Data tables are text files, each text file contains (up to) 256 characters.
To find the encoding for character 0x000000E4 (german umlaut ae) the table file basedir/00/00/00.dat is read. The first subdirectory name ``00'' is the hex representation of the most significant byte in the character, the second ``00'' is for the second byte, the file name ``00.dat'' is derived from the third significant byte.
The file is read line by line. Each line can be either a comment line (the first non-whitespace is ``#'') or a data line.
A data line consists of a mode descriptor, the original character and the LaTeX encoding.
The mode descriptor can bei either ``text'', ``math'' or ``all'' (``*''). The encoding is valid in LaTeXs text mode, math mode or in both modes.
The original character can be specified as a character enclosed in single quotes, in hexadecimal notation started by ``0x'' or in decimal notation (without ``0x''). As three significant bytes of the character are part of the path name only the last byte of the character must be specified.
Example (excerpt from .../00/00/00.dat):
# percent 0x25 \% # asterisk t 0x2A \textasteriskcentered{} m 0x2A *