[ Sorry for the duplicate! I did not realize the tiny
little checkbox
had to be selected in order for the file upload I had
laboriously specified to actually get included in the
report! Would someone please delete the earlier one
and replace it by this? ]
Report by Joe Wells (e-mail: local part: jbw; domain:
macs.hw.ac.uk).
The DVI-to-UTF-8 conversion (used by cut-and-paste and
save-as-text)
is a marvelous feature, but could be improved.
This report is for this version of Xdvi (from teTeX
2.99.9.20050111-beta):
xdvik version 22.84.8 (@(#)Motif Version 2.1.30,
runtime version 2.1)
Libraries: kpathsea version 3.5.4, T1lib version 5.0.1
The attached LaTeX file:
1. Shows a case where Xdvi currently gets the wrong Unicode
characters. It swaps \phi and \varphi. (This is
actually a bug,
but most of this message is a feature request.)
2. Shows a couple of cases (\ldots and \cdots) where
Xdvi maps symbols
to multiple Latin-1 characters even though there is
a single
Unicode character for each.
3. Illustrates a number of symbols generated by LaTeX
which are not
mapped to Unicode characters, and indicates which
characters these
symbols could be mapped to.
4. Illustrates that it is not always obvious what the
correct Unicode
character (or character sequence) is, and that
therefore perhaps
there ought to be separate configuration files which
Xdvi
consults in making the choice.
To see the problems, run LaTeX on the file, then run
Xdvi on the DVI
output, then save as plain text in UTF-8, then compare
the display in
Xdvi with the contents of the plain text file.
Warning: There are actual non-Latin-1 Unicode
characters in the
comments of the LaTeX file. I have no idea if these
will be uploaded
correctly by my browser or if SourceForge will receive
them correctly.
Just ignore the comments if they appear to contain
gibberish. The
actual LaTeX code uses only ASCII characters.
sample LaTeX code as mentioned in Detailed Description