pdftotext has improved to generate ASCII instead of ligatures (like ff)
but it still generates ligatures in some cases. For instance, if I run
pdflatex on the following file:
------------------------------------------------------------
\documentclass[11pt]{article}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage{lmodern}
\begin{document}
\thispagestyle{empty}
chiffre
ê
\end{document}
------------------------------------------------------------
I get some ff-lig1.pdf PDF file (attached), and running pdftotext on
it gives:
chiffre
ê
This is OK. But if I run ps2pdf on this PDF file, I get some other
ff-lig1-gs.pdf PDF file (attached), and running pdftotext on it gives:
chiffre
ê
i.e. with a ligature, which makes searching text such as "chiffre"
unpredictable (the output is also less readable in a terminal with
a monospace font).
This problem doesn't occur if I replace "ê" by "é" in the LaTeX file!
Note that xpdf finds "chiffre" in both cases, so that the bug seems
to be in pdftotext. Moreover pdftotext from poppler-utils 0.18.4-6
(wheezy) gives the ligature on the 4 attached PDF files, so that it
seems that pdftotext has been improved except in the particular case
mentioned above.