Date: 17 Dec 87 Message No: 002
To: TeX implementors and distributors
From: Barbara Beeton
Subject: Announcing TeX 2.8 and other matters
Yet another bug, albeit of limited effect, has been detected and fixed
in TeX; the two messages below from Don Knuth describe the bug and another
problem that has also been called to his attention. The details from the
bug file and a source file .diff list (2.8 to 2.7) follow after that.
I'd like to thank everyone for suggestions on useful topics for future
communications. I have the following on the list:
Metafont -- details of changes after version 1.0
Trip and Trap tests -- changes to the tests and test results,
since TeX 2.0 and Metafont 1.0
LaTeX -- cumulative changes since the earliest version 2.09, with
context to permit their installation
AMSTeX -- a new version, fully documented, will appear early
next year
Fonts originating at the AMS -- a full set of 300dpi versions will
be made available to all distributors, and I will try to
arrange for installation at Score. This includes Euler
(Fraktur, script and cursive, in both lightface and bold
versions), cyrillic, and 2 math symbol fonts. Projected
timeframe, May-June 1988.
Of course, any new messages from Don Knuth will be forwarded post-haste.
And all additional suggestions are welcome.
--------------------
Date: 12 Dec 87 1639 PST
From: Don Knuth <DEK@SAIL.Stanford.EDU>
Subject: a truly insignificant TeX bug
To: BB@SAIL.Stanford.EDU, DRF@SAIL.Stanford.EDU
Peter Breitenlohner in Munich just reported a glitch that would cause
trouble only if all of the following were true:
1) The user is running INITEX with no hyphenation \patterns;
2) INITEX tries to hyphenate a word containing a character
whose \lccode is equal to 1 (the minimum value);
3) The implementation has min_quarterword unequal to its
normal value 0.
I should have set trie_op to min_quarterword in module 952, but I set
it to zero by mistake. If no \patterns are given, the circumstances
above will get past the test "hc[l]=trie_char(z)" in module 923,
and it looks like module 924 will loop endlessly.
I made version 2.8 just for this, but people with version 2.7 need not
feel behind the times.
-------
Date: 12 Dec 87 1712 PST
From: Don Knuth <DEK@SAIL.Stanford.EDU>
Subject: a non-bug
To: BB@SAIL.Stanford.EDU, DRF@SAIL.Stanford.EDU
I received a note from Joachim Schrod at Darmstadt, who observes that
\dump does save the value of mag_set. Therefore TeX does not
complain if you try to reset the magnification after dumping a program
that already "froze" it.
My reply to him was that I don't think it wise to change TeX at this
late date just to disallow magnification. If a user really wants to
dump a format for which magnification is to be disallowed, he can
do so with \everyjob.
--------------------
Extract from TeX82.BUG
336. If there are no \patterns and some \lccode is 1 (Breitenlohner, 12Dec87)
@x module 952
trie:=link(0)_0; trie:=char(0)_0; trie:=op(0)_0;
@y
trie:=link(0)_0; trie:=char(0)_0; trie:=op(0)_min:=quarterword;
@z
337. (I sincerely hope that there won't be any more)
--------------------
Extracts from ERRATA.TeX (differences as compared with previous version)
**** FILE PS:<BEETON.BB>ERRATA.NEW.2, 2-2 (1508)
\bugonpage A326, line 12 (12/12/87)
\ninepoint\noindent
its natural width. The |\hbox| version also invokes |\everymath|.
**** FILE PS:<BEETON.BB>ERRATA.TEX.72, 2-2 (1508)
\bugonpage A326, line 12 (9/20/87)
\ninepoint\noindent
its natural width. The |\hbox| version also invokes |\everypar|.
***************
**** FILE PS:<BEETON.BB>ERRATA.NEW.2, 2-34 (2772)
\bugonpage A465, entry for {\tt\char`\\everymath} (12/12/87)
\eightpoint\indent[Include also a reference to page 326.]
**** FILE PS:<BEETON.BB>ERRATA.TEX.72, 2-34 (2770)
***************
**** FILE PS:<BEETON.BB>ERRATA.NEW.2, 3-10 (3410)
{\tt This\]is\]TeX,\]Version\]2.8\char'23}\quad
$\{\,$printed when \TeX\ starts$\,\}$
**** FILE PS:<BEETON.BB>ERRATA.TEX.72, 3-10 (3283)
{\tt This\]is\]TeX,\]Version\]2.7\char'23}\quad
$\{\,$printed when \TeX\ starts$\,\}$
***************
**** FILE PS:<BEETON.BB>ERRATA.NEW.2, 3-143 (7866)
\bugonpage B396, bottom line (12/12/87)
\ninepoint\noindent\hskip10pt
$\\{trie\_link}(0)\gets0$; $\\{trie\_char}(0)\gets0$;
$\\{trie\_op}(0)\gets\\{min\_quarterword}$;
**** FILE PS:<BEETON.BB>ERRATA.TEX.72, 3-143 (7739)
***************
--------------------
Differences between TeX.WEB versions 2.8 and 2.7
**** FILE TX:<BNB.WORK>TEX-2-8.WEB.1, 1-26 (1735)
% Version 2.8 fixes weird bug if no \patterns are used (December 1987).
% A reward of $40.96 will be paid to the first finder of any remaining bug.
**** FILE TX:<TEX82.NEW>TEX-2-7.WEB.1, 1-27 (1737)
% A reward of $40.96 will be paid to the first finder of any remaining bug.
***************
**** FILE TX:<BNB.WORK>TEX-2-8.WEB.1, 1-167 (8995)
@d banner=='This is TeX, Version 2.8' {printed when \TeX\ starts}
@ Different \PASCAL s have slightly different conventions, and the present
**** FILE TX:<TEX82.NEW>TEX-2-7.WEB.1, 1-166 (8922)
@d banner=='This is TeX, Version 2.7' {printed when \TeX\ starts}
@ Different \PASCAL s have slightly different conventions, and the present
***************
**** FILE TX:<BNB.WORK>TEX-2-8.WEB.1, 1-17868 (772952)
trie_link(0):=0; trie_char(0):=0; trie_op(0):=min_quarterword;
for k:=1 to 127 do trie[k]:=trie[0];
**** FILE TX:<TEX82.NEW>TEX-2-7.WEB.1, 1-17868 (772881)
trie_link(0):=0; trie_char(0):=0; trie_op(0):=0;
for k:=1 to 127 do trie[k]:=trie[0];
***************
-------