Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Lucida fonts (T1) and dvipdfm.

25 views
Skip to first unread message

statuya

unread,
Dec 4, 2001, 2:16:00 AM12/4/01
to
Hello.

Under the following conditions, I get wrong pdf files:

(1) Use fonts in Lucida font family.
(2) Use the fonts in T1 encoding.
(3) Write accents, like \"{O}.
(4) Generate pdf file with dvipdfm.

% This is a sample.
%------------------------------
\documentclass{article}
\usepackage[T1]{fontenc}
\usepackage{lucidabr}
\begin{document}
\"{O}
\end{document}
%------------------------------


Then, the generated pdf file something wrong. For example, xpdf says

Char "Odieresis": Invalid accent ('310) in SEAC

How can I get a correct pdf file?

Thanks.
--
Tattsan

Thierry Bouche

unread,
Dec 4, 2001, 3:19:34 AM12/4/01
to
statuya @ seg.co.jp writes:


It seems that there is a bug in dvipdfm, that pdftex had at a very
early stage, but which is now corrected.

The problem is with font subsetting: a narrow-minded approach is to
download only the relevant charstring definition from the font, but
this charstring may be a simple reference to subroutines or other
glyphs, especially in the case of composite characters with
diacritics.

SEAC is a type 1 instructions meant to asemble an accent and a base
glyph. If dvipdfm only downloads the def of Ö to the PDF, the PDF
won't be complete as it will need to have O and ¨ as well before being
able to do any use of that Ö.

The cure is to have dvipdfm fixed. In the meantime, you can do
something like
\makebox{0pt}[l]{\hspace*{100cm} \"{}O} (put here any glyph part of
accenter letters you could use, in any font...)


--
Thierry Bouche
--
« Elle revint au bout de neuf jours, munie des mêmes jambes. »
Eduardo Mendoza, _la Aventura del tocador de señoras_.

statuya

unread,
Dec 4, 2001, 7:24:14 AM12/4/01
to
Thierry Bouche <Thierry...@ujf-grenoble.fr> writes:
>
> SEAC is a type 1 instructions meant to asemble an accent and a base
> glyph. If dvipdfm only downloads the def of Ö to the PDF, the PDF
> won't be complete as it will need to have O and ¨ as well before being
> able to do any use of that Ö.

Thanks for explaining the bug. Downloading subglyph data precedingly,
I could print out my document with no error. So your explanation seems
to be true.

--
Tattsan

Thierry Bouche

unread,
Dec 4, 2001, 9:10:54 AM12/4/01
to
Thierry Bouche <Thierry...@ujf-grenoble.fr> writes:

> The cure is to have dvipdfm fixed. In the meantime, you can do
> something like
> \makebox{0pt}[l]{\hspace*{100cm} \"{}O} (put here any glyph part of
> accenter letters you could use, in any font...)

or of course stop subsetting, but I'm afraid that lucida's © doesn't
allow you to do it...

Th.

statuya

unread,
Dec 4, 2001, 7:58:03 PM12/4/01
to
Thierry Bouche <Thierry...@ujf-grenoble.fr> writes:
>
> or of course stop subsetting, but I'm afraid that lucida's © doesn't
> allow you to do it...

My documents contain Japanese characters (iso-2022-jp or euc-jp),
incompatible with isolatin1 package. Accents such as \" are only
needed to write European proper name.

--
Tattsan

0 new messages