In the document at http://code.google.com/p/sumatrapdf/issues/detail?id=1014 , a Font's BaseFont differs from its FontDescriptor's FontName, even though the PDF Spec says they should be identical. While MuPDF currently prefers FontName over BaseFont, Adobe Reader seems to do the opposite. Note: I do have one document (I may not share) where BaseFont is "Arial,Bold+000041" and FontName is "Arial,Bold" and which misrenders with the above change unless either an exception is made or the trailing "+<digits>" is ignored.
Would you mind attaching the .pdf with "Arial,Bold+000041" as BaseFont?
Created attachment 6823 [details] sampletest.pdf Sure, even though I don't quite see why you can't just download it from our own issue tracker.
This was a misunderstanding. :) I did download sampletest.pdf from your bugtracker, I meant the one you mentioned in your note. Upon re-reading the comment I noticed that you already state that you may not share the document. Do you have the possibility to run pdfshow on the file and pass on the output of the Creator/Producer entries from Info-object in the trailer and of course the output of the font object named "Arial,Bold+000041" so that we have some debuginfo at least?
Ah, that document, that does indeed make more sense: /Producer (PCL2PDF400 from RJS Software Systems Inc.) 23 0 obj << /Type /FontDescriptor /FontName /Arial,Bold /Flags 262176 /FontBBox [ -200 -300 1200 900 ] /Ascent 720 /CapHeight 660 /Descent -270 /ItalicAngle 0 /StemV 145 >> endobj 24 0 obj << /Type /Font /Subtype /TrueType /Name /Fo41 /BaseFont /Arial,Bold+000041 /FontDescriptor 23 0 R /Encoding /WinAnsiEncoding /FirstChar 32 /LastChar 255 /Widths 22 0 R >> endobj 27 0 obj << /Type /Font /Subtype /TrueType /Name /Fo43 /BaseFont /Arial,BoldItalic+000043 /FontDescriptor 26 0 R /Encoding /WinAnsiEncoding /FirstChar 32 /LastChar 255 /Widths 25 0 R >> endobj 26 0 obj << /Type /FontDescriptor /FontName /Arial,BoldItalic /Flags 262176 /FontBBox [ -200 -300 1200 900 ] /Ascent 720 /CapHeight 660 /Descent -270 /ItalicAngle -15 /StemV 145 >> endobj Anything else that might help you?
I believe this has already been fixed. The sampletest.pdf file renders as expected with the current code.
(In reply to comment #5) > I believe this has already been fixed. I still have files where your code (current git HEAD) wrongly prefers FontName over BaseFont. Unfortunately, these files are confidential and may not be shared.