Bug 91377

Summary: All characters in this PDF are replaced by boxes
Product: poppler Reporter: Michael Fiedler <michael.fiedler87>
Component: generalAssignee: poppler-bugs <poppler-bugs>
Status: RESOLVED MOVED QA Contact:
Severity: normal    
Priority: medium    
Version: unspecified   
Hardware: x86 (IA32)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments: Screenshot
PDF file that is displayed with boxes
original PDF which is displayed correctly (before being converted to the broken one)
output of pdftocairo -pdf, using the "broken" PDF as input (the one that is displayed correctly using other viewers)
minimal test case (character A)

Description Michael Fiedler 2015-07-17 21:30:53 UTC
Created attachment 117205 [details]
Screenshot

When displaying the attached PDF document using a Poppler-based viewer such as Evince, all characters are replaced by boxes (see screenshot).  As other PDF viewers such as Firefox/PDF.js, Google Chrome, MuPDF, Adobe Reader display it correctly, I suppose it is a poppler bug.  The bug can also be reproduced by using utils/pdftocairo -pdf and then viewing the output with a non-poppler-based viewer.

The affected PDF file was created by the online interface of the German self-publishing platform epubli.de from my original document, so I do not exactly know which software they used for that.


My test environment:

Debian GNU/Linux 7.8 (wheezy) (x86)
Evince/GNOME Document Viewer 3.4.0
Poppler 0.18.4-6 (from Debian)
Poppler 0.34.0
Poppler Git commit 1aa2f6e8a41a6a86dc02bf7c5cbc62355e780961

All tested poppler versions are affected.  I verified it using

  utils/pdftocairo -pdf broken.pdf out.pdf
  $other_pdf_viewer out.pdf
Comment 1 Michael Fiedler 2015-07-17 21:32:16 UTC
Created attachment 117206 [details]
PDF file that is displayed with boxes
Comment 2 Michael Fiedler 2015-07-17 21:33:25 UTC
Created attachment 117207 [details]
original PDF which is displayed correctly (before being converted to the broken one)
Comment 3 Michael Fiedler 2015-07-17 21:34:55 UTC
Created attachment 117208 [details]
output of pdftocairo -pdf, using the "broken" PDF as input (the one that is displayed correctly using other viewers)
Comment 4 Albert Astals Cid 2015-07-18 14:22:29 UTC
Adobe Reader displays it wrong here and so does mupdf. which versions of them are you using?
Comment 5 Michael Fiedler 2015-07-18 14:38:32 UTC
(In reply to Albert Astals Cid from comment #4)
> Adobe Reader displays it wrong here and so does mupdf. which versions of
> them are you using?

The "broken" document is attachment 2 [details] [review] (just to be sure).

I am using

Adobe Reader 9.5.5
MuPDF 1.7a (i cannot find a --version option to mupdf-x11, but I built it in June from source and probably used the latest version)
Comment 6 Michael Fiedler 2015-07-20 11:03:40 UTC
Created attachment 117253 [details]
minimal test case (character A)
Comment 7 Adrian Johnson 2015-07-20 12:25:50 UTC
The problem is spaces in the Type 1 font dictionary have been replaced with newlines as shown below. This does not comply with the rules in section 10 of the Type 1 font spec. FoFiType1::parse() would need to be changed to handle this sort of damage.

%!FontType1-1.0: EBGaramondSmallCaps12Regular 001.003
11
dict
begin
/FontInfo
2
dict
dup
begin
/FullName
(EBGaramondSmallCaps12Regular)
readonly
def
/FamilyName
(EBGaramondSmallCaps12Regular)
readonly
def
end
readonly
def
/FontName
/EBGaramondSmallCaps12Regular
def
/Encoding
256
array
dup
0
/.notdef
put
dup
1
/A
put
dup
2
/.notdef
put
Comment 8 GitLab Migration User 2018-08-21 10:40:26 UTC
-- GitLab Migration Automatic Message --

This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.

You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/poppler/poppler/issues/317.

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.