Bug#847462: printer-driver-cups-pdf: make very ugly and uselss pdf which are not searchables
On Thu 08 Dec 2016 at 19:25:20 -0200, Till Kamppeter wrote:
> On 12/08/2016 05:54 PM, Brian Potkin wrote:
> >The printing system is intended to put ink or toner on paper. If it does
> >not do that it has failed. Can we agree on that?
> >
> >Does it matter for the primary purpose of printing whether a PDF produced
> >during the process is searchable or not?
>
> Note that the cups-pdf driver is an exception here. It is meant to be used
> out of applications without "Save as PDF", "Export to PDF", or similar
> functionality. So "printing" a file to a cups-pdf "printer" is not
> necessarily done with the intention of getting the file onto paper. And for
> a PDF file it makes sense to have it searchable.
It seems to me that most users see printer-driver-cups-pdf as a general
purpose printer for converting to PDF from any application and not
confined in the way you describe. The Debian package description would
encourage this view by mentioning "virtual printer in a paperless
network". The OpenSuse description has
CUPS-PDF is a PDF writer backend for CUPS. It provides a virtual
CUPS-PDF printer which produces PDF files so that application
programs which have no built-in support to "Save as PDF" could
print to CUPS-PDF to get a PDF file.
Searchable PDFs are produced from submitted PostScript files. Text files
fare less well with texttopdf, as the documentation for cups-filters
relates. When I asked Tobias Hoffmann (the author of texttopdf) about
future plans for searchable PDFs he replied "...this feature is strictly
optional for texttopdf, it has a very low priority..." It would be nice
if it were done but it helped form my view about the primary purpose of
printing.
> Problem is most probably in cups-pdf upstream, but it can be that the
> application is the culprit sending non-searchable PDF to CUPS (then
> CUPS/cups-filters cannot turn the PDF back searchable).
>
> You need to check the PDF which is sent to CUPS, following
>
> Capturing print job data
>
> on
>
> https://wiki.ubuntu.com/DebuggingPrintingProblems
>
> and tell us from which application you have printed to cups-pdf.
Apart from seeming pointless, I thought feeding cups-pdf with PDF input
was not supported by upstream.
--
Brian.
Reply to: