<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<div class="moz-cite-prefix">David Carlisle wrote:<br>
</div>
<blockquote type="cite"
cite="mid:CAEW6iOgGq7+g=fTfnBMUxbE+ptn9Xcm3tH6JOTdq=PYhGKaZgQ@mail.gmail.com">
<pre class="moz-quote-pre" wrap="">The hyphenation in luatex is different in many ways to that of classic
tex even when they use the same \pattern data so it's not that
surprising that you get different results for some constructs.
Here the main issue is that apart from some legacy compatibility
luatex has moved away from the (frankly weird) reliance on lowercase
codes for determining which characters take part in hyphenation. So
luatex sees the whole construct as a single word and looks it up using
the patterns, pdftex sees the * as a word boundary and doesn't start
the next word until after it sees some white space so skips most of
this.
If you set the lccode of * to itself then you get the same result as luatex</pre>
</blockquote>
<br>
Ah yes, <i>mea culpa</i> — it is \lccode, not \catcode, that
matters here (see below). But I would, with respect, beg to differ
with your assertion that TeX has a "frankly weird" reliance on
\lccodes; at the time that TeX was written, every last byte counted,
and if overloading the significance of \lccode allowed other more
important features to be included, then I would suggest that the
decision was a wise one at the time.<br>
<br>
<blockquote type="cite"> TEX looks for potentially hyphenatable
words by searching ahead from each glue<br>
item that is not in a math formula. The search bypasses characters
whose<br>
\lccode is zero, or ligatures that begin with such characters; it
also bypasses whatsits<br>
and implicit kern items, i.e., kerns that were inserted by TEX
itself because of information<br>
stored with the font. If the search finds a character with nonzero
\lccode, or if it finds a ligature<br>
that begins with such a character, that character is called the
starting letter. But if any<br>
other type of item occurs before a suitable starting letter is
found, hyphenation is abandoned<br>
(until after the next glue item). Thus, a box or rule or mark, or
a kern that was explicitly inserted<br>
by \kern or \/, must not intervene between glue and a hyphenatable
word. If the starting<br>
letter is not lowercase (i.e., if it doesn’t equal its own
\lccode), hyphenation is abandoned<br>
unless \uchyph is positive.<br>
<br>
If a suitable starting letter is found, let it be in font f.
Hyphenation is abandoned unless<br>
the \hyphenchar of f is a number between 0 and 255, inclusive. If
this test is<br>
passed, TEX continues to scan forward until coming to something
that’s not one of the following<br>
three “admissible items”: (1) a character in font f whose \lccode
is nonzero; (2) a ligature<br>
formed entirely from characters of type (1); (3) an implicit kern.
The first inadmissible<br>
item terminates this part of the process; the trial word consists
of all the letters found in admissible<br>
items. Notice that all of these letters are in font f.</blockquote>
<br>
-- <br>
<i>** Phil.</i><br>
</body>
</html>