[PyX-user] encoding of TeX log

[Gert-Ludwig I. <ger...@ph...>]

Hello,

according to the docs, texenc fixes the encoding used in the
communication with (La)TeX. Unfortunately, in general, the log
messages of (La)TeX are not returned in a well-defined encoding.
Specifically, messages containing input from the (La)TeX file are
returned in the fontenc, e.g. T1 (cf. e.g.
https://tex.stackexchange.com/questions/131238/what-controls-the-encoding-of-the-latex-log-file-and-how-to-change-it).
In the following example, this will lead to a UnicodeDecodeError:

from pyx import canvas, text

text.set(text.LatexRunner, texenc="utf-8")
text.preamble(r"""\usepackage[T1]{fontenc}
                  \usepackage[utf8x]{inputenc}""")
c = canvas.canvas()
c.text(0, 0, r'foo föo föo f\"oo f\"oo', [text.parbox(2)])
c.writePDFfile()

The parbox is sufficiently narrow to produce an overfull hbox and
the error message will contain the input string. Independently of
whether the umlaut ö is given as Unicode character or TeX code, it
will be translated into a byte 0xf6 which cannot be interpreted as
UTF-8 encoded character.

As a temporary solution, I set the encoding of MonitorOutput in my
copy of text.py to latin1 because I want to be able to handle input
in utf8. Most likely, however, this is not the best solution...

Best regards,
Gert

-- 
 Gert-Ludwig Ingold     email: Ger...@Ph...
 Institut für Physik    Phone: +49-821-598-3234
 Universität Augsburg   Fax  : +49-821-598-3222
 D-86135 Augsburg       WWW  : www.physik.uni-augsburg.de/theo1/ingold
 Germany                PGP  : 86FF5A93, key available from homepage