_state *,gs_font *,gs_char,gs_glyph)' to data pointer 'void *' 1>..\pl\plchar.c(1535) : warning C4153: function/data pointer conversion in expression 1>f:\afpl\ghostpdl\xps\ghostxps.h(92) : warning C4117: macro name '__FUNCTION__' is reserved, '#define' ignored 1>..\xps\xpscff.c(863) : warning C4028: formal parameter 1 different from declaration I recommend to fix all warnings because I'm hard to find important ones between the huge noize.
Created attachment 4299 [details] log.txt Still can't reproduce even with the full test. My command line is few different due to no md5 on Windows : ..\..\ghostpdl\xps\obj\gxps.exe -dBATCH -dNOPAUSE -r300 -sDEVICE=ppmraw -sOutputFile=out%%d.ppm fts_30xx.xps 2>log.txt I noticed runtime error in the log (attached) : + ..\xps\xpsimage.c:206: xps_paint_image_brush_imp(): gs_image_init failed The problem looks to be compiler dependent. Please fix build warnings. I'm not able to debug when I don't know what the compiler generates.
Created attachment 4300 [details] log.txt Still can't reproduce even with the full test. My command line is few different due to no md5 on Windows : ..\..\ghostpdl\xps\obj\gxps.exe -dBATCH -dNOPAUSE -r300 -sDEVICE=ppmraw -sOutputFile=out%%d.ppm fts_30xx.xps 2>log.txt Used revision 9000 debug build. I noticed runtime error in the log (attached) : + ..\xps\xpsimage.c:206: xps_paint_image_brush_imp(): gs_image_init failed The problem looks to be compiler dependent. Please fix build warnings. I'm not able to debug when I don't know what the compiler generates.
Created attachment 4301 [details] log.txt Still can't reproduce even with the full test. My command line is few different due to no md5 on Windows : ..\..\ghostpdl\xps\obj\gxps.exe -dBATCH -dNOPAUSE -r300 -sDEVICE=ppmraw -sOutputFile=out%%d.ppm fts_30xx.xps 2>log.txt Used revision 9000 debug build. I noticed runtime error in the log (attached) : + ..\xps\xpsimage.c:206: xps_paint_image_brush_imp(): gs_image_init failed The problem looks to be compiler dependent. Please fix build warnings. I'm not able to debug when I don't know what the compiler generates.
I just did "svn update" to revision 9000 and now I can't reproduce either. I recommend closing this bug, and open another bug report for fixing pl and gxcht.c warnings.