Release FreeBSD 1.1upstream/1.1.0_cvsrelease/1.1.0_cvs

This commit was manufactured to restore the state of the 1.1-RELEASE image.
Releases prior to 5.3-RELEASE are omitting the secure/ and crypto/ subdirs.

12 files changed, 7672 insertions, 0 deletions

diff --git a/gnu/usr.bin/dc/COPYING b/gnu/usr.bin/dc/COPYING
new file mode 100644
index 000000000000..e77696ae8ddf
--- /dev/null
+++ b/gnu/usr.bin/dc/COPYING
@@ -0,0 +1,339 @@
+		    GNU GENERAL PUBLIC LICENSE
+		       Version 2, June 1991
+
+ Copyright (C) 1989, 1991 Free Software Foundation, Inc.
+                          675 Mass Ave, Cambridge, MA 02139, USA
+ Everyone is permitted to copy and distribute verbatim copies
+ of this license document, but changing it is not allowed.
+
+			    Preamble
+
+  The licenses for most software are designed to take away your
+freedom to share and change it.  By contrast, the GNU General Public
+License is intended to guarantee your freedom to share and change free
+software--to make sure the software is free for all its users.  This
+General Public License applies to most of the Free Software
+Foundation's software and to any other program whose authors commit to
+using it.  (Some other Free Software Foundation software is covered by
+the GNU Library General Public License instead.)  You can apply it to
+your programs, too.
+
+  When we speak of free software, we are referring to freedom, not
+price.  Our General Public Licenses are designed to make sure that you
+have the freedom to distribute copies of free software (and charge for
+this service if you wish), that you receive source code or can get it
+if you want it, that you can change the software or use pieces of it
+in new free programs; and that you know you can do these things.
+
+  To protect your rights, we need to make restrictions that forbid
+anyone to deny you these rights or to ask you to surrender the rights.
+These restrictions translate to certain responsibilities for you if you
+distribute copies of the software, or if you modify it.
+
+  For example, if you distribute copies of such a program, whether
+gratis or for a fee, you must give the recipients all the rights that
+you have.  You must make sure that they, too, receive or can get the
+source code.  And you must show them these terms so they know their
+rights.
+
+  We protect your rights with two steps: (1) copyright the software, and
+(2) offer you this license which gives you legal permission to copy,
+distribute and/or modify the software.
+
+  Also, for each author's protection and ours, we want to make certain
+that everyone understands that there is no warranty for this free
+software.  If the software is modified by someone else and passed on, we
+want its recipients to know that what they have is not the original, so
+that any problems introduced by others will not reflect on the original
+authors' reputations.
+
+  Finally, any free program is threatened constantly by software
+patents.  We wish to avoid the danger that redistributors of a free
+program will individually obtain patent licenses, in effect making the
+program proprietary.  To prevent this, we have made it clear that any
+patent must be licensed for everyone's free use or not licensed at all.
+
+  The precise terms and conditions for copying, distribution and
+modification follow.
+
+		    GNU GENERAL PUBLIC LICENSE
+   TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
+
+  0. This License applies to any program or other work which contains
+a notice placed by the copyright holder saying it may be distributed
+under the terms of this General Public License.  The "Program", below,
+refers to any such program or work, and a "work based on the Program"
+means either the Program or any derivative work under copyright law:
+that is to say, a work containing the Program or a portion of it,
+either verbatim or with modifications and/or translated into another
+language.  (Hereinafter, translation is included without limitation in
+the term "modification".)  Each licensee is addressed as "you".
+
+Activities other than copying, distribution and modification are not
+covered by this License; they are outside its scope.  The act of
+running the Program is not restricted, and the output from the Program
+is covered only if its contents constitute a work based on the
+Program (independent of having been made by running the Program).
+Whether that is true depends on what the Program does.
+
+  1. You may copy and distribute verbatim copies of the Program's
+source code as you receive it, in any medium, provided that you
+conspicuously and appropriately publish on each copy an appropriate
+copyright notice and disclaimer of warranty; keep intact all the
+notices that refer to this License and to the absence of any warranty;
+and give any other recipients of the Program a copy of this License
+along with the Program.
+
+You may charge a fee for the physical act of transferring a copy, and
+you may at your option offer warranty protection in exchange for a fee.
+
+  2. You may modify your copy or copies of the Program or any portion
+of it, thus forming a work based on the Program, and copy and
+distribute such modifications or work under the terms of Section 1
+above, provided that you also meet all of these conditions:
+
+    a) You must cause the modified files to carry prominent notices
+    stating that you changed the files and the date of any change.
+
+    b) You must cause any work that you distribute or publish, that in
+    whole or in part contains or is derived from the Program or any
+    part thereof, to be licensed as a whole at no charge to all third
+    parties under the terms of this License.
+
+    c) If the modified program normally reads commands interactively
+    when run, you must cause it, when started running for such
+    interactive use in the most ordinary way, to print or display an
+    announcement including an appropriate copyright notice and a
+    notice that there is no warranty (or else, saying that you provide
+    a warranty) and that users may redistribute the program under
+    these conditions, and telling the user how to view a copy of this
+    License.  (Exception: if the Program itself is interactive but
+    does not normally print such an announcement, your work based on
+    the Program is not required to print an announcement.)
+
+These requirements apply to the modified work as a whole.  If
+identifiable sections of that work are not derived from the Program,
+and can be reasonably considered independent and separate works in
+themselves, then this License, and its terms, do not apply to those
+sections when you distribute them as separate works.  But when you
+distribute the same sections as part of a whole which is a work based
+on the Program, the distribution of the whole must be on the terms of
+this License, whose permissions for other licensees extend to the
+entire whole, and thus to each and every part regardless of who wrote it.
+
+Thus, it is not the intent of this section to claim rights or contest
+your rights to work written entirely by you; rather, the intent is to
+exercise the right to control the distribution of derivative or
+collective works based on the Program.
+
+In addition, mere aggregation of another work not based on the Program
+with the Program (or with a work based on the Program) on a volume of
+a storage or distribution medium does not bring the other work under
+the scope of this License.
+
+  3. You may copy and distribute the Program (or a work based on it,
+under Section 2) in object code or executable form under the terms of
+Sections 1 and 2 above provided that you also do one of the following:
+
+    a) Accompany it with the complete corresponding machine-readable
+    source code, which must be distributed under the terms of Sections
+    1 and 2 above on a medium customarily used for software interchange; or,
+
+    b) Accompany it with a written offer, valid for at least three
+    years, to give any third party, for a charge no more than your
+    cost of physically performing source distribution, a complete
+    machine-readable copy of the corresponding source code, to be
+    distributed under the terms of Sections 1 and 2 above on a medium
+    customarily used for software interchange; or,
+
+    c) Accompany it with the information you received as to the offer
+    to distribute corresponding source code.  (This alternative is
+    allowed only for noncommercial distribution and only if you
+    received the program in object code or executable form with such
+    an offer, in accord with Subsection b above.)
+
+The source code for a work means the preferred form of the work for
+making modifications to it.  For an executable work, complete source
+code means all the source code for all modules it contains, plus any
+associated interface definition files, plus the scripts used to
+control compilation and installation of the executable.  However, as a
+special exception, the source code distributed need not include
+anything that is normally distributed (in either source or binary
+form) with the major components (compiler, kernel, and so on) of the
+operating system on which the executable runs, unless that component
+itself accompanies the executable.
+
+If distribution of executable or object code is made by offering
+access to copy from a designated place, then offering equivalent
+access to copy the source code from the same place counts as
+distribution of the source code, even though third parties are not
+compelled to copy the source along with the object code.
+
+  4. You may not copy, modify, sublicense, or distribute the Program
+except as expressly provided under this License.  Any attempt
+otherwise to copy, modify, sublicense or distribute the Program is
+void, and will automatically terminate your rights under this License.
+However, parties who have received copies, or rights, from you under
+this License will not have their licenses terminated so long as such
+parties remain in full compliance.
+
+  5. You are not required to accept this License, since you have not
+signed it.  However, nothing else grants you permission to modify or
+distribute the Program or its derivative works.  These actions are
+prohibited by law if you do not accept this License.  Therefore, by
+modifying or distributing the Program (or any work based on the
+Program), you indicate your acceptance of this License to do so, and
+all its terms and conditions for copying, distributing or modifying
+the Program or works based on it.
+
+  6. Each time you redistribute the Program (or any work based on the
+Program), the recipient automatically receives a license from the
+original licensor to copy, distribute or modify the Program subject to
+these terms and conditions.  You may not impose any further
+restrictions on the recipients' exercise of the rights granted herein.
+You are not responsible for enforcing compliance by third parties to
+this License.
+
+  7. If, as a consequence of a court judgment or allegation of patent
+infringement or for any other reason (not limited to patent issues),
+conditions are imposed on you (whether by court order, agreement or
+otherwise) that contradict the conditions of this License, they do not
+excuse you from the conditions of this License.  If you cannot
+distribute so as to satisfy simultaneously your obligations under this
+License and any other pertinent obligations, then as a consequence you
+may not distribute the Program at all.  For example, if a patent
+license would not permit royalty-free redistribution of the Program by
+all those who receive copies directly or indirectly through you, then
+the only way you could satisfy both it and this License would be to
+refrain entirely from distribution of the Program.
+
+If any portion of this section is held invalid or unenforceable under
+any particular circumstance, the balance of the section is intended to
+apply and the section as a whole is intended to apply in other
+circumstances.
+
+It is not the purpose of this section to induce you to infringe any
+patents or other property right claims or to contest validity of any
+such claims; this section has the sole purpose of protecting the
+integrity of the free software distribution system, which is
+implemented by public license practices.  Many people have made
+generous contributions to the wide range of software distributed
+through that system in reliance on consistent application of that
+system; it is up to the author/donor to decide if he or she is willing
+to distribute software through any other system and a licensee cannot
+impose that choice.
+
+This section is intended to make thoroughly clear what is believed to
+be a consequence of the rest of this License.
+
+  8. If the distribution and/or use of the Program is restricted in
+certain countries either by patents or by copyrighted interfaces, the
+original copyright holder who places the Program under this License
+may add an explicit geographical distribution limitation excluding
+those countries, so that distribution is permitted only in or among
+countries not thus excluded.  In such case, this License incorporates
+the limitation as if written in the body of this License.
+
+  9. The Free Software Foundation may publish revised and/or new versions
+of the General Public License from time to time.  Such new versions will
+be similar in spirit to the present version, but may differ in detail to
+address new problems or concerns.
+
+Each version is given a distinguishing version number.  If the Program
+specifies a version number of this License which applies to it and "any
+later version", you have the option of following the terms and conditions
+either of that version or of any later version published by the Free
+Software Foundation.  If the Program does not specify a version number of
+this License, you may choose any version ever published by the Free Software
+Foundation.
+
+  10. If you wish to incorporate parts of the Program into other free
+programs whose distribution conditions are different, write to the author
+to ask for permission.  For software which is copyrighted by the Free
+Software Foundation, write to the Free Software Foundation; we sometimes
+make exceptions for this.  Our decision will be guided by the two goals
+of preserving the free status of all derivatives of our free software and
+of promoting the sharing and reuse of software generally.
+
+			    NO WARRANTY
+
+  11. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
+FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW.  EXCEPT WHEN
+OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
+PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
+OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
+MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.  THE ENTIRE RISK AS
+TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU.  SHOULD THE
+PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
+REPAIR OR CORRECTION.
+
+  12. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
+WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
+REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
+INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
+OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
+TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
+YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
+PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
+POSSIBILITY OF SUCH DAMAGES.
+
+		     END OF TERMS AND CONDITIONS
+
+	    How to Apply These Terms to Your New Programs
+
+  If you develop a new program, and you want it to be of the greatest
+possible use to the public, the best way to achieve this is to make it
+free software which everyone can redistribute and change under these terms.
+
+  To do so, attach the following notices to the program.  It is safest
+to attach them to the start of each source file to most effectively
+convey the exclusion of warranty; and each file should have at least
+the "copyright" line and a pointer to where the full notice is found.
+
+    <one line to give the program's name and a brief idea of what it does.>
+    Copyright (C) 19yy  <name of author>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License
+    along with this program; if not, write to the Free Software
+    Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+
+Also add information on how to contact you by electronic and paper mail.
+
+If the program is interactive, make it output a short notice like this
+when it starts in an interactive mode:
+
+    Gnomovision version 69, Copyright (C) 19yy name of author
+    Gnomovision comes with ABSOLUTELY NO WARRANTY; for details type `show w'.
+    This is free software, and you are welcome to redistribute it
+    under certain conditions; type `show c' for details.
+
+The hypothetical commands `show w' and `show c' should show the appropriate
+parts of the General Public License.  Of course, the commands you use may
+be called something other than `show w' and `show c'; they could even be
+mouse-clicks or menu items--whatever suits your program.
+
+You should also get your employer (if you work as a programmer) or your
+school, if any, to sign a "copyright disclaimer" for the program, if
+necessary.  Here is a sample; alter the names:
+
+  Yoyodyne, Inc., hereby disclaims all copyright interest in the program
+  `Gnomovision' (which makes passes at compilers) written by James Hacker.
+
+  <signature of Ty Coon>, 1 April 1989
+  Ty Coon, President of Vice
+
+This General Public License does not permit incorporating your program into
+proprietary programs.  If your program is a subroutine library, you may
+consider it more useful to permit linking proprietary applications with the
+library.  If this is what you want to do, use the GNU Library General
+Public License instead of this License.
diff --git a/gnu/usr.bin/dc/ChangeLog b/gnu/usr.bin/dc/ChangeLog
new file mode 100644
index 000000000000..09aaf4706884
--- /dev/null
+++ b/gnu/usr.bin/dc/ChangeLog
@@ -0,0 +1,77 @@
+Fri May 21 15:02:52 1993  Noah Friedman  (friedman@nutrimat.gnu.ai.mit.edu)
+
+        * Version 0.2 released.
+
+Fri May 21 11:48:11 1993  Richard Stallman  (rms@mole.gnu.ai.mit.edu)
+
+	* decimal.c (decimal_rem): Update to match fixes in decimal_div.
+
+Thu May 20 03:12:41 1993  Noah Friedman  (friedman@nutrimat.gnu.ai.mit.edu)
+
+        * Makefile.in (realclean): Delete dc.info* and configure. 
+        (DISTFILES): Add `texinfo.tex' and `NEWS'.
+        texinfo.tex: New file (symlink to canonical source).
+        NEWS: New file. 
+
+Wed May 19 11:30:09 1993  Richard Stallman  (rms@mole.gnu.ai.mit.edu)
+
+	* dc.c (dec_read): Accept only A through F.
+
+Tue May 18 12:35:54 1993  Richard Stallman  (rms@mole.gnu.ai.mit.edu)
+
+	* dc.c (read_string): New arg STARTC to handle nested brackets.
+	(execute): Change calls to read_string.
+	(condop): Don't assume result of decimal_compare has abs value <= 1.
+	(popmacro): If no macro in progress, exit.
+
+Sun May  2 00:42:47 1993  Richard Stallman  (rms@mole.gnu.ai.mit.edu)
+
+	* decimal.c (decimal_div): Include in trial_dividend the digit
+	at length2 + i - 2, if there is one.
+
+Sat May  1 09:54:35 1993  Richard Stallman  (rms@mole.gnu.ai.mit.edu)
+
+	* decimal.c (decimal_parse): Don't use digits without recalculation
+	if some digit exceeds the radix.
+
+	* dc.c (execute): Treat A...F as digits.
+	(dec_read): Treat A...F as digits.
+
+Thu Apr 29 14:17:30 1993  Richard Stallman  (rms@mole.gnu.ai.mit.edu)
+
+	* decimal.h (bcopy): Use memcpy, not memmove.
+
+	* decimal.c (flush_trailing_digits): Use explicit loop, not bcopy.
+
+Tue Apr 20 17:21:27 1993  Noah Friedman  (friedman@nutrimat.gnu.ai.mit.edu)
+
+        * dc.c (pushsqrt): `precision' is an argument to `decimal_sqrt', not
+        `push'. 
+
+Sat Apr 17 15:47:55 1993  Noah Friedman  (friedman@nutrimat.gnu.ai.mit.edu)
+
+        * All files: Updated GPL version number.
+
+        * decimal.c: Include decimal.h and delete duplicate declarations.
+
+        * decimal.h [!HAVE_BCOPY]: #define bcopy.
+        [!HAVE_BZERO]: #define bzero.
+
+Sun Feb 10 22:06:15 1991  Richard Stallman  (rms at mole.ai.mit.edu)
+
+	* dc.c (execute): Insert break; in \n case.
+
+Sun Jul 29 17:50:14 1990  Richard Stallman  (rms at sugar-bombs.ai.mit.edu)
+
+	* decimal.c (decimal_neg): New function.
+
+Fri Jul 27 04:11:34 1990  David J. MacKenzie  (djm at albert.ai.mit.edu)
+
+	* bceval.c, bclex.c, bcprint.c, bcsym.c: Declare some functions
+	static. 
+
+Mon Dec 25 03:01:49 1989  David J. MacKenzie  (djm at hobbes.ai.mit.edu)
+
+	* Makefile: add some missing rules.
+
+	* decimal.c: change a 'max' to 'MAX'.
diff --git a/gnu/usr.bin/dc/Makefile b/gnu/usr.bin/dc/Makefile
new file mode 100644
index 000000000000..937c7d9f2d65
--- /dev/null
+++ b/gnu/usr.bin/dc/Makefile
@@ -0,0 +1,7 @@
+PROG=	dc
+SRCS=	dc.c decimal.c
+CFLAGS+=-I${.CURDIR} -DHAVE_BCOPY=1 -DHAVE_BZERO=1
+DPADD=	${LIBM}
+LDADD=	-lm
+
+.include <bsd.prog.mk>
diff --git a/gnu/usr.bin/dc/NEWS b/gnu/usr.bin/dc/NEWS
new file mode 100644
index 000000000000..6486afb3ec86
--- /dev/null
+++ b/gnu/usr.bin/dc/NEWS
@@ -0,0 +1,7 @@
+Changes between version 0.2 and 0.1:
+
+* You can now have nested square bracket pairs within a string.
+
+* The letters A-F can now be part of a number when the input radix is
+large enough to make them meaningful.
+
diff --git a/gnu/usr.bin/dc/README b/gnu/usr.bin/dc/README
new file mode 100644
index 000000000000..c23cc6699202
--- /dev/null
+++ b/gnu/usr.bin/dc/README
@@ -0,0 +1,13 @@
+This is a preliminary release of GNU `dc', since people asked for it.  GNU
+`bc' (which doesn't rely on a separate `dc') has been available separately
+for a couple of years.  Eventually this version of `dc' will be merged with
+the bc package. 
+
+See comments in the file decimal.c for some limitations in the arbitrary
+precision library.  It's questionable whether it's worth fixing these
+problems since the merged dc will probably use bc's math library instead.
+However, you might want to be aware of known problems. 
+
+See the file `INSTALL' for instructions on building and installing dc.
+
+Please report bugs to bug-gnu-utils@prep.ai.mit.edu.
diff --git a/gnu/usr.bin/dc/dc.1 b/gnu/usr.bin/dc/dc.1
new file mode 100644
index 000000000000..17d9356a8fa0
--- /dev/null
+++ b/gnu/usr.bin/dc/dc.1
@@ -0,0 +1,278 @@
+.TH DC 1 "03 Aug 1993" "GNU Project"
+.SH NAME
+dc, An Arbitrary Precision Calculator
+.SH SYNOPSIS
+.B dc
+.SH DESCRIPTION
+.PP
+DC is a reverse-polish desk calculator which supports unlimited
+precision arithmetic.  It also allows you to define and call macros.
+Normally DC reads from the standard input; if any command arguments
+are given to it, they are filenames, and DC reads and executes the
+contents of the files before reading from standard input.  All output
+is to standard output.
+
+A reverse-polish calculator stores numbers on a stack.  Entering a
+number pushes it on the stack.  Arithmetic operations pop arguments off
+the stack and push the results.
+
+To enter a number in DC, type the digits, with an optional decimal
+point.  Exponential notation is not supported.  To enter a negative
+number, begin the number with `_'.  `-' cannot be used for this, as it
+is a binary operator for subtraction instead.  To enter two numbers in
+succession, separate them with spaces or newlines.  These have no
+meaning as commands.
+.PD
+.SH "Printing Commands"
+.PP
+.B p
+Prints the value on the top of the stack,
+without altering the stack.  A newline is printed
+after the value.
+.PP
+.B P
+Prints the value on the top of the stack,
+popping it off, and does not print a newline after.
+.PP
+.B f
+Prints the entire contents of the stack
+and the contents of all of the registers,
+without altering anything.  This is a good command
+to use if you are lost or want to figure out
+what the effect of some command has been.
+.PD
+.SH "Arithmetic"
+.PP
+.B +
+Pops two values off the stack, adds them,
+and pushes the result.  The precision of the result
+is determined only by the values of the arguments,
+and is enough to be exact.
+.PP
+.B -
+Pops two values, subtracts the first one popped
+from the second one popped, and pushes the result.
+.PP
+.B *
+Pops two values, multiplies them, and pushes the result.
+The number of fraction digits in the result is controlled
+by the current precision flag (see below) and does not
+depend on the values being multiplied.
+.PP
+.B /
+Pops two values, divides the second one popped from
+the first one popped, and pushes the result.
+The number of fraction digits is specified by the precision flag.
+.PP
+.B %
+Pops two values, computes the remainder of the division
+that the \fB/\fR command would do, and pushes that.
+The division is done with as many fraction digits
+as the precision flag specifies, and the remainder
+is also computed with that many fraction digits.
+.PP
+.B ^
+Pops two values and exponentiates, using the first
+value popped as the exponent and the second popped as the base.
+The fraction part of the exponent is ignored.
+The precision flag specifies the number of fraction
+digits in the result.
+.PP
+.B v
+Pops one value, computes its square root, and pushes that.
+The precision flag specifies the number of fraction digits
+in the result.
+.PP
+Most arithmetic operations are affected by the "precision flag",
+which you can set with the
+.BR k
+command.  The default precision
+value is zero, which means that all arithmetic except for
+addition and subtraction produces integer results.
+.PP
+The remainder operation
+.BR %
+requires some explanation: applied to
+arguments `a' and `b' it produces `a - (b * (a / b))',
+where `a / b' is computed in the current precision.
+.PP
+.SH "Stack Control"
+.PP
+.B c
+Clears the stack, rendering it empty.
+.PP
+.B d
+Duplicates the value on the top of the stack,
+pushing another copy of it.  Thus,
+`4d*p' computes 4 squared and prints it.
+.SH "Registers"
+.PP
+DC provides 128 memory registers, each named by a single
+ASCII character.  You can store a number in a register
+and retrieve it later.
+.PP
+.B s\fIr\fR
+Pop the value off the top of the stack and store
+it into register \fIr\fR.
+.PP
+.B l\fIr\fR
+Copy the value in register \fIr\fR and push it onto the stack.  This
+does not alter the contents of \fIr\fR.
+.PP
+Each register also contains its own stack.  The current
+register value is the top of the register's stack.
+.PP
+.B S\fIr\fR
+Pop the value off the top of the (main) stack and
+push it onto the stack of register \fIr\fR.
+The previous value of the register becomes inaccessible.
+.PP
+.B L\fIr\fR
+Pop the value off the top of register \fIr\fR's stack
+and push it onto the main stack.  The previous value
+in register \fIr\fR's stack, if any, is now accessible
+via the 
+.BR Ir
+command.
+.PP
+The
+.BR f
+command prints a list of all registers that have contents
+stored in them, together with their contents.  Only the
+current contents of each register (the top of its stack)
+is printed.
+.PP
+.SH "Parameters"
+.PP
+DC has three parameters that control its operation: the precision, the
+input radix, and the output radix.  The precision specifies the number
+of fraction digits to keep in the result of most arithmetic operations.
+The input radix controls the interpretation of numbers typed in;
+allnumbers typed in use this radix.  The output radix is used
+for printing numbers.
+.PP
+The input and output radices are separate parameters; you can make them
+unequal, which can be useful or confusing.  Each radix must be between 2
+and 36 inclusive.  The precision must be zero or greater.  The precision
+is always measured in decimal digits, regardless of the current input or
+output radix.
+.PP
+.B i
+Pops the value off the top of the stack
+and uses it to set the input radix.
+.PP
+.B o
+.PP
+.B k
+Similarly set the output radix and the precision.
+.PP
+.B I
+Pushes the current input radix on the stack.
+.PP
+.B O
+.PP
+.B K
+Similarly push the current output radix and the current precision.
+.PP
+.SH "Strings"
+.PP
+DC can operate on strings as well as on numbers.  The only things you
+can do with strings are print them and execute them as macros (which
+means that the contents of the string are processed as DC commands).
+Both registers and the stack can hold strings, and DC always knows
+whether any given object is a string or a number.  Some commands such as
+arithmetic operations demand numbers as arguments and print errors if
+given strings.  Other commands can accept either a number or a string;
+for example, the
+.BR p
+command can accept either and prints the object
+according to its type.
+.PP
+.B [characters]
+Makes a string containing
+.BR characters
+and pushes it
+on the stack.  For example,
+.BR [foo]p
+prints the
+characters \fBfoo\fR (with no newline).
+.PP
+.B x
+Pops a value off the stack and executes it as a macro.
+Normally it should be a string; if it is a number,
+it is simply pushed back onto the stack.
+For example,
+.BR [1p]x
+executes the macro
+.BR 1p
+which pushes \fB1\fR on the stack and prints \fB1\fR
+on a separate line.
+.PP
+Macros are most often stored in registers;
+\fB[1p]sa\fR stores a macro to print \fB1\fR into register \fBa\fR,
+and \fBlax\fR invokes the macro.
+.PP
+.B >\fIr\fR
+Pops two values off the stack and compares them
+assuming they are numbers, executing the contents
+of register \fIr\fR as a macro if the original top-of-stack
+is greater.  Thus, \fB1 2>a\fR will invoke register \fBa\fR's contents
+and \fB2 1>a\fR will not.
+.PP
+.B <\fIr\fB
+Similar but invokes the macro if the original top-of-stack
+is less.
+.PP
+.B =\fIr\fR
+Similar but invokes the macro if the two numbers popped
+are equal.  This can also be validly used to compare two
+strings for equality.
+.PP
+.B ?
+Reads a line from the terminal and executes it.
+This command allows a macro to request input from the user.
+.PP
+.B q
+During the execution of a macro, this comand
+does not exit DC.  Instead, it exits from that
+macro and also from the macro which invoked it (if any).
+.PP
+.B Q
+Pops a value off the stack and uses it as a count
+of levels of macro execution to be exited.  Thus,
+\fB3Q\fR exits three levels.
+.SH "Status Inquiry"
+.PP
+.B Z
+Pops a value off the stack, calculates the number of
+digits it has (or number of characters, if it is a string)
+and pushes that number.
+.PP
+.B X
+Pops a value off the stack, calculates the number of
+fraction digits it has, and pushes that number.
+For a string, the value pushed is -1.
+.PP
+.B z
+Pushes the current stack depth; the number of
+objects on the stack before the execution of the \fBz\fR command.
+.PP
+.B I
+Pushes the current value of the input radix.
+.PP
+.B O
+Pushes the current value of the output radix.
+.PP
+.B K
+Pushes the current value of the precision.
+.SH "Notes"
+.PP
+The \fB:\fR and \fB;\fR commands of the Unix DC program are
+not supported, as the documentation does not say what they do.
+The \fB!\fR command is not supported, but will be supported
+as soon as a library for executing a line as a command exists.
+.SH BUGS
+.PP
+Email bug reports to
+.BR bug-gnu-utils@prep.ai.mit.edu .
+Be sure to include the word ``dc'' somewhere in the ``Subject:'' field.
diff --git a/gnu/usr.bin/dc/dc.c b/gnu/usr.bin/dc/dc.c
new file mode 100644
index 000000000000..933b24b916cd
--- /dev/null
+++ b/gnu/usr.bin/dc/dc.c
@@ -0,0 +1,909 @@
+/* 
+ * `dc' desk calculator utility.
+ *
+ * Copyright (C) 1984, 1993 Free Software Foundation, Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, you can either send email to this
+ * program's author (see below) or write to: The Free Software Foundation,
+ * Inc.; 675 Mass Ave. Cambridge, MA 02139, USA.
+ */
+
+#include <stdio.h>
+#include "decimal.h"  /* definitions for our decimal arithmetic package */
+
+FILE *open_file;	/* input file now open */
+int file_count;		/* Number of input files not yet opened */
+char **next_file;	/* Pointer to vector of names of input files left */
+
+struct regstack
+  {
+    decimal value;	/* Saved value of register */
+    struct regstack *rest;	/* Tail of list */
+  };
+
+typedef struct regstack *regstack;
+
+regstack freeregstacks;	/* Chain of free regstack structures for fast realloc */
+
+decimal regs[128];	/* "registers", with single-character names */
+regstack regstacks[128]; /* For each register, a stack of previous values */
+
+int stacktop;		/* index of last used element in stack */
+int stacksize;		/* Current allocates size of stack */
+decimal *stack;	/* Pointer to computation stack */
+
+/* A decimal number can be regarded as a string by
+ treating its contents as characters and ignoring the
+ position of its decimal point.
+ Decimal numbers are marked as strings by having an `after' field of -1
+ One use of strings is to execute them as macros.
+*/
+
+#define STRING -1
+
+int macrolevel;	/* Current macro nesting; 0 if taking keyboard input */
+int macrostacksize;	/* Current allocated size of macrostack and macroindex */
+decimal *macrostack;	/* Pointer to macro stack array */
+int *macroindex;	/* Pointer to index-within-macro stack array */
+	/* Note that an empty macro is popped from the stack
+	   only when an trying to read a character from it
+	   or trying to push another macro.  */
+
+int ibase;	/* Radix for numeric input.  */
+int obase;	/* Radix for numeric output.  */
+int precision;	/* Number of digits to keep in multiply and divide.  */
+
+char *buffer;	/* Address of buffer used for reading numbers */
+int bufsize;	/* Current size of buffer (made bigger when nec) */
+
+decimal dec_read ();
+regstack get_regstack ();
+int fetch ();
+int fgetchar ();
+char *concat ();
+void pushsqrt ();
+void condop ();
+void setibase ();
+void setobase ();
+void setprecision ();
+void pushmacro ();
+decimal read_string ();
+void pushlength ();
+void pushscale ();
+void unfetch ();
+void popmacros ();
+void popmacro ();
+void popstack ();
+void print_obj ();
+void print_string ();
+void free_regstack ();
+void pushreg ();
+void execute ();
+void fputchar ();
+void push ();
+void incref ();
+void decref ();
+void binop ();
+
+main (argc, argv, env)
+     int argc;
+     char **argv, **env;
+{
+
+  ibase = 10;
+  obase = 10;
+  precision = 0;
+
+  freeregstacks = 0;
+
+  bzero (regs, sizeof regs);
+  bzero (regstacks, sizeof regstacks);
+
+  bufsize = 40;
+  buffer = (char *) xmalloc (40);
+
+  stacksize = 40;
+  stack = (decimal *) xmalloc (stacksize * sizeof (decimal));
+  stacktop = -1;
+
+  macrostacksize = 40;
+  macrostack = (decimal *) xmalloc (macrostacksize * sizeof (decimal));
+  macroindex = (int *) xmalloc (macrostacksize * sizeof (int));
+  macrolevel = 0;
+  /* Initialize for reading input files if any */
+
+  open_file = 0;
+
+  file_count = argc - 1;
+  next_file = argv + 1;
+
+
+  while (1)
+    {
+      execute ();
+    }
+}
+
+/* Read and execute one command from the current source of input */
+
+void
+execute ()
+{
+  int c = fetch ();
+
+  if (c < 0) exit (0);
+
+    {
+      switch (c)
+	{
+	case '+':		/* Arithmetic operators... */
+	  binop (decimal_add);
+	  break;
+
+	case '-':
+	  binop (decimal_sub);
+	  break;
+
+	case '*':
+	  binop (decimal_mul_dc);  /* Like decimal_mul but hairy
+				      way of deciding precision to keep */
+	  break;
+
+	case '/':
+	  binop (decimal_div);
+	  break;
+
+	case '%':
+	  binop (decimal_rem);
+	  break;
+
+	case '^':
+	  binop (decimal_expt);
+	  break;
+
+	case '_':		/* Begin a negative decimal constant */
+	  {
+	    decimal tem = dec_read (stdin);
+	    tem->sign = !tem->sign;
+	    push (tem);
+	  }
+	  break;
+
+	case '.':
+	case '0':
+	case '1':
+	case '2':
+	case '3':
+	case '4':
+	case '5':
+	case '6':
+	case '7':
+	case '8':
+	case '9':		/* All these begin decimal constants */
+	  unfetch (c);
+	  push (dec_read (stdin));
+	  break;
+
+	case 'A':
+	case 'B':
+	case 'C':
+	case 'D':
+	case 'E':
+	case 'F':
+	  unfetch (c);
+	  push (dec_read (stdin));
+	  break;
+
+	case 'c':		/* Clear the stack */
+	  while (stacktop >= 0)
+	    decref (stack[stacktop--]);
+	  break;
+
+	case 'd':		/* Duplicate top of stack */
+	  if (stacktop < 0)
+	    error ("stack empty", 0);
+	  else push (stack[stacktop]);
+	  break;
+
+	case 'f':		/* Describe all registers and stack contents */
+	  {
+	    int regno;
+	    int somereg = 0;	/* set to 1 if we print any registers */
+	    for (regno = 0; regno < 128; regno++)
+	      {
+		if (regs[regno])
+		  {
+		    printf ("register %c: ", regno);
+		    print_obj (regs[regno]);
+		    somereg = 1;
+		    printf ("\n");
+		  }
+	      }
+	    if (somereg)
+	      printf ("\n");
+	    if (stacktop < 0)
+	      printf ("stack empty\n");
+	    else
+	      {
+		int i;
+		printf ("stack:\n");
+		for (i = 0; i <= stacktop; i++)
+		  {
+		    print_obj (stack[stacktop - i]);
+		    printf ("\n");
+		  }
+	      }
+	  }
+	  break;
+
+	case 'i':		/* ibase <- top of stack */
+	  popstack (setibase);
+	  break;
+
+	case 'I':		/* Push current ibase */
+	  push (decimal_from_int (ibase));
+	  break;
+
+	case 'k':		/* like i, I but for precision instead of ibase */
+	  popstack (setprecision);
+	  break;
+
+	case 'K':
+	  push (decimal_from_int (precision));
+	  break;
+
+	case 'l':		/* l<x> load register <x> onto stack */
+	  {
+	    char c1 = fetch ();
+	    if (c1 < 0) exit (0);
+	    if (!regs[c1])
+	      error ("register %c empty", c1);
+	    else
+	      push (regs[c1]);
+	  }
+	  break;
+
+	case 'L':		/* L<x> load register <x> to stack, pop <x>'s own stack */
+	  {
+	    char c1 = fetch ();
+	    if (c1 < 0) exit (0);
+	    if (!regstacks[c1])
+	      error ("nothing pushed on register %c", c1);
+	    else
+	      {
+		regstack r = regstacks[c1];
+		if (!regs[c1])
+		  error ("register %c empty after pop", c1);
+		else
+		  push (regs[c1]);
+		regs[c1] = r->value;
+		regstacks[c1] = r->rest;
+		free_regstack (r);
+	      }
+	  }
+	  break;
+
+	case 'o':		/* o, O like i, I but for obase instead of ibase */
+	  popstack (setobase);
+	  break;
+
+	case 'O':
+	  push (decimal_from_int (obase));
+	  break;
+
+	case 'p':		/* Print tos, don't pop, do print newline afterward */
+	  if (stacktop < 0)
+	    error ("stack empty", 0);
+	  else
+	    {
+	      print_obj (stack[stacktop]);
+	      printf ("\n");
+	    }
+	  break;
+
+	case 'P':		/* Print tos, do pop, no newline afterward */
+	  popstack (print_obj);
+	  break;
+
+	case 'q':		/* Exit */
+	  if (macrolevel)
+	    { popmacro (); popmacro (); }  /* decrease recursion level by 2 */
+	  else
+	    exit (0);		/* If not in a macro, exit the program.  */
+
+	  break;
+
+	case 'Q':		/* Tos says how many levels to exit */
+	  popstack (popmacros);
+	  break;
+
+	case 's':		/* s<x> -- Pop stack and set register <x> */
+	  if (stacktop < 0)
+	    empty ();
+	  else
+	    {
+	      int c1 = fetch ();
+	      if (c1 < 0) exit (0);
+	      if (regs[c1]) decref (regs[c1]);
+	      regs[c1] = stack[stacktop--];
+	    }
+	  break;
+
+	case 'S':		/* S<x> -- pop stack and push as new value of register <x> */
+	  if (stacktop < 0)
+	    empty ();
+	  else
+	    {
+	      int c1 = fetch ();
+	      if (c1 < 0) exit (0);
+	      pushreg (c1);
+	      regs[c1] = stack[stacktop--];
+	    }
+	  break;
+
+	case 'v':		/* tos gets square root of tos */
+	  popstack (pushsqrt);
+	  break;
+
+	case 'x':		/* pop stack , call as macro */
+	  popstack (pushmacro);
+	  break;
+
+	case 'X':		/* Pop stack, get # fraction digits, push that */
+	  popstack (pushscale);
+	  break;
+
+	case 'z':		/* Compute depth of stack, push that */
+	  push (decimal_from_int (stacktop + 1));
+	  break;
+
+	case 'Z':		/* Pop stack, get # digits, push that */
+	  popstack (pushlength);
+	  break;
+
+	case '<':		/* Conditional: pop two numbers, compare, maybe execute register */
+	  /* Note: for no obvious reason, the standard Unix `dc'
+	     considers < to be true if the top of stack is less
+	     than the next-to-top of stack,
+	     and vice versa for >.
+	     This seems backwards to me, but I am preserving compatibility.  */
+	  condop (1);
+	  break;
+
+	case '>':
+	  condop (-1);
+	  break;
+
+	case '=':
+	  condop (0);
+	  break;
+
+	case '?':		/* Read expression from terminal and execute it */
+	  /* First ignore any leading newlines */
+	  {
+	    int c1;
+	    while ((c1 = getchar ()) == '\n');
+	    ungetc (c1, stdin);
+	  }
+	  /* Read a line from the terminal and execute it.  */
+	  pushmacro (read_string ('\n', fgetchar, 0));
+	  break;
+
+	case '[':		/* Begin string constant */
+	  push (read_string (']', fetch, '['));
+	  break;
+
+        case ' ':
+	case '\n':
+	  break;
+
+	default:
+	  error ("undefined command %c", c);
+	}
+    }
+}
+
+/* Functionals for performing arithmetic, etc */
+
+/* Call the function `op', with the top of stack value as argument,
+ and then pop the stack.
+ If the stack is empty, print a message and do not call `op'.  */
+
+void
+popstack (op)
+     void (*op) ();
+{
+  if (stacktop < 0)
+    empty ();
+  else
+    {
+      decimal value = stack[stacktop--];
+      op (value);
+      decref (value);
+    }
+}
+
+/* Call the function `op' with two arguments taken from the stack top,
+ then pop those arguments and push the value returned by `op'.
+ `op' is assumed to return a decimal number.
+ If there are not two values on the stack, print a message
+ and do not call `op'.  */
+
+void
+binop (op)
+     decimal (*op) ();
+{
+  if (stacktop < 1)
+    error ("stack empty", 0);
+  else if (stack[stacktop]->after == STRING || stack[stacktop - 1]->after == STRING)
+    error ("operands not both numeric");
+  else
+    {
+      decimal arg2 = stack [stacktop--];
+      decimal arg1 = stack [stacktop--];
+
+      push (op (arg1, arg2, precision));
+
+      decref (arg1);
+      decref (arg2);
+    }
+}
+
+void
+condop (cond)
+     int cond;
+{
+  int regno = fetch ();
+  if (!regs[regno])
+    error ("register %c is empty", regno);
+  else if (stacktop < 1)
+    empty ();
+  else
+    {
+      decimal arg2 = stack[stacktop--];
+      decimal arg1 = stack[stacktop--];
+      int relation = decimal_compare (arg1, arg2);
+      decref (arg1);
+      decref (arg2);
+      if (cond == relation
+	  || (cond < 0 && relation < 0)
+	  || (cond > 0 && relation > 0))
+	pushmacro (regs[regno]);
+    }
+}
+
+/* Handle the command input source */
+
+/* Fetch the next command character from a macro or from the terminal */
+
+int
+fetch()
+{
+  int c = -1;
+
+  while (macrolevel &&
+		LENGTH (macrostack[macrolevel-1]) == macroindex[macrolevel-1])
+    popmacro();
+  if (macrolevel)
+    return macrostack[macrolevel - 1]->contents[macroindex[macrolevel-1]++];
+  while (1)
+    {
+      if (open_file)
+	{
+	  c = getc (open_file);
+	  if (c >= 0) break;
+	  fclose (open_file);
+	  open_file = 0;
+	}
+      else if (file_count)
+	{
+	  open_file = fopen (*next_file++, "r");
+	  file_count--;
+	  if (!open_file)
+	    perror_with_name (*(next_file - 1));
+	}
+      else break;
+    }
+  if (c >= 0) return c;
+  return getc (stdin);
+}
+
+/* Unread character c on command input stream, whatever it is */
+
+void
+unfetch (c)
+     char c;
+{
+  if (macrolevel)
+    macroindex[macrolevel-1]--;
+  else if (open_file)
+    ungetc (c, open_file);
+  else
+    ungetc (c, stdin);
+}
+
+/* Begin execution of macro m.  */
+
+void
+pushmacro (m)
+     decimal m;
+{
+  while (macrolevel &&
+		LENGTH (macrostack[macrolevel-1]) == macroindex[macrolevel-1])
+    popmacro();
+  if (m->after == STRING)
+    {
+      if (macrolevel == macrostacksize)
+	{
+	  macrostacksize *= 2;
+	  macrostack = (decimal *) xrealloc (macrostack, macrostacksize * sizeof (decimal));
+	  macroindex = (int *) xrealloc (macroindex, macrostacksize * sizeof (int));
+	}
+      macroindex[macrolevel] = 0;
+      macrostack[macrolevel++] = m;
+      incref (m);
+    }
+  else
+    {   /* Number supplied as a macro!  */
+      push (m);   /* Its effect wouyld be to push the number.  */
+    }
+}
+
+/* Pop a specified number of levels of macro execution.
+ The number of levels is specified by a decimal number d.  */
+
+void
+popmacros (d)
+     decimal d;
+{
+  int num_pops = decimal_to_int (d);
+  int i;
+  for (i = 0; i < num_pops; i++)
+    popmacro ();
+}
+/* Exit one level of macro execution.  */
+
+void
+popmacro ()
+{
+  if (!macrolevel)
+    exit (0);
+  else
+    {
+      decref (macrostack[--macrolevel]);
+    }
+}
+
+void
+push (d)
+     decimal d;
+{
+  if (stacktop == stacksize - 1)
+    stack = (decimal *) xrealloc (stack, (stacksize *= 2) * sizeof (decimal));
+
+    incref (d);
+
+    stack[++stacktop] = d;
+}
+
+/* Reference counting and storage freeing */
+
+void
+decref (d)
+     decimal d;
+{
+  if (!--d->refcnt)
+    free (d);
+}
+
+void
+incref (d)
+     decimal d;
+{
+  d->refcnt++;
+}
+
+empty ()
+{
+  error ("stack empty", 0);
+}
+
+regstack
+get_regstack ()
+{
+  if (freeregstacks)
+    {
+      regstack r = freeregstacks;
+      freeregstacks = r ->rest;
+      return r;
+    }
+  else
+    return (regstack) xmalloc (sizeof (struct regstack));
+}
+
+void
+free_regstack (r)
+     regstack r;
+{
+  r->rest = freeregstacks;
+  freeregstacks = r;
+}
+
+void
+pushreg (c)
+     char c;
+{
+  regstack r = get_regstack ();
+
+  r->rest = regstacks[c];
+  r->value = regs[c];
+  regstacks[c] = r;
+  regs[c] = 0;
+}
+
+/* Input of numbers and strings */
+
+/* Return a character read from the terminal.  */
+
+fgetchar ()
+{
+  return getchar ();
+}
+
+void
+fputchar (c)
+     char (c);
+{
+  putchar (c);
+}
+
+/* Read text from command input source up to a close-bracket,
+   make a string out of it, and return it.
+   If STARTC is nonzero, then it and STOPC must balance when nested.  */
+
+decimal
+read_string (stopc, inputfn, startc)
+     char stopc;
+     int (*inputfn) ();
+     int startc;
+{
+  int c;
+  decimal result;
+  int i = 0;
+  int count = 0;
+
+  while (1)
+    {
+      c = inputfn ();
+      if (c < 0 || (c == stopc && count == 0))
+	{
+	  if (count != 0)
+	    error ("Unmatched `%c'", startc);
+	  break;
+	}
+      if (c == stopc)
+	count--;
+      if (c == startc)
+	count++;
+      if (i + 1 >= bufsize)
+	buffer = (char *) xrealloc (buffer, bufsize *= 2);
+      buffer[i++] = c;
+    }
+  result = make_decimal (i, 0);
+  result->after = -1;	/* Mark it as a string */
+  result->before++;	/* but keep the length unchanged */
+  bcopy (buffer, result->contents, i);
+  return result;
+}
+
+/* Read a number from the current input source */
+
+decimal
+dec_read ()
+{
+  int c;
+  int i = 0;
+
+  while (1)
+    {
+      c = fetch ();
+      if (! ((c >= '0' && c <= '9')
+	     || (c >= 'A' && c <= 'F')
+	     || c == '.'))
+        break;
+      if (i + 1 >= bufsize)
+	buffer = (char *) xrealloc (buffer, bufsize *= 2);
+      buffer[i++] = c;
+    }
+  buffer[i++] = 0;
+  unfetch (c);
+
+  return decimal_parse (buffer, ibase);
+}
+
+/* Output of numbers and strings */
+
+/* Print the contents of obj, either numerically or as a string,
+ according to what obj says it is.  */
+
+void
+print_obj (obj)
+     decimal obj;
+{
+  if (obj->after == STRING)
+    print_string (obj);
+  else
+    decimal_print (obj, fputchar, obase);
+}
+
+/* Print the contents of the decimal number `string', treated as a string.  */
+
+void
+print_string (string)
+     decimal string;
+{
+  char *p = string->contents;
+  int len = LENGTH (string);
+  int i;
+
+  for (i = 0; i < len; i++)
+    {
+      putchar (*p++);
+    }
+}
+
+/* Set the input radix from the value of the decimal number d, if valid.  */
+
+void
+setibase (d)
+     decimal d;
+{
+  int value = decimal_to_int (d);
+  if (value < 2 || value > 36)
+    error ("input radix must be from 2 to 36", 0);
+  else
+    ibase = value;
+}
+
+/* Set the output radix from the value of the decimal number d, if valid.  */
+
+void
+setobase (d)
+     decimal d;
+{
+  int value = decimal_to_int (d);
+  if (value < 2 || value > 36)
+    error ("output radix must be from 2 to 36", 0);
+  else
+    obase = value;
+}
+
+/* Set the precision for mul and div from the value of the decimal number d, if valid.  */
+
+void
+setprecision (d)
+     decimal d;
+{
+  int value = decimal_to_int (d);
+  if (value < 0 || value > 30000)
+    error ("precision must be nonnegative and < 30000", 0);
+  else
+    precision = value;
+}
+
+/* Push the number of digits in decimal number d, as a decimal number.  */
+
+void
+pushlength (d)
+     decimal d;
+{
+  push (decimal_from_int (LENGTH (d)));
+}
+
+/* Push the number of fraction digits in d.  */
+
+void
+pushscale (d)
+     decimal d;
+{
+  push (decimal_from_int (d->after));
+}
+
+/* Push the square root of decimal number d.  */
+
+void
+pushsqrt (d)
+     decimal d;
+{
+  push (decimal_sqrt (d, precision));
+}
+
+/* Print error message and exit.  */
+
+fatal (s1, s2)
+     char *s1, *s2;
+{
+  error (s1, s2);
+  exit (1);
+}
+
+/* Print error message.  `s1' is printf control string, `s2' is arg for it. */
+
+error (s1, s2)
+     char *s1, *s2;
+{
+  printf ("dc: ");
+  printf (s1, s2);
+  printf ("\n");
+}
+
+decimal_error (s1, s2)
+     char *s1, *s2;
+{
+  error (s1, s2);
+}
+
+perror_with_name (name)
+     char *name;
+{
+  extern int errno, sys_nerr;
+  extern char *sys_errlist[];
+  char *s;
+
+  if (errno < sys_nerr)
+    s = concat ("", sys_errlist[errno], " for %s");
+  else
+    s = "cannot open %s";
+  error (s, name);
+}
+
+/* Return a newly-allocated string whose contents concatenate those of s1, s2, s3.  */
+
+char *
+concat (s1, s2, s3)
+     char *s1, *s2, *s3;
+{
+  int len1 = strlen (s1), len2 = strlen (s2), len3 = strlen (s3);
+  char *result = (char *) xmalloc (len1 + len2 + len3 + 1);
+
+  strcpy (result, s1);
+  strcpy (result + len1, s2);
+  strcpy (result + len1 + len2, s3);
+  *(result + len1 + len2 + len3) = 0;
+
+  return result;
+}
+
+/* Like malloc but get fatal error if memory is exhausted.  */
+
+int
+xmalloc (size)
+     int size;
+{
+  int result = malloc (size);
+  if (!result)
+    fatal ("virtual memory exhausted", 0);
+  return result;
+}
+
+int
+xrealloc (ptr, size)
+     char *ptr;
+     int size;
+{
+  int result = realloc (ptr, size);
+  if (!result)
+    fatal ("virtual memory exhausted");
+  return result;
+}
diff --git a/gnu/usr.bin/dc/dc.info b/gnu/usr.bin/dc/dc.info
new file mode 100644
index 000000000000..a30fea9f5555
--- /dev/null
+++ b/gnu/usr.bin/dc/dc.info
@@ -0,0 +1,330 @@
+This is Info file dc.info, produced by Makeinfo-1.52 from the input
+file dc.texinfo.
+
+   This file documents DC, an arbitrary precision calculator.
+
+   Published by the Free Software Foundation, 675 Massachusetts Avenue,
+Cambridge, MA 02139 USA
+
+   Copyright (C) 1984 Free Software Foundation, Inc.
+
+   Permission is granted to make and distribute verbatim copies of this
+manual provided the copyright notice and this permission notice are
+preserved on all copies.
+
+   Permission is granted to copy and distribute modified versions of
+this manual under the conditions for verbatim copying, provided that
+the entire resulting derived work is distributed under the terms of a
+permission notice identical to this one.
+
+   Permission is granted to copy and distribute translations of this
+manual into another language, under the above conditions for modified
+versions, except that this permission notice may be stated in a
+translation approved by the Foundation.
+
+
+File: dc.info,  Node: Top,  Next: Introduction,  Prev: (dir),  Up: (dir)
+
+* Menu:
+
+* Introduction::                Introduction
+* Printing Commands::           Printing Commands
+* Arithmetic::                  Arithmetic
+* Stack Control::               Stack Control
+* Registers::                   Registers
+* Parameters::                  Parameters
+* Strings::                     Strings
+* Status Inquiry::              Status Inquiry
+* Notes::                       Notes
+
+
+File: dc.info,  Node: Introduction,  Next: Printing Commands,  Prev: Top,  Up: Top
+
+Introduction
+************
+
+   DC is a reverse-polish desk calculator which supports unlimited
+precision arithmetic.  It also allows you to define and call macros.
+Normally DC reads from the standard input; if any command arguments are
+given to it, they are filenames, and DC reads and executes the contents
+of the files before reading from standard input.  All output is to
+standard output.
+
+   To exit, use `q'.  `C-c' does not exit; it is used to abort macros
+that are looping, etc.  (Currently this is not true; `C-c' does exit.)
+
+   A reverse-polish calculator stores numbers on a stack.  Entering a
+number pushes it on the stack.  Arithmetic operations pop arguments off
+the stack and push the results.
+
+   To enter a number in DC, type the digits, with an optional decimal
+point.  Exponential notation is not supported.  To enter a negative
+number, begin the number with `_'.  `-' cannot be used for this, as it
+is a binary operator for subtraction instead.  To enter two numbers in
+succession, separate them with spaces or newlines.  These have no
+meaning as commands.
+
+
+File: dc.info,  Node: Printing Commands,  Next: Arithmetic,  Prev: Introduction,  Up: Top
+
+Printing Commands
+*****************
+
+`p'
+     Prints the value on the top of the stack, without altering the
+     stack.  A newline is printed after the value.
+
+`P'
+     Prints the value on the top of the stack, popping it off, and does
+     not print a newline after.
+
+`f'
+     Prints the entire contents of the stack and the contents of all of
+     the registers, without altering anything.  This is a good command
+     to use if you are lost or want to figure out what the effect of
+     some command has been.
+
+
+File: dc.info,  Node: Arithmetic,  Next: Stack Control,  Prev: Printing Commands,  Up: Top
+
+Arithmetic
+**********
+
+`+'
+     Pops two values off the stack, adds them, and pushes the result.
+     The precision of the result is determined only by the values of
+     the arguments, and is enough to be exact.
+
+`-'
+     Pops two values, subtracts the first one popped from the second
+     one popped, and pushes the result.
+
+`*'
+     Pops two values, multiplies them, and pushes the result.  The
+     number of fraction digits in the result is controlled by the
+     current precision flag (see below) and does not depend on the
+     values being multiplied.
+
+`/'
+     Pops two values, divides the second one popped from the first one
+     popped, and pushes the result.  The number of fraction digits is
+     specified by the precision flag.
+
+`%'
+     Pops two values, computes the remainder of the division that the
+     `/' command would do, and pushes that.  The division is done with
+     as many fraction digits as the precision flag specifies, and the
+     remainder is also computed with that many fraction digits.
+
+`^'
+     Pops two values and exponentiates, using the first value popped as
+     the exponent and the second popped as the base.  The fraction part
+     of the exponent is ignored.  The precision flag specifies the
+     number of fraction digits in the result.
+
+`v'
+     Pops one value, computes its square root, and pushes that.  The
+     precision flag specifies the number of fraction digits in the
+     result.
+
+   Most arithmetic operations are affected by the "precision flag",
+which you can set with the `k' command.  The default precision value is
+zero, which means that all arithmetic except for addition and
+subtraction produces integer results.
+
+   The remainder operation (`%') requires some explanation: applied to
+arguments `a' and `b' it produces `a - (b * (a / b))', where `a / b' is
+computed in the current precision.
+
+
+File: dc.info,  Node: Stack Control,  Next: Registers,  Prev: Arithmetic,  Up: Top
+
+Stack Control
+*************
+
+`c'
+     Clears the stack, rendering it empty.
+
+`d'
+     Duplicates the value on the top of the stack, pushing another copy
+     of it.  Thus, `4d*p' computes 4 squared and prints it.
+
+
+File: dc.info,  Node: Registers,  Next: Parameters,  Prev: Stack Control,  Up: Top
+
+Registers
+*********
+
+   DC provides 128 memory registers, each named by a single ASCII
+character.  You can store a number in a register and retrieve it later.
+
+`sR'
+     Pop the value off the top of the stack and store it into register
+     R.
+
+`lR'
+     Copy the value in register R, and push it onto the stack.  This
+     does not alter the contents of R.
+
+     Each register also contains its own stack.  The current register
+     value is the top of the register's stack.
+
+`SR'
+     Pop the value off the top of the (main) stack and push it onto the
+     stack of register R.  The previous value of the register becomes
+     inaccessible.
+
+`LR'
+     Pop the value off the top of register R's stack and push it onto
+     the main stack.  The previous value in register R's stack, if any,
+     is now accessible via the `lR' command.
+
+   The `f' command prints a list of all registers that have contents
+stored in them, together with their contents.  Only the current
+contents of each register (the top of its stack) is printed.
+
+
+File: dc.info,  Node: Parameters,  Next: Strings,  Prev: Registers,  Up: Top
+
+Parameters
+**********
+
+   DC has three parameters that control its operation: the precision,
+the input radix, and the output radix.  The precision specifies the
+number of fraction digits to keep in the result of most arithmetic
+operations.  The input radix controls the interpretation of numbers
+typed in; *all* numbers typed in use this radix.  The output radix is
+used for printing numbers.
+
+   The input and output radices are separate parameters; you can make
+them unequal, which can be useful or confusing.  Each radix must be
+between 2 and 36 inclusive.  The precision must be zero or greater.
+The precision is always measured in decimal digits, regardless of the
+current input or output radix.
+
+`i'
+     Pops the value off the top of the stack and uses it to set the
+     input radix.
+
+`o'
+`k'
+     Similarly set the output radix and the precision.
+
+`I'
+     Pushes the current input radix on the stack.
+
+`O'
+`K'
+     Similarly push the current output radix and the current precision.
+
+
+File: dc.info,  Node: Strings,  Next: Status Inquiry,  Prev: Parameters,  Up: Top
+
+Strings
+*******
+
+   DC can operate on strings as well as on numbers.  The only things you
+can do with strings are print them and execute them as macros (which
+means that the contents of the string are processed as DC commands).
+Both registers and the stack can hold strings, and DC always knows
+whether any given object is a string or a number.  Some commands such as
+arithmetic operations demand numbers as arguments and print errors if
+given strings.  Other commands can accept either a number or a string;
+for example, the `p' command can accept either and prints the object
+according to its type.
+
+`[CHARACTERS]'
+     Makes a string containing CHARACTERS and pushes it on the stack.
+     For example, `[foo]P' prints the characters `foo' (with no
+     newline).
+
+`x'
+     Pops a value off the stack and executes it as a macro.  Normally
+     it should be a string; if it is a number, it is simply pushed back
+     onto the stack.  For example, `[1p]x' executes the macro `1p',
+     which pushes 1 on the stack and prints `1' on a separate line.
+
+     Macros are most often stored in registers; `[1p]sa' stores a macro
+     to print `1' into register `a', and `lax' invokes the macro.
+
+`>R'
+     Pops two values off the stack and compares them assuming they are
+     numbers, executing the contents of register R as a macro if the
+     original top-of-stack is greater.  Thus, `1 2>a' will invoke
+     register `a''s contents and `2 1>a' will not.
+
+`<R'
+     Similar but invokes the macro if the original top-of-stack is less.
+
+`=R'
+     Similar but invokes the macro if the two numbers popped are equal.
+     This can also be validly used to compare two strings for equality.
+
+`?'
+     Reads a line from the terminal and executes it.  This command
+     allows a macro to request input from the user.
+
+`q'
+     During the execution of a macro, this comand does not exit DC.
+     Instead, it exits from that macro and also from the macro which
+     invoked it (if any).
+
+`Q'
+     Pops a value off the stack and uses it as a count of levels of
+     macro execution to be exited.  Thus, `3Q' exits three levels.
+
+
+File: dc.info,  Node: Status Inquiry,  Next: Notes,  Prev: Strings,  Up: Top
+
+Status Inquiry
+**************
+
+`Z'
+     Pops a value off the stack, calculates the number of digits it has
+     (or number of characters, if it is a string) and pushes that
+     number.
+
+`X'
+     Pops a value off the stack, calculates the number of fraction
+     digits it has, and pushes that number.  For a string, the value
+     pushed is -1.
+
+`z'
+     Pushes the current stack depth; the number of objects on the stack
+     before the execution of the `z' command.
+
+`I'
+     Pushes the current value of the input radix.
+
+`O'
+     Pushes the current value of the output radix.
+
+`K'
+     Pushes the current value of the precision.
+
+
+File: dc.info,  Node: Notes,  Prev: Status Inquiry,  Up: Top
+
+Notes
+*****
+
+   The `:' and `;' commands of the Unix DC program are not supported,
+as the documentation does not say what they do.  The `!' command is not
+supported, but will be supported as soon as a library for executing a
+line as a command exists.
+
+
+
+Tag Table:
+Node: Top960
+Node: Introduction1440
+Node: Printing Commands2603
+Node: Arithmetic3211
+Node: Stack Control5168
+Node: Registers5468
+Node: Parameters6586
+Node: Strings7659
+Node: Status Inquiry9857
+Node: Notes10571
+
+End Tag Table
diff --git a/gnu/usr.bin/dc/dc.texinfo b/gnu/usr.bin/dc/dc.texinfo
new file mode 100644
index 000000000000..15b285fbd579
--- /dev/null
+++ b/gnu/usr.bin/dc/dc.texinfo
@@ -0,0 +1,381 @@
+\input texinfo  @c -*-texinfo-*-
+@c %**start of header
+@setfilename dc.info
+@settitle DC, An Arbitrary Precision Calculator
+@c %**end of header
+
+@c This file has the new style title page commands.
+@c Run `makeinfo' rather than `texinfo-format-buffer'.
+
+@c smallbook
+
+@c tex
+@c \overfullrule=0pt
+@c end tex
+
+@c Combine indices.
+@synindex cp fn
+@syncodeindex vr fn
+@syncodeindex ky fn
+@syncodeindex pg fn
+@syncodeindex tp fn
+
+@ifinfo
+This file documents DC, an arbitrary precision calculator.
+
+Published by the Free Software Foundation,
+675 Massachusetts Avenue,
+Cambridge, MA 02139 USA
+
+Copyright (C) 1984 Free Software Foundation, Inc.
+
+Permission is granted to make and distribute verbatim copies of
+this manual provided the copyright notice and this permission notice
+are preserved on all copies.
+
+@ignore
+Permission is granted to process this file through TeX and print the
+results, provided the printed document carries copying permission
+notice identical to this one except for the removal of this paragraph
+(this paragraph not being relevant to the printed manual).
+
+@end ignore
+Permission is granted to copy and distribute modified versions of this
+manual under the conditions for verbatim copying, provided that the entire
+resulting derived work is distributed under the terms of a permission
+notice identical to this one.
+
+Permission is granted to copy and distribute translations of this manual
+into another language, under the above conditions for modified versions,
+except that this permission notice may be stated in a translation approved
+by the Foundation.
+@end ifinfo
+
+@setchapternewpage odd
+
+@titlepage
+@title DC, An Arbitrary Precision Calculator
+
+@author by Richard Stallman
+@page
+@vskip 0pt plus 1filll
+Copyright @copyright{} 1984 Free Software Foundation, Inc.
+
+@sp 2
+Published by the Free Software Foundation, @*
+675 Massachusetts Avenue, @*
+Cambridge, MA 02139 USA
+
+Permission is granted to make and distribute verbatim copies of
+this manual provided the copyright notice and this permission notice
+are preserved on all copies.
+
+Permission is granted to copy and distribute modified versions of this
+manual under the conditions for verbatim copying, provided that the entire
+resulting derived work is distributed under the terms of a permission
+notice identical to this one.
+
+Permission is granted to copy and distribute translations of this manual
+into another language, under the above conditions for modified versions,
+except that this permission notice may be stated in a translation approved
+by the Foundation.
+
+@end titlepage
+@page
+
+@node Top, Introduction, (dir), (dir)
+
+@menu
+* Introduction::                Introduction
+* Printing Commands::           Printing Commands
+* Arithmetic::                  Arithmetic
+* Stack Control::               Stack Control
+* Registers::                   Registers
+* Parameters::                  Parameters
+* Strings::                     Strings
+* Status Inquiry::              Status Inquiry
+* Notes::                       Notes
+@end menu
+
+@node Introduction, Printing Commands, Top, Top
+@comment  node-name,  next,  previous,  up
+@chapter Introduction
+
+DC is a reverse-polish desk calculator which supports unlimited
+precision arithmetic.  It also allows you to define and call macros.
+Normally DC reads from the standard input; if any command arguments
+are given to it, they are filenames, and DC reads and executes the
+contents of the files before reading from standard input.  All output
+is to standard output.
+
+To exit, use @samp{q}.  @kbd{C-c} does not exit; it is used to abort
+macros that are looping, etc.  (Currently this is not true; @kbd{C-c}
+does exit.)
+
+A reverse-polish calculator stores numbers on a stack.  Entering a
+number pushes it on the stack.  Arithmetic operations pop arguments off
+the stack and push the results.
+
+To enter a number in DC, type the digits, with an optional decimal
+point.  Exponential notation is not supported.  To enter a negative
+number, begin the number with @samp{_}.  @samp{-} cannot be used for
+this, as it is a binary operator for subtraction instead.
+To enter two numbers in succession, separate them with spaces or
+newlines.  These have no meaning as commands.
+
+@node Printing Commands, Arithmetic, Introduction, Top
+@chapter Printing Commands
+
+@table @samp
+@item p
+Prints the value on the top of the stack,
+without altering the stack.  A newline is printed
+after the value.
+
+@item P
+Prints the value on the top of the stack,
+popping it off, and does not print a newline after.
+
+@item f
+Prints the entire contents of the stack
+and the contents of all of the registers,
+without altering anything.  This is a good command
+to use if you are lost or want to figure out
+what the effect of some command has been.
+@end table
+
+@node Arithmetic, Stack Control, Printing Commands, Top
+@chapter Arithmetic
+
+@table @samp
+@item +
+Pops two values off the stack, adds them,
+and pushes the result.  The precision of the result
+is determined only by the values of the arguments,
+and is enough to be exact.
+
+@item -
+Pops two values, subtracts the first one popped
+from the second one popped, and pushes the result.
+
+@item *
+Pops two values, multiplies them, and pushes the result.
+The number of fraction digits in the result is controlled
+by the current precision flag (see below) and does not
+depend on the values being multiplied.
+
+@item /
+Pops two values, divides the second one popped from
+the first one popped, and pushes the result.
+The number of fraction digits is specified by the precision flag.
+
+@item %
+Pops two values, computes the remainder of the division
+that the @samp{/} command would do, and pushes that.
+The division is done with as many fraction digits
+as the precision flag specifies, and the remainder
+is also computed with that many fraction digits.
+
+@item ^
+Pops two values and exponentiates, using the first
+value popped as the exponent and the second popped as the base.
+The fraction part of the exponent is ignored.
+The precision flag specifies the number of fraction
+digits in the result.
+
+@item v
+Pops one value, computes its square root, and pushes that.
+The precision flag specifies the number of fraction digits
+in the result.
+@end table
+
+Most arithmetic operations are affected by the "precision flag",
+which you can set with the @samp{k} command.  The default precision
+value is zero, which means that all arithmetic except for
+addition and subtraction produces integer results.
+
+The remainder operation (@samp{%}) requires some explanation: applied to
+arguments @samp{a} and @samp{b} it produces @samp{a - (b * (a / b))},
+where @samp{a / b} is computed in the current precision.
+
+@node Stack Control, Registers, Arithmetic, Top
+@chapter Stack Control
+
+@table @samp
+@item c
+Clears the stack, rendering it empty.
+
+@item d
+Duplicates the value on the top of the stack,
+pushing another copy of it.  Thus,
+`4d*p' computes 4 squared and prints it.
+@end table
+
+@node Registers, Parameters, Stack Control, Top
+@chapter Registers
+
+DC provides 128 memory registers, each named by a single
+ASCII character.  You can store a number in a register
+and retrieve it later.
+
+@table @samp
+@item s@var{r}
+Pop the value off the top of the stack and store
+it into register @var{r}.
+
+@item l@var{r}
+Copy the value in register @var{r}, and push it onto
+the stack.  This does not alter the contents of @var{r}.
+
+Each register also contains its own stack.  The current
+register value is the top of the register's stack.
+
+@item S@var{r}
+Pop the value off the top of the (main) stack and
+push it onto the stack of register @var{r}.
+The previous value of the register becomes inaccessible.
+
+@item L@var{r}
+Pop the value off the top of register @var{r}'s stack
+and push it onto the main stack.  The previous value
+in register @var{r}'s stack, if any, is now accessible
+via the `l@var{r}' command.
+@end table
+
+The @samp{f} command prints a list of all registers that have contents
+stored in them, together with their contents.  Only the
+current contents of each register (the top of its stack)
+is printed.
+
+@node Parameters, Strings, Registers, Top
+@chapter Parameters
+
+DC has three parameters that control its operation: the precision, the
+input radix, and the output radix.  The precision specifies the number
+of fraction digits to keep in the result of most arithmetic operations.
+The input radix controls the interpretation of numbers typed in;
+@emph{all} numbers typed in use this radix.  The output radix is used
+for printing numbers.
+
+The input and output radices are separate parameters; you can make them
+unequal, which can be useful or confusing.  Each radix must be between 2
+and 36 inclusive.  The precision must be zero or greater.  The precision
+is always measured in decimal digits, regardless of the current input or
+output radix.
+
+@table @samp
+@item i
+Pops the value off the top of the stack
+and uses it to set the input radix.
+
+@item o
+@itemx k
+Similarly set the output radix and the precision.
+
+@item I
+Pushes the current input radix on the stack.
+
+@item O
+@itemx K
+Similarly push the current output radix and the current precision.
+@end table
+
+@node Strings, Status Inquiry, Parameters, Top
+@chapter Strings
+
+DC can operate on strings as well as on numbers.  The only things you
+can do with strings are print them and execute them as macros (which
+means that the contents of the string are processed as DC commands).
+Both registers and the stack can hold strings, and DC always knows
+whether any given object is a string or a number.  Some commands such as
+arithmetic operations demand numbers as arguments and print errors if
+given strings.  Other commands can accept either a number or a string;
+for example, the @samp{p} command can accept either and prints the object
+according to its type.
+
+@table @samp
+@item [@var{characters}]
+Makes a string containing @var{characters} and pushes it
+on the stack.  For example, @samp{[foo]P} prints the
+characters @samp{foo} (with no newline).
+
+@item x
+Pops a value off the stack and executes it as a macro.
+Normally it should be a string; if it is a number,
+it is simply pushed back onto the stack.
+For example, @samp{[1p]x} executes the macro @samp{1p}, which
+pushes 1 on the stack and prints @samp{1} on a separate line.
+
+Macros are most often stored in registers;
+@samp{[1p]sa} stores a macro to print @samp{1} into register @samp{a},
+and @samp{lax} invokes the macro.
+
+@item >@var{r}
+Pops two values off the stack and compares them
+assuming they are numbers, executing the contents
+of register @var{r} as a macro if the original top-of-stack
+is greater.  Thus, @samp{1 2>a} will invoke register @samp{a}'s contents
+and @samp{2 1>a} will not.
+
+@item <@var{r}
+Similar but invokes the macro if the original top-of-stack
+is less.
+
+@item =@var{r}
+Similar but invokes the macro if the two numbers popped
+are equal.  This can also be validly used to compare two
+strings for equality.
+
+@item ?
+Reads a line from the terminal and executes it.
+This command allows a macro to request input from the user.
+
+@item q
+During the execution of a macro, this comand
+does not exit DC.  Instead, it exits from that
+macro and also from the macro which invoked it (if any).
+
+@item Q
+Pops a value off the stack and uses it as a count
+of levels of macro execution to be exited.  Thus,
+@samp{3Q} exits three levels.
+@end table
+
+@node Status Inquiry, Notes, Strings, Top
+@chapter Status Inquiry
+
+@table @samp
+@item Z
+Pops a value off the stack, calculates the number of
+digits it has (or number of characters, if it is a string)
+and pushes that number.
+
+@item X
+Pops a value off the stack, calculates the number of
+fraction digits it has, and pushes that number.
+For a string, the value pushed is -1.
+
+@item z
+Pushes the current stack depth; the number of
+objects on the stack before the execution of the @samp{z} command.
+
+@item I
+Pushes the current value of the input radix.
+
+@item O
+Pushes the current value of the output radix.
+
+@item K
+Pushes the current value of the precision.
+@end table
+
+@node Notes,  , Status Inquiry, Top
+@chapter Notes
+
+The @samp{:} and @samp{;} commands of the Unix DC program are
+not supported, as the documentation does not say what they do.
+The @samp{!} command is not supported, but will be supported
+as soon as a library for executing a line as a command exists.
+
+@contents
+@bye
diff --git a/gnu/usr.bin/dc/decimal.c b/gnu/usr.bin/dc/decimal.c
new file mode 100644
index 000000000000..1fb95c18aa14
--- /dev/null
+++ b/gnu/usr.bin/dc/decimal.c
@@ -0,0 +1,1235 @@
+/* 
+ * Arbitrary precision decimal arithmetic.
+ *
+ * Copyright (C) 1984 Free Software Foundation, Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, you can either send email to this
+ * program's author (see below) or write to: The Free Software Foundation,
+ * Inc.; 675 Mass Ave. Cambridge, MA 02139, USA.
+ */
+
+/* Some known problems:
+
+	Another problem with decimal_div is found when you try to
+	divide a number with > scale fraction digits by 1.  The
+	expected result is simply truncation, but all sorts of things
+	happen instead.  An example is that the result of .99999998/1
+	with scale set to 6 is .000001
+		
+	There are some problems in the behavior of the decimal package
+	related to printing and parsing.  The
+	printer is weird about very large output radices, tending to want
+	to output single ASCII characters for any and all digits (even
+	in radices > 127).  The UNIX bc approach is to print digit groups
+	separated by spaces.  There is a rather overwrought workaround in
+	the function decputc() in bcmisc.c, but it would be better if
+	decimal.c got a fix for this.  */
+
+/* For stand-alone testing, compile with -DTEST.
+   This DTESTable feature defines a `main' function
+   which is a simple loop that accepts input of the form
+   number space op space number newline
+   where op is +, -, *, /, %, p or r,
+   and performs the operation and prints the operands and result.
+   `p' means print the first number in the radix spec'd by the second.
+   `r' means read the first one in the radix specified by the second
+   (and print the result in decimal).
+   Divide in this test keeps three fraction digits. */
+
+#include "decimal.h"
+
+#define MAX(a, b) (((a) > (b) ? (a) : (b)))
+
+/* Some constant decimal numbers */
+
+struct decimal decimal_zero = {0, 0, 0, 0, 0};
+
+struct decimal decimal_one = {0, 0, 1, 0, 1};
+
+/*** Assumes RADIX is even ***/
+struct decimal decimal_half = {0, 1, 0, 0, RADIX / 2};
+
+decimal static decimal_add1 (), decimal_sub1 ();
+static void add_scaled ();
+static int subtract_scaled ();
+
+/* Create and return a decimal number that has `before' digits before
+   the decimal point and `after' digits after.  The digits themselves are
+   initialized to zero.  */
+
+decimal
+make_decimal (before, after)
+     int before, after;
+{
+  decimal result;
+  if (before >= 1<<16)
+    {
+      decimal_error ("%d too many decimal digits", before);
+      return 0;
+    }
+  if (after >= 1<<15)
+    {
+      decimal_error ("%d too many decimal digits", after);
+      return 0;
+    }
+  result = (decimal) malloc (sizeof (struct decimal) + before + after - 1);
+  result->sign = 0;
+  result->before = before;
+  result->after = after;
+  result->refcnt = 0;
+  bzero (result->contents, before + after);
+  return result;
+}
+
+/* Create a copy of the decimal number `b' and return it.  */
+
+decimal
+decimal_copy (b)
+     decimal b;
+{
+  decimal result = make_decimal (b->before, b->after);
+  bcopy (b->contents, result->contents, LENGTH(b));
+  result->sign = b->sign;
+  return result;
+}
+
+/* Copy a decimal number `b' but extend or truncate to exactly
+   `digits' fraction digits. */
+
+static decimal
+decimal_copy_1 (b, digits)
+     decimal b;
+     int digits;
+{
+  if (digits > b->after)
+    {
+      decimal result = make_decimal (b->before, digits);
+      bcopy (b->contents, result->contents + (digits - (int) b->after), LENGTH(b));
+      return result;
+    }
+  else
+    return decimal_round_digits (b, digits);
+}
+
+/* flush specified number `digits' of trailing fraction digits,
+   and flush any trailing fraction zero digits exposed after they are gone.
+   The number `b' is actually modified; no new storage is allocated.
+   That is why this is not global.  */
+
+static void
+flush_trailing_digits (b, digits)
+     decimal b;
+     int digits;
+{
+  int flush = digits;
+  int maxdig = b->after;
+
+  while (flush < maxdig && !b->contents [flush])
+    flush++;
+
+  if (flush)
+    {
+      int i;
+
+      b->after -= flush;
+      for (i = 0; i < LENGTH (b); i++)
+	b->contents[i] = b->contents[flush + i];
+    }
+
+}
+
+/* Return nonzero integer if the value of decimal number `b' is zero.  */
+
+int
+decimal_zerop (b)
+     decimal b;
+{
+  return !LENGTH(b);
+}
+
+/* Compare two decimal numbers arithmetically.
+   The value is < 0 if b1 < b2, > 0 if b1 > b2, 0 if b1 = b2.
+   This is the same way that `strcmp' reports the result of comparing
+   strings.  */ 
+
+int
+decimal_compare (b1, b2)
+     decimal b1, b2;
+{
+  int l1, l2;
+  char *p1, *p2, *s1, *s2;
+  int i;
+
+  /* If signs differ, deduce result from the signs */
+
+  if (b2->sign && !b1->sign) return 1;
+  if (b1->sign && !b2->sign) return -1;
+
+  /* If same sign but number of nonfraction digits differs,
+     the one with more of them is farther from zero.  */
+
+  if (b1->before != b2->before)
+    if (b1->sign)
+      return (int) (b2->before - b1->before);
+    else
+      return (int) (b1->before - b2->before);
+
+  /* Else compare the numbers digit by digit from high end */
+  l1 = LENGTH(b1);
+  l2 = LENGTH(b2);  
+  s1 = b1->contents;		/* Start of number -- don't back up digit pointer past here */
+  s2 = b2->contents;
+  p1 = b1->contents + l1;	/* Scanning pointer, for fetching digits.  */
+  p2 = b2->contents + l2;
+  for (i = MAX(l1, l2); i >= 0; i--)
+    {
+      int r = ((p1 != s1) ? *--p1 : 0) - ((p2 != s2) ? *--p2 : 0);
+      if (r)
+	return b1->sign ? -r : r;
+    }
+  return 0;
+}
+
+/* Return the number of digits stored in decimal number `b' */
+
+int
+decimal_length (b)
+     decimal b;
+{
+  return LENGTH(b);
+}
+
+/* Return the number of fraction digits stored in decimal number `b'.  */
+
+int
+decimal_after (b)
+     decimal b;
+{
+  return b->after;
+}
+
+/* Round decimal number `b' to have only `digits' fraction digits.
+   Result is rounded to nearest unit in the last remaining digit.
+   Return the result, another decimal number.  */
+
+decimal
+decimal_round_digits (b, digits)
+     decimal b;
+     int digits;
+{
+  decimal result;
+  int old;
+
+  if (b->after <= digits) return decimal_copy (b);
+
+  if (digits < 0)
+    {
+      decimal_error ("request to keep negative number of digits %d", digits);
+      return decimal_copy (b);
+    }
+
+  result = make_decimal (b->before + 1, b->after);
+  result->sign = b->sign;
+  bcopy (b->contents, result->contents, LENGTH(b));
+
+  old = result->after;
+
+  /* Add .5 * last place to keep, so that we round rather than truncate */
+  /* Note this ignores sign of result, so if result is negative
+     it is subtracting */
+
+  add_scaled (result, DECIMAL_HALF, 1, old - digits - 1);
+
+  /* Flush desired digits, and any trailing zeros exposed by them.  */
+
+  flush_trailing_digits (result, old - digits);
+
+  /* Flush leading digits -- always is one, unless was a carry into it */
+
+  while (result->before > 0
+	 && result->contents[LENGTH(result) - 1] == 0)
+    result->before--;
+
+  return result;
+}
+
+/* Truncate decimal number `b' to have only `digits' fraction digits.
+   Any fraction digits in `b' beyond that are dropped and ignored.
+   Truncation is toward zero.
+   Return the result, another decimal number.  */
+
+decimal
+decimal_trunc_digits (b, digits)
+     decimal b;
+     int digits;
+{
+  decimal result = decimal_copy (b);
+  int old = result->after;
+
+  if (old <= digits) return result;
+
+  if (digits < 0)
+    {
+      decimal_error ("request to keep negative number of digits %d", digits);
+      return result;
+    }
+
+  flush_trailing_digits (result, old - digits);
+
+  return result;
+}
+
+/* Return the fractional part of decimal number `b':
+   that is, `b' - decimal_trunc_digits (`b') */
+
+decimal
+decimal_fraction (b)
+     decimal b;
+{
+  decimal result = make_decimal (0, b->after);
+  bcopy (b->contents, result->contents, b->after);
+  return result;
+}
+
+/* return an integer whose value is that of decimal `b', sans its fraction.  */
+
+int
+decimal_to_int (b)
+     decimal b;
+{
+  int result = 0;
+  int i;
+  int end = b->after;
+
+  for (i = LENGTH(b) - 1; i >= end; i--)
+    {
+      result *= RADIX;
+      result += b->contents[i];
+    }
+  return result;
+}
+
+/* return a decimal whose value is the integer i.  */
+
+decimal
+decimal_from_int (i)
+     int i;
+{
+  int log, tem;
+  decimal result;
+
+  for (log = 0, tem = (i > 0 ? i : - i); tem; log++, tem /= RADIX);
+
+  result = make_decimal (log, 0);
+
+  for (log = 0, tem = (i > 0 ? i : - i); tem; log++, tem /= RADIX)
+    result->contents[log] = tem % RADIX;
+
+  if (i < 0) result->sign = 1;
+  return result;
+}
+
+/* Return (as an integer) the result of dividing decimal number `b' by
+   integer `divisor'. 
+   This is used in printing decimal numbers in other radices. */
+
+int
+decimal_int_rem (b, divisor)
+     decimal b;
+     int divisor;
+{
+  int len = LENGTH(b);
+  int end = b->after;
+  int accum = 0;
+  int i;
+
+  for (i = len - 1; i >= end; i--)
+    {
+      accum %= divisor;
+      accum *= RADIX;
+      accum += b->contents[i];
+    }
+  return accum % divisor;
+}
+
+/* Convert digit `digit' to a character and output it by calling
+   `charout' with it as arg. */
+
+static void
+print_digit (digit, charout)
+     int digit;
+     void (*charout) ();
+{
+  if (digit < 10)
+    charout ('0' + digit);
+  else
+    charout ('A' + digit - 10);
+}
+
+/* print decimal number `b' in radix `radix', assuming it is an integer.
+   `r' is `radix' expressed as a decimal number. */
+
+static
+decimal_print_1 (b, r, radix, charout)
+     decimal b, r;
+     int radix;
+     void (*charout) ();
+{
+  int digit = decimal_int_rem (b, radix);
+  decimal rest = decimal_div (b, r, 0);
+
+  if (!decimal_zerop (rest))
+    decimal_print_1 (rest, r, radix, charout);
+
+  print_digit (digit, charout);
+
+  free (rest);
+}
+
+/* User entry: print decimal number `b' in radix `radix' (an integer),
+   outputting characters by calling `charout'.  */
+
+void
+decimal_print (b, charout, radix)
+     decimal b;
+     void (*charout) ();
+     int radix;
+{
+  if (b->sign) charout ('-');
+
+  if (radix == RADIX)
+    {
+      /* decimal output => just print the digits, inserting a point in
+	 the proper place.  */ 
+      int i;
+      int before = b->before;
+      int len = before + b->after;
+      for (i = 0; i < len; i++)
+	{
+	  if (i == before) charout ('.');
+	  /* Broken if RADIX /= 10
+	     charout ('0' + b->contents [len - 1 - i]); */
+	  print_digit (b->contents [len - 1 - i], charout);
+	}
+      if (!len)
+	charout ('0');
+    }
+  else
+    {
+      /* nonstandard radix: must use multiply and divide to determine the
+	 digits of the number in that radix.  */
+
+      int i;
+      extern double log10 ();
+      /* Compute the number of fraction digits we want to have in the
+         new radix.  They should contain the same amount of
+         information as the decimal digits we have.  */
+      int nfrac = (b->after / log10 ((double) radix) + .99);
+      decimal r = decimal_from_int (radix);
+      decimal intpart = decimal_trunc_digits (b, 0);
+
+      /* print integer part */
+      decimal_print_1 (intpart, r, radix, charout);
+      free (intpart);
+
+      /* print fraction part */
+      if (nfrac)
+	{
+          decimal tem1, tem2;
+	  tem1 = decimal_fraction (b);
+	  charout ('.');
+	  /* repeatedly multiply by `radix', print integer part as one digit,
+	     and flush the integer part.  */
+	  for (i = 0; i < nfrac; i++)
+	    {
+	      tem2 = decimal_mul (tem1, r);
+	      free (tem1);
+	      print_digit (decimal_to_int (tem2), charout);
+	      tem1 = decimal_fraction (tem2);
+	      free (tem2);
+	    }
+	  free (tem1);
+	}
+      free (r);
+    }
+}
+
+static int
+decode_digit (digitchar)
+     char digitchar;
+{
+  if ('0' <= digitchar && digitchar <= '9')
+    return digitchar - '0';
+  if ('a' <= digitchar && digitchar <= 'z')
+    return digitchar - 'a' + 10;
+  if ('A' <= digitchar && digitchar <= 'Z')
+    return digitchar - 'A' + 10;
+  return -1;
+}
+
+/* Parse string `s' into a number using radix `radix'
+   and return result as a decimal number.  */
+
+decimal
+decimal_parse (s, radix)
+     char *s;
+     int radix;
+{
+  int i, len, before = -1;
+  char *p;
+  char c;
+  decimal result;
+  int negative = 0;
+  int excess_digit = 0;
+
+  if (*s == '-')
+    {
+      s++;
+      negative = 1;
+    }
+
+  /* First scan for valid characters.
+     Count total num digits, and count num before the decimal point.  */
+
+  p = s;
+  i = 0;
+  while (c = *p++)
+    {
+      if (c == '.')
+        {
+	  if (before >= 0)
+	    decimal_error ("two decimal points in %s", s);
+          before = i;
+	}
+      else if (c == '0' && !i && before < 0)
+	s++;   /* Discard leading zeros */
+      else if (decode_digit (c) >= 0)
+	{
+	  i++;
+	  if (decode_digit (c) > RADIX)
+	    excess_digit = 1;
+	}
+      else
+	decimal_error ("invalid number %s", s);
+    }
+
+  len = i;
+  if (before < 0) before = i;
+
+  p = s;
+
+  /* Now parse those digits */
+
+  if (radix != RADIX || excess_digit)
+    {
+      decimal r = decimal_from_int (radix);
+      extern double log10 ();
+      int digits = (len - before) * log10 ((double) radix) + .99;
+      result = decimal_copy (DECIMAL_ZERO);
+
+      /* Parse all the digits into an integer, ignoring decimal point,
+	 by multiplying by `radix'.  */
+
+      while (i > 0 && (c = *p++))
+	{
+	  if (c != '.')
+	    {
+	      decimal newdig = decimal_from_int (decode_digit (c));
+	      decimal prod = decimal_mul (result, r);
+	      decimal newresult = decimal_add (newdig, prod);
+
+	      free (newdig);  free (prod);  free (result);
+	      result = newresult;
+	      i--;
+	    }
+	}
+
+      /* Now put decimal point in right place
+	 by dividing by `radix' once for each digit
+	 that really should have followed the decimal point.  */
+
+      for (i = before; i < len; i++)
+	{
+	  decimal newresult = decimal_div (result, r, digits);
+	  free (result);
+	  result = newresult;
+	}
+      free (r);
+    }
+  else
+    {
+      /* radix is standard - just copy the digits into a decimal number.  */
+
+      int tem;
+      result = make_decimal (before, len - before);
+
+      while (i > 0 && (c = *p++))
+	{
+	  if ((c != '.') &&
+	      ((tem = decode_digit (c)) >= 0))
+	    result->contents [--i] = tem;
+	}
+    }
+
+  if (negative) result->sign = 1;
+  flush_trailing_digits (result, 0);
+  return result;
+}
+
+/* Add b1 and b2, considering their signs */
+
+decimal
+decimal_add (b1, b2)
+     decimal b1, b2;
+{
+  decimal v;
+
+  if (b1->sign != b2->sign)
+    v = decimal_sub1 (b1, b2);
+  else
+    v = decimal_add1 (b1, b2);
+  if (b1->sign && !decimal_zerop (v))
+    v->sign = !v->sign;
+  return v;
+}
+
+/* Add b1 and minus b2, considering their signs */
+
+decimal
+decimal_sub (b1, b2)
+     decimal b1, b2;
+{
+  decimal v;
+
+  if (b1->sign != b2->sign)
+    v = decimal_add1 (b1, b2);
+  else
+    v = decimal_sub1 (b1, b2);
+  if (b1->sign && !decimal_zerop (v))
+    v->sign = !v->sign;
+  return v;
+}
+
+/* Return the negation of b2.  */
+
+decimal
+decimal_neg (b2)
+     decimal b2;
+{
+  decimal v = decimal_copy (b2);
+
+  if (!decimal_zerop (v))
+    v->sign = !v->sign;
+  return v;
+}
+
+/* add magnitudes of b1 and b2, ignoring their signs. */
+
+static decimal
+decimal_add1 (b1, b2)
+     decimal b1, b2;
+{
+  int before = MAX (b1->before, b2->before);
+  int after = MAX (b1->after, b2->after);
+
+  int len = before+after+1;
+  decimal result = make_decimal (before+1, after);
+
+  int i;
+  char *s1 = b1->contents;
+  char *s2 = b2->contents;
+  char *p1 = s1 + b1->after - after;
+  char *p2 = s2 + b2->after - after;
+  char *e1 = s1 + b1->before + b1->after;
+  char *e2 = s2 + b2->before + b2->after;
+  char *pr = result->contents;
+  int accum = 0;
+
+  for (i = 0; i < len; i++, p1++, p2++)
+    {
+      accum /= RADIX;
+      if (p1 >= s1 && p1 < e1) accum += *p1;
+      if (p2 >= s2 && p2 < e2) accum += *p2;
+      *pr++ = accum % RADIX;
+    }
+  if (!accum)
+    (result->before)--;
+
+  flush_trailing_digits (result, 0);
+
+  return result;
+}
+
+/* subtract magnitude of b2 from that or b1, returning signed decimal
+   number. */ 
+
+static decimal
+decimal_sub1 (b1, b2)
+     decimal b1, b2;
+{
+  int before = MAX (b1->before, b2->before);
+  int after = MAX (b1->after, b2->after);
+
+  int len = before+after;
+  decimal result = make_decimal (before, after);
+
+  int i;
+  char *s1 = b1->contents;
+  char *s2 = b2->contents;
+  char *p1 = s1 + b1->after - after;
+  char *p2 = s2 + b2->after - after;
+  char *e1 = s1 + b1->before + b1->after;
+  char *e2 = s2 + b2->before + b2->after;
+  char *pr = result->contents;
+  int accum = 0;
+
+  for (i = 0; i < len; i++, p1++, p2++)
+    {
+      if (p1 >= s1 && p1 < e1) accum += *p1;
+      if (p2 >= s2 && p2 < e2) accum -= *p2;
+      if (accum < 0 && accum % RADIX)
+        *pr = RADIX - (- accum) % RADIX;
+      else
+	*pr = accum % RADIX;
+      accum -= *pr++;
+      accum /= RADIX;
+    }
+
+  /* If result is negative, subtract it from RADIX**length
+     so that we get the right digits for sign-magnitude
+     rather than RADIX-complement */
+
+  if (accum)
+    {
+      result->sign = 1;
+      pr = result->contents;
+      accum = 0;
+      for (i = 0; i < len; i++)
+	{
+	  accum -= *pr;
+	  if (accum)
+	    *pr = accum + RADIX;
+	  else
+	    *pr = 0;
+	  accum -= *pr++;
+	  accum /= RADIX;
+	}
+    }
+
+  /* flush leading nonfraction zero digits */
+
+  while (result->before && *--pr == 0)
+    (result->before)--;
+
+  flush_trailing_digits (result, 0);
+
+  return result;
+}
+
+/* multiply b1 and b2 keeping `digits' fraction digits */
+
+decimal
+decimal_mul_rounded (b1, b2, digits)
+     decimal b1, b2;
+     int digits;
+{
+  decimal tem = decimal_mul (b1, b2);
+  decimal result = decimal_round_digits (tem, digits);
+  free (tem);
+  return result;
+}
+
+/* multiply b1 and b2 keeping the right number of fraction digits
+   for the `dc' program with precision = `digits'.  */
+
+decimal
+decimal_mul_dc (b1, b2, digits)
+     decimal b1, b2;
+     int digits;
+{
+  decimal tem = decimal_mul (b1, b2);
+  decimal result
+    = decimal_round_digits (tem, MAX (digits, MAX (b1->after, b2->after)));
+  free (tem);
+  return result;
+}
+
+/* multiply b1 and b2 as decimal error-free values;
+   keep LENGTH(b1) plus LENGTH(b2) significant figures. */
+
+decimal
+decimal_mul (b1, b2)
+     decimal b1, b2;
+{
+  decimal result = make_decimal (b1->before + b2->before, b1->after + b2->after);
+  int i;
+  int length2 = LENGTH(b2);
+  char *pr;
+
+  for (i = 0; i < length2; i++)
+    add_scaled (result, b1, b2->contents[i], i);
+
+  /* flush leading nonfraction zero digits */
+
+  pr = result->contents + LENGTH(result);
+  while (result->before && *--pr == 0)
+    (result->before)--;
+
+  flush_trailing_digits (result, 0);   /* flush trailing zeros */
+
+  /* Set sign properly */
+
+  if (b1->sign != b2->sign && LENGTH(result))
+    result->sign = 1;
+
+  return result;
+}
+
+/* Modify decimal number `into' by adding `from',
+   multiplied by `factor' (which should be nonnegative and less than RADIX)
+   and shifted left `scale' digits at the least significant end. */
+
+static void
+add_scaled (into, from, factor, scale)
+     decimal into, from;
+     int factor, scale;
+{
+  char *pf = from->contents;
+  char *pi = into->contents + scale;
+  int lengthf = LENGTH(from);
+  int lengthi = LENGTH(into) - scale;
+  
+  int accum = 0;
+  int i;
+
+  for (i = 0; i < lengthi; i++)
+    {
+      accum /= RADIX;
+      if (i < lengthf)
+        accum += *pf++ * factor;
+      accum += *pi;
+      *pi++ = accum % RADIX;
+    }
+}
+ 
+/* Divide decimal number `b1' by `b2', keeping at most `digits'
+   fraction digits. 
+   Returns the result as a decimal number.
+
+   When division is not exact, the quotient is truncated toward zero.  */
+
+decimal
+decimal_div (b1, b2, digits)
+     decimal b1, b2;
+     int digits;
+{
+  decimal result = make_decimal (MAX(1, (int) (1 + b1->before - b2->before)), digits);
+
+  /* b1copy holds what is left of the dividend,
+     that is not accounted for by the quotient digits already known */
+
+  decimal b1copy = decimal_copy_1 (b1, b2->after + digits);
+  int length1 = LENGTH(b1copy);
+  int length2 = LENGTH(b2);
+  int lengthr = LENGTH(result);
+  int i;
+
+  /* leading_divisor_digits contains the first two divisor digits, as
+     an integer */ 
+
+  int leading_divisor_digits = b2->contents[length2-1]*RADIX;
+  if (length2 > 1)
+    leading_divisor_digits += b2->contents[length2-2];
+
+  if (decimal_zerop (b2))
+    {
+      decimal_error ("divisor is zero", 0);
+      return decimal_copy (DECIMAL_ZERO);
+    }
+
+  if (lengthr <= (length1 - length2))
+    abort();		 /* My reasoning says this cannot happen, I hope */
+
+  for (i = length1 - length2; i >= 0; i--)
+    {
+      /* Guess the next quotient digit (in order of decreasing significance)
+	 using integer division */
+
+      int guess;
+      int trial_dividend = b1copy->contents[length2+i-1]*RADIX;
+      if (i != length1 - length2)
+	trial_dividend += b1copy->contents[length2+i]*RADIX*RADIX;
+      if (length2 + i > 1)
+	trial_dividend += b1copy->contents[length2+i-2];
+
+      guess = trial_dividend / leading_divisor_digits;
+
+      /* Remove the quotient times this digit from the dividend left */
+      /* We may find that the quotient digit is too large,
+	 when we consider the entire divisor.
+	 Then we decrement the quotient digit and add the divisor back in */
+
+      if (guess && 0 > subtract_scaled (b1copy, b2, guess, i))
+	{
+	  guess--;
+	  add_scaled (b1copy, b2, 1, i);
+	}
+
+      if (guess >= RADIX)
+	{
+	  result->contents[i + 1] += guess / RADIX;
+	  guess %= RADIX;
+	}
+      result->contents[i] = guess;
+    }
+
+  free (b1copy);
+
+  result->sign = (b1->sign != b2->sign);
+
+  /* flush leading nonfraction zero digits */
+
+  {
+    char *pr = result->contents + lengthr;
+    while (result->before && *--pr == 0)
+      (result->before)--;
+  }
+
+  flush_trailing_digits (result, 0);	/* Flush trailing zero fraction digits */
+
+  return result;
+}
+
+/* The remainder for the above division.
+   Same as `b1' - (`b1' / `b2') * 'b2'.
+   Note that the value depends on the number of fraction digits
+   that were kept in computing `b1' / `b2';
+   the argument `digits' specifies this.
+
+   The remainder has the same sign as the dividend.
+   The divisor's sign is ignored.  */
+
+decimal
+decimal_rem (b1, b2, digits)
+     decimal b1, b2;
+     int digits;
+{
+  decimal b1copy = decimal_copy_1 (b1, b2->after + digits);
+  int length1 = LENGTH(b1copy);
+  int length2 = LENGTH(b2);
+  int i;
+
+  int leading_divisor_digits = b2->contents[length2-1]*RADIX;
+
+  if (length2 > 1)
+    leading_divisor_digits += b2->contents[length2-2];
+
+  if (decimal_zerop (b2))
+    {
+      decimal_error ("divisor is zero", 0);
+      return decimal_copy (DECIMAL_ZERO);
+    }
+
+  /* Do like division, above, but throw away the quotient.
+     Keep only the final `rest of dividend', which becomes the remainder.  */
+
+  for (i = length1 - length2; i >= 0; i--)
+    {
+      int guess;
+      int trial_dividend = b1copy->contents[length2+i-1]*RADIX;
+      if (i != length1 - length2)
+	trial_dividend += b1copy->contents[length2+i]*RADIX*RADIX;
+      if (length2 + i > 1)
+	trial_dividend += b1copy->contents[length2+i-2];
+
+      guess = trial_dividend / leading_divisor_digits;
+
+      if (guess && 0 > subtract_scaled (b1copy, b2, guess, i))
+	{
+	  guess--;
+	  add_scaled (b1copy, b2, 1, i);
+	}
+      /* No need to check whether guess exceeds RADIX
+	 since we are not saving guess.  */
+    }
+
+  /* flush leading nonfraction zero digits */
+
+  {
+    char *pr = b1copy->contents + length1;
+    while (b1copy->before && *--pr == 0)
+      (b1copy->before)--;
+  }
+
+  flush_trailing_digits (b1copy, 0);
+  return b1copy;
+}
+
+/* returns negative number if we chose factor too large */
+
+static int
+subtract_scaled (into, from, factor, scale)
+     decimal into, from;
+     int factor, scale;
+{
+  char *pf = from->contents;
+  char *pi = into->contents + scale;
+  int lengthf = LENGTH(from);
+  int lengthi = LENGTH(into) - scale;
+  int accum = 0;
+  int i;
+
+  for (i = 0; i < lengthi && i <= lengthf; i++)
+    {
+      if (i < lengthf)
+        accum -= *pf++ * factor;
+      accum += *pi;
+      if (accum < 0 && accum % RADIX)
+        *pi = RADIX - (- accum) % RADIX;
+      else
+	*pi = accum % RADIX;
+      accum -= *pi++;
+      accum /= RADIX;
+    }
+  return accum;
+}
+
+/* Return the square root of decimal number D, using Newton's method.
+   Number of fraction digits returned is max of FRAC_DIGITS
+   and D's number of fraction digits.  */
+
+decimal
+decimal_sqrt (d, frac_digits)
+     decimal d;
+     int frac_digits;
+{
+  decimal guess;
+  int notdone = 1;
+
+  if (decimal_zerop (d)) return d;
+  if (d->sign)
+    {
+      decimal_error ("square root argument negative", 0);
+      return decimal_copy (DECIMAL_ZERO);
+    }
+
+  frac_digits = MAX (frac_digits, d->after);
+
+  /* Compute an initial guess by taking the square root 
+     of a nearby power of RADIX.  */
+
+  if (d->before)
+    {
+      guess = make_decimal ((d->before + 1) / 2, 0);
+      guess->contents[guess->before - 1] = 1;
+    }
+  else
+    {
+      /* Arg is less than 1; compute nearest power of RADIX */
+      char *p = d->contents + LENGTH(d);
+      char *sp = p;
+
+      while (!*--p);	/* Find most significant nonzero digit */
+      if (sp - p == 1)
+	{
+	  /* Arg is bigger than 1/RADIX; use 1 as a guess */
+	  guess = decimal_copy (DECIMAL_ONE);
+	}
+      else
+	{
+	  guess = make_decimal (0, (sp - p) / 2);
+	  guess->contents[0] = 1;
+	}
+    }
+
+  /* Iterate doing guess = (guess + d/guess) / 2  */
+
+  while (notdone)
+    {
+      decimal tem1 = decimal_div (d, guess, frac_digits + 1);
+      decimal tem2 = decimal_add (guess, tem1);
+      decimal tem3 = decimal_mul_rounded (tem2, DECIMAL_HALF, frac_digits);
+      notdone = decimal_compare (guess, tem3);
+      free (tem1);
+      free (tem2);
+      free (guess);
+      guess = tem3;
+      if (decimal_zerop (guess)) return guess;  /* Avoid divide-by-zero */
+    }
+
+  return guess;
+}
+
+/* Raise decimal number `base' to power of integer part of decimal
+   number `expt'.
+   This function depends on using radix 10.
+   It is too hard to write it to work for any value of RADIX,
+   so instead it is simply not available if RADIX is not ten.  */
+
+#if !(RADIX - 10)
+
+decimal
+decimal_expt (base, expt, frac_digits)
+     decimal base, expt;
+     int frac_digits;
+{
+  decimal accum = decimal_copy (DECIMAL_ONE);
+  decimal basis1 = base;
+  int digits = expt->before;
+  int dig = 0;				/* Expt digit being processed */
+
+  if (expt->sign)
+  /* If negative power, take reciprocal first thing
+     so that fraction digit truncation won't destroy
+     what will ultimately be nonfraction digits.  */
+    basis1 = decimal_div (DECIMAL_ONE, base, frac_digits);
+  while (dig < digits)
+    {
+      decimal basis2, basis4, basis8, basis10;
+      int thisdigit = expt->contents[expt->after + dig];
+
+      /* Compute factors to multiply in for each bit of this digit */
+
+      basis2 = decimal_mul_rounded (basis1, basis1, frac_digits);
+      basis4 = decimal_mul_rounded (basis2, basis2, frac_digits);
+      basis8 = decimal_mul_rounded (basis4, basis4, frac_digits);
+
+      /* Now accumulate the factors this digit value selects */
+
+      if (thisdigit & 1)
+	{
+	  decimal accum1 = decimal_mul_rounded (accum, basis1, frac_digits);
+	  free (accum);
+	  accum = accum1;
+	}
+
+      if (thisdigit & 2)
+	{
+	  decimal accum1 = decimal_mul_rounded (accum, basis2, frac_digits);
+	  free (accum);
+	  accum = accum1;
+	}
+
+      if (thisdigit & 4)
+	{
+	  decimal accum1 = decimal_mul_rounded (accum, basis4, frac_digits);
+	  free (accum);
+	  accum = accum1;
+	}
+
+      if (thisdigit & 8)
+	{
+	  decimal accum1 = decimal_mul_rounded (accum, basis8, frac_digits);
+	  free (accum);
+	  accum = accum1;
+	}
+
+      /* If there are further digits, compute the basis1 for the next digit */
+
+      if (++dig < digits)
+	basis10 = decimal_mul_rounded (basis2, basis8, frac_digits);
+
+      /* Free intermediate results */
+
+      if (basis1 != base) free (basis1);
+      free (basis2);
+      free (basis4);
+      free (basis8);
+      basis1 = basis10;
+    }
+  return accum;
+}
+#endif
+
+#ifdef TEST
+
+fputchar (c)
+     char c;
+{
+  putchar (c);
+}
+
+/* Top level that can be used to test the arithmetic functions */
+
+main ()
+{
+  char s1[40], s2[40];
+  decimal b1, b2, b3;
+  char c;
+
+  while (1)
+    {
+      scanf ("%s %c %s", s1, &c, s2);
+      b1 = decimal_parse (s1, RADIX);
+      b2 = decimal_parse (s2, RADIX);
+      switch (c)
+	{
+	default:
+	  c = '+';
+	case '+':
+	  b3 = decimal_add (b1, b2);
+	  break;
+	case '*':
+	  b3 = decimal_mul (b1, b2);
+	  break;
+        case '/':
+	  b3 = decimal_div (b1, b2, 3);
+	  break;
+	case '%':
+	  b3 = decimal_rem (b1, b2, 3);
+	  break;
+        case 'p':
+	  decimal_print (b1, fputchar, RADIX);
+	  printf (" printed in base %d is ", decimal_to_int (b2));
+	  decimal_print (b1, fputchar, decimal_to_int (b2));
+	  printf ("\n");
+	  continue;
+	case 'r':
+	  printf ("%s read in base %d is ", s1, decimal_to_int (b2));
+	  decimal_print (decimal_parse (s1, decimal_to_int (b2)), fputchar, RADIX);
+	  printf ("\n");
+	  continue;
+	}
+      decimal_print (b1, fputchar, RADIX);
+      printf (" %c ", c);
+      decimal_print (b2, fputchar, RADIX);
+      printf (" = ");
+      decimal_print (b3, fputchar, RADIX);
+      printf ("\n");
+    }
+}
+
+decimal_error (s1, s2)
+     char *s1, *s2;
+{
+  printf ("\n");
+  printf (s1, s2);
+  printf ("\n");
+}
+
+static void
+pbi (b)
+      int b;
+{
+  decimal_print ((decimal) b, fputchar, RADIX);
+}
+
+static void
+pb (b)
+      decimal b;
+{
+  decimal_print (b, fputchar, RADIX);
+}
+
+#endif
diff --git a/gnu/usr.bin/dc/decimal.h b/gnu/usr.bin/dc/decimal.h
new file mode 100644
index 000000000000..2b41158166a2
--- /dev/null
+++ b/gnu/usr.bin/dc/decimal.h
@@ -0,0 +1,93 @@
+/* 
+ * Header file for decimal.c (arbitrary precision decimal arithmetic)
+ *
+ * Copyright (C) 1984 Free Software Foundation, Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, you can either send email to this
+ * program's author (see below) or write to: The Free Software Foundation,
+ * Inc.; 675 Mass Ave. Cambridge, MA 02139, USA.
+ */
+
+/* Autoconf stuff */
+#ifndef HAVE_BCOPY
+#undef bcopy
+#define bcopy(s2, s1, n) memcpy (s1, s2, n)
+#endif
+
+#ifndef HAVE_BZERO
+#undef bzero
+#define bzero(b, l) memset (b, 0, l)
+#endif
+
+/* Define the radix to use by default, and for representing the
+   numbers internally.  This does not need to be decimal; that is just
+   the default for it.  */
+
+/* Currently, this is required to be even for this program to work. */
+
+#ifndef RADIX
+#define RADIX 10
+#endif
+
+/* The user must define the external function `decimal_error'
+   which is called with two arguments to report errors in this package.
+   The two arguments may be passed to `printf' to print a message. */
+
+/* Structure that represents a decimal number */
+
+struct decimal
+{
+  unsigned int sign: 1;		/* One for negative number */
+				/* The sign should always be zero for the number 0 */
+  int after: 15;		/* number of fraction digits */
+  unsigned short before;	/* number of non-fraction digits */
+  unsigned short refcnt;	/* number of pointers to this number */
+				/* (used by calling program) */
+  char contents[1];		/* the digits themselves, least significant first. */
+				/* digits are just numbers 0 .. RADIX-1 */
+};
+
+/* There may never be leading nonfraction zeros or trailing fraction
+   zeros in a number.  They must be removed by all the arithmetic
+   functions.  Therefore, the number zero always has no digits stored. */
+
+typedef struct decimal *decimal;
+
+/* Decimal numbers are always passed around as pointers.
+   All the external entries in this file allocate new numbers
+   using `malloc' to store values in.
+   They never modify their arguments or any existing numbers. */
+
+/* Return the total number of digits stored in the number `b' */
+#define LENGTH(b) ((b)->before + (b)->after)
+
+/* Some constant decimal numbers */
+
+
+#define DECIMAL_ZERO &decimal_zero
+
+
+#define DECIMAL_ONE &decimal_one
+
+#define DECIMAL_HALF &decimal_half
+
+decimal decimal_add (), decimal_sub (), decimal_mul (), decimal_div ();
+decimal decimal_mul_dc (), decimal_mul_rounded (), decimal_rem ();
+decimal decimal_round_digits ();
+decimal make_decimal (), decimal_copy (), decimal_parse ();
+decimal decimal_sqrt (), decimal_expt ();
+
+void decimal_print ();
+
+/* End of decimal.h */
diff --git a/gnu/usr.bin/dc/texinfo.tex b/gnu/usr.bin/dc/texinfo.tex
new file mode 100644
index 000000000000..ce8124ecb7bd
--- /dev/null
+++ b/gnu/usr.bin/dc/texinfo.tex
@@ -0,0 +1,4003 @@
+%% TeX macros to handle texinfo files
+
+%   Copyright (C) 1985, 86, 88, 90, 91, 92, 1993 Free Software Foundation, Inc.
+
+%This texinfo.tex file is free software; you can redistribute it and/or
+%modify it under the terms of the GNU General Public License as
+%published by the Free Software Foundation; either version 2, or (at
+%your option) any later version.
+
+%This texinfo.tex file is distributed in the hope that it will be
+%useful, but WITHOUT ANY WARRANTY; without even the implied warranty
+%of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+%General Public License for more details.
+
+%You should have received a copy of the GNU General Public License
+%along with this texinfo.tex file; see the file COPYING.  If not, write
+%to the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139,
+%USA.
+
+
+%In other words, you are welcome to use, share and improve this program.
+%You are forbidden to forbid anyone else to use, share and improve
+%what you give them.   Help stamp out software-hoarding!
+
+\def\texinfoversion{2.108}
+\message{Loading texinfo package [Version \texinfoversion]:}
+
+% Print the version number if in a .fmt file.
+\everyjob{\message{[Texinfo version \texinfoversion]}\message{}}
+
+% Save some parts of plain tex whose names we will redefine.
+
+\let\ptexlbrace=\{
+\let\ptexrbrace=\}
+\let\ptexdots=\dots
+\let\ptexdot=\.
+\let\ptexstar=\*
+\let\ptexend=\end
+\let\ptexbullet=\bullet
+\let\ptexb=\b
+\let\ptexc=\c
+\let\ptexi=\i
+\let\ptext=\t
+\let\ptexl=\l
+\let\ptexL=\L
+
+\def\tie{\penalty 10000\ }     % Save plain tex definition of ~.
+
+\message{Basics,}
+\chardef\other=12
+
+% If this character appears in an error message or help string, it
+% starts a new line in the output.
+\newlinechar = `^^J
+
+% Ignore a token.
+%
+\def\gobble#1{}
+
+\hyphenation{ap-pen-dix}
+\hyphenation{mini-buf-fer mini-buf-fers}
+\hyphenation{eshell}
+
+% Margin to add to right of even pages, to left of odd pages.
+\newdimen \bindingoffset  \bindingoffset=0pt
+\newdimen \normaloffset   \normaloffset=\hoffset
+\newdimen\pagewidth \newdimen\pageheight
+\pagewidth=\hsize \pageheight=\vsize
+
+% Sometimes it is convenient to have everything in the transcript file
+% and nothing on the terminal.  We don't just call \tracingall here,
+% since that produces some useless output on the terminal.
+%
+\def\gloggingall{\begingroup \globaldefs = 1 \loggingall \endgroup}%
+\def\loggingall{\tracingcommands2 \tracingstats2
+   \tracingpages1 \tracingoutput1 \tracinglostchars1
+   \tracingmacros2 \tracingparagraphs1 \tracingrestores1
+   \showboxbreadth\maxdimen\showboxdepth\maxdimen
+}%
+
+%---------------------Begin change-----------------------
+%
+%%%% For @cropmarks command.
+% Dimensions to add cropmarks at corners Added by P. A. MacKay, 12 Nov. 1986
+%
+\newdimen\cornerlong \newdimen\cornerthick
+\newdimen \topandbottommargin
+\newdimen \outerhsize \newdimen \outervsize
+\cornerlong=1pc\cornerthick=.3pt	% These set size of cropmarks
+\outerhsize=7in
+%\outervsize=9.5in
+% Alternative @smallbook page size is 9.25in
+\outervsize=9.25in
+\topandbottommargin=.75in
+%
+%---------------------End change-----------------------
+
+% \onepageout takes a vbox as an argument.  Note that \pagecontents
+% does insertions itself, but you have to call it yourself.
+\chardef\PAGE=255  \output={\onepageout{\pagecontents\PAGE}}
+\def\onepageout#1{\hoffset=\normaloffset
+\ifodd\pageno  \advance\hoffset by \bindingoffset
+\else \advance\hoffset by -\bindingoffset\fi
+{\escapechar=`\\\relax % makes sure backslash is used in output files.
+\shipout\vbox{{\let\hsize=\pagewidth \makeheadline} \pagebody{#1}%
+{\let\hsize=\pagewidth \makefootline}}}%
+\advancepageno \ifnum\outputpenalty>-20000 \else\dosupereject\fi}
+
+%%%% For @cropmarks command %%%%
+
+% Here is a modification of the main output routine for Near East Publications
+% This provides right-angle cropmarks at all four corners.
+% The contents of the page are centerlined into the cropmarks,
+% and any desired binding offset is added as an \hskip on either
+% site of the centerlined box.  (P. A. MacKay, 12 November, 1986)
+%
+\def\croppageout#1{\hoffset=0pt % make sure this doesn't mess things up
+{\escapechar=`\\\relax % makes sure backslash is used in output files.
+		 \shipout
+		 \vbox to \outervsize{\hsize=\outerhsize
+                 \vbox{\line{\ewtop\hfill\ewtop}}
+                 \nointerlineskip
+                 \line{\vbox{\moveleft\cornerthick\nstop}
+                       \hfill
+                       \vbox{\moveright\cornerthick\nstop}}
+                 \vskip \topandbottommargin
+                 \centerline{\ifodd\pageno\hskip\bindingoffset\fi
+			\vbox{
+			{\let\hsize=\pagewidth \makeheadline}
+			\pagebody{#1}
+			{\let\hsize=\pagewidth \makefootline}}
+			\ifodd\pageno\else\hskip\bindingoffset\fi}
+		 \vskip \topandbottommargin plus1fill minus1fill
+                 \boxmaxdepth\cornerthick
+                 \line{\vbox{\moveleft\cornerthick\nsbot}
+                       \hfill
+                       \vbox{\moveright\cornerthick\nsbot}}
+                 \nointerlineskip
+                 \vbox{\line{\ewbot\hfill\ewbot}}
+	}}
+  \advancepageno
+  \ifnum\outputpenalty>-20000 \else\dosupereject\fi}
+%
+% Do @cropmarks to get crop marks
+\def\cropmarks{\let\onepageout=\croppageout }
+
+\def\pagebody#1{\vbox to\pageheight{\boxmaxdepth=\maxdepth #1}}
+{\catcode`\@ =11
+\gdef\pagecontents#1{\ifvoid\topins\else\unvbox\topins\fi
+\dimen@=\dp#1 \unvbox#1
+\ifvoid\footins\else\vskip\skip\footins\footnoterule \unvbox\footins\fi
+\ifr@ggedbottom \kern-\dimen@ \vfil \fi}
+}
+
+%
+% Here are the rules for the cropmarks.  Note that they are
+% offset so that the space between them is truly \outerhsize or \outervsize
+% (P. A. MacKay, 12 November, 1986)
+%
+\def\ewtop{\vrule height\cornerthick depth0pt width\cornerlong}
+\def\nstop{\vbox
+  {\hrule height\cornerthick depth\cornerlong width\cornerthick}}
+\def\ewbot{\vrule height0pt depth\cornerthick width\cornerlong}
+\def\nsbot{\vbox
+  {\hrule height\cornerlong depth\cornerthick width\cornerthick}}
+
+% Parse an argument, then pass it to #1.  The argument is the rest of
+% the input line (except we remove a trailing comment).  #1 should be a
+% macro which expects an ordinary undelimited TeX argument.
+%
+\def\parsearg#1{%
+  \let\next = #1%
+  \begingroup
+    \obeylines
+    \futurelet\temp\parseargx
+}
+
+% If the next token is an obeyed space (from an @example environment or
+% the like), remove it and recurse.  Otherwise, we're done.
+\def\parseargx{%
+  % \obeyedspace is defined far below, after the definition of \sepspaces.
+  \ifx\obeyedspace\temp
+    \expandafter\parseargdiscardspace
+  \else
+    \expandafter\parseargline
+  \fi
+}
+
+% Remove a single space (as the delimiter token to the macro call).
+{\obeyspaces %
+ \gdef\parseargdiscardspace {\futurelet\temp\parseargx}}
+
+{\obeylines %
+  \gdef\parseargline#1^^M{%
+    \endgroup % End of the group started in \parsearg.
+    %
+    % First remove any @c comment, then any @comment.
+    % Result of each macro is put in \toks0.
+    \argremovec #1\c\relax %
+    \expandafter\argremovecomment \the\toks0 \comment\relax %
+    %
+    % Call the caller's macro, saved as \next in \parsearg.
+    \expandafter\next\expandafter{\the\toks0}%
+  }%
+}
+
+% Since all \c{,omment} does is throw away the argument, we can let TeX
+% do that for us.  The \relax here is matched by the \relax in the call
+% in \parseargline; it could be more or less anything, its purpose is
+% just to delimit the argument to the \c.
+\def\argremovec#1\c#2\relax{\toks0 = {#1}}
+\def\argremovecomment#1\comment#2\relax{\toks0 = {#1}}
+
+% \argremovec{,omment} might leave us with trailing spaces, though; e.g.,
+%    @end itemize  @c foo
+% will have two active spaces as part of the argument with the
+% `itemize'.  Here we remove all active spaces from #1, and assign the
+% result to \toks0.
+%
+% This loses if there are any *other* active characters besides spaces
+% in the argument -- _ ^ +, for example -- since they get expanded.
+% Fortunately, Texinfo does not define any such commands.  (If it ever
+% does, the catcode of the characters in questionwill have to be changed
+% here.)  But this means we cannot call \removeactivespaces as part of
+% \argremovec{,omment}, since @c uses \parsearg, and thus the argument
+% that \parsearg gets might well have any character at all in it.
+%
+\def\removeactivespaces#1{%
+  \begingroup
+    \ignoreactivespaces
+    \edef\temp{#1}%
+    \global\toks0 = \expandafter{\temp}%
+  \endgroup
+}
+
+% Change the active space to expand to nothing.
+%
+\begingroup
+  \obeyspaces
+  \gdef\ignoreactivespaces{\obeyspaces\let =\empty}
+\endgroup
+
+
+\def\flushcr{\ifx\par\lisppar \def\next##1{}\else \let\next=\relax \fi \next}
+
+%% These are used to keep @begin/@end levels from running away
+%% Call \inENV within environments (after a \begingroup)
+\newif\ifENV \ENVfalse \def\inENV{\ifENV\relax\else\ENVtrue\fi}
+\def\ENVcheck{%
+\ifENV\errmessage{Still within an environment.  Type Return to continue.}
+\endgroup\fi} % This is not perfect, but it should reduce lossage
+
+% @begin foo  is the same as @foo, for now.
+\newhelp\EMsimple{Type <Return> to continue.}
+
+\outer\def\begin{\parsearg\beginxxx}
+
+\def\beginxxx #1{%
+\expandafter\ifx\csname #1\endcsname\relax
+{\errhelp=\EMsimple \errmessage{Undefined command @begin #1}}\else
+\csname #1\endcsname\fi}
+
+% @end foo executes the definition of \Efoo.
+%
+\def\end{\parsearg\endxxx}
+\def\endxxx #1{%
+  \removeactivespaces{#1}%
+  \edef\endthing{\the\toks0}%
+  %
+  \expandafter\ifx\csname E\endthing\endcsname\relax
+    \expandafter\ifx\csname \endthing\endcsname\relax
+      % There's no \foo, i.e., no ``environment'' foo.
+      \errhelp = \EMsimple
+      \errmessage{Undefined command `@end \endthing'}%
+    \else
+      \unmatchedenderror\endthing
+    \fi
+  \else
+    % Everything's ok; the right environment has been started.
+    \csname E\endthing\endcsname
+  \fi
+}
+
+% There is an environment #1, but it hasn't been started.  Give an error.
+%
+\def\unmatchedenderror#1{%
+  \errhelp = \EMsimple
+  \errmessage{This `@end #1' doesn't have a matching `@#1'}%
+}
+
+% Define the control sequence \E#1 to give an unmatched @end error.
+%
+\def\defineunmatchedend#1{%
+  \expandafter\def\csname E#1\endcsname{\unmatchedenderror{#1}}%
+}
+
+
+% Single-spacing is done by various environments (specifically, in
+% \nonfillstart and \quotations).
+\newskip\singlespaceskip \singlespaceskip = \baselineskip
+\def\singlespace{%
+% Why was this kern here?  It messes up equalizing space above and below
+% environments.  --karl, 6may93
+%{\advance \baselineskip by -\singlespaceskip
+%\kern \baselineskip}%
+\baselineskip=\singlespaceskip
+}
+
+%% Simple single-character @ commands
+
+% @@ prints an @
+% Kludge this until the fonts are right (grr).
+\def\@{{\tt \char '100}}
+
+% This is turned off because it was never documented
+% and you can use @w{...} around a quote to suppress ligatures.
+%% Define @` and @' to be the same as ` and '
+%% but suppressing ligatures.
+%\def\`{{`}}
+%\def\'{{'}}
+
+% Used to generate quoted braces.
+
+\def\mylbrace {{\tt \char '173}}
+\def\myrbrace {{\tt \char '175}}
+\let\{=\mylbrace
+\let\}=\myrbrace
+
+% @: forces normal size whitespace following.
+\def\:{\spacefactor=1000 }
+
+% @* forces a line break.
+\def\*{\hfil\break\hbox{}\ignorespaces}
+
+% @. is an end-of-sentence period.
+\def\.{.\spacefactor=3000 }
+
+% @w prevents a word break.  Without the \leavevmode, @w at the
+% beginning of a paragraph, when TeX is still in vertical mode, would
+% produce a whole line of output instead of starting the paragraph.
+\def\w#1{\leavevmode\hbox{#1}}
+
+% @group ... @end group forces ... to be all on one page, by enclosing
+% it in a TeX vbox.  We use \vtop instead of \vbox to construct the box
+% to keep its height that of a normal line.  According to the rules for
+% \topskip (p.114 of the TeXbook), the glue inserted is
+% max (\topskip - \ht (first item), 0).  If that height is large,
+% therefore, no glue is inserted, and the space between the headline and
+% the text is small, which looks bad.
+%
+\def\group{\begingroup
+  \ifnum\catcode13=\active \else
+    \errhelp = \groupinvalidhelp
+    \errmessage{@group invalid in context where filling is enabled}%
+  \fi
+  %
+  % The \vtop we start below produces a box with normal height and large
+  % depth; thus, TeX puts \baselineskip glue before it, and (when the
+  % next line of text is done) \lineskip glue after it.  (See p.82 of
+  % the TeXbook.)  Thus, space below is not quite equal to space
+  % above.  But it's pretty close.
+  \def\Egroup{%
+    \egroup           % End the \vtop.
+    \endgroup         % End the \group.
+  }%
+  %
+  \vtop\bgroup
+    % We have to put a strut on the last line in case the @group is in
+    % the midst of an example, rather than completely enclosing it.
+    % Otherwise, the interline space between the last line of the group
+    % and the first line afterwards is too small.  But we can't put the
+    % strut in \Egroup, since there it would be on a line by itself.
+    % Hence this just inserts a strut at the beginning of each line.
+    \everypar = {\strut}%
+    %
+    % Since we have a strut on every line, we don't need any of TeX's
+    % normal interline spacing.
+    \offinterlineskip
+    %
+    % OK, but now we have to do something about blank
+    % lines in the input in @example-like environments, which normally
+    % just turn into \lisppar, which will insert no space now that we've
+    % turned off the interline space.  Simplest is to make them be an
+    % empty paragraph.
+    \ifx\par\lisppar
+      \edef\par{\leavevmode \par}%
+      %
+      % Reset ^^M's definition to new definition of \par.
+      \obeylines
+    \fi
+    %
+    % We do @comment here in case we are called inside an environment,
+    % such as @example, where each end-of-line in the input causes an
+    % end-of-line in the output.  We don't want the end-of-line after
+    % the `@group' to put extra space in the output.  Since @group
+    % should appear on a line by itself (according to the Texinfo
+    % manual), we don't worry about eating any user text.
+    \comment
+}
+%
+% TeX puts in an \escapechar (i.e., `@') at the beginning of the help
+% message, so this ends up printing `@group can only ...'.
+%
+\newhelp\groupinvalidhelp{%
+group can only be used in environments such as @example,^^J%
+where each line of input produces a line of output.}
+
+% @need space-in-mils
+% forces a page break if there is not space-in-mils remaining.
+
+\newdimen\mil  \mil=0.001in
+
+\def\need{\parsearg\needx}
+
+% Old definition--didn't work.
+%\def\needx #1{\par %
+%% This method tries to make TeX break the page naturally
+%% if the depth of the box does not fit.
+%{\baselineskip=0pt%
+%\vtop to #1\mil{\vfil}\kern -#1\mil\penalty 10000
+%\prevdepth=-1000pt
+%}}
+
+\def\needx#1{%
+  % Go into vertical mode, so we don't make a big box in the middle of a
+  % paragraph.
+  \par
+  %
+  % Don't add any leading before our big empty box, but allow a page
+  % break, since the best break might be right here.
+  \allowbreak
+  \nointerlineskip
+  \vtop to #1\mil{\vfil}%
+  %
+  % TeX does not even consider page breaks if a penalty added to the
+  % main vertical list is 10000 or more.  But in order to see if the
+  % empty box we just added fits on the page, we must make it consider
+  % page breaks.  On the other hand, we don't want to actually break the
+  % page after the empty box.  So we use a penalty of 9999.
+  %
+  % There is an extremely small chance that TeX will actually break the
+  % page at this \penalty, if there are no other feasible breakpoints in
+  % sight.  (If the user is using lots of big @group commands, which
+  % almost-but-not-quite fill up a page, TeX will have a hard time doing
+  % good page breaking, for example.)  However, I could not construct an
+  % example where a page broke at this \penalty; if it happens in a real
+  % document, then we can reconsider our strategy.
+  \penalty9999
+  %
+  % Back up by the size of the box, whether we did a page break or not.
+  \kern -#1\mil
+  %
+  % Do not allow a page break right after this kern.
+  \nobreak
+}
+
+% @br   forces paragraph break
+
+\let\br = \par
+
+% @dots{}  output some dots
+
+\def\dots{$\ldots$}
+
+% @page    forces the start of a new page
+
+\def\page{\par\vfill\supereject}
+
+% @exdent text....
+% outputs text on separate line in roman font, starting at standard page margin
+
+% This records the amount of indent in the innermost environment.
+% That's how much \exdent should take out.
+\newskip\exdentamount
+
+% This defn is used inside fill environments such as @defun.
+\def\exdent{\parsearg\exdentyyy}
+\def\exdentyyy #1{{\hfil\break\hbox{\kern -\exdentamount{\rm#1}}\hfil\break}}
+
+% This defn is used inside nofill environments such as @example.
+\def\nofillexdent{\parsearg\nofillexdentyyy}
+\def\nofillexdentyyy #1{{\advance \leftskip by -\exdentamount
+\leftline{\hskip\leftskip{\rm#1}}}}
+
+%\hbox{{\rm#1}}\hfil\break}}
+
+% @include file    insert text of that file as input.
+
+\def\include{\parsearg\includezzz}
+%Use \input\thisfile to avoid blank after \input, which may be an active
+%char (in which case the blank would become the \input argument).
+%The grouping keeps the value of \thisfile correct even when @include
+%is nested.
+\def\includezzz #1{\begingroup
+\def\thisfile{#1}\input\thisfile
+\endgroup}
+
+\def\thisfile{}
+
+% @center line   outputs that line, centered
+
+\def\center{\parsearg\centerzzz}
+\def\centerzzz #1{{\advance\hsize by -\leftskip
+\advance\hsize by -\rightskip
+\centerline{#1}}}
+
+% @sp n   outputs n lines of vertical space
+
+\def\sp{\parsearg\spxxx}
+\def\spxxx #1{\par \vskip #1\baselineskip}
+
+% @comment ...line which is ignored...
+% @c is the same as @comment
+% @ignore ... @end ignore  is another way to write a comment
+
+\def\comment{\catcode 64=\other \catcode 123=\other \catcode 125=\other%
+\parsearg \commentxxx}
+
+\def\commentxxx #1{\catcode 64=0 \catcode 123=1 \catcode 125=2 }
+
+\let\c=\comment
+
+% Prevent errors for section commands.
+% Used in @ignore and in failing conditionals.
+\def\ignoresections{%
+\let\chapter=\relax
+\let\unnumbered=\relax
+\let\top=\relax
+\let\unnumberedsec=\relax
+\let\unnumberedsection=\relax
+\let\unnumberedsubsec=\relax
+\let\unnumberedsubsection=\relax
+\let\unnumberedsubsubsec=\relax
+\let\unnumberedsubsubsection=\relax
+\let\section=\relax
+\let\subsec=\relax
+\let\subsubsec=\relax
+\let\subsection=\relax
+\let\subsubsection=\relax
+\let\appendix=\relax
+\let\appendixsec=\relax
+\let\appendixsection=\relax
+\let\appendixsubsec=\relax
+\let\appendixsubsection=\relax
+\let\appendixsubsubsec=\relax
+\let\appendixsubsubsection=\relax
+\let\contents=\relax
+\let\smallbook=\relax
+\let\titlepage=\relax
+}
+
+% Used in nested conditionals, where we have to parse the Texinfo source
+% and so want to turn off most commands, in case they are used
+% incorrectly.
+%
+\def\ignoremorecommands{%
+  \let\defcv = \relax
+  \let\deffn = \relax
+  \let\deffnx = \relax
+  \let\defindex = \relax
+  \let\defivar = \relax
+  \let\defmac = \relax
+  \let\defmethod = \relax
+  \let\defop = \relax
+  \let\defopt = \relax
+  \let\defspec = \relax
+  \let\deftp = \relax
+  \let\deftypefn = \relax
+  \let\deftypefun = \relax
+  \let\deftypevar = \relax
+  \let\deftypevr = \relax
+  \let\defun = \relax
+  \let\defvar = \relax
+  \let\defvr = \relax
+  \let\ref = \relax
+  \let\xref = \relax
+  \let\printindex = \relax
+  \let\pxref = \relax
+  \let\settitle = \relax
+  \let\include = \relax
+  \let\lowersections = \relax
+  \let\down = \relax
+  \let\raisesections = \relax
+  \let\up = \relax
+  \let\set = \relax
+  \let\clear = \relax
+}
+
+% Ignore @ignore ... @end ignore.
+%
+\def\ignore{\doignore{ignore}}
+
+% Also ignore @ifinfo, @menu, and @direntry text.
+%
+\def\ifinfo{\doignore{ifinfo}}
+\def\menu{\doignore{menu}}
+\def\direntry{\doignore{direntry}}
+
+% Ignore text until a line `@end #1'.
+%
+\def\doignore#1{\begingroup
+  % Don't complain about control sequences we have declared \outer.
+  \ignoresections
+  %
+  % Define a command to swallow text until we reach `@end #1'.
+  \long\def\doignoretext##1\end #1{\enddoignore}%
+  %
+  % Make sure that spaces turn into tokens that match what \doignoretext wants.
+  \catcode32 = 10
+  %
+  % And now expand that command.
+  \doignoretext
+}
+
+% What we do to finish off ignored text.
+%
+\def\enddoignore{\endgroup\ignorespaces}%
+
+\newif\ifwarnedobs\warnedobsfalse
+\def\obstexwarn{%
+  \ifwarnedobs\relax\else
+  % We need to warn folks that they may have trouble with TeX 3.0.
+  % This uses \immediate\write16 rather than \message to get newlines.
+    \immediate\write16{}
+    \immediate\write16{***WARNING*** for users of Unix TeX 3.0!}
+    \immediate\write16{This manual trips a bug in TeX version 3.0 (tex hangs).}
+    \immediate\write16{If you are running another version of TeX, relax.}
+    \immediate\write16{If you are running Unix TeX 3.0, kill this TeX process.}
+    \immediate\write16{  Then upgrade your TeX installation if you can.}
+    \immediate\write16{If you are stuck with version 3.0, run the}
+    \immediate\write16{  script ``tex3patch'' from the Texinfo distribution}
+    \immediate\write16{  to use a workaround.}
+    \immediate\write16{}
+    \warnedobstrue
+    \fi
+}
+
+% **In TeX 3.0, setting text in \nullfont hangs tex.  For a
+% workaround (which requires the file ``dummy.tfm'' to be installed),
+% uncomment the following line:
+%%%%%\font\nullfont=dummy\let\obstexwarn=\relax
+
+% Ignore text, except that we keep track of conditional commands for
+% purposes of nesting, up to an `@end #1' command.
+%
+\def\nestedignore#1{%
+  \obstexwarn
+  % We must actually expand the ignored text to look for the @end
+  % command, so that nested ignore constructs work.  Thus, we put the
+  % text into a \vbox and then do nothing with the result.  To minimize
+  % the change of memory overflow, we follow the approach outlined on
+  % page 401 of the TeXbook: make the current font be a dummy font.
+  %
+  \setbox0 = \vbox\bgroup
+    % Don't complain about control sequences we have declared \outer.
+    \ignoresections
+    %
+    % Define `@end #1' to end the box, which will in turn undefine the
+    % @end command again.
+    \expandafter\def\csname E#1\endcsname{\egroup\ignorespaces}%
+    %
+    % We are going to be parsing Texinfo commands.  Most cause no
+    % trouble when they are used incorrectly, but some commands do
+    % complicated argument parsing or otherwise get confused, so we
+    % undefine them.
+    %
+    % We can't do anything about stray @-signs, unfortunately;
+    % they'll produce `undefined control sequence' errors.
+    \ignoremorecommands
+    %
+    % Set the current font to be \nullfont, a TeX primitive, and define
+    % all the font commands to also use \nullfont.  We don't use
+    % dummy.tfm, as suggested in the TeXbook, because not all sites
+    % might have that installed.  Therefore, math mode will still
+    % produce output, but that should be an extremely small amount of
+    % stuff compared to the main input.
+    %
+    \nullfont
+    \let\tenrm = \nullfont  \let\tenit = \nullfont  \let\tensl = \nullfont
+    \let\tenbf = \nullfont  \let\tentt = \nullfont  \let\smallcaps = \nullfont
+    \let\tensf = \nullfont
+    %
+    % Don't complain when characters are missing from the fonts.
+    \tracinglostchars = 0
+    %
+    % Don't bother to do space factor calculations.
+    \frenchspacing
+    %
+    % Don't report underfull hboxes.
+    \hbadness = 10000
+    %
+    % Do minimal line-breaking.
+    \pretolerance = 10000
+    %
+    % Do not execute instructions in @tex
+    \def\tex{\doignore{tex}}
+}
+
+% @set VAR sets the variable VAR to an empty value.
+% @set VAR REST-OF-LINE sets VAR to the value REST-OF-LINE.
+%
+% Since we want to separate VAR from REST-OF-LINE (which might be
+% empty), we can't just use \parsearg; we have to insert a space of our
+% own to delimit the rest of the line, and then take it out again if we
+% didn't need it.
+%
+\def\set{\parsearg\setxxx}
+\def\setxxx#1{\setyyy#1 \endsetyyy}
+\def\setyyy#1 #2\endsetyyy{%
+  \def\temp{#2}%
+  \ifx\temp\empty \global\expandafter\let\csname SET#1\endcsname = \empty
+  \else \setzzz{#1}#2\endsetzzz % Remove the trailing space \setxxx inserted.
+  \fi
+}
+\def\setzzz#1#2 \endsetzzz{\expandafter\xdef\csname SET#1\endcsname{#2}}
+
+% @clear VAR clears (i.e., unsets) the variable VAR.
+%
+\def\clear{\parsearg\clearxxx}
+\def\clearxxx#1{\global\expandafter\let\csname SET#1\endcsname=\relax}
+
+% @value{foo} gets the text saved in variable foo.
+%
+\def\value#1{\expandafter
+		\ifx\csname SET#1\endcsname\relax
+			{\{No value for ``#1''\}}
+		\else \csname SET#1\endcsname \fi}
+
+% @ifset VAR ... @end ifset reads the `...' iff VAR has been defined
+% with @set.
+%
+\def\ifset{\parsearg\ifsetxxx}
+\def\ifsetxxx #1{%
+  \expandafter\ifx\csname SET#1\endcsname\relax
+    \expandafter\ifsetfail
+  \else
+    \expandafter\ifsetsucceed
+  \fi
+}
+\def\ifsetsucceed{\conditionalsucceed{ifset}}
+\def\ifsetfail{\nestedignore{ifset}}
+\defineunmatchedend{ifset}
+
+% @ifclear VAR ... @end ifclear reads the `...' iff VAR has never been
+% defined with @set, or has been undefined with @clear.
+%
+\def\ifclear{\parsearg\ifclearxxx}
+\def\ifclearxxx #1{%
+  \expandafter\ifx\csname SET#1\endcsname\relax
+    \expandafter\ifclearsucceed
+  \else
+    \expandafter\ifclearfail
+  \fi
+}
+\def\ifclearsucceed{\conditionalsucceed{ifclear}}
+\def\ifclearfail{\nestedignore{ifclear}}
+\defineunmatchedend{ifclear}
+
+% @iftex always succeeds; we read the text following, through @end
+% iftex).  But `@end iftex' should be valid only after an @iftex.
+%
+\def\iftex{\conditionalsucceed{iftex}}
+\defineunmatchedend{iftex}
+
+% We can't just want to start a group at @iftex (for example) and end it
+% at @end iftex, since then @set commands inside the conditional have no
+% effect (they'd get reverted at the end of the group).  So we must
+% define \Eiftex to redefine itself to be its previous value.  (We can't
+% just define it to fail again with an ``unmatched end'' error, since
+% the @ifset might be nested.)
+%
+\def\conditionalsucceed#1{%
+  \edef\temp{%
+    % Remember the current value of \E#1.
+    \let\nece{prevE#1} = \nece{E#1}%
+    %
+    % At the `@end #1', redefine \E#1 to be its previous value.
+    \def\nece{E#1}{\let\nece{E#1} = \nece{prevE#1}}%
+  }%
+  \temp
+}
+
+% We need to expand lots of \csname's, but we don't want to expand the
+% control sequences after we've constructed them.
+%
+\def\nece#1{\expandafter\noexpand\csname#1\endcsname}
+
+% @asis just yields its argument.  Used with @table, for example.
+%
+\def\asis#1{#1}
+
+% @math means output in math mode.
+% We don't use $'s directly in the definition of \math because control
+% sequences like \math are expanded when the toc file is written.  Then,
+% we read the toc file back, the $'s will be normal characters (as they
+% should be, according to the definition of Texinfo).  So we must use a
+% control sequence to switch into and out of math mode.
+%
+% This isn't quite enough for @math to work properly in indices, but it
+% seems unlikely it will ever be needed there.
+%
+\let\implicitmath = $
+\def\math#1{\implicitmath #1\implicitmath}
+
+% @bullet and @minus need the same treatment as @math, just above.
+\def\bullet{\implicitmath\ptexbullet\implicitmath}
+\def\minus{\implicitmath-\implicitmath}
+
+\def\node{\ENVcheck\parsearg\nodezzz}
+\def\nodezzz#1{\nodexxx [#1,]}
+\def\nodexxx[#1,#2]{\gdef\lastnode{#1}}
+\let\nwnode=\node
+\let\lastnode=\relax
+
+\def\donoderef{\ifx\lastnode\relax\else
+\expandafter\expandafter\expandafter\setref{\lastnode}\fi
+\let\lastnode=\relax}
+
+\def\unnumbnoderef{\ifx\lastnode\relax\else
+\expandafter\expandafter\expandafter\unnumbsetref{\lastnode}\fi
+\let\lastnode=\relax}
+
+\def\appendixnoderef{\ifx\lastnode\relax\else
+\expandafter\expandafter\expandafter\appendixsetref{\lastnode}\fi
+\let\lastnode=\relax}
+
+\let\refill=\relax
+
+% @setfilename is done at the beginning of every texinfo file.
+% So open here the files we need to have open while reading the input.
+% This makes it possible to make a .fmt file for texinfo.
+\def\setfilename{%
+   \readauxfile
+   \opencontents
+   \openindices
+   \fixbackslash  % Turn off hack to swallow `\input texinfo'.
+   \global\let\setfilename=\comment % Ignore extra @setfilename cmds.
+   \comment % Ignore the actual filename.
+}
+
+\outer\def\bye{\pagealignmacro\tracingstats=1\ptexend}
+
+\def\inforef #1{\inforefzzz #1,,,,**}
+\def\inforefzzz #1,#2,#3,#4**{See Info file \file{\ignorespaces #3{}},
+  node \samp{\ignorespaces#1{}}}
+
+\message{fonts,}
+
+% Font-change commands.
+
+% Texinfo supports the sans serif font style, which plain TeX does not.
+% So we set up a \sf analogous to plain's \rm, etc.
+\newfam\sffam
+\def\sf{\fam=\sffam \tensf}
+\let\li = \sf % Sometimes we call it \li, not \sf.
+
+%% Try out Computer Modern fonts at \magstephalf
+\let\mainmagstep=\magstephalf
+
+\ifx\bigger\relax
+\let\mainmagstep=\magstep1
+\font\textrm=cmr12
+\font\texttt=cmtt12
+\else
+\font\textrm=cmr10 scaled \mainmagstep
+\font\texttt=cmtt10 scaled \mainmagstep
+\fi
+% Instead of cmb10, you many want to use cmbx10.
+% cmbx10 is a prettier font on its own, but cmb10
+% looks better when embedded in a line with cmr10.
+\font\textbf=cmb10 scaled \mainmagstep
+\font\textit=cmti10 scaled \mainmagstep
+\font\textsl=cmsl10 scaled \mainmagstep
+\font\textsf=cmss10 scaled \mainmagstep
+\font\textsc=cmcsc10 scaled \mainmagstep
+\font\texti=cmmi10 scaled \mainmagstep
+\font\textsy=cmsy10 scaled \mainmagstep
+
+% A few fonts for @defun, etc.
+\font\defbf=cmbx10 scaled \magstep1 %was 1314
+\font\deftt=cmtt10 scaled \magstep1
+\def\df{\let\tentt=\deftt \let\tenbf = \defbf \bf}
+
+% Fonts for indices and small examples.
+% We actually use the slanted font rather than the italic,
+% because texinfo normally uses the slanted fonts for that.
+% Do not make many font distinctions in general in the index, since they
+% aren't very useful.
+\font\ninett=cmtt9
+\font\indrm=cmr9
+\font\indit=cmsl9
+\let\indsl=\indit
+\let\indtt=\ninett
+\let\indsf=\indrm
+\let\indbf=\indrm
+\let\indsc=\indrm
+\font\indi=cmmi9
+\font\indsy=cmsy9
+
+% Fonts for headings
+\font\chaprm=cmbx12 scaled \magstep2
+\font\chapit=cmti12 scaled \magstep2
+\font\chapsl=cmsl12 scaled \magstep2
+\font\chaptt=cmtt12 scaled \magstep2
+\font\chapsf=cmss12 scaled \magstep2
+\let\chapbf=\chaprm
+\font\chapsc=cmcsc10 scaled\magstep3
+\font\chapi=cmmi12 scaled \magstep2
+\font\chapsy=cmsy10 scaled \magstep3
+
+\font\secrm=cmbx12 scaled \magstep1
+\font\secit=cmti12 scaled \magstep1
+\font\secsl=cmsl12 scaled \magstep1
+\font\sectt=cmtt12 scaled \magstep1
+\font\secsf=cmss12 scaled \magstep1
+\font\secbf=cmbx12 scaled \magstep1
+\font\secsc=cmcsc10 scaled\magstep2
+\font\seci=cmmi12 scaled \magstep1
+\font\secsy=cmsy10 scaled \magstep2
+
+% \font\ssecrm=cmbx10 scaled \magstep1    % This size an font looked bad.
+% \font\ssecit=cmti10 scaled \magstep1    % The letters were too crowded.
+% \font\ssecsl=cmsl10 scaled \magstep1
+% \font\ssectt=cmtt10 scaled \magstep1
+% \font\ssecsf=cmss10 scaled \magstep1
+
+%\font\ssecrm=cmb10 scaled 1315	% Note the use of cmb rather than cmbx.
+%\font\ssecit=cmti10 scaled 1315	% Also, the size is a little larger than
+%\font\ssecsl=cmsl10 scaled 1315	% being scaled magstep1.
+%\font\ssectt=cmtt10 scaled 1315
+%\font\ssecsf=cmss10 scaled 1315
+
+%\let\ssecbf=\ssecrm
+
+\font\ssecrm=cmbx12 scaled \magstephalf
+\font\ssecit=cmti12 scaled \magstephalf
+\font\ssecsl=cmsl12 scaled \magstephalf
+\font\ssectt=cmtt12 scaled \magstephalf
+\font\ssecsf=cmss12 scaled \magstephalf
+\font\ssecbf=cmbx12 scaled \magstephalf
+\font\ssecsc=cmcsc10 scaled \magstep1
+\font\sseci=cmmi12 scaled \magstephalf
+\font\ssecsy=cmsy10 scaled \magstep1
+% The smallcaps and symbol fonts should actually be scaled \magstep1.5,
+% but that is not a standard magnification.
+
+% Fonts for title page:
+\font\titlerm = cmbx12 scaled \magstep3
+\let\authorrm = \secrm
+
+% In order for the font changes to affect most math symbols and letters,
+% we have to define the \textfont of the standard families.  Since
+% texinfo doesn't allow for producing subscripts and superscripts, we
+% don't bother to reset \scriptfont and \scriptscriptfont (which would
+% also require loading a lot more fonts).
+%
+\def\resetmathfonts{%
+  \textfont0 = \tenrm \textfont1 = \teni \textfont2 = \tensy
+  \textfont\itfam = \tenit \textfont\slfam = \tensl \textfont\bffam = \tenbf
+  \textfont\ttfam = \tentt \textfont\sffam = \tensf
+}
+
+
+% The font-changing commands redefine the meanings of \tenSTYLE, instead
+% of just \STYLE.  We do this so that font changes will continue to work
+% in math mode, where it is the current \fam that is relevant in most
+% cases, not the current.  Plain TeX does, for example,
+% \def\bf{\fam=\bffam \tenbf}  By redefining \tenbf, we obviate the need
+% to redefine \bf itself.
+\def\textfonts{%
+  \let\tenrm=\textrm \let\tenit=\textit \let\tensl=\textsl
+  \let\tenbf=\textbf \let\tentt=\texttt \let\smallcaps=\textsc
+  \let\tensf=\textsf \let\teni=\texti \let\tensy=\textsy
+  \resetmathfonts}
+\def\chapfonts{%
+  \let\tenrm=\chaprm \let\tenit=\chapit \let\tensl=\chapsl
+  \let\tenbf=\chapbf \let\tentt=\chaptt \let\smallcaps=\chapsc
+  \let\tensf=\chapsf \let\teni=\chapi \let\tensy=\chapsy
+  \resetmathfonts}
+\def\secfonts{%
+  \let\tenrm=\secrm \let\tenit=\secit \let\tensl=\secsl
+  \let\tenbf=\secbf \let\tentt=\sectt \let\smallcaps=\secsc
+  \let\tensf=\secsf \let\teni=\seci \let\tensy=\secsy
+  \resetmathfonts}
+\def\subsecfonts{%
+  \let\tenrm=\ssecrm \let\tenit=\ssecit \let\tensl=\ssecsl
+  \let\tenbf=\ssecbf \let\tentt=\ssectt \let\smallcaps=\ssecsc
+  \let\tensf=\ssecsf \let\teni=\sseci \let\tensy=\ssecsy
+  \resetmathfonts}
+\def\indexfonts{%
+  \let\tenrm=\indrm \let\tenit=\indit \let\tensl=\indsl
+  \let\tenbf=\indbf \let\tentt=\indtt \let\smallcaps=\indsc
+  \let\tensf=\indsf \let\teni=\indi \let\tensy=\indsy
+  \resetmathfonts}
+
+% Set up the default fonts, so we can use them for creating boxes.
+%
+\textfonts
+
+% Count depth in font-changes, for error checks
+\newcount\fontdepth \fontdepth=0
+
+% Fonts for short table of contents.
+\font\shortcontrm=cmr12
+\font\shortcontbf=cmbx12
+\font\shortcontsl=cmsl12
+
+%% Add scribe-like font environments, plus @l for inline lisp (usually sans
+%% serif) and @ii for TeX italic
+
+% \smartitalic{ARG} outputs arg in italics, followed by an italic correction
+% unless the following character is such as not to need one.
+\def\smartitalicx{\ifx\next,\else\ifx\next-\else\ifx\next.\else\/\fi\fi\fi}
+\def\smartitalic#1{{\sl #1}\futurelet\next\smartitalicx}
+
+\let\i=\smartitalic
+\let\var=\smartitalic
+\let\dfn=\smartitalic
+\let\emph=\smartitalic
+\let\cite=\smartitalic
+
+\def\b#1{{\bf #1}}
+\let\strong=\b
+
+% We can't just use \exhyphenpenalty, because that only has effect at
+% the end of a paragraph.  Restore normal hyphenation at the end of the
+% group within which \nohyphenation is presumably called.
+%
+\def\nohyphenation{\hyphenchar\font = -1  \aftergroup\restorehyphenation}
+\def\restorehyphenation{\hyphenchar\font = `- }
+
+\def\t#1{%
+  {\tt \nohyphenation \rawbackslash \frenchspacing #1}%
+  \null
+}
+\let\ttfont = \t
+%\def\samp #1{`{\tt \rawbackslash \frenchspacing #1}'\null}
+\def\samp #1{`\tclose{#1}'\null}
+\def\key #1{{\tt \nohyphenation \uppercase{#1}}\null}
+\def\ctrl #1{{\tt \rawbackslash \hat}#1}
+
+\let\file=\samp
+
+% @code is a modification of @t,
+% which makes spaces the same size as normal in the surrounding text.
+\def\tclose#1{%
+  {%
+    % Change normal interword space to be same as for the current font.
+    \spaceskip = \fontdimen2\font
+    %
+    % Switch to typewriter.
+    \tt
+    %
+    % But `\ ' produces the large typewriter interword space.
+    \def\ {{\spaceskip = 0pt{} }}%
+    %
+    % Turn off hyphenation.
+    \nohyphenation
+    %
+    \rawbackslash
+    \frenchspacing
+    #1%
+  }%
+  \null
+}
+
+% We *must* turn on hyphenation at `-' and `_' in \code.
+% Otherwise, it is too hard to avoid overful hboxes
+% in the Emacs manual, the Library manual, etc.
+
+% Unfortunately, TeX uses one parameter (\hyphenchar) to control
+% both hyphenation at - and hyphenation within words.
+% We must therefore turn them both off (\tclose does that)
+% and arrange explicitly to hyphenate an a dash.
+%  -- rms.
+{
+\catcode `\-=\active
+\catcode `\_=\active
+\global\def\code{\begingroup \catcode `\-=\active \let-\codedash \let_\codeunder \codex}
+}
+\def\codedash{-\discretionary{}{}{}}
+\def\codeunder{\normalunderscore\discretionary{}{}{}}
+\def\codex #1{\tclose{#1}\endgroup}
+
+%\let\exp=\tclose  %Was temporary
+
+% @kbd is like @code, except that if the argument is just one @key command,
+% then @kbd has no effect.
+
+\def\xkey{\key}
+\def\kbdfoo#1#2#3\par{\def\one{#1}\def\three{#3}\def\threex{??}%
+\ifx\one\xkey\ifx\threex\three \key{#2}%
+\else\tclose{\look}\fi
+\else\tclose{\look}\fi}
+
+% Typeset a dimension, e.g., `in' or `pt'.  The only reason for the
+% argument is to make the input look right: @dmn{pt} instead of
+% @dmn{}pt.
+%
+\def\dmn#1{\thinspace #1}
+
+\def\kbd#1{\def\look{#1}\expandafter\kbdfoo\look??\par}
+
+\def\l#1{{\li #1}\null}		%
+
+\def\r#1{{\rm #1}}		% roman font
+% Use of \lowercase was suggested.
+\def\sc#1{{\smallcaps#1}}	% smallcaps font
+\def\ii#1{{\it #1}}		% italic font
+
+\message{page headings,}
+
+\newskip\titlepagetopglue \titlepagetopglue = 1.5in
+\newskip\titlepagebottomglue \titlepagebottomglue = 2pc
+
+% First the title page.  Must do @settitle before @titlepage.
+\def\titlefont#1{{\titlerm #1}}
+
+\newif\ifseenauthor
+\newif\iffinishedtitlepage
+
+\def\shorttitlepage{\parsearg\shorttitlepagezzz}
+\def\shorttitlepagezzz #1{\begingroup\hbox{}\vskip 1.5in \chaprm \centerline{#1}%
+	\endgroup\page\hbox{}\page}
+
+\def\titlepage{\begingroup \parindent=0pt \textfonts
+   \let\subtitlerm=\tenrm
+% I deinstalled the following change because \cmr12 is undefined.
+% This change was not in the ChangeLog anyway.  --rms.
+%   \let\subtitlerm=\cmr12
+   \def\subtitlefont{\subtitlerm \normalbaselineskip = 13pt \normalbaselines}%
+   %
+   \def\authorfont{\authorrm \normalbaselineskip = 16pt \normalbaselines}%
+   %
+   % Leave some space at the very top of the page.
+   \vglue\titlepagetopglue
+   %
+   % Now you can print the title using @title.
+   \def\title{\parsearg\titlezzz}%
+   \def\titlezzz##1{\leftline{\titlefont{##1}}
+		    % print a rule at the page bottom also.
+		    \finishedtitlepagefalse
+		    \vskip4pt \hrule height 4pt \vskip4pt}%
+   % No rule at page bottom unless we print one at the top with @title.
+   \finishedtitlepagetrue
+   %
+   % Now you can put text using @subtitle.
+   \def\subtitle{\parsearg\subtitlezzz}%
+   \def\subtitlezzz##1{{\subtitlefont \rightline{##1}}}%
+   %
+   % @author should come last, but may come many times.
+   \def\author{\parsearg\authorzzz}%
+   \def\authorzzz##1{\ifseenauthor\else\vskip 0pt plus 1filll\seenauthortrue\fi
+      {\authorfont \leftline{##1}}}%
+   %
+   % Most title ``pages'' are actually two pages long, with space
+   % at the top of the second.  We don't want the ragged left on the second.
+   \let\oldpage = \page
+   \def\page{%
+      \iffinishedtitlepage\else
+	 \finishtitlepage
+      \fi
+      \oldpage
+      \let\page = \oldpage
+      \hbox{}}%
+%   \def\page{\oldpage \hbox{}}
+}
+
+\def\Etitlepage{%
+   \iffinishedtitlepage\else
+      \finishtitlepage
+   \fi
+   % It is important to do the page break before ending the group,
+   % because the headline and footline are only empty inside the group.
+   % If we use the new definition of \page, we always get a blank page
+   % after the title page, which we certainly don't want.
+   \oldpage
+   \endgroup
+   \HEADINGSon
+}
+
+\def\finishtitlepage{%
+   \vskip4pt \hrule height 2pt
+   \vskip\titlepagebottomglue
+   \finishedtitlepagetrue
+}
+
+%%% Set up page headings and footings.
+
+\let\thispage=\folio
+
+\newtoks \evenheadline    % Token sequence for heading line of even pages
+\newtoks \oddheadline     % Token sequence for heading line of odd pages
+\newtoks \evenfootline    % Token sequence for footing line of even pages
+\newtoks \oddfootline     % Token sequence for footing line of odd pages
+
+% Now make Tex use those variables
+\headline={{\textfonts\rm \ifodd\pageno \the\oddheadline
+                            \else \the\evenheadline \fi}}
+\footline={{\textfonts\rm \ifodd\pageno \the\oddfootline
+                            \else \the\evenfootline \fi}\HEADINGShook}
+\let\HEADINGShook=\relax
+
+% Commands to set those variables.
+% For example, this is what  @headings on  does
+% @evenheading @thistitle|@thispage|@thischapter
+% @oddheading @thischapter|@thispage|@thistitle
+% @evenfooting @thisfile||
+% @oddfooting ||@thisfile
+
+\def\evenheading{\parsearg\evenheadingxxx}
+\def\oddheading{\parsearg\oddheadingxxx}
+\def\everyheading{\parsearg\everyheadingxxx}
+
+\def\evenfooting{\parsearg\evenfootingxxx}
+\def\oddfooting{\parsearg\oddfootingxxx}
+\def\everyfooting{\parsearg\everyfootingxxx}
+
+{\catcode`\@=0 %
+
+\gdef\evenheadingxxx #1{\evenheadingyyy #1@|@|@|@|\finish}
+\gdef\evenheadingyyy #1@|#2@|#3@|#4\finish{%
+\global\evenheadline={\rlap{\centerline{#2}}\line{#1\hfil#3}}}
+
+\gdef\oddheadingxxx #1{\oddheadingyyy #1@|@|@|@|\finish}
+\gdef\oddheadingyyy #1@|#2@|#3@|#4\finish{%
+\global\oddheadline={\rlap{\centerline{#2}}\line{#1\hfil#3}}}
+
+\gdef\everyheadingxxx #1{\everyheadingyyy #1@|@|@|@|\finish}
+\gdef\everyheadingyyy #1@|#2@|#3@|#4\finish{%
+\global\evenheadline={\rlap{\centerline{#2}}\line{#1\hfil#3}}
+\global\oddheadline={\rlap{\centerline{#2}}\line{#1\hfil#3}}}
+
+\gdef\evenfootingxxx #1{\evenfootingyyy #1@|@|@|@|\finish}
+\gdef\evenfootingyyy #1@|#2@|#3@|#4\finish{%
+\global\evenfootline={\rlap{\centerline{#2}}\line{#1\hfil#3}}}
+
+\gdef\oddfootingxxx #1{\oddfootingyyy #1@|@|@|@|\finish}
+\gdef\oddfootingyyy #1@|#2@|#3@|#4\finish{%
+\global\oddfootline={\rlap{\centerline{#2}}\line{#1\hfil#3}}}
+
+\gdef\everyfootingxxx #1{\everyfootingyyy #1@|@|@|@|\finish}
+\gdef\everyfootingyyy #1@|#2@|#3@|#4\finish{%
+\global\evenfootline={\rlap{\centerline{#2}}\line{#1\hfil#3}}
+\global\oddfootline={\rlap{\centerline{#2}}\line{#1\hfil#3}}}
+%
+}% unbind the catcode of @.
+
+% @headings double	turns headings on for double-sided printing.
+% @headings single	turns headings on for single-sided printing.
+% @headings off		turns them off.
+% @headings on		same as @headings double, retained for compatibility.
+% @headings after	turns on double-sided headings after this page.
+% @headings doubleafter	turns on double-sided headings after this page.
+% @headings singleafter turns on single-sided headings after this page.
+% By default, they are off.
+
+\def\headings #1 {\csname HEADINGS#1\endcsname}
+
+\def\HEADINGSoff{
+\global\evenheadline={\hfil} \global\evenfootline={\hfil}
+\global\oddheadline={\hfil} \global\oddfootline={\hfil}}
+\HEADINGSoff
+% When we turn headings on, set the page number to 1.
+% For double-sided printing, put current file name in lower left corner,
+% chapter name on inside top of right hand pages, document
+% title on inside top of left hand pages, and page numbers on outside top
+% edge of all pages.
+\def\HEADINGSdouble{
+%\pagealignmacro
+\global\pageno=1
+\global\evenfootline={\hfil}
+\global\oddfootline={\hfil}
+\global\evenheadline={\line{\folio\hfil\thistitle}}
+\global\oddheadline={\line{\thischapter\hfil\folio}}
+}
+% For single-sided printing, chapter title goes across top left of page,
+% page number on top right.
+\def\HEADINGSsingle{
+%\pagealignmacro
+\global\pageno=1
+\global\evenfootline={\hfil}
+\global\oddfootline={\hfil}
+\global\evenheadline={\line{\thischapter\hfil\folio}}
+\global\oddheadline={\line{\thischapter\hfil\folio}}
+}
+\def\HEADINGSon{\HEADINGSdouble}
+
+\def\HEADINGSafter{\let\HEADINGShook=\HEADINGSdoublex}
+\let\HEADINGSdoubleafter=\HEADINGSafter
+\def\HEADINGSdoublex{%
+\global\evenfootline={\hfil}
+\global\oddfootline={\hfil}
+\global\evenheadline={\line{\folio\hfil\thistitle}}
+\global\oddheadline={\line{\thischapter\hfil\folio}}
+}
+
+\def\HEADINGSsingleafter{\let\HEADINGShook=\HEADINGSsinglex}
+\def\HEADINGSsinglex{%
+\global\evenfootline={\hfil}
+\global\oddfootline={\hfil}
+\global\evenheadline={\line{\thischapter\hfil\folio}}
+\global\oddheadline={\line{\thischapter\hfil\folio}}
+}
+
+% Subroutines used in generating headings
+% Produces Day Month Year style of output.
+\def\today{\number\day\space
+\ifcase\month\or
+January\or February\or March\or April\or May\or June\or
+July\or August\or September\or October\or November\or December\fi
+\space\number\year}
+
+% Use this if you want the Month Day, Year style of output.
+%\def\today{\ifcase\month\or
+%January\or February\or March\or April\or May\or June\or
+%July\or August\or September\or October\or November\or December\fi
+%\space\number\day, \number\year}
+
+% @settitle line...  specifies the title of the document, for headings
+% It generates no output of its own
+
+\def\thistitle{No Title}
+\def\settitle{\parsearg\settitlezzz}
+\def\settitlezzz #1{\gdef\thistitle{#1}}
+
+\message{tables,}
+
+% @tabs -- simple alignment
+
+% These don't work.  For one thing, \+ is defined as outer.
+% So these macros cannot even be defined.
+
+%\def\tabs{\parsearg\tabszzz}
+%\def\tabszzz #1{\settabs\+#1\cr}
+%\def\tabline{\parsearg\tablinezzz}
+%\def\tablinezzz #1{\+#1\cr}
+%\def\&{&}
+
+% Tables -- @table, @ftable, @vtable, @item(x), @kitem(x), @xitem(x).
+
+% default indentation of table text
+\newdimen\tableindent \tableindent=.8in
+% default indentation of @itemize and @enumerate text
+\newdimen\itemindent  \itemindent=.3in
+% margin between end of table item and start of table text.
+\newdimen\itemmargin  \itemmargin=.1in
+
+% used internally for \itemindent minus \itemmargin
+\newdimen\itemmax
+
+% Note @table, @vtable, and @vtable define @item, @itemx, etc., with
+% these defs.
+% They also define \itemindex
+% to index the item name in whatever manner is desired (perhaps none).
+
+\def\internalBitem{\smallbreak \parsearg\itemzzz}
+\def\internalBitemx{\par \parsearg\itemzzz}
+
+\def\internalBxitem "#1"{\def\xitemsubtopix{#1} \smallbreak \parsearg\xitemzzz}
+\def\internalBxitemx "#1"{\def\xitemsubtopix{#1} \par \parsearg\xitemzzz}
+
+\def\internalBkitem{\smallbreak \parsearg\kitemzzz}
+\def\internalBkitemx{\par \parsearg\kitemzzz}
+
+\def\kitemzzz #1{\dosubind {kw}{\code{#1}}{for {\bf \lastfunction}}%
+                 \itemzzz {#1}}
+
+\def\xitemzzz #1{\dosubind {kw}{\code{#1}}{for {\bf \xitemsubtopic}}%
+                 \itemzzz {#1}}
+
+\def\itemzzz #1{\begingroup %
+  \advance\hsize by -\rightskip
+  \advance\hsize by -\tableindent
+  \setbox0=\hbox{\itemfont{#1}}%
+  \itemindex{#1}%
+  \nobreak % This prevents a break before @itemx.
+  %
+  % Be sure we are not still in the middle of a paragraph.
+  {\parskip = 0in
+  \par
+  }%
+  %
+  % If the item text does not fit in the space we have, put it on a line
+  % by itself, and do not allow a page break either before or after that
+  % line.  We do not start a paragraph here because then if the next
+  % command is, e.g., @kindex, the whatsit would get put into the
+  % horizontal list on a line by itself, resulting in extra blank space.
+  \ifdim \wd0>\itemmax
+    \setbox0=\hbox{\hskip \leftskip \hskip -\tableindent \unhbox0}\box0
+    %
+    % We're going to be starting a paragraph, but we don't want the
+    % \parskip glue -- logically it's part of the @item we just started.
+    \nobreak \vskip-\parskip
+    %
+    % Stop a page break at the \parskip glue coming up.  Unfortunately
+    % we can't prevent a possible page break at the following
+    % \baselineskip glue.
+    \nobreak
+  \else
+    % The item text fits into the space.  Start a paragraph, so that the
+    % following text (if any) will end up on the same line.  Since that
+    % text will be indented by \tableindent, we make the item text be in
+    % a zero-width box.
+    \noindent
+    \rlap{\hskip -\tableindent\box0}%
+  \fi
+  \endgroup
+}
+
+\def\item{\errmessage{@item while not in a table}}
+\def\itemx{\errmessage{@itemx while not in a table}}
+\def\kitem{\errmessage{@kitem while not in a table}}
+\def\kitemx{\errmessage{@kitemx while not in a table}}
+\def\xitem{\errmessage{@xitem while not in a table}}
+\def\xitemx{\errmessage{@xitemx while not in a table}}
+
+%% Contains a kludge to get @end[description] to work
+\def\description{\tablez{\dontindex}{1}{}{}{}{}}
+
+\def\table{\begingroup\inENV\obeylines\obeyspaces\tablex}
+{\obeylines\obeyspaces%
+\gdef\tablex #1^^M{%
+\tabley\dontindex#1        \endtabley}}
+
+\def\ftable{\begingroup\inENV\obeylines\obeyspaces\ftablex}
+{\obeylines\obeyspaces%
+\gdef\ftablex #1^^M{%
+\tabley\fnitemindex#1        \endtabley
+\def\Eftable{\endgraf\afterenvbreak\endgroup}%
+\let\Etable=\relax}}
+
+\def\vtable{\begingroup\inENV\obeylines\obeyspaces\vtablex}
+{\obeylines\obeyspaces%
+\gdef\vtablex #1^^M{%
+\tabley\vritemindex#1        \endtabley
+\def\Evtable{\endgraf\afterenvbreak\endgroup}%
+\let\Etable=\relax}}
+
+\def\dontindex #1{}
+\def\fnitemindex #1{\doind {fn}{\code{#1}}}%
+\def\vritemindex #1{\doind {vr}{\code{#1}}}%
+
+{\obeyspaces %
+\gdef\tabley#1#2 #3 #4 #5 #6 #7\endtabley{\endgroup%
+\tablez{#1}{#2}{#3}{#4}{#5}{#6}}}
+
+\def\tablez #1#2#3#4#5#6{%
+\aboveenvbreak %
+\begingroup %
+\def\Edescription{\Etable}% Neccessary kludge.
+\let\itemindex=#1%
+\ifnum 0#3>0 \advance \leftskip by #3\mil \fi %
+\ifnum 0#4>0 \tableindent=#4\mil \fi %
+\ifnum 0#5>0 \advance \rightskip by #5\mil \fi %
+\def\itemfont{#2}%
+\itemmax=\tableindent %
+\advance \itemmax by -\itemmargin %
+\advance \leftskip by \tableindent %
+\exdentamount=\tableindent
+\parindent = 0pt
+\parskip = \smallskipamount
+\ifdim \parskip=0pt \parskip=2pt \fi%
+\def\Etable{\endgraf\afterenvbreak\endgroup}%
+\let\item = \internalBitem %
+\let\itemx = \internalBitemx %
+\let\kitem = \internalBkitem %
+\let\kitemx = \internalBkitemx %
+\let\xitem = \internalBxitem %
+\let\xitemx = \internalBxitemx %
+}
+
+% This is the counter used by @enumerate, which is really @itemize
+
+\newcount \itemno
+
+\def\itemize{\parsearg\itemizezzz}
+
+\def\itemizezzz #1{%
+  \begingroup % ended by the @end itemsize
+  \itemizey {#1}{\Eitemize}
+}
+
+\def\itemizey #1#2{%
+\aboveenvbreak %
+\itemmax=\itemindent %
+\advance \itemmax by -\itemmargin %
+\advance \leftskip by \itemindent %
+\exdentamount=\itemindent
+\parindent = 0pt %
+\parskip = \smallskipamount %
+\ifdim \parskip=0pt \parskip=2pt \fi%
+\def#2{\endgraf\afterenvbreak\endgroup}%
+\def\itemcontents{#1}%
+\let\item=\itemizeitem}
+
+% Set sfcode to normal for the chars that usually have another value.
+% These are `.?!:;,'
+\def\frenchspacing{\sfcode46=1000 \sfcode63=1000 \sfcode33=1000
+  \sfcode58=1000 \sfcode59=1000 \sfcode44=1000 }
+
+% \splitoff TOKENS\endmark defines \first to be the first token in
+% TOKENS, and \rest to be the remainder.
+%
+\def\splitoff#1#2\endmark{\def\first{#1}\def\rest{#2}}%
+
+% Allow an optional argument of an uppercase letter, lowercase letter,
+% or number, to specify the first label in the enumerated list.  No
+% argument is the same as `1'.
+%
+\def\enumerate{\parsearg\enumeratezzz}
+\def\enumeratezzz #1{\enumeratey #1  \endenumeratey}
+\def\enumeratey #1 #2\endenumeratey{%
+  \begingroup % ended by the @end enumerate
+  %
+  % If we were given no argument, pretend we were given `1'.
+  \def\thearg{#1}%
+  \ifx\thearg\empty \def\thearg{1}\fi
+  %
+  % Detect if the argument is a single token.  If so, it might be a
+  % letter.  Otherwise, the only valid thing it can be is a number.
+  % (We will always have one token, because of the test we just made.
+  % This is a good thing, since \splitoff doesn't work given nothing at
+  % all -- the first parameter is undelimited.)
+  \expandafter\splitoff\thearg\endmark
+  \ifx\rest\empty
+    % Only one token in the argument.  It could still be anything.
+    % A ``lowercase letter'' is one whose \lccode is nonzero.
+    % An ``uppercase letter'' is one whose \lccode is both nonzero, and
+    %   not equal to itself.
+    % Otherwise, we assume it's a number.
+    %
+    % We need the \relax at the end of the \ifnum lines to stop TeX from
+    % continuing to look for a <number>.
+    %
+    \ifnum\lccode\expandafter`\thearg=0\relax
+      \numericenumerate % a number (we hope)
+    \else
+      % It's a letter.
+      \ifnum\lccode\expandafter`\thearg=\expandafter`\thearg\relax
+        \lowercaseenumerate % lowercase letter
+      \else
+        \uppercaseenumerate % uppercase letter
+      \fi
+    \fi
+  \else
+    % Multiple tokens in the argument.  We hope it's a number.
+    \numericenumerate
+  \fi
+}
+
+% An @enumerate whose labels are integers.  The starting integer is
+% given in \thearg.
+%
+\def\numericenumerate{%
+  \itemno = \thearg
+  \startenumeration{\the\itemno}%
+}
+
+% The starting (lowercase) letter is in \thearg.
+\def\lowercaseenumerate{%
+  \itemno = \expandafter`\thearg
+  \startenumeration{%
+    % Be sure we're not beyond the end of the alphabet.
+    \ifnum\itemno=0
+      \errmessage{No more lowercase letters in @enumerate; get a bigger
+                  alphabet}%
+    \fi
+    \char\lccode\itemno
+  }%
+}
+
+% The starting (uppercase) letter is in \thearg.
+\def\uppercaseenumerate{%
+  \itemno = \expandafter`\thearg
+  \startenumeration{%
+    % Be sure we're not beyond the end of the alphabet.
+    \ifnum\itemno=0
+      \errmessage{No more uppercase letters in @enumerate; get a bigger
+                  alphabet}
+    \fi
+    \char\uccode\itemno
+  }%
+}
+
+% Call itemizey, adding a period to the first argument and supplying the
+% common last two arguments.  Also subtract one from the initial value in
+% \itemno, since @item increments \itemno.
+%
+\def\startenumeration#1{%
+  \advance\itemno by -1
+  \itemizey{#1.}\Eenumerate\flushcr
+}
+
+% @alphaenumerate and @capsenumerate are abbreviations for giving an arg
+% to @enumerate.
+%
+\def\alphaenumerate{\enumerate{a}}
+\def\capsenumerate{\enumerate{A}}
+\def\Ealphaenumerate{\Eenumerate}
+\def\Ecapsenumerate{\Eenumerate}
+
+% Definition of @item while inside @itemize.
+
+\def\itemizeitem{%
+\advance\itemno by 1
+{\let\par=\endgraf \smallbreak}%
+\ifhmode \errmessage{\in hmode at itemizeitem}\fi
+{\parskip=0in \hskip 0pt
+\hbox to 0pt{\hss \itemcontents\hskip \itemmargin}%
+\vadjust{\penalty 1200}}%
+\flushcr}
+
+\message{indexing,}
+% Index generation facilities
+
+% Define \newwrite to be identical to plain tex's \newwrite
+% except not \outer, so it can be used within \newindex.
+{\catcode`\@=11
+\gdef\newwrite{\alloc@7\write\chardef\sixt@@n}}
+
+% \newindex {foo} defines an index named foo.
+% It automatically defines \fooindex such that
+% \fooindex ...rest of line... puts an entry in the index foo.
+% It also defines \fooindfile to be the number of the output channel for
+% the file that	accumulates this index.  The file's extension is foo.
+% The name of an index should be no more than 2 characters long
+% for the sake of vms.
+
+\def\newindex #1{
+\expandafter\newwrite \csname#1indfile\endcsname% Define number for output file
+\openout \csname#1indfile\endcsname \jobname.#1	% Open the file
+\expandafter\xdef\csname#1index\endcsname{%	% Define \xxxindex
+\noexpand\doindex {#1}}
+}
+
+% @defindex foo  ==  \newindex{foo}
+
+\def\defindex{\parsearg\newindex}
+
+% Define @defcodeindex, like @defindex except put all entries in @code.
+
+\def\newcodeindex #1{
+\expandafter\newwrite \csname#1indfile\endcsname% Define number for output file
+\openout \csname#1indfile\endcsname \jobname.#1	% Open the file
+\expandafter\xdef\csname#1index\endcsname{%	% Define \xxxindex
+\noexpand\docodeindex {#1}}
+}
+
+\def\defcodeindex{\parsearg\newcodeindex}
+
+% @synindex foo bar    makes index foo feed into index bar.
+% Do this instead of @defindex foo if you don't want it as a separate index.
+\def\synindex #1 #2 {%
+\expandafter\let\expandafter\synindexfoo\expandafter=\csname#2indfile\endcsname
+\expandafter\let\csname#1indfile\endcsname=\synindexfoo
+\expandafter\xdef\csname#1index\endcsname{%	% Define \xxxindex
+\noexpand\doindex {#2}}%
+}
+
+% @syncodeindex foo bar   similar, but put all entries made for index foo
+% inside @code.
+\def\syncodeindex #1 #2 {%
+\expandafter\let\expandafter\synindexfoo\expandafter=\csname#2indfile\endcsname
+\expandafter\let\csname#1indfile\endcsname=\synindexfoo
+\expandafter\xdef\csname#1index\endcsname{%	% Define \xxxindex
+\noexpand\docodeindex {#2}}%
+}
+
+% Define \doindex, the driver for all \fooindex macros.
+% Argument #1 is generated by the calling \fooindex macro,
+%  and it is "foo", the name of the index.
+
+% \doindex just uses \parsearg; it calls \doind for the actual work.
+% This is because \doind is more useful to call from other macros.
+
+% There is also \dosubind {index}{topic}{subtopic}
+% which makes an entry in a two-level index such as the operation index.
+
+\def\doindex#1{\edef\indexname{#1}\parsearg\singleindexer}
+\def\singleindexer #1{\doind{\indexname}{#1}}
+
+% like the previous two, but they put @code around the argument.
+\def\docodeindex#1{\edef\indexname{#1}\parsearg\singlecodeindexer}
+\def\singlecodeindexer #1{\doind{\indexname}{\code{#1}}}
+
+\def\indexdummies{%
+\def\_{{\realbackslash _}}%
+\def\w{\realbackslash w }%
+\def\bf{\realbackslash bf }%
+\def\rm{\realbackslash rm }%
+\def\sl{\realbackslash sl }%
+\def\sf{\realbackslash sf}%
+\def\tt{\realbackslash tt}%
+\def\gtr{\realbackslash gtr}%
+\def\less{\realbackslash less}%
+\def\hat{\realbackslash hat}%
+\def\char{\realbackslash char}%
+\def\TeX{\realbackslash TeX}%
+\def\dots{\realbackslash dots }%
+\def\copyright{\realbackslash copyright }%
+\def\tclose##1{\realbackslash tclose {##1}}%
+\def\code##1{\realbackslash code {##1}}%
+\def\samp##1{\realbackslash samp {##1}}%
+\def\t##1{\realbackslash r {##1}}%
+\def\r##1{\realbackslash r {##1}}%
+\def\i##1{\realbackslash i {##1}}%
+\def\b##1{\realbackslash b {##1}}%
+\def\cite##1{\realbackslash cite {##1}}%
+\def\key##1{\realbackslash key {##1}}%
+\def\file##1{\realbackslash file {##1}}%
+\def\var##1{\realbackslash var {##1}}%
+\def\kbd##1{\realbackslash kbd {##1}}%
+\def\dfn##1{\realbackslash dfn {##1}}%
+\def\emph##1{\realbackslash emph {##1}}%
+}
+
+% \indexnofonts no-ops all font-change commands.
+% This is used when outputting the strings to sort the index by.
+\def\indexdummyfont#1{#1}
+\def\indexdummytex{TeX}
+\def\indexdummydots{...}
+
+\def\indexnofonts{%
+\let\w=\indexdummyfont
+\let\t=\indexdummyfont
+\let\r=\indexdummyfont
+\let\i=\indexdummyfont
+\let\b=\indexdummyfont
+\let\emph=\indexdummyfont
+\let\strong=\indexdummyfont
+\let\cite=\indexdummyfont
+\let\sc=\indexdummyfont
+%Don't no-op \tt, since it isn't a user-level command
+% and is used in the definitions of the active chars like <, >, |...
+%\let\tt=\indexdummyfont
+\let\tclose=\indexdummyfont
+\let\code=\indexdummyfont
+\let\file=\indexdummyfont
+\let\samp=\indexdummyfont
+\let\kbd=\indexdummyfont
+\let\key=\indexdummyfont
+\let\var=\indexdummyfont
+\let\TeX=\indexdummytex
+\let\dots=\indexdummydots
+}
+
+% To define \realbackslash, we must make \ not be an escape.
+% We must first make another character (@) an escape
+% so we do not become unable to do a definition.
+
+{\catcode`\@=0 \catcode`\\=\other
+@gdef@realbackslash{\}}
+
+\let\indexbackslash=0  %overridden during \printindex.
+
+\def\doind #1#2{%
+{\count10=\lastpenalty %
+{\indexdummies % Must do this here, since \bf, etc expand at this stage
+\escapechar=`\\%
+{\let\folio=0% Expand all macros now EXCEPT \folio
+\def\rawbackslashxx{\indexbackslash}% \indexbackslash isn't defined now
+% so it will be output as is; and it will print as backslash in the indx.
+%
+% Now process the index-string once, with all font commands turned off,
+% to get the string to sort the index by.
+{\indexnofonts
+\xdef\temp1{#2}%
+}%
+% Now produce the complete index entry.  We process the index-string again,
+% this time with font commands expanded, to get what to print in the index.
+\edef\temp{%
+\write \csname#1indfile\endcsname{%
+\realbackslash entry {\temp1}{\folio}{#2}}}%
+\temp }%
+}\penalty\count10}}
+
+\def\dosubind #1#2#3{%
+{\count10=\lastpenalty %
+{\indexdummies % Must do this here, since \bf, etc expand at this stage
+\escapechar=`\\%
+{\let\folio=0%
+\def\rawbackslashxx{\indexbackslash}%
+%
+% Now process the index-string once, with all font commands turned off,
+% to get the string to sort the index by.
+{\indexnofonts
+\xdef\temp1{#2 #3}%
+}%
+% Now produce the complete index entry.  We process the index-string again,
+% this time with font commands expanded, to get what to print in the index.
+\edef\temp{%
+\write \csname#1indfile\endcsname{%
+\realbackslash entry {\temp1}{\folio}{#2}{#3}}}%
+\temp }%
+}\penalty\count10}}
+
+% The index entry written in the file actually looks like
+%  \entry {sortstring}{page}{topic}
+% or
+%  \entry {sortstring}{page}{topic}{subtopic}
+% The texindex program reads in these files and writes files
+% containing these kinds of lines:
+%  \initial {c}
+%     before the first topic whose initial is c
+%  \entry {topic}{pagelist}
+%     for a topic that is used without subtopics
+%  \primary {topic}
+%     for the beginning of a topic that is used with subtopics
+%  \secondary {subtopic}{pagelist}
+%     for each subtopic.
+
+% Define the user-accessible indexing commands
+% @findex, @vindex, @kindex, @cindex.
+
+\def\findex {\fnindex}
+\def\kindex {\kyindex}
+\def\cindex {\cpindex}
+\def\vindex {\vrindex}
+\def\tindex {\tpindex}
+\def\pindex {\pgindex}
+
+\def\cindexsub {\begingroup\obeylines\cindexsub}
+{\obeylines %
+\gdef\cindexsub "#1" #2^^M{\endgroup %
+\dosubind{cp}{#2}{#1}}}
+
+% Define the macros used in formatting output of the sorted index material.
+
+% This is what you call to cause a particular index to get printed.
+% Write
+% @unnumbered Function Index
+% @printindex fn
+
+\def\printindex{\parsearg\doprintindex}
+
+\def\doprintindex#1{%
+  \tex
+  \dobreak \chapheadingskip {10000}
+  \catcode`\%=\other\catcode`\&=\other\catcode`\#=\other
+  \catcode`\$=\other\catcode`\_=\other
+  \catcode`\~=\other
+  %
+  % The following don't help, since the chars were translated
+  % when the raw index was written, and their fonts were discarded
+  % due to \indexnofonts.
+  %\catcode`\"=\active
+  %\catcode`\^=\active
+  %\catcode`\_=\active
+  %\catcode`\|=\active
+  %\catcode`\<=\active
+  %\catcode`\>=\active
+  % %
+  \def\indexbackslash{\rawbackslashxx}
+  \indexfonts\rm \tolerance=9500 \advance\baselineskip -1pt
+  \begindoublecolumns
+  %
+  % See if the index file exists and is nonempty.
+  \openin 1 \jobname.#1s
+  \ifeof 1
+    % \enddoublecolumns gets confused if there is no text in the index,
+    % and it loses the chapter title and the aux file entries for the
+    % index.  The easiest way to prevent this problem is to make sure
+    % there is some text.
+    (Index is nonexistent)
+    \else
+    %
+    % If the index file exists but is empty, then \openin leaves \ifeof
+    % false.  We have to make TeX try to read something from the file, so
+    % it can discover if there is anything in it.
+    \read 1 to \temp
+    \ifeof 1
+      (Index is empty)
+    \else
+      \input \jobname.#1s
+    \fi
+  \fi
+  \closein 1
+  \enddoublecolumns
+  \Etex
+}
+
+% These macros are used by the sorted index file itself.
+% Change them to control the appearance of the index.
+
+% Same as \bigskipamount except no shrink.
+% \balancecolumns gets confused if there is any shrink.
+\newskip\initialskipamount \initialskipamount 12pt plus4pt
+
+\def\initial #1{%
+{\let\tentt=\sectt \let\tt=\sectt \let\sf=\sectt
+\ifdim\lastskip<\initialskipamount
+\removelastskip \penalty-200 \vskip \initialskipamount\fi
+\line{\secbf#1\hfill}\kern 2pt\penalty10000}}
+
+% This typesets a paragraph consisting of #1, dot leaders, and then #2
+% flush to the right margin.  It is used for index and table of contents
+% entries.  The paragraph is indented by \leftskip.
+%
+\def\entry #1#2{\begingroup
+  %
+  % Start a new paragraph if necessary, so our assignments below can't
+  % affect previous text.
+  \par
+  %
+  % Do not fill out the last line with white space.
+  \parfillskip = 0in
+  %
+  % No extra space above this paragraph.
+  \parskip = 0in
+  %
+  % Do not prefer a separate line ending with a hyphen to fewer lines.
+  \finalhyphendemerits = 0
+  %
+  % \hangindent is only relevant when the entry text and page number
+  % don't both fit on one line.  In that case, bob suggests starting the
+  % dots pretty far over on the line.  Unfortunately, a large
+  % indentation looks wrong when the entry text itself is broken across
+  % lines.  So we use a small indentation and put up with long leaders.
+  %
+  % \hangafter is reset to 1 (which is the value we want) at the start
+  % of each paragraph, so we need not do anything with that.
+  \hangindent=2em
+  %
+  % When the entry text needs to be broken, just fill out the first line
+  % with blank space.
+  \rightskip = 0pt plus1fil
+  %
+  % Start a ``paragraph'' for the index entry so the line breaking
+  % parameters we've set above will have an effect.
+  \noindent
+  %
+  % Insert the text of the index entry.  TeX will do line-breaking on it.
+  #1%
+  %
+  % If we must, put the page number on a line of its own, and fill out
+  % this line with blank space.  (The \hfil is overwhelmed with the
+  % fill leaders glue in \indexdotfill if the page number does fit.)
+  \hfil\penalty50
+  \null\nobreak\indexdotfill % Have leaders before the page number.
+  %
+  % The `\ ' here is removed by the implicit \unskip that TeX does as
+  % part of (the primitive) \par.  Without it, a spurious underfull
+  % \hbox ensues.
+  \ #2% The page number ends the paragraph.
+  \par
+\endgroup}
+
+% Like \dotfill except takes at least 1 em.
+\def\indexdotfill{\cleaders
+  \hbox{$\mathsurround=0pt \mkern1.5mu . \mkern1.5mu$}\hskip 1em plus 1fill}
+
+\def\primary #1{\line{#1\hfil}}
+
+\newskip\secondaryindent \secondaryindent=0.5cm
+
+\def\secondary #1#2{
+{\parfillskip=0in \parskip=0in
+\hangindent =1in \hangafter=1
+\noindent\hskip\secondaryindent\hbox{#1}\indexdotfill #2\par
+}}
+
+%% Define two-column mode, which is used in indexes.
+%% Adapted from the TeXbook, page 416.
+\catcode `\@=11
+
+\newbox\partialpage
+
+\newdimen\doublecolumnhsize
+
+\def\begindoublecolumns{\begingroup
+  % Grab any single-column material above us.
+  \output = {\global\setbox\partialpage
+    =\vbox{\unvbox255\kern -\topskip \kern \baselineskip}}%
+  \eject
+  %
+  % Now switch to the double-column output routine.
+  \output={\doublecolumnout}%
+  %
+  % Change the page size parameters.  We could do this once outside this
+  % routine, in each of @smallbook, @afourpaper, and the default 8.5x11
+  % format, but then we repeat the same computation.  Repeating a couple
+  % of assignments once per index is clearly meaningless for the
+  % execution time, so we may as well do it once.
+  %
+  % First we halve the line length, less a little for the gutter between
+  % the columns.  We compute the gutter based on the line length, so it
+  % changes automatically with the paper format.  The magic constant
+  % below is chosen so that the gutter has the same value (well, +- <
+  % 1pt) as it did when we hard-coded it.
+  %
+  % We put the result in a separate register, \doublecolumhsize, so we
+  % can restore it in \pagesofar, after \hsize itself has (potentially)
+  % been clobbered.
+  %
+  \doublecolumnhsize = \hsize
+    \advance\doublecolumnhsize by -.04154\hsize
+    \divide\doublecolumnhsize by 2
+  \hsize = \doublecolumnhsize
+  %
+  % Double the \vsize as well.  (We don't need a separate register here,
+  % since nobody clobbers \vsize.)
+  \vsize = 2\vsize
+  \doublecolumnpagegoal
+}
+
+\def\enddoublecolumns{\eject \endgroup \pagegoal=\vsize \unvbox\partialpage}
+
+\def\doublecolumnsplit{\splittopskip=\topskip \splitmaxdepth=\maxdepth
+  \global\dimen@=\pageheight \global\advance\dimen@ by-\ht\partialpage
+  \global\setbox1=\vsplit255 to\dimen@ \global\setbox0=\vbox{\unvbox1}
+  \global\setbox3=\vsplit255 to\dimen@ \global\setbox2=\vbox{\unvbox3}
+  \ifdim\ht0>\dimen@ \setbox255=\vbox{\unvbox0\unvbox2} \global\setbox255=\copy5 \fi
+  \ifdim\ht2>\dimen@ \setbox255=\vbox{\unvbox0\unvbox2} \global\setbox255=\copy5 \fi
+}
+\def\doublecolumnpagegoal{%
+  \dimen@=\vsize \advance\dimen@ by-2\ht\partialpage \global\pagegoal=\dimen@
+}
+\def\pagesofar{\unvbox\partialpage %
+  \hsize=\doublecolumnhsize % have to restore this since output routine
+  \wd0=\hsize \wd2=\hsize \hbox to\pagewidth{\box0\hfil\box2}}
+\def\doublecolumnout{%
+  \setbox5=\copy255
+  {\vbadness=10000 \doublecolumnsplit}
+  \ifvbox255
+    \setbox0=\vtop to\dimen@{\unvbox0}
+    \setbox2=\vtop to\dimen@{\unvbox2}
+    \onepageout\pagesofar \unvbox255 \penalty\outputpenalty
+  \else
+    \setbox0=\vbox{\unvbox5}
+    \ifvbox0
+      \dimen@=\ht0 \advance\dimen@ by\topskip \advance\dimen@ by-\baselineskip
+      \divide\dimen@ by2 \splittopskip=\topskip \splitmaxdepth=\maxdepth
+      {\vbadness=10000
+	\loop \global\setbox5=\copy0
+          \setbox1=\vsplit5 to\dimen@
+          \setbox3=\vsplit5 to\dimen@
+          \ifvbox5 \global\advance\dimen@ by1pt \repeat
+        \setbox0=\vbox to\dimen@{\unvbox1}
+        \setbox2=\vbox to\dimen@{\unvbox3}
+        \global\setbox\partialpage=\vbox{\pagesofar}
+        \doublecolumnpagegoal
+      }
+    \fi
+  \fi
+}
+
+\catcode `\@=\other
+\message{sectioning,}
+% Define chapters, sections, etc.
+
+\newcount \chapno
+\newcount \secno        \secno=0
+\newcount \subsecno     \subsecno=0
+\newcount \subsubsecno  \subsubsecno=0
+
+% This counter is funny since it counts through charcodes of letters A, B, ...
+\newcount \appendixno  \appendixno = `\@
+\def\appendixletter{\char\the\appendixno}
+
+\newwrite \contentsfile
+% This is called from \setfilename.
+\def\opencontents{\openout \contentsfile = \jobname.toc}
+
+% Each @chapter defines this as the name of the chapter.
+% page headings and footings can use it.  @section does likewise
+
+\def\thischapter{} \def\thissection{}
+\def\seccheck#1{\if \pageno<0 %
+\errmessage{@#1 not allowed after generating table of contents}\fi
+%
+}
+
+\def\chapternofonts{%
+\let\rawbackslash=\relax%
+\let\frenchspacing=\relax%
+\def\result{\realbackslash result}
+\def\equiv{\realbackslash equiv}
+\def\expansion{\realbackslash expansion}
+\def\print{\realbackslash print}
+\def\TeX{\realbackslash TeX}
+\def\dots{\realbackslash dots}
+\def\copyright{\realbackslash copyright}
+\def\tt{\realbackslash tt}
+\def\bf{\realbackslash bf }
+\def\w{\realbackslash w}
+\def\less{\realbackslash less}
+\def\gtr{\realbackslash gtr}
+\def\hat{\realbackslash hat}
+\def\char{\realbackslash char}
+\def\tclose##1{\realbackslash tclose {##1}}
+\def\code##1{\realbackslash code {##1}}
+\def\samp##1{\realbackslash samp {##1}}
+\def\r##1{\realbackslash r {##1}}
+\def\b##1{\realbackslash b {##1}}
+\def\key##1{\realbackslash key {##1}}
+\def\file##1{\realbackslash file {##1}}
+\def\kbd##1{\realbackslash kbd {##1}}
+% These are redefined because @smartitalic wouldn't work inside xdef.
+\def\i##1{\realbackslash i {##1}}
+\def\cite##1{\realbackslash cite {##1}}
+\def\var##1{\realbackslash var {##1}}
+\def\emph##1{\realbackslash emph {##1}}
+\def\dfn##1{\realbackslash dfn {##1}}
+}
+
+\newcount\absseclevel % used to calculate proper heading level
+\newcount\secbase\secbase=0 % @raise/lowersections modify this count
+
+% @raisesections: treat @section as chapter, @subsection as section, etc.
+\def\raisesections{\global\advance\secbase by -1}
+\let\up=\raisesections % original BFox name
+
+% @lowersections: treat @chapter as section, @section as subsection, etc.
+\def\lowersections{\global\advance\secbase by 1}
+\let\down=\lowersections % original BFox name
+
+% Choose a numbered-heading macro
+% #1 is heading level if unmodified by @raisesections or @lowersections
+% #2 is text for heading
+\def\numhead#1#2{\absseclevel=\secbase\advance\absseclevel by #1
+\ifcase\absseclevel
+  \chapterzzz{#2}
+\or
+  \seczzz{#2}
+\or
+  \numberedsubseczzz{#2}
+\or
+  \numberedsubsubseczzz{#2}
+\else
+  \ifnum \absseclevel<0
+    \chapterzzz{#2}
+  \else
+    \numberedsubsubseczzz{#2}
+  \fi
+\fi
+}
+
+% like \numhead, but chooses appendix heading levels
+\def\apphead#1#2{\absseclevel=\secbase\advance\absseclevel by #1
+\ifcase\absseclevel
+  \appendixzzz{#2}
+\or
+  \appendixsectionzzz{#2}
+\or
+  \appendixsubseczzz{#2}
+\or
+  \appendixsubsubseczzz{#2}
+\else
+  \ifnum \absseclevel<0
+    \appendixzzz{#2}
+  \else
+    \appendixsubsubseczzz{#2}
+  \fi
+\fi
+}
+
+% like \numhead, but chooses numberless heading levels
+\def\unnmhead#1#2{\absseclevel=\secbase\advance\absseclevel by #1
+\ifcase\absseclevel
+  \unnumberedzzz{#2}
+\or
+  \unnumberedseczzz{#2}
+\or
+  \unnumberedsubseczzz{#2}
+\or
+  \unnumberedsubsubseczzz{#2}
+\else
+  \ifnum \absseclevel<0
+    \unnumberedzzz{#2}
+  \else
+    \unnumberedsubsubseczzz{#2}
+  \fi
+\fi
+}
+
+
+\def\thischaptername{No Chapter Title}
+\outer\def\chapter{\parsearg\chapteryyy}
+\def\chapteryyy #1{\numhead0{#1}} % normally numhead0 calls chapterzzz
+\def\chapterzzz #1{\seccheck{chapter}%
+\secno=0 \subsecno=0 \subsubsecno=0
+\global\advance \chapno by 1 \message{Chapter \the\chapno}%
+\chapmacro {#1}{\the\chapno}%
+\gdef\thissection{#1}%
+\gdef\thischaptername{#1}%
+% We don't substitute the actual chapter name into \thischapter
+% because we don't want its macros evaluated now.
+\xdef\thischapter{Chapter \the\chapno: \noexpand\thischaptername}%
+{\chapternofonts%
+\edef\temp{{\realbackslash chapentry {#1}{\the\chapno}{\noexpand\folio}}}%
+\escapechar=`\\%
+\write \contentsfile \temp  %
+\donoderef %
+\global\let\section = \numberedsec
+\global\let\subsection = \numberedsubsec
+\global\let\subsubsection = \numberedsubsubsec
+}}
+
+\outer\def\appendix{\parsearg\appendixyyy}
+\def\appendixyyy #1{\apphead0{#1}} % normally apphead0 calls appendixzzz
+\def\appendixzzz #1{\seccheck{appendix}%
+\secno=0 \subsecno=0 \subsubsecno=0
+\global\advance \appendixno by 1 \message{Appendix \appendixletter}%
+\chapmacro {#1}{Appendix \appendixletter}%
+\gdef\thissection{#1}%
+\gdef\thischaptername{#1}%
+\xdef\thischapter{Appendix \appendixletter: \noexpand\thischaptername}%
+{\chapternofonts%
+\edef\temp{{\realbackslash chapentry
+  {#1}{Appendix \appendixletter}{\noexpand\folio}}}%
+\escapechar=`\\%
+\write \contentsfile \temp  %
+\appendixnoderef %
+\global\let\section = \appendixsec
+\global\let\subsection = \appendixsubsec
+\global\let\subsubsection = \appendixsubsubsec
+}}
+
+\outer\def\top{\parsearg\unnumberedyyy}
+\outer\def\unnumbered{\parsearg\unnumberedyyy}
+\def\unnumberedyyy #1{\unnmhead0{#1}} % normally unnmhead0 calls unnumberedzzz
+\def\unnumberedzzz #1{\seccheck{unnumbered}%
+\secno=0 \subsecno=0 \subsubsecno=0
+%
+% This used to be simply \message{#1}, but TeX fully expands the
+% argument to \message.  Therefore, if #1 contained @-commands, TeX
+% expanded them.  For example, in `@unnumbered The @cite{Book}', TeX
+% expanded @cite (which turns out to cause errors because \cite is meant
+% to be executed, not expanded).
+%
+% Anyway, we don't want the fully-expanded definition of @cite to appear
+% as a result of the \message, we just want `@cite' itself.  We use
+% \the<toks register> to achieve this: TeX expands \the<toks> only once,
+% simply yielding the contents of the <toks register>.
+\toks0 = {#1}\message{(\the\toks0)}%
+%
+\unnumbchapmacro {#1}%
+\gdef\thischapter{#1}\gdef\thissection{#1}%
+{\chapternofonts%
+\edef\temp{{\realbackslash unnumbchapentry {#1}{\noexpand\folio}}}%
+\escapechar=`\\%
+\write \contentsfile \temp  %
+\unnumbnoderef %
+\global\let\section = \unnumberedsec
+\global\let\subsection = \unnumberedsubsec
+\global\let\subsubsection = \unnumberedsubsubsec
+}}
+
+\outer\def\numberedsec{\parsearg\secyyy}
+\def\secyyy #1{\numhead1{#1}} % normally calls seczzz
+\def\seczzz #1{\seccheck{section}%
+\subsecno=0 \subsubsecno=0 \global\advance \secno by 1 %
+\gdef\thissection{#1}\secheading {#1}{\the\chapno}{\the\secno}%
+{\chapternofonts%
+\edef\temp{{\realbackslash secentry %
+{#1}{\the\chapno}{\the\secno}{\noexpand\folio}}}%
+\escapechar=`\\%
+\write \contentsfile \temp %
+\donoderef %
+\penalty 10000 %
+}}
+
+\outer\def\appenixsection{\parsearg\appendixsecyyy}
+\outer\def\appendixsec{\parsearg\appendixsecyyy}
+\def\appendixsecyyy #1{\apphead1{#1}} % normally calls appendixsectionzzz
+\def\appendixsectionzzz #1{\seccheck{appendixsection}%
+\subsecno=0 \subsubsecno=0 \global\advance \secno by 1 %
+\gdef\thissection{#1}\secheading {#1}{\appendixletter}{\the\secno}%
+{\chapternofonts%
+\edef\temp{{\realbackslash secentry %
+{#1}{\appendixletter}{\the\secno}{\noexpand\folio}}}%
+\escapechar=`\\%
+\write \contentsfile \temp %
+\appendixnoderef %
+\penalty 10000 %
+}}
+
+\outer\def\unnumberedsec{\parsearg\unnumberedsecyyy}
+\def\unnumberedsecyyy #1{\unnmhead1{#1}} % normally calls unnumberedseczzz
+\def\unnumberedseczzz #1{\seccheck{unnumberedsec}%
+\plainsecheading {#1}\gdef\thissection{#1}%
+{\chapternofonts%
+\edef\temp{{\realbackslash unnumbsecentry{#1}{\noexpand\folio}}}%
+\escapechar=`\\%
+\write \contentsfile \temp %
+\unnumbnoderef %
+\penalty 10000 %
+}}
+
+\outer\def\numberedsubsec{\parsearg\numberedsubsecyyy}
+\def\numberedsubsecyyy #1{\numhead2{#1}} % normally calls numberedsubseczzz
+\def\numberedsubseczzz #1{\seccheck{subsection}%
+\gdef\thissection{#1}\subsubsecno=0 \global\advance \subsecno by 1 %
+\subsecheading {#1}{\the\chapno}{\the\secno}{\the\subsecno}%
+{\chapternofonts%
+\edef\temp{{\realbackslash subsecentry %
+{#1}{\the\chapno}{\the\secno}{\the\subsecno}{\noexpand\folio}}}%
+\escapechar=`\\%
+\write \contentsfile \temp %
+\donoderef %
+\penalty 10000 %
+}}
+
+\outer\def\appendixsubsec{\parsearg\appendixsubsecyyy}
+\def\appendixsubsecyyy #1{\apphead2{#1}} % normally calls appendixsubseczzz
+\def\appendixsubseczzz #1{\seccheck{appendixsubsec}%
+\gdef\thissection{#1}\subsubsecno=0 \global\advance \subsecno by 1 %
+\subsecheading {#1}{\appendixletter}{\the\secno}{\the\subsecno}%
+{\chapternofonts%
+\edef\temp{{\realbackslash subsecentry %
+{#1}{\appendixletter}{\the\secno}{\the\subsecno}{\noexpand\folio}}}%
+\escapechar=`\\%
+\write \contentsfile \temp %
+\appendixnoderef %
+\penalty 10000 %
+}}
+
+\outer\def\unnumberedsubsec{\parsearg\unnumberedsubsecyyy}
+\def\unnumberedsubsecyyy #1{\unnmhead2{#1}} %normally calls unnumberedsubseczzz
+\def\unnumberedsubseczzz #1{\seccheck{unnumberedsubsec}%
+\plainsecheading {#1}\gdef\thissection{#1}%
+{\chapternofonts%
+\edef\temp{{\realbackslash unnumbsubsecentry{#1}{\noexpand\folio}}}%
+\escapechar=`\\%
+\write \contentsfile \temp %
+\unnumbnoderef %
+\penalty 10000 %
+}}
+
+\outer\def\numberedsubsubsec{\parsearg\numberedsubsubsecyyy}
+\def\numberedsubsubsecyyy #1{\numhead3{#1}} % normally numberedsubsubseczzz
+\def\numberedsubsubseczzz #1{\seccheck{subsubsection}%
+\gdef\thissection{#1}\global\advance \subsubsecno by 1 %
+\subsubsecheading {#1}
+  {\the\chapno}{\the\secno}{\the\subsecno}{\the\subsubsecno}%
+{\chapternofonts%
+\edef\temp{{\realbackslash subsubsecentry %
+  {#1}
+  {\the\chapno}{\the\secno}{\the\subsecno}{\the\subsubsecno}
+  {\noexpand\folio}}}%
+\escapechar=`\\%
+\write \contentsfile \temp %
+\donoderef %
+\penalty 10000 %
+}}
+
+\outer\def\appendixsubsubsec{\parsearg\appendixsubsubsecyyy}
+\def\appendixsubsubsecyyy #1{\apphead3{#1}} % normally appendixsubsubseczzz
+\def\appendixsubsubseczzz #1{\seccheck{appendixsubsubsec}%
+\gdef\thissection{#1}\global\advance \subsubsecno by 1 %
+\subsubsecheading {#1}
+  {\appendixletter}{\the\secno}{\the\subsecno}{\the\subsubsecno}%
+{\chapternofonts%
+\edef\temp{{\realbackslash subsubsecentry{#1}%
+  {\appendixletter}
+  {\the\secno}{\the\subsecno}{\the\subsubsecno}{\noexpand\folio}}}%
+\escapechar=`\\%
+\write \contentsfile \temp %
+\appendixnoderef %
+\penalty 10000 %
+}}
+
+\outer\def\unnumberedsubsubsec{\parsearg\unnumberedsubsubsecyyy}
+\def\unnumberedsubsubsecyyy #1{\unnmhead3{#1}} %normally unnumberedsubsubseczzz
+\def\unnumberedsubsubseczzz #1{\seccheck{unnumberedsubsubsec}%
+\plainsecheading {#1}\gdef\thissection{#1}%
+{\chapternofonts%
+\edef\temp{{\realbackslash unnumbsubsubsecentry{#1}{\noexpand\folio}}}%
+\escapechar=`\\%
+\write \contentsfile \temp %
+\unnumbnoderef %
+\penalty 10000 %
+}}
+
+% These are variants which are not "outer", so they can appear in @ifinfo.
+% Actually, they should now be obsolete; ordinary section commands should work.
+\def\infotop{\parsearg\unnumberedzzz}
+\def\infounnumbered{\parsearg\unnumberedzzz}
+\def\infounnumberedsec{\parsearg\unnumberedseczzz}
+\def\infounnumberedsubsec{\parsearg\unnumberedsubseczzz}
+\def\infounnumberedsubsubsec{\parsearg\unnumberedsubsubseczzz}
+
+\def\infoappendix{\parsearg\appendixzzz}
+\def\infoappendixsec{\parsearg\appendixseczzz}
+\def\infoappendixsubsec{\parsearg\appendixsubseczzz}
+\def\infoappendixsubsubsec{\parsearg\appendixsubsubseczzz}
+
+\def\infochapter{\parsearg\chapterzzz}
+\def\infosection{\parsearg\sectionzzz}
+\def\infosubsection{\parsearg\subsectionzzz}
+\def\infosubsubsection{\parsearg\subsubsectionzzz}
+
+% These macros control what the section commands do, according
+% to what kind of chapter we are in (ordinary, appendix, or unnumbered).
+% Define them by default for a numbered chapter.
+\global\let\section = \numberedsec
+\global\let\subsection = \numberedsubsec
+\global\let\subsubsection = \numberedsubsubsec
+
+% Define @majorheading, @heading and @subheading
+
+% NOTE on use of \vbox for chapter headings, section headings, and
+% such:
+%	1) We use \vbox rather than the earlier \line to permit
+%	   overlong headings to fold.
+%	2) \hyphenpenalty is set to 10000 because hyphenation in a
+%	   heading is obnoxious; this forbids it.
+%       3) Likewise, headings look best if no \parindent is used, and
+%          if justification is not attempted.  Hence \raggedright.
+
+
+\def\majorheading{\parsearg\majorheadingzzz}
+\def\majorheadingzzz #1{%
+{\advance\chapheadingskip by 10pt \chapbreak }%
+{\chapfonts \vbox{\hyphenpenalty=10000\tolerance=5000
+                  \parindent=0pt\raggedright
+                  \rm #1\hfill}}\bigskip \par\penalty 200}
+
+\def\chapheading{\parsearg\chapheadingzzz}
+\def\chapheadingzzz #1{\chapbreak %
+{\chapfonts \vbox{\hyphenpenalty=10000\tolerance=5000
+                  \parindent=0pt\raggedright
+                  \rm #1\hfill}}\bigskip \par\penalty 200}
+
+\def\heading{\parsearg\secheadingi}
+
+\def\subheading{\parsearg\subsecheadingi}
+
+\def\subsubheading{\parsearg\subsubsecheadingi}
+
+% These macros generate a chapter, section, etc. heading only
+% (including whitespace, linebreaking, etc. around it),
+% given all the information in convenient, parsed form.
+
+%%% Args are the skip and penalty (usually negative)
+\def\dobreak#1#2{\par\ifdim\lastskip<#1\removelastskip\penalty#2\vskip#1\fi}
+
+\def\setchapterstyle #1 {\csname CHAPF#1\endcsname}
+
+%%% Define plain chapter starts, and page on/off switching for it
+% Parameter controlling skip before chapter headings (if needed)
+
+\newskip \chapheadingskip \chapheadingskip = 30pt plus 8pt minus 4pt
+
+\def\chapbreak{\dobreak \chapheadingskip {-4000}}
+\def\chappager{\par\vfill\supereject}
+\def\chapoddpage{\chappager \ifodd\pageno \else \hbox to 0pt{} \chappager\fi}
+
+\def\setchapternewpage #1 {\csname CHAPPAG#1\endcsname}
+
+\def\CHAPPAGoff{
+\global\let\pchapsepmacro=\chapbreak
+\global\let\pagealignmacro=\chappager}
+
+\def\CHAPPAGon{
+\global\let\pchapsepmacro=\chappager
+\global\let\pagealignmacro=\chappager
+\global\def\HEADINGSon{\HEADINGSsingle}}
+
+\def\CHAPPAGodd{
+\global\let\pchapsepmacro=\chapoddpage
+\global\let\pagealignmacro=\chapoddpage
+\global\def\HEADINGSon{\HEADINGSdouble}}
+
+\CHAPPAGon
+
+\def\CHAPFplain{
+\global\let\chapmacro=\chfplain
+\global\let\unnumbchapmacro=\unnchfplain}
+
+\def\chfplain #1#2{%
+  \pchapsepmacro
+  {%
+    \chapfonts \vbox{\hyphenpenalty=10000\tolerance=5000
+                     \parindent=0pt\raggedright
+                     \rm #2\enspace #1}%
+  }%
+  \bigskip
+  \penalty5000
+}
+
+\def\unnchfplain #1{%
+\pchapsepmacro %
+{\chapfonts \vbox{\hyphenpenalty=10000\tolerance=5000
+                  \parindent=0pt\raggedright
+                  \rm #1\hfill}}\bigskip \par\penalty 10000 %
+}
+\CHAPFplain % The default
+
+\def\unnchfopen #1{%
+\chapoddpage {\chapfonts \vbox{\hyphenpenalty=10000\tolerance=5000
+                       \parindent=0pt\raggedright
+                       \rm #1\hfill}}\bigskip \par\penalty 10000 %
+}
+
+\def\chfopen #1#2{\chapoddpage {\chapfonts
+\vbox to 3in{\vfil \hbox to\hsize{\hfil #2} \hbox to\hsize{\hfil #1} \vfil}}%
+\par\penalty 5000 %
+}
+
+\def\CHAPFopen{
+\global\let\chapmacro=\chfopen
+\global\let\unnumbchapmacro=\unnchfopen}
+
+% Parameter controlling skip before section headings.
+
+\newskip \subsecheadingskip  \subsecheadingskip = 17pt plus 8pt minus 4pt
+\def\subsecheadingbreak{\dobreak \subsecheadingskip {-500}}
+
+\newskip \secheadingskip  \secheadingskip = 21pt plus 8pt minus 4pt
+\def\secheadingbreak{\dobreak \secheadingskip {-1000}}
+
+% @paragraphindent  is defined for the Info formatting commands only.
+\let\paragraphindent=\comment
+
+% Section fonts are the base font at magstep2, which produces
+% a size a bit more than 14 points in the default situation.
+
+\def\secheading #1#2#3{\secheadingi {#2.#3\enspace #1}}
+\def\plainsecheading #1{\secheadingi {#1}}
+\def\secheadingi #1{{\advance \secheadingskip by \parskip %
+\secheadingbreak}%
+{\secfonts \vbox{\hyphenpenalty=10000\tolerance=5000
+                 \parindent=0pt\raggedright
+                 \rm #1\hfill}}%
+\ifdim \parskip<10pt \kern 10pt\kern -\parskip\fi \penalty 10000 }
+
+
+% Subsection fonts are the base font at magstep1,
+% which produces a size of 12 points.
+
+\def\subsecheading #1#2#3#4{\subsecheadingi {#2.#3.#4\enspace #1}}
+\def\subsecheadingi #1{{\advance \subsecheadingskip by \parskip %
+\subsecheadingbreak}%
+{\subsecfonts \vbox{\hyphenpenalty=10000\tolerance=5000
+                     \parindent=0pt\raggedright
+                     \rm #1\hfill}}%
+\ifdim \parskip<10pt \kern 10pt\kern -\parskip\fi \penalty 10000 }
+
+\def\subsubsecfonts{\subsecfonts} % Maybe this should change:
+				  % Perhaps make sssec fonts scaled
+				  % magstep half
+\def\subsubsecheading #1#2#3#4#5{\subsubsecheadingi {#2.#3.#4.#5\enspace #1}}
+\def\subsubsecheadingi #1{{\advance \subsecheadingskip by \parskip %
+\subsecheadingbreak}%
+{\subsubsecfonts \vbox{\hyphenpenalty=10000\tolerance=5000
+                       \parindent=0pt\raggedright
+                       \rm #1\hfill}}%
+\ifdim \parskip<10pt \kern 10pt\kern -\parskip\fi \penalty 10000}
+
+
+\message{toc printing,}
+
+% Finish up the main text and prepare to read what we've written
+% to \contentsfile.
+
+\newskip\contentsrightmargin \contentsrightmargin=1in
+\def\startcontents#1{%
+   \pagealignmacro
+   \immediate\closeout \contentsfile
+   \ifnum \pageno>0
+      \pageno = -1		% Request roman numbered pages.
+   \fi
+   % Don't need to put `Contents' or `Short Contents' in the headline.
+   % It is abundantly clear what they are.
+   \unnumbchapmacro{#1}\def\thischapter{}%
+   \begingroup   		% Set up to handle contents files properly.
+      \catcode`\\=0  \catcode`\{=1  \catcode`\}=2  \catcode`\@=11
+      \raggedbottom             % Worry more about breakpoints than the bottom.
+      \advance\hsize by -\contentsrightmargin % Don't use the full line length.
+}
+
+
+% Normal (long) toc.
+\outer\def\contents{%
+   \startcontents{Table of Contents}%
+      \input \jobname.toc
+   \endgroup
+   \vfill \eject
+}
+
+% And just the chapters.
+\outer\def\summarycontents{%
+   \startcontents{Short Contents}%
+      %
+      \let\chapentry = \shortchapentry
+      \let\unnumbchapentry = \shortunnumberedentry
+      % We want a true roman here for the page numbers.
+      \secfonts
+      \let\rm=\shortcontrm \let\bf=\shortcontbf \let\sl=\shortcontsl
+      \rm
+      \advance\baselineskip by 1pt % Open it up a little.
+      \def\secentry ##1##2##3##4{}
+      \def\unnumbsecentry ##1##2{}
+      \def\subsecentry ##1##2##3##4##5{}
+      \def\unnumbsubsecentry ##1##2{}
+      \def\subsubsecentry ##1##2##3##4##5##6{}
+      \def\unnumbsubsubsecentry ##1##2{}
+      \input \jobname.toc
+   \endgroup
+   \vfill \eject
+}
+\let\shortcontents = \summarycontents
+
+% These macros generate individual entries in the table of contents.
+% The first argument is the chapter or section name.
+% The last argument is the page number.
+% The arguments in between are the chapter number, section number, ...
+
+% Chapter-level things, for both the long and short contents.
+\def\chapentry#1#2#3{\dochapentry{#2\labelspace#1}{#3}}
+
+% See comments in \dochapentry re vbox and related settings
+\def\shortchapentry#1#2#3{%
+  \tocentry{\shortchaplabel{#2}\labelspace #1}{\doshortpageno{#3}}%
+}
+
+% Typeset the label for a chapter or appendix for the short contents.
+% The arg is, e.g. `Appendix A' for an appendix, or `3' for a chapter.
+% We could simplify the code here by writing out an \appendixentry
+% command in the toc file for appendices, instead of using \chapentry
+% for both, but it doesn't seem worth it.
+\setbox0 = \hbox{\shortcontrm Appendix }
+\newdimen\shortappendixwidth \shortappendixwidth = \wd0
+
+\def\shortchaplabel#1{%
+  % We typeset #1 in a box of constant width, regardless of the text of
+  % #1, so the chapter titles will come out aligned.
+  \setbox0 = \hbox{#1}%
+  \dimen0 = \ifdim\wd0 > \shortappendixwidth \shortappendixwidth \else 0pt \fi
+  %
+  % This space should be plenty, since a single number is .5em, and the
+  % widest letter (M) is 1em, at least in the Computer Modern fonts.
+  % (This space doesn't include the extra space that gets added after
+  % the label; that gets put in in \shortchapentry above.)
+  \advance\dimen0 by 1.1em
+  \hbox to \dimen0{#1\hfil}%
+}
+
+\def\unnumbchapentry#1#2{\dochapentry{#1}{#2}}
+\def\shortunnumberedentry#1#2{\tocentry{#1}{\doshortpageno{#2}}}
+
+% Sections.
+\def\secentry#1#2#3#4{\dosecentry{#2.#3\labelspace#1}{#4}}
+\def\unnumbsecentry#1#2{\dosecentry{#1}{#2}}
+
+% Subsections.
+\def\subsecentry#1#2#3#4#5{\dosubsecentry{#2.#3.#4\labelspace#1}{#5}}
+\def\unnumbsubsecentry#1#2{\dosubsecentry{#1}{#2}}
+
+% And subsubsections.
+\def\subsubsecentry#1#2#3#4#5#6{%
+  \dosubsubsecentry{#2.#3.#4.#5\labelspace#1}{#6}}
+\def\unnumbsubsubsecentry#1#2{\dosubsubsecentry{#1}{#2}}
+
+
+% This parameter controls the indentation of the various levels.
+\newdimen\tocindent \tocindent = 3pc
+
+% Now for the actual typesetting. In all these, #1 is the text and #2 is the
+% page number.
+%
+% If the toc has to be broken over pages, we would want to be at chapters
+% if at all possible; hence the \penalty.
+\def\dochapentry#1#2{%
+   \penalty-300 \vskip\baselineskip
+   \begingroup
+     \chapentryfonts
+     \tocentry{#1}{\dopageno{#2}}%
+   \endgroup
+   \nobreak\vskip .25\baselineskip
+}
+
+\def\dosecentry#1#2{\begingroup
+  \secentryfonts \leftskip=\tocindent
+  \tocentry{#1}{\dopageno{#2}}%
+\endgroup}
+
+\def\dosubsecentry#1#2{\begingroup
+  \subsecentryfonts \leftskip=2\tocindent
+  \tocentry{#1}{\dopageno{#2}}%
+\endgroup}
+
+\def\dosubsubsecentry#1#2{\begingroup
+  \subsubsecentryfonts \leftskip=3\tocindent
+  \tocentry{#1}{\dopageno{#2}}%
+\endgroup}
+
+% Final typesetting of a toc entry; we use the same \entry macro as for
+% the index entries, but we want to suppress hyphenation here.  (We
+% can't do that in the \entry macro, since index entries might consist
+% of hyphenated-identifiers-that-do-not-fit-on-a-line-and-nothing-else.)
+%
+\def\tocentry#1#2{\begingroup
+  \hyphenpenalty = 10000
+  \entry{#1}{#2}%
+\endgroup}
+
+% Space between chapter (or whatever) number and the title.
+\def\labelspace{\hskip1em \relax}
+
+\def\dopageno#1{{\rm #1}}
+\def\doshortpageno#1{{\rm #1}}
+
+\def\chapentryfonts{\secfonts \rm}
+\def\secentryfonts{\textfonts}
+\let\subsecentryfonts = \textfonts
+\let\subsubsecentryfonts = \textfonts
+
+
+\message{environments,}
+
+% Since these characters are used in examples, it should be an even number of
+% \tt widths. Each \tt character is 1en, so two makes it 1em.
+% Furthermore, these definitions must come after we define our fonts.
+\newbox\dblarrowbox    \newbox\longdblarrowbox
+\newbox\pushcharbox    \newbox\bullbox
+\newbox\equivbox       \newbox\errorbox
+
+\let\ptexequiv = \equiv
+
+%{\tentt
+%\global\setbox\dblarrowbox = \hbox to 1em{\hfil$\Rightarrow$\hfil}
+%\global\setbox\longdblarrowbox = \hbox to 1em{\hfil$\mapsto$\hfil}
+%\global\setbox\pushcharbox = \hbox to 1em{\hfil$\dashv$\hfil}
+%\global\setbox\equivbox = \hbox to 1em{\hfil$\ptexequiv$\hfil}
+% Adapted from the manmac format (p.420 of TeXbook)
+%\global\setbox\bullbox = \hbox to 1em{\kern.15em\vrule height .75ex width .85ex
+%                                      depth .1ex\hfil}
+%}
+
+\def\point{$\star$}
+
+\def\result{\leavevmode\raise.15ex\hbox to 1em{\hfil$\Rightarrow$\hfil}}
+\def\expansion{\leavevmode\raise.1ex\hbox to 1em{\hfil$\mapsto$\hfil}}
+\def\print{\leavevmode\lower.1ex\hbox to 1em{\hfil$\dashv$\hfil}}
+
+\def\equiv{\leavevmode\lower.1ex\hbox to 1em{\hfil$\ptexequiv$\hfil}}
+
+% Adapted from the TeXbook's \boxit.
+{\tentt \global\dimen0 = 3em}% Width of the box.
+\dimen2 = .55pt % Thickness of rules
+% The text. (`r' is open on the right, `e' somewhat less so on the left.)
+\setbox0 = \hbox{\kern-.75pt \tensf error\kern-1.5pt}
+
+\global\setbox\errorbox=\hbox to \dimen0{\hfil
+   \hsize = \dimen0 \advance\hsize by -5.8pt % Space to left+right.
+   \advance\hsize by -2\dimen2 % Rules.
+   \vbox{
+      \hrule height\dimen2
+      \hbox{\vrule width\dimen2 \kern3pt          % Space to left of text.
+         \vtop{\kern2.4pt \box0 \kern2.4pt}% Space above/below.
+         \kern3pt\vrule width\dimen2}% Space to right.
+      \hrule height\dimen2}
+    \hfil}
+
+% The @error{} command.
+\def\error{\leavevmode\lower.7ex\copy\errorbox}
+
+% @tex ... @end tex    escapes into raw Tex temporarily.
+% One exception: @ is still an escape character, so that @end tex works.
+% But \@ or @@ will get a plain tex @ character.
+
+\def\tex{\begingroup
+\catcode `\\=0 \catcode `\{=1 \catcode `\}=2
+\catcode `\$=3 \catcode `\&=4 \catcode `\#=6
+\catcode `\^=7 \catcode `\_=8 \catcode `\~=13 \let~=\tie
+\catcode `\%=14
+\catcode 43=12
+\catcode`\"=12
+\catcode`\==12
+\catcode`\|=12
+\catcode`\<=12
+\catcode`\>=12
+\escapechar=`\\
+%
+\let\{=\ptexlbrace
+\let\}=\ptexrbrace
+\let\.=\ptexdot
+\let\*=\ptexstar
+\let\dots=\ptexdots
+\def\@{@}%
+\let\bullet=\ptexbullet
+\let\b=\ptexb \let\c=\ptexc \let\i=\ptexi \let\t=\ptext \let\l=\ptexl
+\let\L=\ptexL
+%
+\let\Etex=\endgroup}
+
+% Define @lisp ... @endlisp.
+% @lisp does a \begingroup so it can rebind things,
+% including the definition of @endlisp (which normally is erroneous).
+
+% Amount to narrow the margins by for @lisp.
+\newskip\lispnarrowing \lispnarrowing=0.4in
+
+% This is the definition that ^^M gets inside @lisp, @example, and other
+% such environments.  \null is better than a space, since it doesn't
+% have any width.
+\def\lisppar{\null\endgraf}
+
+% Make each space character in the input produce a normal interword
+% space in the output.  Don't allow a line break at this space, as this
+% is used only in environments like @example, where each line of input
+% should produce a line of output anyway.
+%
+{\obeyspaces %
+\gdef\sepspaces{\obeyspaces\let =\tie}}
+
+% Define \obeyedspace to be our active space, whatever it is.  This is
+% for use in \parsearg.
+{\sepspaces %
+\global\let\obeyedspace= }
+
+% This space is always present above and below environments.
+\newskip\envskipamount \envskipamount = 0pt
+
+% Make spacing and below environment symmetrical.  We use \parskip here
+% to help in doing that, since in @example-like environments \parskip
+% is reset to zero; thus the \afterenvbreak inserts no space -- but the
+% start of the next paragraph will insert \parskip
+%
+\def\aboveenvbreak{{\advance\envskipamount by \parskip
+\endgraf \ifdim\lastskip<\envskipamount
+\removelastskip \penalty-50 \vskip\envskipamount \fi}}
+
+\let\afterenvbreak = \aboveenvbreak
+
+% \nonarrowing is a flag.  If "set", @lisp etc don't narrow margins.
+\let\nonarrowing=\relax
+
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+% \cartouche: draw rectangle w/rounded corners around argument
+\font\circle=lcircle10
+\newdimen\circthick
+\newdimen\cartouter\newdimen\cartinner
+\newskip\normbskip\newskip\normpskip\newskip\normlskip
+\circthick=\fontdimen8\circle
+%
+\def\ctl{{\circle\char'013\hskip -6pt}}% 6pt from pl file: 1/2charwidth
+\def\ctr{{\hskip 6pt\circle\char'010}}
+\def\cbl{{\circle\char'012\hskip -6pt}}
+\def\cbr{{\hskip 6pt\circle\char'011}}
+\def\carttop{\hbox to \cartouter{\hskip\lskip
+	\ctl\leaders\hrule height\circthick\hfil\ctr
+	\hskip\rskip}}
+\def\cartbot{\hbox to \cartouter{\hskip\lskip
+	\cbl\leaders\hrule height\circthick\hfil\cbr
+	\hskip\rskip}}
+%
+\newskip\lskip\newskip\rskip
+
+\long\def\cartouche{%
+\begingroup
+	\lskip=\leftskip \rskip=\rightskip
+	\leftskip=0pt\rightskip=0pt %we want these *outside*.
+	\cartinner=\hsize \advance\cartinner by-\lskip
+		 	  \advance\cartinner by-\rskip
+	\cartouter=\hsize
+	\advance\cartouter by 18pt % allow for 3pt kerns on either
+%				     side, and for 6pt waste from
+%				     each corner char
+	\normbskip=\baselineskip \normpskip=\parskip \normlskip=\lineskip
+	% Flag to tell @lisp, etc., not to narrow margin.
+	\let\nonarrowing=\comment
+	\vbox\bgroup
+		\baselineskip=0pt\parskip=0pt\lineskip=0pt
+		\carttop
+		\hbox\bgroup
+			\hskip\lskip
+			\vrule\kern3pt
+			\vbox\bgroup
+				\hsize=\cartinner
+				\kern3pt
+				\begingroup
+					\baselineskip=\normbskip
+					\lineskip=\normlskip
+					\parskip=\normpskip
+					\vskip -\parskip
+\def\Ecartouche{%
+				\endgroup
+				\kern3pt
+			\egroup
+			\kern3pt\vrule
+			\hskip\rskip
+		\egroup
+		\cartbot
+	\egroup
+\endgroup
+}}
+
+
+% This macro is called at the beginning of all the @example variants,
+% inside a group.
+\def\nonfillstart{%
+  \aboveenvbreak
+  \inENV % This group ends at the end of the body
+  \hfuzz = 12pt % Don't be fussy
+  \sepspaces % Make spaces be word-separators rather than space tokens.
+  \singlespace
+  \let\par = \lisppar % don't ignore blank lines
+  \obeylines % each line of input is a line of output
+  \parskip = 0pt
+  \parindent = 0pt
+  \emergencystretch = 0pt % don't try to avoid overfull boxes
+  % @cartouche defines \nonarrowing to inhibit narrowing
+  % at next level down.
+  \ifx\nonarrowing\relax
+    \advance \leftskip by \lispnarrowing
+    \exdentamount=\lispnarrowing
+    \let\exdent=\nofillexdent
+    \let\nonarrowing=\relax
+  \fi
+}
+
+% To ending an @example-like environment, we first end the paragraph
+% (via \afterenvbreak's vertical glue), and then the group.  That way we
+% keep the zero \parskip that the environments set -- \parskip glue
+% will be inserted at the beginning of the next paragraph in the
+% document, after the environment.
+%
+\def\nonfillfinish{\afterenvbreak\endgroup}%
+
+% This macro is
+\def\lisp{\begingroup
+  \nonfillstart
+  \let\Elisp = \nonfillfinish
+  \tt
+  \rawbackslash % have \ input char produce \ char from current font
+  \gobble
+}
+
+% Define the \E... control sequence only if we are inside the
+% environment, so the error checking in \end will work.
+%
+% We must call \lisp last in the definition, since it reads the
+% return following the @example (or whatever) command.
+%
+\def\example{\begingroup \def\Eexample{\nonfillfinish\endgroup}\lisp}
+\def\smallexample{\begingroup \def\Esmallexample{\nonfillfinish\endgroup}\lisp}
+\def\smalllisp{\begingroup \def\Esmalllisp{\nonfillfinish\endgroup}\lisp}
+
+% @smallexample and @smalllisp.  This is not used unless the @smallbook
+% command is given.  Originally contributed by Pavel@xerox.
+%
+\def\smalllispx{\begingroup
+  \nonfillstart
+  \let\Esmalllisp = \nonfillfinish
+  \let\Esmallexample = \nonfillfinish
+  %
+  % Smaller interline space and fonts for small examples.
+  \baselineskip 10pt
+  \indexfonts \tt
+  \rawbackslash % output the \ character from the current font
+  \gobble
+}
+
+% This is @display; same as @lisp except use roman font.
+%
+\def\display{\begingroup
+  \nonfillstart
+  \let\Edisplay = \nonfillfinish
+  \gobble
+}
+
+% This is @format; same as @display except don't narrow margins.
+%
+\def\format{\begingroup
+  \let\nonarrowing = t
+  \nonfillstart
+  \let\Eformat = \nonfillfinish
+  \gobble
+}
+
+% @flushleft (same as @format) and @flushright.
+%
+\def\flushleft{\begingroup
+  \let\nonarrowing = t
+  \nonfillstart
+  \let\Eflushleft = \nonfillfinish
+  \gobble
+}
+\def\flushright{\begingroup
+  \let\nonarrowing = t
+  \nonfillstart
+  \let\Eflushright = \nonfillfinish
+  \advance\leftskip by 0pt plus 1fill
+  \gobble}
+
+% @quotation does normal linebreaking and narrows the margins.
+%
+\def\quotation{%
+\begingroup\inENV %This group ends at the end of the @quotation body
+{\parskip=0pt  % because we will skip by \parskip too, later
+\aboveenvbreak}%
+\singlespace
+\parindent=0pt
+\let\Equotation = \nonfillfinish
+% @cartouche defines \nonarrowing to inhibit narrowing
+% at next level down.
+\ifx\nonarrowing\relax
+\advance \leftskip by \lispnarrowing
+\advance \rightskip by \lispnarrowing
+\exdentamount=\lispnarrowing
+\let\nonarrowing=\relax
+\fi}
+
+\message{defuns,}
+% Define formatter for defuns
+% First, allow user to change definition object font (\df) internally
+\def\setdeffont #1 {\csname DEF#1\endcsname}
+
+\newskip\defbodyindent \defbodyindent=.4in
+\newskip\defargsindent \defargsindent=50pt
+\newskip\deftypemargin \deftypemargin=12pt
+\newskip\deflastargmargin \deflastargmargin=18pt
+
+\newcount\parencount
+% define \functionparens, which makes ( and ) and & do special things.
+% \functionparens affects the group it is contained in.
+\def\activeparens{%
+\catcode`\(=\active \catcode`\)=\active \catcode`\&=\active
+\catcode`\[=\active \catcode`\]=\active}
+
+% Make control sequences which act like normal parenthesis chars.
+\let\lparen = ( \let\rparen = )
+
+{\activeparens % Now, smart parens don't turn on until &foo (see \amprm)
+
+% Be sure that we always have a definition for `(', etc.  For example,
+% if the fn name has parens in it, \boldbrax will not be in effect yet,
+% so TeX would otherwise complain about undefined control sequence.
+\global\let(=\lparen \global\let)=\rparen
+\global\let[=\lbrack \global\let]=\rbrack
+
+\gdef\functionparens{\boldbrax\let&=\amprm\parencount=0 }
+\gdef\boldbrax{\let(=\opnr\let)=\clnr\let[=\lbrb\let]=\rbrb}
+
+% Definitions of (, ) and & used in args for functions.
+% This is the definition of ( outside of all parentheses.
+\gdef\oprm#1 {{\rm\char`\(}#1 \bf \let(=\opnested %
+\global\advance\parencount by 1 }
+%
+% This is the definition of ( when already inside a level of parens.
+\gdef\opnested{\char`\(\global\advance\parencount by 1 }
+%
+\gdef\clrm{% Print a paren in roman if it is taking us back to depth of 0.
+% also in that case restore the outer-level definition of (.
+\ifnum \parencount=1 {\rm \char `\)}\sl \let(=\oprm \else \char `\) \fi
+\global\advance \parencount by -1 }
+% If we encounter &foo, then turn on ()-hacking afterwards
+\gdef\amprm#1 {{\rm\&#1}\let(=\oprm \let)=\clrm\ }
+%
+\gdef\normalparens{\boldbrax\let&=\ampnr}
+} % End of definition inside \activeparens
+%% These parens (in \boldbrax) actually are a little bolder than the
+%% contained text.  This is especially needed for [ and ]
+\def\opnr{{\sf\char`\(}} \def\clnr{{\sf\char`\)}} \def\ampnr{\&}
+\def\lbrb{{\bf\char`\[}} \def\rbrb{{\bf\char`\]}}
+
+% First, defname, which formats the header line itself.
+% #1 should be the function name.
+% #2 should be the type of definition, such as "Function".
+
+\def\defname #1#2{%
+% Get the values of \leftskip and \rightskip as they were
+% outside the @def...
+\dimen2=\leftskip
+\advance\dimen2 by -\defbodyindent
+\dimen3=\rightskip
+\advance\dimen3 by -\defbodyindent
+\noindent        %
+\setbox0=\hbox{\hskip \deflastargmargin{\rm #2}\hskip \deftypemargin}%
+\dimen0=\hsize \advance \dimen0 by -\wd0 % compute size for first line
+\dimen1=\hsize \advance \dimen1 by -\defargsindent %size for continuations
+\parshape 2 0in \dimen0 \defargsindent \dimen1     %
+% Now output arg 2 ("Function" or some such)
+% ending at \deftypemargin from the right margin,
+% but stuck inside a box of width 0 so it does not interfere with linebreaking
+{% Adjust \hsize to exclude the ambient margins,
+% so that \rightline will obey them.
+\advance \hsize by -\dimen2 \advance \hsize by -\dimen3
+\rlap{\rightline{{\rm #2}\hskip \deftypemargin}}}%
+% Make all lines underfull and no complaints:
+\tolerance=10000 \hbadness=10000
+\advance\leftskip by -\defbodyindent
+\exdentamount=\defbodyindent
+{\df #1}\enskip        % Generate function name
+}
+
+% Actually process the body of a definition
+% #1 should be the terminating control sequence, such as \Edefun.
+% #2 should be the "another name" control sequence, such as \defunx.
+% #3 should be the control sequence that actually processes the header,
+%    such as \defunheader.
+
+\def\defparsebody #1#2#3{\begingroup\inENV% Environment for definitionbody
+\medbreak %
+% Define the end token that this defining construct specifies
+% so that it will exit this group.
+\def#1{\endgraf\endgroup\medbreak}%
+\def#2{\begingroup\obeylines\activeparens\spacesplit#3}%
+\parindent=0in
+\advance\leftskip by \defbodyindent \advance \rightskip by \defbodyindent
+\exdentamount=\defbodyindent
+\begingroup %
+\catcode 61=\active %
+\obeylines\activeparens\spacesplit#3}
+
+\def\defmethparsebody #1#2#3#4 {\begingroup\inENV %
+\medbreak %
+% Define the end token that this defining construct specifies
+% so that it will exit this group.
+\def#1{\endgraf\endgroup\medbreak}%
+\def#2##1 {\begingroup\obeylines\activeparens\spacesplit{#3{##1}}}%
+\parindent=0in
+\advance\leftskip by \defbodyindent \advance \rightskip by \defbodyindent
+\exdentamount=\defbodyindent
+\begingroup\obeylines\activeparens\spacesplit{#3{#4}}}
+
+\def\defopparsebody #1#2#3#4#5 {\begingroup\inENV %
+\medbreak %
+% Define the end token that this defining construct specifies
+% so that it will exit this group.
+\def#1{\endgraf\endgroup\medbreak}%
+\def#2##1 ##2 {\def#4{##1}%
+\begingroup\obeylines\activeparens\spacesplit{#3{##2}}}%
+\parindent=0in
+\advance\leftskip by \defbodyindent \advance \rightskip by \defbodyindent
+\exdentamount=\defbodyindent
+\begingroup\obeylines\activeparens\spacesplit{#3{#5}}}
+
+% These parsing functions are similar to the preceding ones
+% except that they do not make parens into active characters.
+% These are used for "variables" since they have no arguments.
+
+\def\defvarparsebody #1#2#3{\begingroup\inENV% Environment for definitionbody
+\medbreak %
+% Define the end token that this defining construct specifies
+% so that it will exit this group.
+\def#1{\endgraf\endgroup\medbreak}%
+\def#2{\begingroup\obeylines\spacesplit#3}%
+\parindent=0in
+\advance\leftskip by \defbodyindent \advance \rightskip by \defbodyindent
+\exdentamount=\defbodyindent
+\begingroup %
+\catcode 61=\active %
+\obeylines\spacesplit#3}
+
+\def\defvrparsebody #1#2#3#4 {\begingroup\inENV %
+\medbreak %
+% Define the end token that this defining construct specifies
+% so that it will exit this group.
+\def#1{\endgraf\endgroup\medbreak}%
+\def#2##1 {\begingroup\obeylines\spacesplit{#3{##1}}}%
+\parindent=0in
+\advance\leftskip by \defbodyindent \advance \rightskip by \defbodyindent
+\exdentamount=\defbodyindent
+\begingroup\obeylines\spacesplit{#3{#4}}}
+
+% This seems to work right in all cases.
+\let\deftpparsebody=\defvrparsebody
+% This fails to work.  When given `@deftp {Data Type} foo_t',
+% it thinks the type name is just `f'.
+%%% This is the same as all the others except for the last line.  We need
+%%% to parse the arguments differently for @deftp, since the ``attributes''
+%%% there are optional.
+%%%
+%%\def\deftpparsebody #1#2#3#4 {\begingroup\inENV %
+%%\medbreak %
+%%% Define the end token that this defining construct specifies
+%%% so that it will exit this group.
+%%\def#1{\endgraf\endgroup\medbreak}%
+%%\def#2##1 {\begingroup\obeylines\spacesplit{#3{##1}}}%
+%%\parindent=0in
+%%\advance\leftskip by \defbodyindent \advance \rightskip by \defbodyindent
+%%\exdentamount=\defbodyindent
+%%\begingroup\obeylines\parsetpheaderline{#3{#4}}}
+
+%%{\obeylines %
+%%  % Parse the type name and any attributes (field names, etc.).
+%%  % #1 is the beginning of the macro call that will produce the output,
+%%  %   i.e., \deftpheader{CLASS}; this is passed from \deftpparsebody.
+%%  % #2 is the type name, e.g., `struct termios'.
+%%  % #3 is the (possibly empty) attribute list.
+%%  %
+%%  \gdef\parsetpheaderline#1#2#3^^M{%
+%%    \endgroup % Started in \deftpparsebody.
+%%    %
+%%    % If the attribute list is in fact empty, there will be no space after
+%%    % #2; so we can't put a space in our TeX parameter list.  But if it
+%%    % isn't empty, then #3 will begin with an unwanted space.
+%%    \def\theargs{\ignorespaces #3}%
+%%    %
+%%    % Call the macro to produce the output.
+%%    #1{#2}\theargs %
+%%  }%
+%%}
+
+\def\defopvarparsebody #1#2#3#4#5 {\begingroup\inENV %
+\medbreak %
+% Define the end token that this defining construct specifies
+% so that it will exit this group.
+\def#1{\endgraf\endgroup\medbreak}%
+\def#2##1 ##2 {\def#4{##1}%
+\begingroup\obeylines\spacesplit{#3{##2}}}%
+\parindent=0in
+\advance\leftskip by \defbodyindent \advance \rightskip by \defbodyindent
+\exdentamount=\defbodyindent
+\begingroup\obeylines\spacesplit{#3{#5}}}
+
+% Split up #2 at the first space token.
+% call #1 with two arguments:
+%  the first is all of #2 before the space token,
+%  the second is all of #2 after that space token.
+% If #2 contains no space token, all of it is passed as the first arg
+% and the second is passed as empty.
+
+{\obeylines
+\gdef\spacesplit#1#2^^M{\endgroup\spacesplitfoo{#1}#2 \relax\spacesplitfoo}%
+\long\gdef\spacesplitfoo#1#2 #3#4\spacesplitfoo{%
+\ifx\relax #3%
+#1{#2}{}\else #1{#2}{#3#4}\fi}}
+
+% So much for the things common to all kinds of definitions.
+
+% Define @defun.
+
+% First, define the processing that is wanted for arguments of \defun
+% Use this to expand the args and terminate the paragraph they make up
+
+\def\defunargs #1{\functionparens \sl
+% Expand, preventing hyphenation at `-' chars.
+% Note that groups don't affect changes in \hyphenchar.
+\hyphenchar\tensl=0
+#1%
+\hyphenchar\tensl=45
+\ifnum\parencount=0 \else \errmessage{unbalanced parens in @def arguments}\fi%
+\interlinepenalty=10000
+\advance\rightskip by 0pt plus 1fil
+\endgraf\penalty 10000\vskip -\parskip\penalty 10000%
+}
+
+\def\deftypefunargs #1{%
+% Expand, preventing hyphenation at `-' chars.
+% Note that groups don't affect changes in \hyphenchar.
+\functionparens
+\code{#1}%
+\interlinepenalty=10000
+\advance\rightskip by 0pt plus 1fil
+\endgraf\penalty 10000\vskip -\parskip\penalty 10000%
+}
+
+% Do complete processing of one @defun or @defunx line already parsed.
+
+% @deffn Command forward-char nchars
+
+\def\deffn{\defmethparsebody\Edeffn\deffnx\deffnheader}
+
+\def\deffnheader #1#2#3{\doind {fn}{\code{#2}}%
+\begingroup\defname {#2}{#1}\defunargs{#3}\endgroup %
+\catcode 61=\other % Turn off change made in \defparsebody
+}
+
+% @defun == @deffn Function
+
+\def\defun{\defparsebody\Edefun\defunx\defunheader}
+
+\def\defunheader #1#2{\doind {fn}{\code{#1}}% Make entry in function index
+\begingroup\defname {#1}{Function}%
+\defunargs {#2}\endgroup %
+\catcode 61=\other % Turn off change made in \defparsebody
+}
+
+% @deftypefun int foobar (int @var{foo}, float @var{bar})
+
+\def\deftypefun{\defparsebody\Edeftypefun\deftypefunx\deftypefunheader}
+
+% #1 is the data type.  #2 is the name and args.
+\def\deftypefunheader #1#2{\deftypefunheaderx{#1}#2 \relax}
+% #1 is the data type, #2 the name, #3 the args.
+\def\deftypefunheaderx #1#2 #3\relax{%
+\doind {fn}{\code{#2}}% Make entry in function index
+\begingroup\defname {\code{#1} #2}{Function}%
+\deftypefunargs {#3}\endgroup %
+\catcode 61=\other % Turn off change made in \defparsebody
+}
+
+% @deftypefn {Library Function} int foobar (int @var{foo}, float @var{bar})
+
+\def\deftypefn{\defmethparsebody\Edeftypefn\deftypefnx\deftypefnheader}
+
+% #1 is the classification.  #2 is the data type.  #3 is the name and args.
+\def\deftypefnheader #1#2#3{\deftypefnheaderx{#1}{#2}#3 \relax}
+% #1 is the classification, #2 the data type, #3 the name, #4 the args.
+\def\deftypefnheaderx #1#2#3 #4\relax{%
+\doind {fn}{\code{#3}}% Make entry in function index
+\begingroup\defname {\code{#2} #3}{#1}%
+\deftypefunargs {#4}\endgroup %
+\catcode 61=\other % Turn off change made in \defparsebody
+}
+
+% @defmac == @deffn Macro
+
+\def\defmac{\defparsebody\Edefmac\defmacx\defmacheader}
+
+\def\defmacheader #1#2{\doind {fn}{\code{#1}}% Make entry in function index
+\begingroup\defname {#1}{Macro}%
+\defunargs {#2}\endgroup %
+\catcode 61=\other % Turn off change made in \defparsebody
+}
+
+% @defspec == @deffn Special Form
+
+\def\defspec{\defparsebody\Edefspec\defspecx\defspecheader}
+
+\def\defspecheader #1#2{\doind {fn}{\code{#1}}% Make entry in function index
+\begingroup\defname {#1}{Special Form}%
+\defunargs {#2}\endgroup %
+\catcode 61=\other % Turn off change made in \defparsebody
+}
+
+% This definition is run if you use @defunx
+% anywhere other than immediately after a @defun or @defunx.
+
+\def\deffnx #1 {\errmessage{@deffnx in invalid context}}
+\def\defunx #1 {\errmessage{@defunx in invalid context}}
+\def\defmacx #1 {\errmessage{@defmacx in invalid context}}
+\def\defspecx #1 {\errmessage{@defspecx in invalid context}}
+\def\deftypefnx #1 {\errmessage{@deftypefnx in invalid context}}
+\def\deftypeunx #1 {\errmessage{@deftypeunx in invalid context}}
+
+% @defmethod, and so on
+
+% @defop {Funny Method} foo-class frobnicate argument
+
+\def\defop #1 {\def\defoptype{#1}%
+\defopparsebody\Edefop\defopx\defopheader\defoptype}
+
+\def\defopheader #1#2#3{%
+\dosubind {fn}{\code{#2}}{on #1}% Make entry in function index
+\begingroup\defname {#2}{\defoptype{} on #1}%
+\defunargs {#3}\endgroup %
+}
+
+% @defmethod == @defop Method
+
+\def\defmethod{\defmethparsebody\Edefmethod\defmethodx\defmethodheader}
+
+\def\defmethodheader #1#2#3{%
+\dosubind {fn}{\code{#2}}{on #1}% entry in function index
+\begingroup\defname {#2}{Method on #1}%
+\defunargs {#3}\endgroup %
+}
+
+% @defcv {Class Option} foo-class foo-flag
+
+\def\defcv #1 {\def\defcvtype{#1}%
+\defopvarparsebody\Edefcv\defcvx\defcvarheader\defcvtype}
+
+\def\defcvarheader #1#2#3{%
+\dosubind {vr}{\code{#2}}{of #1}% Make entry in var index
+\begingroup\defname {#2}{\defcvtype{} of #1}%
+\defvarargs {#3}\endgroup %
+}
+
+% @defivar == @defcv {Instance Variable}
+
+\def\defivar{\defvrparsebody\Edefivar\defivarx\defivarheader}
+
+\def\defivarheader #1#2#3{%
+\dosubind {vr}{\code{#2}}{of #1}% Make entry in var index
+\begingroup\defname {#2}{Instance Variable of #1}%
+\defvarargs {#3}\endgroup %
+}
+
+% These definitions are run if you use @defmethodx, etc.,
+% anywhere other than immediately after a @defmethod, etc.
+
+\def\defopx #1 {\errmessage{@defopx in invalid context}}
+\def\defmethodx #1 {\errmessage{@defmethodx in invalid context}}
+\def\defcvx #1 {\errmessage{@defcvx in invalid context}}
+\def\defivarx #1 {\errmessage{@defivarx in invalid context}}
+
+% Now @defvar
+
+% First, define the processing that is wanted for arguments of @defvar.
+% This is actually simple: just print them in roman.
+% This must expand the args and terminate the paragraph they make up
+\def\defvarargs #1{\normalparens #1%
+\interlinepenalty=10000
+\endgraf\penalty 10000\vskip -\parskip\penalty 10000}
+
+% @defvr Counter foo-count
+
+\def\defvr{\defvrparsebody\Edefvr\defvrx\defvrheader}
+
+\def\defvrheader #1#2#3{\doind {vr}{\code{#2}}%
+\begingroup\defname {#2}{#1}\defvarargs{#3}\endgroup}
+
+% @defvar == @defvr Variable
+
+\def\defvar{\defvarparsebody\Edefvar\defvarx\defvarheader}
+
+\def\defvarheader #1#2{\doind {vr}{\code{#1}}% Make entry in var index
+\begingroup\defname {#1}{Variable}%
+\defvarargs {#2}\endgroup %
+}
+
+% @defopt == @defvr {User Option}
+
+\def\defopt{\defvarparsebody\Edefopt\defoptx\defoptheader}
+
+\def\defoptheader #1#2{\doind {vr}{\code{#1}}% Make entry in var index
+\begingroup\defname {#1}{User Option}%
+\defvarargs {#2}\endgroup %
+}
+
+% @deftypevar int foobar
+
+\def\deftypevar{\defvarparsebody\Edeftypevar\deftypevarx\deftypevarheader}
+
+% #1 is the data type.  #2 is the name.
+\def\deftypevarheader #1#2{%
+\doind {vr}{\code{#2}}% Make entry in variables index
+\begingroup\defname {\code{#1} #2}{Variable}%
+\interlinepenalty=10000
+\endgraf\penalty 10000\vskip -\parskip\penalty 10000
+\endgroup}
+
+% @deftypevr {Global Flag} int enable
+
+\def\deftypevr{\defvrparsebody\Edeftypevr\deftypevrx\deftypevrheader}
+
+\def\deftypevrheader #1#2#3{\doind {vr}{\code{#3}}%
+\begingroup\defname {\code{#2} #3}{#1}
+\interlinepenalty=10000
+\endgraf\penalty 10000\vskip -\parskip\penalty 10000
+\endgroup}
+
+% This definition is run if you use @defvarx
+% anywhere other than immediately after a @defvar or @defvarx.
+
+\def\defvrx #1 {\errmessage{@defvrx in invalid context}}
+\def\defvarx #1 {\errmessage{@defvarx in invalid context}}
+\def\defoptx #1 {\errmessage{@defoptx in invalid context}}
+\def\deftypevarx #1 {\errmessage{@deftypevarx in invalid context}}
+\def\deftypevrx #1 {\errmessage{@deftypevrx in invalid context}}
+
+% Now define @deftp
+% Args are printed in bold, a slight difference from @defvar.
+
+\def\deftpargs #1{\bf \defvarargs{#1}}
+
+% @deftp Class window height width ...
+
+\def\deftp{\deftpparsebody\Edeftp\deftpx\deftpheader}
+
+\def\deftpheader #1#2#3{\doind {tp}{\code{#2}}%
+\begingroup\defname {#2}{#1}\deftpargs{#3}\endgroup}
+
+% This definition is run if you use @deftpx, etc
+% anywhere other than immediately after a @deftp, etc.
+
+\def\deftpx #1 {\errmessage{@deftpx in invalid context}}
+
+\message{cross reference,}
+% Define cross-reference macros
+\newwrite \auxfile
+
+\newif\ifhavexrefs  % True if xref values are known.
+\newif\ifwarnedxrefs  % True if we warned once that they aren't known.
+
+% \setref{foo} defines a cross-reference point named foo.
+
+\def\setref#1{%
+%\dosetq{#1-title}{Ytitle}%
+\dosetq{#1-pg}{Ypagenumber}%
+\dosetq{#1-snt}{Ysectionnumberandtype}}
+
+\def\unnumbsetref#1{%
+%\dosetq{#1-title}{Ytitle}%
+\dosetq{#1-pg}{Ypagenumber}%
+\dosetq{#1-snt}{Ynothing}}
+
+\def\appendixsetref#1{%
+%\dosetq{#1-title}{Ytitle}%
+\dosetq{#1-pg}{Ypagenumber}%
+\dosetq{#1-snt}{Yappendixletterandtype}}
+
+% \xref, \pxref, and \ref generate cross-references to specified points.
+% For \xrefX, #1 is the node name, #2 the name of the Info
+% cross-reference, #3 the printed node name, #4 the name of the Info
+% file, #5 the name of the printed manual.  All but the node name can be
+% omitted.
+%
+\def\pxref#1{see \xrefX[#1,,,,,,,]}
+\def\xref#1{See \xrefX[#1,,,,,,,]}
+\def\ref#1{\xrefX[#1,,,,,,,]}
+\def\xrefX[#1,#2,#3,#4,#5,#6]{\begingroup%
+\def\printedmanual{\ignorespaces #5}%
+\def\printednodename{\ignorespaces #3}%
+%
+\setbox1=\hbox{\printedmanual}%
+\setbox0=\hbox{\printednodename}%
+\ifdim \wd0=0pt%
+\def\printednodename{\ignorespaces #1}%
+%%% Uncommment the following line to make the actual chapter or section title
+%%% appear inside the square brackets.
+%\def\printednodename{#1-title}%
+\fi%
+%
+%
+% If we use \unhbox0 and \unhbox1 to print the node names, TeX does
+% not insert empty discretionaries after hyphens, which means that it
+% will not find a line break at a hyphen in a node names.  Since some
+% manuals are best written with fairly long node names, containing
+% hyphens, this is a loss.  Therefore, we simply give the text of
+% the node name again, so it is as if TeX is seeing it for the first
+% time.
+\ifdim \wd1>0pt
+section ``\printednodename'' in \cite{\printedmanual}%
+\else%
+\turnoffactive%
+\refx{#1-snt}{} [\printednodename], page\tie\refx{#1-pg}{}%
+\fi
+\endgroup}
+
+% \dosetq is the interface for calls from other macros
+
+% Use \turnoffactive so that punctuation chars such as underscore
+% work in node names.
+\def\dosetq #1#2{{\let\folio=0 \turnoffactive%
+\edef\next{\write\auxfile{\internalsetq {#1}{#2}}}%
+\next}}
+
+% \internalsetq {foo}{page} expands into
+% CHARACTERS 'xrdef {foo}{...expansion of \Ypage...}
+% When the aux file is read, ' is the escape character
+
+\def\internalsetq #1#2{'xrdef {#1}{\csname #2\endcsname}}
+
+% Things to be expanded by \internalsetq
+
+\def\Ypagenumber{\folio}
+
+\def\Ytitle{\thischapter}
+
+\def\Ynothing{}
+
+\def\Ysectionnumberandtype{%
+\ifnum\secno=0 Chapter\xreftie\the\chapno %
+\else \ifnum \subsecno=0 Section\xreftie\the\chapno.\the\secno %
+\else \ifnum \subsubsecno=0 %
+Section\xreftie\the\chapno.\the\secno.\the\subsecno %
+\else %
+Section\xreftie\the\chapno.\the\secno.\the\subsecno.\the\subsubsecno %
+\fi \fi \fi }
+
+\def\Yappendixletterandtype{%
+\ifnum\secno=0 Appendix\xreftie'char\the\appendixno{}%
+\else \ifnum \subsecno=0 Section\xreftie'char\the\appendixno.\the\secno %
+\else \ifnum \subsubsecno=0 %
+Section\xreftie'char\the\appendixno.\the\secno.\the\subsecno %
+\else %
+Section\xreftie'char\the\appendixno.\the\secno.\the\subsecno.\the\subsubsecno %
+\fi \fi \fi }
+
+\gdef\xreftie{'tie}
+
+% Use TeX 3.0's \inputlineno to get the line number, for better error
+% messages, but if we're using an old version of TeX, don't do anything.
+%
+\ifx\inputlineno\thisisundefined
+  \let\linenumber = \empty % Non-3.0.
+\else
+  \def\linenumber{\the\inputlineno:\space}
+\fi
+
+% Define \refx{NAME}{SUFFIX} to reference a cross-reference string named NAME.
+% If its value is nonempty, SUFFIX is output afterward.
+
+\def\refx#1#2{%
+  \expandafter\ifx\csname X#1\endcsname\relax
+    % If not defined, say something at least.
+    $\langle$un\-de\-fined$\rangle$%
+    \ifhavexrefs
+      \message{\linenumber Undefined cross reference `#1'.}%
+    \else
+      \ifwarnedxrefs\else
+        \global\warnedxrefstrue
+        \message{Cross reference values unknown; you must run TeX again.}%
+      \fi
+    \fi
+  \else
+    % It's defined, so just use it.
+    \csname X#1\endcsname
+  \fi
+  #2% Output the suffix in any case.
+}
+
+% Read the last existing aux file, if any.  No error if none exists.
+
+% This is the macro invoked by entries in the aux file.
+\def\xrdef #1#2{
+{\catcode`\'=\other\expandafter \gdef \csname X#1\endcsname {#2}}}
+
+\def\readauxfile{%
+\begingroup
+\catcode `\^^@=\other
+\catcode `\
=\other
+\catcode `\=\other
+\catcode `\^^C=\other
+\catcode `\^^D=\other
+\catcode `\^^E=\other
+\catcode `\^^F=\other
+\catcode `\^^G=\other
+\catcode `\^^H=\other
+\catcode `\=\other
+\catcode `\^^L=\other
+\catcode `\=\other
+\catcode `\=\other
+\catcode `\=\other
+\catcode `\=\other
+\catcode `\=\other
+\catcode `\=\other
+\catcode `\=\other
+\catcode `\=\other
+\catcode `\=\other
+\catcode `\=\other
+\catcode `\=\other
+\catcode `\=\other
+\catcode 26=\other
+\catcode `\^^[=\other
+\catcode `\^^\=\other
+\catcode `\^^]=\other
+\catcode `\^^^=\other
+\catcode `\^^_=\other
+\catcode `\@=\other
+\catcode `\^=\other
+\catcode `\~=\other
+\catcode `\[=\other
+\catcode `\]=\other
+\catcode`\"=\other
+\catcode`\_=\other
+\catcode`\|=\other
+\catcode`\<=\other
+\catcode`\>=\other
+\catcode `\$=\other
+\catcode `\#=\other
+\catcode `\&=\other
+% `\+ does not work, so use 43.
+\catcode 43=\other
+% the aux file uses ' as the escape.
+% Turn off \ as an escape so we do not lose on
+% entries which were dumped with control sequences in their names.
+% For example, 'xrdef {$\leq $-fun}{page ...} made by @defun ^^
+% Reference to such entries still does not work the way one would wish,
+% but at least they do not bomb out when the aux file is read in.
+\catcode `\{=1 \catcode `\}=2
+\catcode `\%=\other
+\catcode `\'=0
+\catcode `\\=\other
+\openin 1 \jobname.aux
+\ifeof 1 \else \closein 1 \input \jobname.aux \global\havexrefstrue
+\global\warnedobstrue
+\fi
+% Open the new aux file.  Tex will close it automatically at exit.
+\openout \auxfile=\jobname.aux
+\endgroup}
+
+
+% Footnotes.
+
+\newcount \footnoteno
+
+% The trailing space in the following definition for supereject is
+% vital for proper filling; pages come out unaligned when you do a
+% pagealignmacro call if that space before the closing brace is
+% removed.
+\def\supereject{\par\penalty -20000\footnoteno =0 }
+
+% @footnotestyle is meaningful for info output only..
+\let\footnotestyle=\comment
+
+\let\ptexfootnote=\footnote
+
+{\catcode `\@=11
+%
+% Auto-number footnotes.  Otherwise like plain.
+\gdef\footnote{%
+  \global\advance\footnoteno by \@ne
+  \edef\thisfootno{$^{\the\footnoteno}$}%
+  %
+  % In case the footnote comes at the end of a sentence, preserve the
+  % extra spacing after we do the footnote number.
+  \let\@sf\empty
+  \ifhmode\edef\@sf{\spacefactor\the\spacefactor}\/\fi
+  %
+  % Remove inadvertent blank space before typesetting the footnote number.
+  \unskip
+  \thisfootno\@sf
+  \footnotezzz
+}%
+
+% Don't bother with the trickery in plain.tex to not require the
+% footnote text as a parameter.  Our footnotes don't need to be so general.
+%
+\long\gdef\footnotezzz#1{\insert\footins{%
+  % We want to typeset this text as a normal paragraph, even if the
+  % footnote reference occurs in (for example) a display environment.
+  % So reset some parameters.
+  \interlinepenalty\interfootnotelinepenalty
+  \splittopskip\ht\strutbox % top baseline for broken footnotes
+  \splitmaxdepth\dp\strutbox
+  \floatingpenalty\@MM
+  \leftskip\z@skip
+  \rightskip\z@skip
+  \spaceskip\z@skip
+  \xspaceskip\z@skip
+  \parindent\defaultparindent
+  %
+  % Hang the footnote text off the number.
+  \hang
+  \textindent{\thisfootno}%
+  %
+  % Don't crash into the line above the footnote text.  Since this
+  % expands into a box, it must come within the paragraph, lest it
+  % provide a place where TeX can split the footnote.
+  \footstrut
+  #1\strut}%
+}
+
+}%end \catcode `\@=11
+
+% Set the baselineskip to #1, and the lineskip and strut size
+% correspondingly.  There is no deep meaning behind these magic numbers
+% used as factors; they just match (closely enough) what Knuth defined.
+%
+\def\lineskipfactor{.08333}
+\def\strutheightpercent{.70833}
+\def\strutdepthpercent {.29167}
+%
+\def\setleading#1{%
+  \normalbaselineskip = #1\relax
+  \normallineskip = \lineskipfactor\normalbaselineskip
+  \normalbaselines
+  \setbox\strutbox =\hbox{%
+    \vrule width0pt height\strutheightpercent\baselineskip
+                    depth \strutdepthpercent \baselineskip
+  }%
+}
+
+% @| inserts a changebar to the left of the current line.  It should
+% surround any changed text.  This approach does *not* work if the
+% change spans more than two lines of output.  To handle that, we would
+% have adopt a much more difficult approach (putting marks into the main
+% vertical list for the beginning and end of each change).
+%
+\def\|{%
+  % \vadjust can only be used in horizontal mode.
+  \leavevmode
+  %
+  % Append this vertical mode material after the current line in the output.
+  \vadjust{%
+    % We want to insert a rule with the height and depth of the current
+    % leading; that is exactly what \strutbox is supposed to record.
+    \vskip-\baselineskip
+    %
+    % \vadjust-items are inserted at the left edge of the type.  So
+    % the \llap here moves out into the left-hand margin.
+    \llap{%
+      %
+      % For a thicker or thinner bar, change the `1pt'.
+      \vrule height\baselineskip width1pt
+      %
+      % This is the space between the bar and the text.
+      \hskip 12pt
+    }%
+  }%
+}
+
+% For a final copy, take out the rectangles
+% that mark overfull boxes (in case you have decided
+% that the text looks ok even though it passes the margin).
+%
+\def\finalout{\overfullrule=0pt}
+
+
+% End of control word definitions.
+
+\message{and turning on texinfo input format.}
+
+\def\openindices{%
+   \newindex{cp}%
+   \newcodeindex{fn}%
+   \newcodeindex{vr}%
+   \newcodeindex{tp}%
+   \newcodeindex{ky}%
+   \newcodeindex{pg}%
+}
+
+% Set some numeric style parameters, for 8.5 x 11 format.
+
+%\hsize = 6.5in
+\newdimen\defaultparindent \defaultparindent = 15pt
+\parindent = \defaultparindent
+\parskip 18pt plus 1pt
+\setleading{15pt}
+\advance\topskip by 1.2cm
+
+% Prevent underfull vbox error messages.
+\vbadness=10000
+
+% Following George Bush, just get rid of widows and orphans.
+\widowpenalty=10000
+\clubpenalty=10000
+
+% Use TeX 3.0's \emergencystretch to help line breaking, but if we're
+% using an old version of TeX, don't do anything.  We want the amount of
+% stretch added to depend on the line length, hence the dependence on
+% \hsize.  This makes it come to about 9pt for the 8.5x11 format.
+%
+\ifx\emergencystretch\thisisundefined
+  % Allow us to assign to \emergencystretch anyway.
+  \def\emergencystretch{\dimen0}%
+\else
+  \emergencystretch = \hsize
+  \divide\emergencystretch by 45
+\fi
+
+% Use @smallbook to reset parameters for 7x9.5 format  (or else 7x9.25)
+\def\smallbook{
+
+% These values for secheadingskip and subsecheadingskip are
+% experiments.  RJC 7 Aug 1992
+\global\secheadingskip = 17pt plus 6pt minus 3pt
+\global\subsecheadingskip = 14pt plus 6pt minus 3pt
+
+\global\lispnarrowing = 0.3in
+\setleading{12pt}
+\advance\topskip by -1cm
+\global\parskip 3pt plus 1pt
+\global\hsize = 5in
+\global\vsize=7.5in
+\global\tolerance=700
+\global\hfuzz=1pt
+\global\contentsrightmargin=0pt
+
+\global\pagewidth=\hsize
+\global\pageheight=\vsize
+
+\global\let\smalllisp=\smalllispx
+\global\let\smallexample=\smalllispx
+\global\def\Esmallexample{\Esmalllisp}
+}
+
+% Use @afourpaper to print on European A4 paper.
+\def\afourpaper{
+\global\tolerance=700
+\global\hfuzz=1pt
+\setleading{12pt}
+\global\parskip 15pt plus 1pt
+
+\global\vsize= 53\baselineskip
+\advance\vsize by \topskip
+%\global\hsize=   5.85in     % A4 wide 10pt
+\global\hsize=  6.5in
+\global\outerhsize=\hsize
+\global\advance\outerhsize by 0.5in
+\global\outervsize=\vsize
+\global\advance\outervsize by 0.6in
+
+\global\pagewidth=\hsize
+\global\pageheight=\vsize
+}
+
+% Define macros to output various characters with catcode for normal text.
+\catcode`\"=\other
+\catcode`\~=\other
+\catcode`\^=\other
+\catcode`\_=\other
+\catcode`\|=\other
+\catcode`\<=\other
+\catcode`\>=\other
+\catcode`\+=\other
+\def\normaldoublequote{"}
+\def\normaltilde{~}
+\def\normalcaret{^}
+\def\normalunderscore{_}
+\def\normalverticalbar{|}
+\def\normalless{<}
+\def\normalgreater{>}
+\def\normalplus{+}
+
+% This macro is used to make a character print one way in ttfont
+% where it can probably just be output, and another way in other fonts,
+% where something hairier probably needs to be done.
+%
+% #1 is what to print if we are indeed using \tt; #2 is what to print
+% otherwise.  Since all the Computer Modern typewriter fonts have zero
+% interword stretch (and shrink), and it is reasonable to expect all
+% typewriter fonts to have this, we can check that font parameter.
+%
+\def\ifusingtt#1#2{\ifdim \fontdimen3\the\font=0pt #1\else #2\fi}
+
+% Turn off all special characters except @
+% (and those which the user can use as if they were ordinary).
+% Most of these we simply print from the \tt font, but for some, we can
+% use math or other variants that look better in normal text.
+
+\catcode`\"=\active
+\def\activedoublequote{{\tt \char '042}}
+\let"=\activedoublequote
+\catcode`\~=\active
+\def~{{\tt \char '176}}
+\chardef\hat=`\^
+\catcode`\^=\active
+\def^{{\tt \hat}}
+
+\catcode`\_=\active
+\def_{\ifusingtt\normalunderscore\_}
+% Subroutine for the previous macro.
+\def\_{\lvvmode \kern.06em \vbox{\hrule width.3em height.1ex}}
+
+% \lvvmode is equivalent in function to \leavevmode.
+% Using \leavevmode runs into trouble when written out to
+% an index file due to the expansion of \leavevmode into ``\unhbox
+% \voidb@x'' ---which looks to TeX like ``\unhbox \voidb\x'' due to our
+% magic tricks with @.
+\def\lvvmode{\vbox to 0pt{}}
+
+\catcode`\|=\active
+\def|{{\tt \char '174}}
+\chardef \less=`\<
+\catcode`\<=\active
+\def<{{\tt \less}}
+\chardef \gtr=`\>
+\catcode`\>=\active
+\def>{{\tt \gtr}}
+\catcode`\+=\active
+\def+{{\tt \char 43}}
+%\catcode 27=\active
+%\def^^[{$\diamondsuit$}
+
+% Used sometimes to turn off (effectively) the active characters
+% even after parsing them.
+\def\turnoffactive{\let"=\normaldoublequote
+\let~=\normaltilde
+\let^=\normalcaret
+\let_=\normalunderscore
+\let|=\normalverticalbar
+\let<=\normalless
+\let>=\normalgreater
+\let+=\normalplus}
+
+% Set up an active definition for =, but don't enable it most of the time.
+{\catcode`\==\active
+\global\def={{\tt \char 61}}}
+
+\catcode`\@=0
+
+% \rawbackslashxx output one backslash character in current font
+\global\chardef\rawbackslashxx=`\\
+%{\catcode`\\=\other
+%@gdef@rawbackslashxx{\}}
+
+% \rawbackslash redefines \ as input to do \rawbackslashxx.
+{\catcode`\\=\active
+@gdef@rawbackslash{@let\=@rawbackslashxx }}
+
+% \normalbackslash outputs one backslash in fixed width font.
+\def\normalbackslash{{\tt\rawbackslashxx}}
+
+% Say @foo, not \foo, in error messages.
+\escapechar=`\@
+
+% \catcode 17=0   % Define control-q
+\catcode`\\=\active
+
+% If a .fmt file is being used, we don't want the `\input texinfo' to show up.
+% That is what \eatinput is for; after that, the `\' should revert to printing
+% a backslash.
+%
+@gdef@eatinput input texinfo{@fixbackslash}
+@global@let\ = @eatinput
+
+% On the other hand, perhaps the file did not have a `\input texinfo'. Then
+% the first `\{ in the file would cause an error. This macro tries to fix
+% that, assuming it is called before the first `\' could plausibly occur.
+%
+@gdef@fixbackslash{@ifx\@eatinput @let\ = @normalbackslash @fi}
+
+%% These look ok in all fonts, so just make them not special.  The @rm below
+%% makes sure that the current font starts out as the newly loaded cmr10
+@catcode`@$=@other @catcode`@%=@other @catcode`@&=@other @catcode`@#=@other
+
+@textfonts
+@rm
+
+@c Local variables:
+@c page-delimiter: "^\\\\message"
+@c End: