ligatures dropped from imported PDF text

[babayasin]

When importing a PDF page with text blocks, ligatures like ff, fi, fl (if present) are lost.

[Alexandre Prokoudine (alexandre-prokoudine)]

Can't reproduce in current trunk with a PDF created by Scribus 1.3.5 and embedded DejaVu Serif font.

[babayasin (serge-timakov)]

Try this one (InDesign CS2).

[Alexandre Prokoudine (alexandre-prokoudine)]

Indeed. I'll ping Miklos who did PDF importing feature.

[Alexandre Prokoudine (alexandre-prokoudine)]

OK, Miklos says it's rather a poppler bug. Will ping them.

[Johan Engelen (johanengelen)]

I am adding a very simple ligature test file here, made with the following pdflatex source:

\documentclass{article}
\usepackage{pslatex}
\begin{document}
ff fi ffl ffi
\end{document}

Hope this helps. Using bzr9458, Inkscape seems to hang when trying to import this document; perhaps another bug?

[su_v (suv-lp)]

'ligatures.pdf' imports without errors (no freeze, nor dropped letters) in Inkscape 0.47+devel r9456 on OS X 10.5.8

(If recreating the text in Inkscape using the same font, no ligatures are used.)

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[koko-ng (kokong)]

bug #218045 seem related

[Marcel Partap (empee584)]

While this has been fixed in poppler, it still remains unsolved in inkscape after 8 years. So difficult to hunt down?

[jazzynico (jazzynico)]

Reproduced Inkscape 0.91 and trunk rev. 14564 when importing with the internal system.
Not reproduced when using Poppler/Cairo import.

Tested with Poppler 0.12.1 (on Windows XP 32-bit) and Poppler 0.33.0 (Xubuntu 15.10 64-bit).

That said, Poppler/Cairo is not the default option.I didn't check the issue tracker and the dev. list, but it should be interesting to compare both options and switch to Poppler/cairo as default option if the overall result is better than with the internal import.

[chag (chagam)]

Problem still present in Inkscape 0.91 r13725 (latest available in mint 18)

[Stanislav Brabec (sbrabec-suse)]

In inkscape 0.92, the problem is still present.

Example: http://www.financnisprava.cz/assets/tiskopisy/5405_23.pdf

It is easy to work-around this problem manually: Save document, open in the text editor, and the carefully replace affected ocurrences of ">f" by ">fi", and repeat the same for ff, fl, ffi, ffl.

[Nathan Lee (nathan.lee)]

Closing as part of the migration to GitLab http://alpha.inkscape.org/bug-migration/

Discussion/progress can be tracked in https://gitlab.com/inkscape/inkscape/issues/593 (at the time of posting, no progress)