cups-pdf produces large pdfs and text can not be searched

[Ryan]

Binary package hint: cups-pdf

When printing PDF files using the 'PDF' printer created from cups-pdf:
    lp -d PDF ~/.bashrc
The resulting PDF files appears to be all images: you can not search nor copy text. (See attached.

Also, when I try printing a larger file (e.g. 437K of text), a 3.9 MB PDF files is created (as opposed to 485K when using gedit's straight-to-pdf functionality).

What I find very weird is that if you look at the cups error log, you'll see three filters running:
- texttopdf
- pdftopdf
- cpdftops
and then finally to cups-pdf.

You'll also see warnings that pdftopdf produces a problematic pdf file.

If there is a texttopdf, why are all the other programs running?

-Ryan

ProblemType: Bug
Architecture: amd64
DistroRelease: Ubuntu 9.04
Lpstat: device for PDF: cups-pdf:/
MachineType: Dell Inc. Precision M6300
NonfreeKernelModules: nvidia
Package: cups 1.3.9-17ubuntu1
Papersize: letter
PpdFiles: PDF: Generic CUPS-PDF Printer
ProcCmdLine: root=UUID=35d8fdfb-7734-4c06-90eb-e5d1c632eb52 ro quiet splash
ProcEnviron:
 PATH=(custom, user)
 LANG=en_US.UTF-8
 SHELL=/bin/bash
ProcVersionSignature: Ubuntu 2.6.28-11.42-generic
SourcePackage: cups

[SAn (san.at)]

I have the same problem: cups-pdf printed embedded text with Intrepid ibex without any problem. After installing Jaunty (no upgrade), I get only graphics instead of text in the pdf documents. Printing application was gedit or open-office.

[emmonsjb (emmons)]

I also have the same problem after installing Jaunty (using cups-pdf 2.5.0-1ubuntu1, cups 1.3.9-17ubuntu3). I only get graphics, no text, when printing to PDF from any program.

The PDF file size is not inordinately large, however. Exporting an OpenOffice file to PDF using the application's "Export to PDF" feature creates a file (with text properly embedded) of 198.0K. Printing the exact same file to PDF creates a file (with only image, and a garbled one at that, no text) of 38.2K; upping the dpi using printer output options only increases the file size by 0.2K, and the image still looks garbled.

Others have reported the same problem at http://ubuntuforums.org/showthread.php?t=1153731.

[emmonsjb (emmons)]

To demonstrate the problem, here is an example of an OO document exported to PDF using the app's native export function: http://oregonstate.edu/~emmonsj/Exported%20to%20PDF%20in%20OO.pdf

Here is the same file printed using cups-pdf: http://oregonstate.edu/~emmonsj/Printed%20using%20cups-pdf%20from%20OO.pdf

[NoOp (glgxg)]

I think the issue is this:
https://bugs.launchpad.net/bugs/381788
[[jaunty] cups-pdf no longer embeds fonts in pdf file]

If you look at the document properties in both files that you've provided, you will find that the cups-pdf file contains no embedded fonts whereas the file printed directly from OOo does.

[Benedikt Wegmann (benny-wegmann)]

As a result of this behaviour Tracker is no longer able to index the files and screen readers will likely fail to read the contents as well.

Neither installing the cups-pdf package from 8.10 or 9.10 through a local repository did improve the situation, PDFs remain only graphic, no searchable text.

Yet again taking a normal Postscript document and running it through Ghostscript's ps2pdf resulted in a nice, searchable PDF document.

This bug really affects archiving of content with PDF and accessibility for users, thus limiting the use of this software which worked fine up until Jaunty.

[Martin-Éric Racine (q-funk)]

Can anyone try compiling 2.5.0 against Hardy and see if it works there? If it works on Hardy, we can immediately rule out CUPS-PDF issues and presume this to be a CUPS issues, probably in which default filters are used by CUPS to produce the final document.