[lucida] Bug in lucida ligature handling

[Jaap-Henk Hoepman]

Hi,

Some more (weird) datapoints. I managed to find an old .pdf with
corresponding .ps that was used to create the .pdf with (back then) AFPL
Ghostscript 7.0. If I view that .pdf in Bluebeam PDF revu the ffl
ligature is displayed correctly.

If I then use the _same_ .ps file to create a .pdf file today, the
resulting .pdf will _not_ show the ffl ligature in Bluebeam PDF revu
(but will in Acrobat). It doesn't matter whether I generate the .pdf
using GPL Ghostscript 9.04, Acrobat, or Distiller...

As the .ps is self contained (full .pfb for each font, with all encoding
information), it seems as if the meaning of 'standard encoding' against
which fonts are apparently reencoded when creating a .pdf file has
somehow changed over the years. The original .ps (and .pdf) file are
from 1999...

Any ideas would be very much appreciated... I only started to see these
issues a year or so ago, when I started to use Bluebeam PDF revu to
annotate pdf's on my tablet.

Jaap-Henk

On 02/08/2012 13:55, Jaap-Henk Hoepman wrote:
> Hi,
> 
> There appears to be a problem in the way ligatures are handled for the
> lucida fonts. In particular, the \ffl ligature is displayed as a white
> box, but _only_ when viewed with Bluebeam Pdf Revu and _only_ when the
> .pdf is produced from the .ps that was produced using dvips. In the
> latter case it doesn't matter whether ps2pdf was used or "Convert to
> Adobe PDF" under windows.
> 
> In other words
> - the .pdf displays fine in Acrobat (and PdfAnnotator)
> - the .pdf also displays fine in Bluebeam Pdf Revu if it is generated
> directly from the latex source using pdflatex (but not when generated
> from the .ps file).
> 
> Note that the .ps file views fine in ghostview.
> 
> The problem seems to be related to the following snippet of .pdf
> 
> 8 0 obj
> <</BaseFont/TVCKVO+LucidaBright/FontDescriptor 9 0 R/Type/Font
> /FirstChar 8/LastChar 121/Widths[ 663 0 0 0 657 0 994 1000
> 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
> 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
> 0 608 0 0 0 0 0 0 0 0 0 0 0 0 0 0
> 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
> 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
> 0 543 0 509 642 530 0 0 0 307 0 0 313 0 633 602
> 642 0 0 504 0 0 0 795 0 551]
> /Encoding 13 0 R/Subtype/Type1>>
> endobj
> 13 0 obj
> <</Type/Encoding/BaseEncoding/WinAnsiEncoding/Differences[
> 8/fl
> 12/fi
> 14/ffi/ffl]>>
> endobj
> 
> If I remove the encoding object (obj 13) or if I remove /Encoding 13
> from the font object (obj 8) the file displays fine in Bluebeam PDF Revu.
> 
> I have not been able to determine which parts of the .ps file lead to
> the above snippet of .pdf.
> 
> What gives?
> 
> Best,
> Jaap-Henk
> 

-- 
Jaap-Henk Hoepman            |  Sure! We've eaten off the silver
TNO, Groningen &             |  (when even food was against us)
Radboud University Nijmegen  |    - The Birthday Party "Just You and Me"
(e) jhh at cs.ru.nl /           |
    jaap-henk.hoepman at tno.nl | (m) +31 6 20619554
(w) www.cs.ru.nl/~jhh        | (t) @xotoxot