Merge the trunk onto the tdfx-2-1 branch. Now up to date with 4.0.1.tdfx-2-1-20000711

27 files changed, 8656 insertions, 0 deletions

diff --git a/xc/extras/rman/ANNOUNCE-rman b/xc/extras/rman/ANNOUNCE-rman
new file mode 100644
index 000000000..209b843a9
--- /dev/null
+++ b/xc/extras/rman/ANNOUNCE-rman
@@ -0,0 +1,78 @@
+======================================================================
+
+PolyglotMan (nee RosettaMan) is a filter for UNIX manual pages.  It
+takes as input man pages for a variety of UNIX flavors and produces as
+output a variety of file formats.  Currently PolyglotMan accepts man
+pages from the following flavors of UNIX: Hewlett-Packard HP-UX, AT&T
+System V, SunOS, Sun Solaris, OSF/1, DEC Ultrix, SGI IRIX, Linux, SCO,
+FreeBSD; and produces output for the following formats: printable
+ASCII only (stripping page headers and footers), section and
+subsection headers only, TkMan, [tn]roff, RTF, SGML (soon--I finally
+found a DTD), HTML, MIME, LaTeX, LaTeX 2e, Perl 5's pod.  Previously
+<I>PolyglotMan</I> required pages to be formatted by nroff prior to
+its processing; with version 3.0, it prefers [tn]roff source and
+usually can produce results that are better yet.
+
+PolyglotMan improves upon other man page filters in several ways: (1) its
+analysis recognizes the structural pieces of man pages, enabling high
+quality output, (2) its modular structure permits easy augmentation of
+output formats, (3) it accepts man pages formatted with the variant
+macros of many different flavors of UNIX, and (4) it doesn't require
+modification of or cooperation with any other program.
+
+PolyglotMan is a rewrite of TkMan's man page filter, called bs2tk.  (If
+you haven't heard about TkMan, a hypertext man page browser, you
+should grab it via anonymous ftp from ftp.cs.berkeley.edu:
+/ucb/people/phelps/tkman.tar.Z.)  Whereas bs2tk generated output only for
+TkMan, PolyglotMan generalizes the process so that the analysis can be
+leveraged to new output formats.  A single analysis engine recognizes
+section heads, subsection heads, body text, lists, references to other
+man pages, boldface, italics, bold italics, special characters (like
+bullets), tables (to a degree) and strips out page headers and
+footers.  The engine sends signals to the selected output functions so
+that an enhancement in the engine improves the quality of output of
+all of them.  Output format functions are easy to add, and thus far
+average about about 75 lines of C code each.
+
+A note for HTML consumers:  This filter does real (heuristic) parsing--
+no <PRE>!  Man page references are turned into hypertext links.  The files 
+<URL:ftp://ftp.cs.berkeley.edu/ucb/people/phelps/tcltk/sgi-ls.1.html>
+and <URL:ftp://ftp.cs.berkeley.edu/ucb/people/phelps/tcltk/ksh.1.html>
+are examples of the quality of output produced entirely automatically
+(no retouching) by PolyglotMan.  These translations were produced by
+PolyglotMan starting with the [tn]roff source (again no retouching).
+Several people have extended World Wide Web servers to format man pages 
+on the fly.  Check the README file in the contrib directory for a list.
+
+
+CHANGES in 3.0
+
+* [tn]roff source preferred for superior results, when roff macros are
+  sufficiently recognized.  Autodetection of source or formatted input.
+* New software license that makes it free for any use
+
+
+CHANGES in 2.5
+
+* SGML output format that adheres to Davenport DocBook v2.3 DTD
+  (NOT READY IN CURRENT VERSION!)
+* MIME output format, for e-mail and Emacs 19.29's enriched mode
+  (Neal Becker)
+* port to Macintosh by Matthias Neeracher
+* list of valid volume names can be given as a parameter (Dag Nygren)
+* updated to LaTeX2e (H. Palme)
+* debugging scaffolding erected (at the end of software's development cycle!)
+
+
+CHANGES in 2.2
+
+* when in SEE ALSO, hyphens would confuse man page-reference finder, 
+  so re-linebreak if necessary to eliminate them (!) (Greg Earle & Uri Guttman)
+
+
+CHANGES in 2.1
+
+* gets() replaced by custom code.  gets() deprecated since it reads until \0, 
+  introducing security problems. (Robert Withrow)
+
+* TkMan module revised for Tk 4.0
diff --git a/xc/extras/rman/CHANGES b/xc/extras/rman/CHANGES
new file mode 100644
index 000000000..e37d026ad
--- /dev/null
+++ b/xc/extras/rman/CHANGES
@@ -0,0 +1,100 @@
+1993
+ 1 Apr	as bs2tk posted to comp.lang.tcl (126 lines)
+ 2	bullets, change bars, copyright symbol
+ 5	boldface, other SGI nicks
+ 7	skip unrecognized escape codes
+10	small caps
+13	underscores considered uppercase so show up in default small caps font
+	screen out Ultrix junk (code getting pretty tangled now)
+14	until Tk text has better tab support, replace tabs by spaces until get to next tab stop (for Ultrix); -t gives tabstop spacing
+20	Solaris support (Larry Tsui)
+ 3 Jun	section subheading parsing (Per-Erik Martin)
+28	hyphenated man pages in SEE ALSO show up correctly in Links (Mike Steele)
+13 Jul	under FILES, fully qualified path names are added to Links, but this taken out immediately because not useful
+14	option to keep changebars on right (Warren Jessop)
+ 5 Aug	search for header, footer dynamically--no need to edit or search large list of patterns
+11	-m kicks in man page formatting beyond nroff backspace kludges
+27	handle double digit numbers better by trying again relative to end of line
+19 Sep	-T gives Tk extras (otherwise ASCII only)
+	-H gives headers only (implies -T off)
+10 Oct	-r reverse compiles to [tn]roff source (as Geoff Collyer's nam and fontch, but leveraging existing analysis so only addition of ~60 lines)  (The code is device-driver obscure now--obfuscated C contest next.)
+13	header and footer optionally available at bottom in Tk view (Marty Leisner)
+19	"reflected" odd and even page headers&footers zapped
+20	keep count of sections and subsections, using smaller font for larger numbers
+ 1 Nov	reverse compiles to Ensemble, except for character ranges
+ 4	started rman rewrite for cleaner support of multiple output targets, including: plain ascii, headers only, TkMan, [nt]roff, Ensemble, SGML, HTML
+ 5	line filtering separated from other logic despite greater sophistication, RosettaMan faster than bs2tk (!)
+28 Dec	man page reference recognition (Michael Harrison)
+
+
+1994
+ 1 Jan	identify descriptive lists by comparing scnt2 with s_avg
+ 3	tail-end table of contents in HTML documents
+ 5	-f <filter> and LaTeX output mode
+24	proof-of-concept RTF output mode
+26	handle man pages that don't have a header on the first page
+28	parse "handwritten" man pages
+22 Feb	alpha version released
+ 6 Mar	various bug fixes
+10	beta version released
+13 Jun	fixed surious generation on <DL>'s (the existence of which pointed out by David Sibley)
+22 Jul	table recognition experiment.  works reasonably well, except for tables with centered headers
+ 3 Aug	allow for off-by-one (and -two) in identification of header and footer
+	fixed problem with recurrent/leftover text with OSF/1 bold bullets (yeesh)
+12 Sep	2.0gamma released
+13	check for *third* header, possibly centered, possibly after blank lines (Charles Anderson)
+	fixed tag ranges for lines following blank lines (just \n) of pages with global indentation (Owen Rees)
+19	fixed two small problems with LaTeX (^ => \^, \bullet => $\bullet$) (Neal Becker)
+24	simple check for erroneously being fed roff source
+26	deal with bold +- as in ksh (ugh)
+30	2.0delta released
+ 9 Oct	special check for OSF to guard against section head interpreted as footer
+ 8 Nov	Perl pod output format (result still needs work, but not much)
+ 7 Dec	2.0epsilon released (last one before final 2.0)
+22	Happy Winter Solstice!  2.0 released
+	deprecated gets() replaced (Robert Withrow)
+25	TkMan module's $w.show => $t, saving about 9% in generated characters
+
+
+1995
+ 1 Jan	experiment with TkMan output to take advantage of my hack to Tk text (i.e., $t insert end "text" => $t insert end "text1" tag1 "text2" tag2 ...); results => output size reduced about 25%, time reduced about 12-15%
+25 Mar	back to old mark command for Tk module
+8 May	hyphens in SEE ALSO section would confuse link-finder, so re-linebreak if necessary(!) (Greg Earle & Uri Guttman)
+ 4 Aug	put formats and options into tables (inspired by Steve Maguire's Writing Solid Code)
+19	-V accepts colon-separated list of valid volume names (Dag Nygren)
+22	MIME output format that's usable in Emacs 19.29 (just three hours at hacking HTML module) (Neal Becker)
+ 9 Sep	nits in HTML and better Solaris code snippets (Drazen Kacar)
+13 Nov	Macintosh port by Matthias Neeracher <neeri@iis.ee.ethz.ch>
+18 Dec	adapted to LaTeX2e, null manRef yields italicized man ref (H. Palme)
+28	allow long option names (Larry Schwimmer)
+
+
+1996
+22 Jan	fixed problem with hyphenation supression and tabs in man page--sheesh! (H. Palme)
+23 May	split TkMan format into Tk and TkMan (which calls Tk)
+25	in TkMan format, initial spaces converted to tabs
+24 Sep	experiment with formatting from source's macros, for better transcription short of full nroff interpreter
+27	commented out Ensemble output format, which nobody used
+ 2 Oct	>4000 lines
+11
+ 8 Nov  released 3.0 alpha.  source code parsing works well for Solaris, SunOS, HP-UX; in generating HTML
+25	recognize URLs (Mic Campanel)
+
+1997
+19 Mar	bug fixes, more special characters (roff expert Larry Jones)
+ 8 Aug  renamed to PolyglotMan
+ 4 Nov	TkMan module: Rebus and NoteMark identification for paragraph lengths and command line options taken over from Tcl (still have search, highlight in Tcl, necessarily) (Chad Loder)
+	>5000 lines, or nearly 40X the lines of code of version 1.0 (which just supported TkMan)
+
+1998
+20 Mar	automatic detection of Tcl/Tk source (within automatic detection of source; already had within formatted)
+	centralize casifying SH and SS for formatted and source
+	group exception table by category
+17 Apr	in source translation, pass along comment lines too
+20	incorporate (RCS) versioning diff's, for HTML
+
+
+Anticipated/Contemplated
+make source translation work on Digital UNIX better--just tweaks now
+fix up other output formats to work well with source parsing
+??-???	SGML/XML output format (DTD found at long last), validated by sgmls/sp
diff --git a/xc/extras/rman/MANIFEST b/xc/extras/rman/MANIFEST
new file mode 100644
index 000000000..a88687fc1
--- /dev/null
+++ b/xc/extras/rman/MANIFEST
@@ -0,0 +1,28 @@
+gcksum crc   length name
+----------   ------ ----
+3783133468 152817 rman.c
+2435527268 1654 getopt.c
+1480364202 71 macish.c
+476345420 3770 ANNOUNCE-rman
+510464631 3311 README-rman
+892917966 4571 Makefile
+236255660 3149 Makefile.mac
+733124117 11877 rman.1
+694318729 13321 rman.html
+1726493852 5810 CHANGES
+3948042252 4732 contrib/README-contrib
+953283015 911 contrib/authried.txt
+641079878 13609 contrib/bennett.txt
+1827405570 1220 contrib/gzip.patch
+4230791356 250 contrib/hman.cgi
+1743159239 359 contrib/hman.ksh
+2919099478 8005 contrib/hman.pl
+1821945783 4978 contrib/http-rman.c
+3091246851 661 contrib/http-rman.html
+2646244070 350 contrib/lewis.pl
+2860919314 2049 contrib/man2html
+3466647040 7531 contrib/rman_html_split
+75383086 2482 contrib/rman_html_split.1
+2032677288 150 contrib/sco-wrapper.sh
+884546947 3806 contrib/sutter.txt
+3112317992 5086 contrib/youki.pl
diff --git a/xc/extras/rman/Makefile b/xc/extras/rman/Makefile
new file mode 100644
index 000000000..1e348a959
--- /dev/null
+++ b/xc/extras/rman/Makefile
@@ -0,0 +1,166 @@
+#
+# Makefile for PolyglotMan
+#   It's helpful to read the README-rman file first.
+#   You should read over all parts of this file,
+#      down to the "you shouldn't modify" line
+#
+# Tom Phelps (phelps@ACM.org)
+#
+
+
+### you need to localize the paths on these lines
+
+# The executable `rman' is placed in BINDIR.
+# If you're also installing TkMan (available separately--see README-rman),
+#    this must be a directory that's in your bin PATH.
+# MANDIR holds the man page.
+
+BINDIR = /private/share/bin
+#BINDIR = /usr/local/bin
+MANDIR = /home/auspex/h/bair/phelps/man/man1
+#MANDIR = /usr/local/man/man1
+
+
+### if you have GNU gcc v2.x, use these definitions
+
+CC = gcc
+# if you don't have gcc v2.x, try -O for CFLAGS
+CFLAGS = -O2 -finline-functions
+
+### if you just have a standard UNIX, use these instead of GNU.
+### CC must be an ANSI C compiler
+
+#CC = cc
+#CFLAGS = -O
+
+# Solaris and SysV people may need this
+#CFLAGS = -O2 -finline-functions
+
+# For HP-UX
+#CC = cc
+#CFLAGS = -Aa -O
+# HP-UX 10.20
+#CFLAGS = -Ae -O
+
+# DEC Alpha and Ultrix, -std1 needed to conform to ANSI C
+#CC = cc
+#CFLAGS = -std1 -O3 -Olimit 1000
+
+
+# list of valid volume numbers and letters
+# you can also set these at runtime with -V
+VOLLIST = "1:2:3:4:5:6:7:8:9:o:l:n:p"
+# SCO Unix has expanded set of volume letters
+#VOLLIST = "1:2:3:4:5:6:7:8:9:o:l:n:p:C:X:S:L:M:F:G:A:H"
+# SGI and UnixWare 2.0
+#VOLLIST = "1:2:3:4:5:6:7:8:9:o:l:n:p:D"
+
+
+# the printf strings used to set the <TITLE> and
+# to set URL hyperlinks to referenced manual pages
+# can be defined at runtime.  The defaults are defined below.
+# The first %s parameter is the manual page name,
+# the second the volume/section number.
+# you can set these at runtime with -l and -r, respectively
+
+MANTITLEPRINTF = "%s(%s) manual page"
+# relative link to pregenerated file in same directory
+MANREFPRINTF = "%s.%s"
+# on-the-fly through a cgi-bin script
+#MANREFPRINTF = "/cgi-bin/man2html?%s&%s"
+
+
+# # # these lines are probably fine
+
+CP = cp
+# or you can use GNU's cp and backup files that are about to be overwritten
+#CP = cp -b
+RM = rm
+
+
+#--------------------------------------------------
+#
+# you shouldn't modify anything below here
+#
+#--------------------------------------------------
+
+version = 3.0.8
+rman = rman-$(version)
+srcs = rman.c getopt.c macish.c
+objs = rman
+defs = -DVOLLIST='$(VOLLIST)' -DMANTITLEPRINTF='$(MANTITLEPRINTF)' -DMANREFPRINTF='$(MANREFPRINTF)'
+libs = 
+aux = ANNOUNCE-rman README-rman Makefile Makefile.mac rman.1 rman.html CHANGES
+distrib = $(srcs) $(libs) $(aux) contrib
+
+
+all:  rman
+	@echo 'Files made in current directory.'
+	@echo 'You should "make install".'
+
+# everyone but me zaps assertions with the -DNDEBUG flag
+rman: rman.c Makefile
+	$(CC) -DNDEBUG $(defs) -DPOLYGLOTMANVERSION=\"$(version)\" $(CFLAGS) -o rman rman.c
+
+dox: rman.1 ANNOUNCE-rman
+
+ANNOUNCE-rman: Announce-rman
+	sed -e /=========/,/bogusline/p -e d Announce-rman > ANNOUNCE-rman
+
+rman.1: rman.html rman
+	echo '.TH PolyglotMan 1' > rman.1
+	cat ../html/html.dtd rman.html | sgmls | ../html/html2man.pl >> rman.1
+
+debug:
+	$(CC) $(defs) -DDEBUG -DPOLYGLOTMANVERSION=\"debug\" -g -Wall -o rman rman.c
+
+prof:
+	quantify -cache-dir=/home/orodruin/h/bair/phelps/spine/rman/cache $(CC) -DNDEBUG $(defs) -DPOLYGLOTMANVERSION=\"QUANTIFY\" -g -o rman rman.c
+
+install: rman
+#	$(INSTALL) -s rman $(BINDIR)
+	$(RM) -f $(BINDIR)/rman
+	$(CP) rman $(BINDIR)
+	$(RM) -f $(MANDIR)/rman.1
+	$(CP) rman.1 $(MANDIR)
+
+# test version includes assertions
+#	ginstall rman $(BINDIR)/`arch`
+test: rman.c Makefile
+	$(CC) $(defs) -DPOLYGLOTMANVERSION=\"$(version)\" $(CFLAGS) -Wall -ansi -pedantic -o rman rman.c
+	ls -l rman
+	ginstall rman $(BINDIR)
+	rman -v
+	rman --help
+	@echo 'Assertion checks:'
+	rman -f html weirdman/hp-tbl.1 > /dev/null
+	rman -f html weirdman/Pnews.1 > /dev/null
+	nroff -man rman.1 | rman -f html > /dev/null
+
+sww:
+	rm -f rman $(wildcard ~/bin/{sun4,snake,alpha}/rman)
+	rman
+
+clean:
+	rm -f $(objs)
+
+dist: dox
+	rm -rf $(rman)*
+	mkdir $(rman)
+	$(CP) -R $(distrib) $(rman)
+#	expand -4 rman.c > $(rman)/rman.c
+	rm -f $(rman)/contrib/*~
+	@echo 'gcksum crc   length name' > MANIFEST
+	@echo '----------   ------ ----' >> MANIFEST
+	@gcksum $(filter-out contrib, $(filter-out %~, $(distrib) $(wildcard contrib/*))) | gtee -a MANIFEST
+	mv MANIFEST $(rman)
+	gtar chvf $(rman).tar $(rman)
+	compress $(rman).tar
+#	gzip -9v $(rman).tar
+	rm -rf $(rman) ANNOUNCE-rman rman.1
+	@echo "*** Did you remember to ci -l first?"
+
+uu: tar
+	gznew $(rman).tar.Z
+	echo 'uudecode, gunzip (from GNU), untar' > $(rman).tar.gz.uu
+	uuencode $(rman).tar.gz $(rman).tar.gz >> $(rman).tar.gz.uu
diff --git a/xc/extras/rman/Makefile.mac b/xc/extras/rman/Makefile.mac
new file mode 100644
index 000000000..2bdd606ba
--- /dev/null
+++ b/xc/extras/rman/Makefile.mac
@@ -0,0 +1,105 @@
+#
+# Makefile for RosettaMan
+#   It's helpful to read the README-rman file first.
+#   You should read over all parts of this file,
+#      down to the "you shouldn't modify" line
+#
+# Tom Phelps (phelps@ACM.org)
+#
+# Mac version: Matthias Neeracher <neeri@iis.ee.ethz.ch>
+#
+
+BINDIR = "{{MPW}}LocalTools:"
+
+version = 2.4
+rman = rman-$(version)
+srcs = rman.c getopt.c macish.c
+defs = -d MANTITLEPRINTF='$(MANTITLEPRINTF)' -d MANREFPRINTF='$(MANREFPRINTF)' \
+		 -d ROSETTAMANVERSION='"$(version)"'
+libs = 
+aux = README-rman Makefile rman.1 rman.html Copyright
+distrib = $(srcs) $(libs) $(aux) contrib
+
+COpt			=	$(defs) -d macintosh -w off -sym on
+C68K			=	MWC68K ${COpt} -mc68020 -model far -mbg on
+CPPC			=	MWCPPC ${COpt} -traceback
+ROptions 	= 	-i :
+Lib68K		=	MWLink68K -xm library -sym on
+LibPPC		=	MWLinkPPC -xm library -sym on
+LOpt			= 	-sym on -w -xm mpwtool
+Link68K		=	MWLink68K ${LOpt} -model far
+LinkPPC		=	MWLinkPPC ${LOpt} 
+
+LibFiles68K	=			"{{MW68KLibraries}}GUSIDispatch.Lib.68K"				\
+							"{{MW68KLibraries}}GUSIMPW.Lib.68K"						\
+							"{{MW68KLibraries}}MPW_Runtime.o.lib"					\
+							"{{MW68KLibraries}}new MPW ANSI (4i/8d) C.68K.Lib"	\
+							"{{MW68KLibraries}}MathLib68K (4i/8d).Lib"			\
+							"{{MW68KLibraries}}MacOS.Lib"								\
+							"{{MW68KLibraries}}GUSI.Lib.68K"							\
+							"{{MW68KLibraries}}CPlusPlus.lib"						\
+							"{{MW68KLibraries}}ToolLibs.o"							\
+							"{{MW68KLibraries}}PLStringFuncs.glue.lib"
+
+LibFilesPPC	=			"{{MWPPCLibraries}}GUSIMPW.Lib.PPC"						\
+							"{{MWPPCLibraries}}GUSI.Lib.PPC"							\
+							"{{MWPPCLibraries}}MWStdCRuntime.Lib"					\
+							"{{MWPPCLibraries}}InterfaceLib"							\
+							"{{MWPPCLibraries}}ANSI (NL) C++.PPC.Lib"				\
+							"{{MWPPCLibraries}}ANSI (NL) C.PPC.Lib"				\
+							"{{MWPPCLibraries}}StdCLib"								\
+							"{{MWPPCLibraries}}MathLib"								\
+							"{{MWPPCLibraries}}PPCToolLibs.o"
+
+# the printf strings used to set the <TITLE> and
+# to set URL hyperlinks to referenced manual pages
+# can be defined at runtime.  The defaults are defined below.
+# The first %s parameter is the manual page name,
+# the second the volume/section number.
+
+MANTITLEPRINTF = "%s(%s) manual page"
+MANREFPRINTF = "http://localhost/cgi-bin/man2html?%s?%s"
+CP = Duplicate -y
+
+OBJ68K		=	$(srcs:s/.c/.c.68K.o/)
+OBJPPC		=	$(srcs:s/.c/.c.PPC.o/)
+
+.SOURCE.o	:	":Obj:"
+
+%.c.68K.o		:	%.c
+	Set Echo 0
+	Set Src68K "{{Src68K}} $<"
+%.c.PPC.o		:	%.c
+	Set Echo 0
+	Set SrcPPC "{{SrcPPC}} $<"
+
+all:  rman
+
+install: all
+	Begin
+		Backup -a -to $(BINDIR) rman 
+	End>"{{TempFolder}}"rman.run
+	"{{TempFolder}}"rman.run
+
+rman: rman.68K rman.PPC
+	Duplicate -y rman.PPC rman
+	Echo 'Include "rman.68K" '¶''CODE'¶'';Include "rman.68K" '¶''DATA'¶'';' ¶
+		| Rez -a -c 'MPS ' -t MPST -o rman
+rman.68K:	Objects68K
+	${Link68K} -o rman.68K :Obj:{${OBJ68K}} ${LibFiles68K}
+rman.PPC:	ObjectsPPC
+	${LinkPPC} -o rman.PPC  :Obj:{${OBJPPC}} ${LibFilesPPC}
+
+Objects68K	:	${OBJ68K}
+	Set Echo 1
+	If "{{Src68K}}" != "" 
+		${C68K} -t -fatext {{Src68K}} -o :Obj:
+	End
+	echo > Objects68K
+	
+ObjectsPPC	:	${OBJPPC}	
+	Set Echo 1
+	If "{{SrcPPC}}" != "" 
+		${CPPC} -t -fatext {{SrcPPC}} -o :Obj:
+	End
+	echo > ObjectsPPC
diff --git a/xc/extras/rman/README-rman b/xc/extras/rman/README-rman
new file mode 100644
index 000000000..b53c34b7b
--- /dev/null
+++ b/xc/extras/rman/README-rman
@@ -0,0 +1,69 @@
+*** INSTALLING ***
+
+Set BINDIR in the Makefile to where you keep your binaries and MANDIR
+to where you keep your man pages (in their source form).  (If you're
+using PolyglotMan with TkMan, BINDIR needs to be a component of your
+bin PATH.)  After properly editing the Makefile, type `make install'.
+Thereafter (perhaps after a `rehash') type `rman' to invoke PolyglotMan.
+PolyglotMan requires an ANSI C compiler.  To compile on a Macintosh
+under MPW, use Makefile.mac.
+
+If you send me bug reports and/or suggestions for new features,
+include the version of PolyglotMan (available by typing `rman -v').
+PolyglotMan doesn't parse every aspect of every man page perfectly, but
+if it blows up spectacularly where it doesn't seem like it should, you
+can send me the man page (or a representative man page if it blows up
+on a class of man pages) in BOTH: (1) [tn]roff source form, from
+.../man/manX and (2) formatted form (as formatted by `nroff -man'),
+uuencoded to preserve the control characters, from .../man/catX.
+
+The home location for PolyglotMan is ftp.cs.berkeley.edu:
+/ucb/people/phelps/tcltk/rman.tar.Z (this is a softlink to the latest,
+numbered version).  If you discover a bug and you obtained PolyglotMan
+at some other site, first grab it from this one to see if the problem
+has been fixed.
+
+Be sure to look in the contrib directory for WWW server interfaces,
+a batch converter, and a wrapper for SCO.
+
+--------------------------------------------------
+
+*** NOTES ON CURRENT VERSION ***
+
+Help!  I'm looking for people to help with the following projects.
+(1) Better RTF output format.  The current one works, but could be
+made better.  (2) Extending the macro sets for source recognition.  If
+you write an output format or otherwise improve PolyglotMan, please
+send in your code so that I may share the wealth in future releases.
+(3) Fixing output for various (accented?) characters in the Latin1
+character set.
+
+--------------------------------------------------
+
+
+Copyright (c) 1993-1997 Regents of the University of California
+
+Permission to use, copy, modify, and distribute this software and its
+documentation for educational, research, internal corporate and
+non-profit purposes, without fee, and without a written agreement is
+hereby granted for all cases that do not conflict with the restriction
+in the first sentence of this paragraph, provided that the above
+copyright notice, this paragraph, and the following three paragraphs
+appear in all copies.
+
+Permission to incorporate this software into commercial products may 
+be obtained from the Office of Technology Licensing, 2150 Shattuck 
+Avenue, Suite 510, Berkeley, CA  94704. 
+
+IN NO EVENT SHALL THE UNIVERSITY OF CALIFORNIA BE LIABLE TO ANY PARTY 
+FOR DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES 
+ARISING OUT OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN IF 
+THE UNIVERSITY OF CALIFORNIA HAS BEEN ADVISED OF THE POSSIBILITY OF 
+SUCH DAMAGE.
+
+THE UNIVERSITY OF CALIFORNIA SPECIFICALLY DISCLAIMS ANY WARRANTIES, 
+INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF 
+MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE SOFTWARE 
+PROVIDED HEREUNDER IS ON AN "AS IS" BASIS, AND THE UNIVERSITY OF 
+CALIFORNIA HAS NO OBLIGATION TO PROVIDE MAINTENANCE, SUPPORT, UPDATES, 
+ENHANCEMENTS, OR MODIFICATIONS.
diff --git a/xc/extras/rman/contrib/README-contrib b/xc/extras/rman/contrib/README-contrib
new file mode 100644
index 000000000..111236989
--- /dev/null
+++ b/xc/extras/rman/contrib/README-contrib
@@ -0,0 +1,147 @@
+This directory contains files contributed by PolyglotMan users.  These
+files are not installed, supported or regularly updated to work with
+new versions of PolyglotMan.  They are included because some may find
+them useful, and if so they prevent duplication of effort.
+
+--
+
+* sutter.txt - "Edward L. Sutter" <els@sage.sage.att.com>
+
+"...converting "man" directories with their typical sub-directories of
+man1, man2, etc... to an equivalently structured "html" directory with
+the same sub-directory heirarchy."
+
+"I added an option to rman that allows it to search for the files over
+a specified set of directories.  This allows (for example) manpages under
+man1 that reference something under man3 to be properly linked."
+
+--
+
+* bennett.txt - Anne Bennett <anne@alcor.concordia.ca>
+
+"... I applied the mods in contrib/sutter.txt, and then amplified
+them.  I append the relevant context diffs.  In short, I added options
+"-P" (a URL prefix for use with the paths found by "-S"), and "-H" and
+"-F" (header and footer HTML files to help adjust the style of the
+resulting page, for example by adding company logos,
+return-to-home-page buttons, etc.)."
+
+--
+
+* authried.txt -  Erwin Authried <eauth@mail4u.atnet.at>
+
+"I wrote a small perl script (it's a quick hack) to view man-pages
+with the web-browser."
+
+--
+
+* gzip.patch - Todd Larason <jtl@molehill.org>
+
+"I keep several manpage hierarchies locally, for easy comparison when
+programming for multiple unix machines, most of which are across a
+relatively slow link.  To save disk space, I keep them gzipped.
+
+"Here's a small patch to make .so handling work against gzipped files.
+It's a late-night-after-a-long-day hack, but it seems clean enough.  If
+you think there might be any interest, feel free to include it either in
+rman itself or the contrib directory.  It's against 3.0a3."
+
+--
+
+* man2html - Marc Donner (donner@fid.morgan.com)
+
+A Korn Shell script that is moderately generic.  It takes all the
+files in a directory or a set of directories and does them one at a
+time, putting the html output somewhere appropriate.  The only
+infelicity in the script is the stuff with localeo (local electronic
+office) which is a hack put together to tweak the URLs not to refer to
+localhost, which is broken right now, but directly to the local
+server's alias.
+
+---
+
+* sco-wrapper.sh - Dag Nygren <dag@newtech.fi>
+
+SCO puts the names of valid man page sections in the file /etc/default/man.
+This wrapper extracts the names from this file and passes them on to 
+PolyglotMan for use in determining man page cross references.
+
+---
+
+* http-rman.{c,html} - Fredrik Lundh <fredrik@kalle.teragon.ivab.se>
+
+An HTTP server using PolyglotMan to produce hypertext renditions of
+man-pages on the fly.
+
+---
+
+* lewis.pl - Herb Lewis <herb@chomps.engr.sgi.com>
+
+CGI bin script written in Perl.
+
+---
+
+* hman.{cgi,ksh,pl} by teto@geocities.com
+A more recent version may be available at http://www.geocities.com/SiliconValley/Lakes/8777/hman.html
+
+    hman.cgi		An example of a cgi-bin wrapper to launch hman.pl
+			Ideally, all customization should occur here
+    hman.ksh		A stand-alone (non-browser) version of hman.cgi
+    hman.pl		The core of the PolyglotMan front-end
+
+"I've written a man to html front-end that appears to have unique features:
+A second attempt to locate related man pages is made if the original
+inquiry fails
+
+"Since PolyglotMan can filter nroff's output or can work with raw
+nroff files, this viewer allows selection of the actual man page
+file.  Thus either raw or cooked man pages can be selected directly.
+
+"The program is basic perl, but the implementation allows all customization
+to be embodied in enviroment variables that should be set in a cgi-bin
+wrapper."
+
+--
+
+* youki.pl - Youki Kadobayashi <youki@center.osaka-u.ac.jp>
+
+CGI bin script written in Perl 5, and using CGI.pm
+and thus "it can be accelerated with mod_perl".
+
+---
+
+rman_html_split - Robert Moniot <moniot@fordham.edu>
+rman_html_split.1
+
+Breaks up rman HTML output in file (or standard input) into separate
+section files, placing them into a directory.  With manual page.
+
+---
+
+Other converters are not included in this directory, but rather are
+available over the net:
+
+* bsdi-man.pl - Fritz Heinrichmeyer
+
+Available at:  http://ES-sun2.fernuni-hagen.de
+
+
+* bbc_man2html - Brooks Cutter (brooks@paradyne.com)
+
+Available at:  http://www.stuff.com/~bcutter/home/programs/bbc_man2html.html
+
+
+* rman.pl - Alex Gagin
+
+Available at: http://www.jinr.dubna.su/~gagin/rman.pl.html
+
+(Be aware that the communications link is slow.)
+
+
+* manner.pl - Ray W. Hiltbrand
+
+Available at:  ftp://ftp.eng.auburn.edu/pub/rayh/man2html/manner.pl
+
+
+The active list kept at http://http.cs.berkeley.edu/~phelps/tcltk/polyglotmen.html
+
diff --git a/xc/extras/rman/contrib/authried.txt b/xc/extras/rman/contrib/authried.txt
new file mode 100644
index 000000000..7bdbc4023
--- /dev/null
+++ b/xc/extras/rman/contrib/authried.txt
@@ -0,0 +1,36 @@
+#!/usr/local/bin/perl
+
+require "common.pl";
+
+$webmaster="webadmin\@domain";
+$man_path="/usr/local/man:/usr/man:/usr/X11/man:/usr/lib/teTeX/man";
+$manpage_and_section = $ENV{"QUERY_STRING"};
+$rman="/usr/local/bin/rman -f HTML";
+
+($manpage,$section)=split("[&]",$manpage_and_section);
+
+
+ split(":",$man_path);
+ foreach(@_){
+   $file=join("", $_, "/man", $section,"/",$manpage,".",$section);
+
+#   printf("try %s\n", $file);
+
+   if (-e $file) {
+ 	exec "echo Content-Type:text/html; echo; cat $file | $rman";
+   }
+   elsif (-e join("",$file,".gz")){
+ 	exec "echo Content-Type:text/html; echo; zcat $file | $rman";
+   }
+
+   $file=join("",$file,"x");
+   if (-e $file) {
+ 	exec "echo Content-Type:text/html; echo; cat $file | $rman";
+   }
+   elsif (-e join("",$file,".gz")){
+ 	exec "echo Content-Type:text/html; echo; zcat $file | $rman";
+   }
+
+ } 
+  
+&return_error(500,"man2html error","could not find manual page.");
diff --git a/xc/extras/rman/contrib/bennett.txt b/xc/extras/rman/contrib/bennett.txt
new file mode 100644
index 000000000..1ce1ab4fe
--- /dev/null
+++ b/xc/extras/rman/contrib/bennett.txt
@@ -0,0 +1,371 @@
+*** rman.c.original	Tue Jul 16 21:19:49 1996
+--- rman.c	Tue Aug 27 15:59:38 1996
+***************
+*** 1,5 ****
+--- 1,9 ----
+  static char rcsid[] = "$Header: /home/ajax/dri-backup/xc/xc/extras/rman/contrib/bennett.txt,v 1.1.2.1 2000/07/11 19:36:05 daryll Exp $";
+  
++ /* ELS: SearchDirs by "Edward L. Sutter" <els@sage.sage.att.com> */
++ /* AB:  HTMLHeader and HTMLFooter by Anne Bennett <anne@alcor.concordia.ca> */
++ /* AB:  SearchDirsPrefix by Anne Bennett <anne@alcor.concordia.ca> */
++ 
+  /*
+    RosettaMan
+  
+***************
+*** 265,271 ****
+--- 269,278 ----
+  
+  /*** utility functions ***/
+  
++ /* 1996/08/26 Anne Bennett patched */
++ int HrefSearch(char *manname, char *section);
+  
++ 
+  /* case insensitive versions of strcmp and strncmp */
+  
+  int
+***************
+*** 1338,1343 ****
+--- 1345,1355 ----
+  char *manTitle = MANTITLEPRINTF;
+  char *manRef = MANREFPRINTF;
+  int fmanRef=1;	/* valid man ref? */
++                           /* 1996/08/26 Anne Bennett patched */
++ char *SearchDirs=0;       /* ELS: added for smarter external reference generator. */
++ char *SearchDirsPrefix=0; /* AB:  added for HREF prefix to manpage path.  */
++ char *HTMLHeader=0;       /* AB:  added for HTML header and footer files. */
++ char *HTMLFooter=0;       /* AB:  added for HTML header and footer files. */
+  
+  /*
+   * HTML
+***************
+*** 1349,1354 ****
+--- 1361,1373 ----
+  	int i,j;
+  	int lasttoc;
+  	char *p, *p0;
++ /* 1996/08/26 Anne Bennett patched */
++ /* AB:  To support HTMLHeader and HTMLFooter */
++ #define SUB_MANNAME "@@MANPAGE_NAME@@"
++ #define SUB_MANSECT "@@MANPAGE_SECTION@@"
++ #define STRING_SIZE 1024
++         char line[STRING_SIZE];
++         FILE *fp;
+  
+  	/* always respond to these signals */
+  	switch (cmd) {
+***************
+*** 1397,1402 ****
+--- 1416,1474 ----
+  		/* better title possible? */
+  		printf("<TITLE>"); printf(manTitle, manName, manSect); printf("</TITLE>\n");
+  		printf("</HEAD>\n<BODY>\n");
++ /* 1996/08/26 Anne Bennett patched */
++ /* AB:  added HTMLHeader file, if -H option is set */
++                 if (HTMLHeader) {
++                       if ((fp = fopen(HTMLHeader,"r")) == NULL) {
++                         fprintf(stderr,"Could not open HTMLHeader file\n");
++                       }
++                       else {
++                         /* Output the header file, replacing strings as needed */
++                         while (fgets(line, STRING_SIZE, fp) != (char *)NULL) {
++                            char *walker, *found;
++                            int i;
++ 
++                            /* replace man section */
++                            walker = line;
++                            while ( ( found = strstr(walker, SUB_MANSECT) ) != NULL ) {
++                               if ( (  strlen(line) + strlen(manSect) 
++                                     - strlen (SUB_MANSECT)          ) > sizeof(line) ) {
++                                  fprintf(stderr,"Cannot substitute %s for %s; "
++                                                 "line would be too long:\n "
++                                                 "%s\n", manSect, SUB_MANSECT, line);
++                               walker = found + strlen(SUB_MANSECT);
++                               }
++                               else {
++                               (void) memmove(found+strlen(manSect),
++                                              found+strlen(SUB_MANSECT),
++                                              strlen(found+strlen(SUB_MANSECT))+1);
++                               memcpy(found, manSect, strlen(manSect));
++                               walker = found + strlen(manSect);
++                               }
++                            }
++                            /* replace man name    */
++                            walker = line;
++                            while ( ( found = strstr(walker, SUB_MANNAME) ) != NULL ) {
++                               if ( (  strlen(line) + strlen(manName) 
++                                     - strlen (SUB_MANNAME)          ) > sizeof(line) ) {
++                                  fprintf(stderr,"Cannot substitute %s for %s; "
++                                                 "line would be too long:\n "
++                                                 "%s\n", manName, SUB_MANNAME, line);
++                               walker = found + strlen(SUB_MANNAME);
++                               }
++                               else {
++                               (void) memmove(found+strlen(manName),
++                                              found+strlen(SUB_MANNAME),
++                                              strlen(found+strlen(SUB_MANNAME))+1);
++                               memcpy(found, manName, strlen(manName));
++                               walker = found + strlen(manName);
++                               }
++                            }
++                            printf("%s", line);
++                         }
++                         fclose(fp);
++                       }
++                 }
+  		printf("<A HREF=\"#toc\">%s</A><P>\n", TABLEOFCONTENTS);
+  		I=0;
+  		break;
+***************
+*** 1424,1429 ****
+--- 1496,1514 ----
+  			if (lasttoc==BEGINSUBSECTION) printf("</UL>");
+  			printf("</UL>\n");
+  		}
++ /* 1996/08/26 Anne Bennett patched */
++ /* AB:  added HTMLFooter file, if -F option is set */
++                 if (HTMLFooter) {
++                       if ((fp = fopen(HTMLFooter,"r")) == NULL) {
++                         fprintf(stderr,"Could not open HTMLFooter file\n");
++                       }
++                       else {
++                         /* Output the footer file */
++                         while (fgets(line, STRING_SIZE, fp) != (char *)NULL)
++                            printf("%s", line);
++                         fclose(fp);
++                       }
++                 }
+  		printf("</BODY></HTML>\n");
+  		break;
+  	   case BEGINBODY:		break;
+***************
+*** 1500,1510 ****
+  		*p++='\0'; p0=p;
+  		for (; *p && *p!=')'; p++) /* empty */;
+  		*p='\0';
+! 		if (fmanRef) { printf("<A HREF=\""); printf(manRef, hitxt, p0); printf("\">"); }
+  		else printf("<I>");
+  		break;
+  	   case ENDMANREF:
+! 		if (fmanRef) printf("</A>"); else printf("</I>");
+  		break;
+  
+  	   case BEGINSC: case ENDSC:
+--- 1585,1604 ----
+  		*p++='\0'; p0=p;
+  		for (; *p && *p!=')'; p++) /* empty */;
+  		*p='\0';
+! /* 1996/08/26 Anne Bennett patched */
+! /* ELS: added a call to HrefSearch() if the -S option is set.. */
+!                 if (SearchDirs)
+!                         HrefSearch(hitxt,p0);
+!                 else if (fmanRef) {
+!                         printf("<A HREF=\"");
+!                         printf(manRef, hitxt, p0);
+!                         printf("\">");
+!                 }
+  		else printf("<I>");
+  		break;
+  	   case ENDMANREF:
+! /* 1996/08/26 Anne Bennett patched */
+! 		if (fmanRef || SearchDirs) printf("</A>"); else printf("</I>");
+  		break;
+  
+  	   case BEGINSC: case ENDSC:
+***************
+*** 2744,2754 ****
+--- 2838,2856 ----
+  	extern char *optarg;
+  	extern int optind, opterr;
+  
++ /* 1996/08/26 Anne Bennett patched */
++ /* ELS: added the -S option: */
++ /* AB:  added the -H and -F options */
++ /* AB:  added the -P option */
+  	char strgetopt[80];
+  	/* options with an arg must have a '<' in the description */
+  	struct { char letter; int arg; char *longnames; char *desc; } option[] = {
+  		{ 'f', 1, "filter", " <ASCII|roff|TkMan|Tk|Ensemble|Sections|HTML|SGML|MIME|LaTeX|LaTeX2e|RTF|pod>" },
+  		{ 'r', 1, "reference:manref:ref", " <man reference printf string>" },
++ 		{ 'S', 1, "searchdirs", " <man reference search directories>" },
++ 		{ 'P', 1, "prefix", " <prefix for manpage paths in -S>" },
++ 		{ 'H', 1, "header", " <HTML header file>" },
++ 		{ 'F', 1, "footer", " <HTML footer file>" },
+  		{ 'l', 1, "title", " <title printf string>" },
+  		{ 'b', 0, "subsections:sub", " (show subsections)" },
+  		{ 'n', 1, "name", "(ame of man page) <string>" },
+***************
+*** 2830,2835 ****
+--- 2932,2953 ----
+  		   case 'r': manRef = optarg;
+  			if (strlen(manRef)==0 || strcmp(manRef,"-")==0 || strcmp(manRef,"off")==0) fmanRef=0;
+  			break;
++ /* 1996/08/26 Anne Bennett patched */
++ /* ELS: added the -S option: */
++ /* AB:  added the -H and -F options */
++ /* AB:  added the -P option */
++ 		   case 'S': SearchDirs = optarg;
++ 			if (strlen(SearchDirs)==0 ) SearchDirs=0;
++ 			break;
++ 		   case 'P': SearchDirsPrefix = optarg;
++ 			if (strlen(SearchDirsPrefix)==0 ) SearchDirsPrefix=0;
++ 			break;
++ 		   case 'H': HTMLHeader = optarg;
++ 			if (strlen(HTMLHeader)==0 ) HTMLHeader=0;
++ 			break;
++ 		   case 'F': HTMLFooter = optarg;
++ 			if (strlen(HTMLFooter)==0 ) HTMLFooter=0;
++ 			break;
+  		   case 't': TabStops=atoi(optarg); break;
+  		   case 'm': fMan=0; break;
+  		   case 'T': fTable=1; break;
+***************
+*** 2904,2906 ****
+--- 3022,3126 ----
+  	filter();
+  	return 0;
+  }
++ 
++ 
++ /* 1996/08/26 Anne Bennett patched */
++ /* ELS... 
++    HrefSearch():
++ 	Active only with command line option -S...
++ 	Called when rman -fHTML has determined that it is going to add a
++ 	hypertext link.  The user tells rman where to search for the hypertext
++ 	links (local machine search only) and if HrefSearch() finds the file
++ 
++ 			SRCHDIR/manname.section
++ 
++ 	where
++ 		SRCHDIR is one of the colon-delimited paths specified with
++ 			the -S option;
++ 		manname is the text that rman found preceding a "manname(##)"
++ 			detection;
++ 		section is the string within the parens of the manname spec;
++ 
++ 
++ 	then it will use that path to build the HREF line.  If not found,
++ 	then <A> is all that is inserted.
++ 	This is generally only helpful when you are simply attempting to
++ 	turn a man directory into an html directory.
++ 
++ 	Note that if the first char of SearchDirs is a colon, then if
++ 	HrefSearch does not find the reference, it defaults to what rman
++ 	used to do (use manRef, -r option); otherwise, it will not add
++ 	a hypertext link at all.
++ */
++ int HrefSearch(char *manname, char *section)
++ {
++ 	char	*dir, *colon, tmp;
++ 	int	DefaultToManRef;
++ 	FILE	*fp;
++ 	static char	path[256];
++ 
++ 	tmp = 0;
++ 
++ again:
++ 	if (SearchDirs[0] == ':') {
++ 		dir = &SearchDirs[1];
++ 		DefaultToManRef = 1;
++ 	}
++ 	else {
++ 		dir = SearchDirs;
++ 		DefaultToManRef = 0;
++ 	}
++ 
++ 	/* Make 2 passes on all search directories... */
++ 	/* First pass is with the path dir/manname.section */
++ 	/* Second pass is with the path dir/manname.section[0] */
++ 	/* This allows the spec manname(3x) to be found as manname.3 */
++ 	/* just in cast manname.3x doesn't exist. */
++ 	/* Note that the second pass is only necessary if the section */
++ 	/* string is more than one character in length. */
++ 	while(1) {
++ 		colon =  strchr(dir,':');
++ 		if (colon) *colon = 0;
++ 		sprintf(path,"%s/%s.%s.html",dir,manname,section); 
++                 /* 1996/08/26 Anne: read-only, not rw, please. */
++ 		if ((fp = fopen(path,"r")) != NULL) {
++ 			printf("<A HREF=\"");
++                         /* AB:  Put in the manpage location prefix */
++                         if (SearchDirsPrefix)
++                                 printf("%s",SearchDirsPrefix);
++ 			printf("%s",path);
++ 			printf("\">");
++ 			fclose(fp);
++ 			if (colon) *colon = ':';
++ 			fprintf(stderr,"HREF @ %s\n",path);
++ 			return(1);
++ 		}
++ 		if (colon) {
++ 			*colon = ':';
++ 			dir = colon+1;
++ 		}
++ 		else
++ 			break;
++ 	}
++ 	if (section[1]) {
++ 		tmp = section[1];
++ 		section[1] = 0;
++ 		dir = SearchDirs;
++ 		goto again;
++ 	}
++ 	if (tmp)
++ 		section[1] = tmp;
++ 
++         /* 1996/08/27 Anne Bennett: use manRef only if set */
++ 	if (DefaultToManRef && manRef) {
++ 		printf("<A HREF=\"");
++ 		printf(manRef, manname, section);
++ 		printf("\">");
++ 	}
++ 	else
++ 		printf("<A>");
++ 	return(1);
++ }
++ 
++ /* End ELS additions. */
++ 
+*** rman.1.original	Tue Jul 16 21:19:49 1996
+--- rman.1	Tue Aug 27 15:16:29 1996
+***************
+*** 75,80 ****
+--- 75,109 ----
+  or `off', then man page references will not be HREFs, just set 
+  in italics. 
+  .TP 15
++ -S|--searchDirs \fIcolon-separated-directories \fR
++ Valid only with \fI-fHTML \fR, this option modifies the behaviour of
++ rman when inserting a hyperlink to a referenced manpage. 
++ The directories indicate where to look
++ for a pre-HTMLized version of the referenced manpage under the name 
++ " \fIname \fR. \fIsection \fR. \fIhtml \fR".  If the file is not
++ found, no hypertext link is added, except if the first character of
++ the directories list is a colon, in which case a link is added using
++ manRef, ( \fI-r \fR option).
++ .TP 15
++ -P|--prefix \fIstring \fR
++ Valid only with \fI-S \fR, the string indicates what prefix to add to
++ the path when constructing the hyperlink for a referenced manpage.
++ The string will usually be
++ similar to that given with \fI-r \fR, except that there will be no
++ printf-style escapes, for example "http://localhost/manpages".
++ .TP 15
++ -H|--header \fIfilename \fR
++ Valid only with \fI-fHTML \fR, the filename refers to a file
++ containing an HTML "header", which might consist of, for example, a
++ company logo, or body color specifications.  If the strings
++ "@@MANPAGE_NAME@@" or "@@MANPAGE_SECTION@@" are found in this file,
++ they are replaced by the appropriate manpage name and section.
++ .TP 15
++ -F|--footer \fIfilename \fR
++ Valid only with \fI-fHTML \fR, the filename refers to a file
++ containing an HTML "footer", which might consist of, for example, a
++ link back to the site's home page.
++ .TP 15
+  -s|section \fI# \fR
+  Set volume (aka section) number of man page (used in roff format). 
+  .TP 15
+
diff --git a/xc/extras/rman/contrib/gzip.patch b/xc/extras/rman/contrib/gzip.patch
new file mode 100644
index 000000000..d10666b6a
--- /dev/null
+++ b/xc/extras/rman/contrib/gzip.patch
@@ -0,0 +1,46 @@
+--- rman.c.~1~	Wed Nov 20 13:33:52 1996
++++ rman.c	Fri Oct 24 00:10:56 1997
+@@ -3583,6 +3583,43 @@
+ 			}
+ 			free(sobuf);
+ 		}
++	} else {
++#define GZIP "/bin/gzip"
++		char * gz = malloc(strlen(p)+3+1);
++		sprintf(gz, "%s.gz", p);
++		if (stat(gz, &fileinfo)==0) {
++			/* first, figure out how big */
++			char * cmd = malloc(strlen(gz) + strlen(GZIP) + 7 + 1);
++			char buffer[512];
++			unsigned long compr, uncomp;
++			FILE * proc;
++			sprintf(cmd, "%s -l \"%s\"", GZIP, gz);
++			proc = popen(cmd, "r");
++			fgets(buffer, sizeof buffer, proc); /* label line */
++			fgets(buffer, sizeof buffer, proc); /* length line */
++			sscanf(buffer, "%lu %lu", &compr, &uncomp);
++			fclose(proc);
++			/* Boy, don't you wish stat would do that? */
++			sobuf = malloc(uncomp + 1);
++			if (sobuf!=NULL) {
++				/* suck in entire file, as above */
++				sprintf(cmd, "%s -dc \"%s\"", GZIP, gz);
++				proc = popen(cmd, "r");
++				if (proc!=NULL) {
++					if(fread(sobuf, 1, uncomp, proc)) {
++						sobuf[uncomp]='\0';
++						for (q=sobuf; (q=strstr(q," \\}"))!=NULL; q+=3) *q='\n';
++						source_subfile(sobuf);
++						err = 0;
++					}
++					fclose(proc);
++				}
++				free(sobuf);
++			}
++			free(cmd);
++		}
++		free(gz);
++
+ 	}
+ 
+ 	if (err) {
diff --git a/xc/extras/rman/contrib/hman.cgi b/xc/extras/rman/contrib/hman.cgi
new file mode 100755
index 000000000..540899c03
--- /dev/null
+++ b/xc/extras/rman/contrib/hman.cgi
@@ -0,0 +1,7 @@
+#!/bin/ksh
+##########
+export MANPATH=/trane/mach/man:/trane/share/man:/usr/man:/usr/X11/man:/usr/openwin/man:/var/man
+export PATH=/trane/mach/bin:/trane/share/bin:$PATH
+export HMANPRG=rman
+export HMANOPT='-b -f html'
+exec /home/teto/dev/hman/hman.pl
diff --git a/xc/extras/rman/contrib/hman.ksh b/xc/extras/rman/contrib/hman.ksh
new file mode 100755
index 000000000..99df60107
--- /dev/null
+++ b/xc/extras/rman/contrib/hman.ksh
@@ -0,0 +1,9 @@
+#!/bin/ksh
+##########
+export MANPATH=/trane/mach/man:/trane/share/man:/usr/man:/usr/X11/man:/usr/openwin/man:/var/man
+export PATH=/trane/mach/bin:/trane/share/bin:$PATH
+export HMANPRG=rman
+export HMANOPT='-b -f html'
+export QUERY_STRING='DirectPath=/usr/man/cat4.Z/cdf.4'
+export QUERY_STRING='ManTopic=test&ManSection=key'
+perl -d /home/teto/dev/hman/hman.pl
diff --git a/xc/extras/rman/contrib/hman.pl b/xc/extras/rman/contrib/hman.pl
new file mode 100755
index 000000000..d12c01fd8
--- /dev/null
+++ b/xc/extras/rman/contrib/hman.pl
@@ -0,0 +1,261 @@
+#!/usr/bin/perl -w
+#!/citi/gtfd/mach/bin/perl -w
+###############
+# $Id: hman.pl,v 1.1.2.1 2000/07/11 19:36:05 daryll Exp $
+# $Source: /home/ajax/dri-backup/xc/xc/extras/rman/contrib/hman.pl,v $
+############################################
+# TODO:
+#	reorganize location of man pages - move 3x stuff from man3 to man3x
+#	recurse - if 'man 3x curses' does not work try 'man 3 curses'
+#	display more STDERR
+#	Fix broken whatis entries - instead of
+#		manpage section - description
+#		manpage description section -
+#	highlite keywords found
+#	pass MANPATH as a command line argument
+############################################
+# Inspired by:
+#	http://www.jinr.dubna.su/~gagin
+#	http://thsun1.jinr.dubna.su/~gagin/rman.pl
+#	http://thsun1.jinr.dubna.su/cgi-bin/rman.pl
+#	http://www.jinr.dubna.su/~gagin/rman.pl.html
+#
+# CGI form interface to PolyglotMan program, which is available as
+#	ftp://ftp.cs.berkeley.edu:/ucb/people/phelps/tcltk/rman.tar.Z
+#
+# The most recent version of this program available as
+#	http://www.geocities.com/SiliconValley/Lakes/8777/hman.html
+#---------------------------------------------------------------------
+#  This 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.
+# 
+#  hman 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 software; see the file COPYING.  If not, write to the
+#  Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+#  Boston, MA 02111-1307, USA.
+# 
+#  THIS SOFTWARE IS NOT DESIGNED OR INTENDED FOR USE OR RESALE AS ON-LINE
+#  CONTROL EQUIPMENT IN HAZARDOUS ENVIRONMENTS REQUIRING FAIL-SAFE
+#  PERFOHMANCE, SUCH AS IN THE OPERATION OF NUCLEAR FACILITIES, AIRCRAFT
+#  NAVIGATION OR COMMUNICATION SYSTEMS, AIR TRAFFIC CONTROL, DIRECT LIFE
+#  SUPPORT MACHINES, OR WEAPONS SYSTEMS, IN WHICH THE FAILURE OF THE
+#  SOFTWARE COULD LEAD DIRECTLY TO DEATH, PERSONAL INJURY, OR SEVERE
+#  PHYSICAL OR ENVIRONMENTAL DAMAGE ("HIGH RISK ACTIVITIES").
+#  ANY EXPRESS OR IMPLIED WARRANTY OF FITNESS FOR HIGH RISK ACTIVITIES
+#  IS SPECIFICALLY DISCLAIMED.
+#---------------------------------------------------------------------
+# Request form: hman.pl?ManTopic=SOMETOPIC&ManSection=SECTION
+# Request form: hman.pl?DirectPath=filename
+#
+#---------------------------------------------------------------------
+# Stuff to change
+
+# path to PolyglotMan program. "-b" is not nessesary
+$hman="$ENV{HMANPRG} $ENV{HMANOPT}";
+
+# path to man program
+$ManPrg='/usr/bin/man';
+
+# path to cat program
+$ENV{PAGER} = '/bin/cat';
+
+# path to man directories
+($ManPathFind = $ENV{MANPATH}) =~ s/:/ /g;
+
+# URL to this program
+$hmanpl='/cgi-bin/hman';
+
+# if man produced number of lines less then follows,
+# I assume that request failed
+$emptyman=5;
+
+# tail of every produced html document
+$HtmlTail='<hr><A HREF="' . $hmanpl . '">Back to Hman</a>';
+# $HtmlTitle="<title>CGI form interface to PolyglotMan</title>\n";
+$HtmlHdr="Content-type: text/html\n\n";
+
+# end changable things
+#----------------------------------------------------------------------
+@ManSections = (
+    '1', 'user commands',
+    '2', 'system calls',
+    '3', 'subroutines',
+    '4', 'devices',
+    '5', 'file formats',
+    '6', 'games',
+    '7', 'miscellanious',
+    '8', 'sys. admin.',
+    '9', 'Linux Internals',
+    'n', 'section n'
+);
+
+#   Set unbuffered I/O.  Prevents buffering problems with
+#   "system()" calls.
+select((select(STDOUT), $| = 1)[0]);
+print $HtmlHdr;
+
+$string = $ENV{QUERY_STRING};
+#
+#	Initial Form
+#
+if($string eq "") { initialForm(); }
+
+#
+#	Generic parameter parsing ...
+#
+$DirectPath = $ManSection = $ManTopic = "";
+$string =~ s/&/'; \$/g;
+$string =~ s/=/='/g;
+$string =~ s/^(.*)$/\$$1';/;
+eval $string;
+
+hmanDirect($DirectPath) if ($DirectPath ne "");
+
+if ($ManTopic eq "") { badness("<code>Topic for man search needed</code>\n"); }
+
+if ($ManSection eq "") { badness("<code>No section specified</code>\n"); }
+
+$ManSection =~ s/all//;
+if ($ManSection =~ /key/) { manKey($ManTopic); }
+
+findIt($ManTopic);
+open(MANOUT, "$ManPrg $ManSection $ManTopic |")
+    || die "$hmanpl: can't run \"$ManPrg Section $ManTopic |\", $!\n";
+for (0..$emptyman) {
+    $temp = <MANOUT> || last;
+    push @temp, $temp;
+}
+#
+if (@temp < $emptyman) {
+    close(MANOUT);
+    print"<strong>Request failed for topic $ManTopic:</strong>\n<code>";
+    for (@temp) {print;}
+    print "</code><br><h3>Let's try a keyword search:</h3><hr>\n";
+    manKey($ManTopic);
+}
+#
+$cmd = "$hman -r \"$hmanpl?ManTopic=%s&ManSection=%s\" -l \"Man page for $ManTopic";
+if ($ManSection eq "") {
+    $cmd .= '"';
+} else {
+    $cmd .= "($ManSection)\"";
+}
+#
+open(HMANIN, "| $cmd") || die "$hmanpl: can't open $cmd: $!\n";
+for (@temp) {print HMANIN;}
+while(<MANOUT>){print HMANIN;}
+close(MANOUT);
+close(HMANIN);
+
+exitIt();
+###############################################################################
+sub initialForm {
+    print <<EOF;
+<h1>Select a manual page:</h1>
+    <FORM METHOD="GET" action="$hmanpl">
+	<table border=0>
+	    <tr>
+		<td align=right>Section:</td>
+		<td>
+		    <select name=ManSection>
+			<OPTION value=all>All</option>
+			<OPTION value=key>Keyword Search</option>
+EOF
+    for ($ndx = 0; $ndx < @ManSections; $ndx += 2) {
+	print '			<OPTION value=' . $ManSections[$ndx] . '>'
+	    . "$ManSections[$ndx] - $ManSections[$ndx + 1] "
+	    . '</option>' . "\n";
+    }
+    print <<EOF;
+		    </SELECT>
+		</td>
+	    </tr>
+	    <tr>
+		<td align=right>Topic:</td>
+		<td><input type="TEXT" name="ManTopic"></td>
+	    </tr>
+	    <tr>
+		<td align=right><INPUT TYPE="submit" VALUE="Search">
+		<td align=left><input type="reset" value="Reset">
+		</td>
+	    </tr>
+	</table>
+    </FORM>
+EOF
+    exitIt();
+}
+sub findIt {
+    my($topic) = ($_[0]);
+    my($cmd, $mixedCase, $navigation, $zcat);
+    $mixedCase = '';
+    foreach (split(/\s*/, $topic)) {
+	$mixedCase .= "[" . $_ . uc($_) . "]";
+    }
+    $cmd = 'find ' . $ManPathFind . ' \( -type f -o -type l \) -name '
+	. $mixedCase .'\* | sort -t. +1';
+    open(FINDIN, "$cmd |") || die "can't open pipe \"$cmd |\": $!";
+    $navigation = 0;
+    while (<FINDIN>) {
+	if ($navigation == 0) {
+	    print "<UL>\n";
+	    $navigation = 1;
+	}
+	$_ =~ s/[\n\r]*$//;
+	print "<li><A HREF=\"$hmanpl?DirectPath=$_\">$_</a>\n";
+    }
+    close(FINDIN);
+    print "</UL><hr>\n" if ($navigation == 1);
+}
+sub hmanDirect {
+    my($path) = ($_[0], $_[1]);
+    my($cmd, $dir, $topic, $section);
+    ($dir = $path) =~ s/\.(gz|z|Z)$//;
+    ($topic = $dir) =~ s,^.*/,,;
+    $dir =~ s,/[^/]*/[^/]*$,,;
+    # $dir =~ s,/[^/]*$,,;
+    ($section=$topic)=~s/^.*\.([^\.]*)$/$1/;
+    $topic =~ s/\.[^\.]*$//;
+    findIt($topic);
+    $cmd = "cd $dir; (gzip -dc < $path 2>/dev/null || cat < $path) | $hman -r '"
+	. $hmanpl . '?ManTopic=%s&ManSection=%s' ."' -l 'Man page for $topic($section)'";
+    system($cmd) || warn "can't run command \"$cmd\": $!";
+    print $HtmlTail;
+    exit 0;
+}
+sub exitIt {
+    print $HtmlTail;
+    exit 0;
+}
+sub badness {
+    my($text) = ($_[0]);
+    print "<strong>Request failed:</strong> $text, Try again<hr>\n";
+    initialForm();
+}
+sub manKey {
+    my($topic) = ($_[0]);
+    open(TMPHTML,"$ManPrg -k $topic | sort -u |")
+	|| die "can't open pipe \"$ManPrg -k $topic | sort -u |\": $!\n";
+    print "<title>Keyword search results for \"$topic\"</title>\n";
+    print "<h1>Keyword search results for \"$topic\"</h1><hr>\n";
+    while(<TMPHTML>) {
+	s/\( \)//g;
+	next if (! /^([^(]+)\s*\(([^)]+)[^-]+-\s(.*)[\n\r]*$/);
+	@topics=split(/, /,$1);
+	next if ($2 eq "");
+	print "<h2>$3:</h2>\n";
+	print "<UL>\n";
+	for $topic (@topics) {
+	    print "<li><A HREF=\"$hmanpl?ManSection=$2&ManTopic=$topic\">$topic</a>($2)\n";
+	}
+	print "</UL>\n";
+    }
+    close(TMPHTML);
+    exitIt();
+}
diff --git a/xc/extras/rman/contrib/http-rman.c b/xc/extras/rman/contrib/http-rman.c
new file mode 100644
index 000000000..6ff91cc65
--- /dev/null
+++ b/xc/extras/rman/contrib/http-rman.c
@@ -0,0 +1,200 @@
+/* $Id: http-rman.c,v 1.1.2.1 2000/07/11 19:36:05 daryll Exp $
+ *
+ * Name:
+ *	http-rman.c -- a rudimentary man-page HTTP server
+ *
+ * Description:
+ *	This is a minimal HTTP server using RosettaMan by T.A. Phelps
+ *	(phelps@ACM.org) to produce hypertext renditions
+ *	of man-pages on the fly.
+ *
+ *	This server processes URLs with the following syntax:
+ *
+ *		[/man] ?<command> [ ?<section> ]
+ *
+ *	For URLs matching this format, it pipes the output of
+ *	man <section> <command> through rman and sends it to
+ *	the HTTP client. For other URLs, it returns the document
+ *	given as argv[1] (using cat(1). The leading /man is
+ *	optional, but is strongly recommended.
+ *
+ *	This server is shipped as two files, the http-rman.c
+ *	sources and the http-rman.html sample frontpage. I have
+ *	not included a Makefile; you can write your own or just
+ *	type [g]cc -o http-rman http-rman.c
+ *
+ * What do I need to run this:
+ *	If you don't have it, pick up RosettaMan by anonymous ftp
+ *	from ftp.cs.berkeley.edu: /ucb/people/phelps/tcl/rman.tar.Z
+ *
+ *	You'll also need an HTTP client such as NCSA Mosaic to talk
+ *	to this server. Mosaic is available by anonymous FTP from
+ *	ftp://ftp.ncsa.uiuc.edu/Mosaic 
+ *
+ *	Both RosettaMan (rman) and Mosaic are available from many
+ *	other sites. Try Archie, or check your local or national net
+ *	archive.
+ *
+ * How do I get it running:
+ *	First, compile the server (see above), and install it
+ *	somewhere.
+ *
+ *	The server runs under inetd(8). Add a service to
+ *	/etc/services, say:
+ *
+ *		http-rman 4080/tcp
+ *
+ *	If you're not about to install another HTTP server on your
+ *	machine, you may use the default HTTP port, 80, instead.
+ *
+ *	Then add an entry to /etc/inetd.conf, such as (on a single line):
+ *
+ *		http-rman stream tcp nowait root /usr/local/bin/http-rman
+ *			http-rman /usr/local/lib/rman/http-rman.html
+ *
+ *	Change /usr/local/bin and /usr/local/lib/rman to where you
+ *	installed the two files. In addition, you may wish to run
+ *	the server as something other than root...
+ *
+ *	Restart inetd(8) (use kill -HUP or kill it and start it again)
+ *	and try the following:
+ *
+ *	$ Mosaic http://localhost:4080
+ *
+ *	If you don't have Mosaic, try the following instead:
+ *
+ *	$ telnet localhost 4080
+ *	Trying 127.0.0.1...
+ *	Connected to localhost.
+ *	Escape character is '^]'.
+ *	GET /man?ls <return>
+ *	<return>
+ *	HTTP/1.0 200 OK
+ *	...
+ *
+ * Portability:
+ *	You'll need an ANSI compiler (or an editor and some patience).
+ *	As it stands right now, this code has been successfully
+ *	compiled on OSF/1 AXP using cc, and on SunOS 4.1 using gcc.
+ *	Might need some tuning for other platforms.
+ *
+ * Legal Issues:
+ *	Check the external visibility of the http-rman service
+ *	you choose. This server gives a user access to ALL man-
+ *	pages on your machine. You may have installed copyrighted
+ *	software (your operating system, for example) with
+ *	man-pages that you are NOT allowed to make visible for
+ *	anyone out there...
+ *
+ * History:
+ *	94-04-30 fl: created
+ *	94-05-13 fl: stripped away everything but rman support
+ *
+ * Copyright (c) Fredrik Lundh 1994 (fredrik_lundh@ivab.se)
+ * All rights reserved.
+ */
+
+
+#include <stdio.h>		/* printf(), fflush(stdio) etc */
+#include <string.h>		/* strrchr(), strcmp() etc */
+
+
+static int
+http_error(int error)
+{
+    char *p;
+
+    switch (error) {
+    case 400:
+	p = "Bad Request";
+	break;
+    case 404:
+	p = "Not Found";
+	break;
+    default:
+	p = "Error";
+    }
+    printf("HTTP/1.0 %d %s\r\n", error, p);
+    printf("MIME-version: 1.0\r\n");
+    printf("Content-Type: text/html\r\n\r\n");
+    printf("<head><title>%d %s</title></head>\r\n", error, p);
+    printf("<body><h1>%d %s</h1></body>\r\n", error, p);
+    return 0;
+}
+
+
+static int
+http_rman(char *url)
+{
+    char *pName;
+    char *pSection;
+    char buf[200];
+
+    /* parse URL: should be /man?command[?section] */
+    pSection = strrchr(url, '?');
+    if (!pSection) {
+	return -1;
+    }
+    pName = pSection-1;
+    *pSection++ = '\0';
+
+    pName = strrchr(url, '?');
+    if (!pName) {
+	pName = pSection;
+	pSection = "";
+    }
+    else
+	pName++;
+
+    sprintf(buf, "man %s %s | rman -r \"man?%%s?%%s\" -n %s -f html",
+	    pSection, pName, pName);
+
+    return system(buf);
+}
+
+
+int
+main(int ac, char **av)
+{
+    char buf[200];
+    char url[200];
+    int status;
+    char *sFrontpage = "/usr/local/lib/rman/http-rman.html";
+
+/* check arguments */
+
+    if (ac > 1)
+	sFrontpage = av[1];
+
+/* just read in one line from stdin and make sure it is a GET
+command */
+
+    if (gets(buf)) {
+
+	/* read rest of command (just for the sake of it) */
+	while (gets(url) && url[0] != '\r')
+	    ;
+
+	/* command should be GET <url> [HTTP/1.0] */
+	if (sscanf(buf, "GET %s", url) == 1) {
+
+	    status  = http_rman(url);
+
+	    if (status < 0) {
+		sprintf(buf, "cat %s", sFrontpage);
+		if (system(buf) == 0)
+		    status = 0;
+	    }
+
+	    if (status < 0)
+		http_error(404);
+
+	} else
+
+	    http_error(400);
+    }
+
+    fflush(stdout);
+
+    exit(0);
+}
diff --git a/xc/extras/rman/contrib/http-rman.html b/xc/extras/rman/contrib/http-rman.html
new file mode 100644
index 000000000..ec4cb5973
--- /dev/null
+++ b/xc/extras/rman/contrib/http-rman.html
@@ -0,0 +1,21 @@
+HTTP/1.0 200 OK
+Server: http-rman 1.0
+MIME-Version: 1.0
+Content-Type: text/html
+Content-Transfer-Encoding: binary
+
+<html>
+<!-- this is a sample frontpage for the http-rman man-page server -->
+<!-- Copyright (c) 1994 Fredrik Lundh (fredrik_lundh@ivab.se) -->
+<head><title>RosettaMan Gateway</title></head>
+<h1>RosettaMan Gateway</h1>
+
+This is a gateway to the man pages on this machine. Please enter the
+function or command you want to know more about.
+<p>
+<isindex>
+<p>
+The hypertext man-pages are generated on the fly using the excellent
+<A HREF="/man?rman">RosettaMan</A> utility by T.A. Phelps. This minimal
+HTTP-server was written by Fredrik Lundh.
+</html>
diff --git a/xc/extras/rman/contrib/lewis.pl b/xc/extras/rman/contrib/lewis.pl
new file mode 100644
index 000000000..116b652c2
--- /dev/null
+++ b/xc/extras/rman/contrib/lewis.pl
@@ -0,0 +1,17 @@
+#!/usr/bin/perl
+
+# Syntax: man2html <topic> <section>
+
+# Print out a content-type for HTTP/1.0 compatibility
+print "Content-type: text/html\n\n";
+
+if( $ENV{'REQUEST_METHOD'} eq "GET" )
+{
+   $buffer=$ENV{'QUERY_STRING'} ;
+   @ARGV = split(/&/,$buffer) ;
+}
+
+@manpage = `man @ARGV[1] @ARGV[0] | rman -f html -n @ARGV[0] -s
+@ARGV[1]`;
+
+print "@manpage";
diff --git a/xc/extras/rman/contrib/man2html b/xc/extras/rman/contrib/man2html
new file mode 100755
index 000000000..3df77a3a8
--- /dev/null
+++ b/xc/extras/rman/contrib/man2html
@@ -0,0 +1,108 @@
+#!/bin/ksh
+
+# Take a man tree and make an html tree out of it
+
+DOMAIN=$(domainname)
+ARCH=$(arch)
+ 
+case ${DOMAIN} in
+   acf)     localeo=reo1 ;;
+   bwfid)   localeo=reo1 ;;
+   chfid)   localeo=neo1 ;;
+   hkfid)   localeo=teo1 ;;
+   lnfid)   localeo=leo1 ;;
+   nyfid)   localeo=neo1 ;;
+   psfid)   localeo=leo1 ;;
+   sffid)   localeo=neo1 ;;
+   tkfid)   localeo=teo1 ;;
+   esac
+
+sections="{1,2,3,4,5,6,7,8}"
+from=/usr/man
+to=/u/eo/repository/system/unix/man
+
+function disambiguate
+{
+newbase=${1}
+newname="${newbase}.1"
+dis=2
+while [ -a "${newname}" ]
+   do
+      newname=$newbase"."$dis
+      dis=$(expr $dis + 1)
+      done
+}
+
+while getopts f:t:l:is:v c
+   do
+      case $c in
+         f) from=$OPTARG ;;
+         t) to=$OPTARG ;;
+         l) localeo=$OPTARG ;;
+         i) indexonly=1 ;;
+         s) sections=$OPTARG ;;
+         v) verbose=1 ;;
+         esac
+      done
+shift OPTIND-1
+
+if [ "${verbose}" ]
+   then
+      print "from: $from"
+      print "to: $to"
+      print "localeo: $localeo"
+      print "sections: $sections"
+      if [ "${indexonly}" ]
+	 then
+	    print "indexonly: 1"
+	 fi
+   fi
+
+cd $from
+
+if [ ! "${indexonly}" ]
+   then
+      print "Processing the man pages ..."
+      for i in man${sections}/*
+	 do
+	    if [ "$verbose" ]
+	       then
+		  print $i
+	       fi
+	    # n=${i%.*}
+	    name=${to}/${i}
+	    if [ -a "${name}" ]
+	       then
+		  oldname=$name
+		  disambiguate $name
+		  name=$newname
+		  print "Collision - ${oldname} will be stored as ${name}"
+	       fi
+	    eqn $i | tbl | nroff -man | rman -f HTML | sed -e "s/MS_LOCAL_HOST/${localeo}/g" > ${name}
+	    done
+   fi
+
+print "Building the index.html files ..."
+cd $to
+for i in man${sections}
+   do
+      if [ "$verbose" ]
+	 then
+	    print $i
+	 fi
+      cd $i
+      rm -f index.html
+      echo '<ul>' > ../new.html
+      for j in *
+	 do
+	    if [ "$verbose" ]
+	       then
+		  print -n "$j "
+	       fi
+	    print
+	    print "<li> <a href=$j>$j</a>" >> ../new.html
+	    done
+      echo '</ul>' >> ../new.html
+      mv ../new.html index.html
+      cd ..
+      done
diff --git a/xc/extras/rman/contrib/rman_html_split b/xc/extras/rman/contrib/rman_html_split
new file mode 100755
index 000000000..aafb023c5
--- /dev/null
+++ b/xc/extras/rman/contrib/rman_html_split
@@ -0,0 +1,324 @@
+#!/bin/sh
+#
+# Script to break up rman html output into separate sections in a directory.
+#
+# Author: Robert Moniot <moniot@fordham.edu>
+# Date:   4 Aug 1998
+#
+
+version="rman_html_split V1.0 (c) 1998 Robert K. Moniot"
+
+	    # Usage function prints synopsis and exits
+Usage(){
+  myname=`echo $0 | sed -e 's,.*/,,'`
+cat <<EOF
+
+Usage: $myname [ options ] [ file ]
+
+Breaks up rman HTML output in file (or standard input) into separate
+section files, placing them into a directory.
+
+Options:
+    -d | --html-dir dir   Place output into directory dir (default = html).
+    -m | --merge #        Merge sections 0 thru # (default 1) into one file.
+    -t | --toc-first      Link index.html to table of contents instead
+                          of to sect0.html.
+    -T | --title "title"  Use "title" as title on top of each page
+                          instead of rman's title.  "title" may be empty.
+    -h | --help           Print synopsis and exit.
+    -V | --version        Print version information and exit.
+EOF
+  exit 0
+}
+
+	    # Version function prints version number and exits
+Version(){
+  echo "$version"
+  exit 0
+}
+
+	    # This function puts out header needed by all but 1st page.
+do_header(){
+cat <<EOF
+<HTML>
+<HEAD>
+<TITLE>$*</TITLE>
+</HEAD>
+<BODY>
+EOF
+do_title
+}
+
+	    # This function puts out the title with name of manpage
+do_title(){
+  if [ -n "$title" ]
+  then
+    echo "<H2>$title</H2>"
+  fi
+}
+
+	    # This function puts out ref to table of contents that
+	    # follows header for all but 1st and toc page.
+do_href_toc(){
+cat <<EOF
+<A HREF="toc.html">Table of Contents</A><P>
+EOF
+}
+
+	    # This function puts out ref to previous section
+do_href_prev(){
+  echo "<P>Previous: <A HREF=\"$1.html\">$2</A><HR><P>"
+}
+
+	    # This function puts out footer needed by all but last page.
+do_footer(){
+cat <<EOF
+</BODY></HTML>
+EOF
+}
+
+	    # This function changes the trailing NAME anchor
+	    # of a section into an href to the next section.
+	    # The first edit is for sections.  The second edit
+	    # handles the toc anchor, which has no href and so
+	    # is missed by the first edit.
+do_href_next(){
+ sed -e '$s,^.*<A NAME="\([^"]*\)" HREF=[^>]*>\([^<]*\).*$,<P><HR><P>Next: <A HREF="\1.html">\2</A>,' \
+     -e '$s,<A NAME="\(toc[^"]*\)",Next: <A HREF="\1.html",'
+}
+
+	    # This function adapts internal refs from one-big-file
+	    # form into multi-file format.
+	    # First edit changes "#tocn" into "toc.html#tocn"
+	    # Second edit changes "#sectn" into "sectn.html#sectn"
+	    # Third edit changes "#toc" into "toc.html"
+change_hrefs(){
+    sed -e 's,HREF="#toc\([0-9][0-9]*\)",HREF="toc.html#toc\1",g' \
+	-e 's,HREF="#sect\([0-9][0-9]*\)",HREF="sect\1.html#sect\1",g' \
+        -e 's,HREF="#\(toc\)",HREF="\1.html",g'
+}
+
+
+#  Execution begins here.
+
+				# Process the command-line options.
+htmldir="html"
+mergesect="1"
+
+while [ $# != 0 ]
+do
+  case $1 in
+
+    -d | --html-dir)
+	shift
+	if [ $# != 0 ]
+	then
+	    htmldir="$1"
+	    shift
+	fi
+	;;
+
+    -m | --merge)
+	shift
+	if [ $# != 0 ]
+	then
+	    mergesect="$1"
+	    shift
+	fi
+	case $mergesect in
+	    [0-9])
+		;;
+	    *)
+		echo "ERROR: merge section must be a digit from 0 to 9."
+		Usage
+		;;
+	esac
+	;;
+
+    -t | --toc-first)
+	shift
+	tocfirst="t"
+	;;
+
+    -T | --title)
+	shift
+	if [ $# != 0 ]
+	then
+	    title="$1"
+	    shift
+	fi
+	if [ -z "$title" ]	# If title is the null string, set notitle
+	then
+	    notitle="t"
+	fi
+	;;
+
+    -h | --help)
+	Usage
+	;;
+
+    -V | --version)
+	Version
+	;;
+
+				# Unrecognized flag.
+    -*)
+	echo "Unrecognized option: $1"
+	Usage
+	;;
+
+				# Non-flag: go to process file.
+    *)
+	break
+	;;
+  esac
+done
+
+				# At this point we should have at most one
+				# argument, the file name.  Check it.
+if [ $# -gt 1 ]
+then
+  Usage
+fi
+
+if [ $# = 1 ]
+then
+  if [ ! -r "$1" ]
+  then
+    echo "ERROR: Cannot open file $1."
+    exit 1
+  fi
+fi
+
+		# Filter the input (file arg or stdin) thru a sed script
+		# that fixes cases where an anchor for a section has
+		# had its section title split across two or more lines.
+		# Put the result in a tmp file so it can be re-read.
+
+filename="/tmp/rman_html_split_$$.html"
+trap "rm -f $filename" 0
+sed -e ':top' \
+    -e '/^<H[1-9]><A NAME="sect.*[^>]$/N' \
+    -e '/^<LI><A NAME="toc.*[^>]$/N' \
+    -e 's/\n/ /' \
+    -e 't top' $1 > $filename
+
+		# If user did not supply title, get title for top of
+		# each page from rman-generated <TITLE> line.  If user
+		# gave a blank title then leave it blank.
+if [ -z "$title" -a -z "$notitle" ]
+then
+  title=`sed -e '/^<TITLE>/q'  $filename |
+         sed -n -e 's,^<TITLE>\([^<]*\).*$,\1,p'`
+fi
+
+		# Get a list of all the sections.  Separate it into a
+		# list of merged sections and a list of all other sections.
+		# Merged sects are combined and get special treatment besides.
+allsects=`sed -n 's,^.*<A NAME="\(sect[0-9][0-9]*\)".*$,\1,p' $filename`
+
+mergesects=`echo $allsects | \
+	    awk '{for(i=1; i<=NF && i<'"$mergesect"+2'; i++) print \$i;}'`
+
+sectlist=`echo $allsects |
+	    awk '{for(i='"$mergesect"'+2; i<=NF; i++) print \$i;}'`
+
+
+# This little bit, copied from GNU configure, sets ac_n and ac_c such
+# that echo $ac_n "stuff $ac_c" yields "stuff " w/o following newline.
+if (echo "testing\c"; echo 1,2,3) | grep c >/dev/null; then
+  # Stardent Vistra SVR4 grep lacks -e, says ghazi@caip.rutgers.edu.
+  if (echo -n testing; echo 1,2,3) | sed s/-n/xn/ | grep xn >/dev/null; then
+    ac_n= ac_c='
+' ac_t='	'
+  else
+    ac_n=-n ac_c= ac_t=
+  fi
+else
+  ac_n= ac_c='\c' ac_t=
+fi
+
+		# Create html directory if it does not exist.
+if [ -d $htmldir ]
+then
+  echo "Re-using directory $htmldir."
+else
+  echo "Creating directory $htmldir..."
+  if mkdir $htmldir
+  then
+    true
+  else
+    echo "Failed!"
+    exit 1
+  fi
+fi
+
+echo "Creating section pages..."
+
+		# Produce index page.  It is special, since it combines
+		# merged sections and uses the rman-generated header.
+nextsect=`echo $sectlist | awk '{print $1;}'`
+echo $ac_n $mergesects "$ac_c"
+(do_title ;
+ sed -n -e "1,/^<H[23]><A NAME=\"$nextsect\"/p"  $filename | \
+ do_href_next | \
+ change_hrefs ;
+ do_footer) > $htmldir/sect0.html
+
+		# Produce pages for all other sections except toc.
+prevsect="sect$mergesect"
+prevtext=`sed -n 's,^<H[23]><A NAME="'$prevsect'" HREF="#toc[0-9][0-9]*">\([^<]*\).*$,\1,p' $filename | sed -e 's/ *$//'`
+for sect in $sectlist;
+do
+    echo $ac_n "$sect $ac_c"
+    headtext=`sed -n 's,^<H[23]><A NAME="'$sect'" HREF="#toc[0-9][0-9]*">\([^<]*\).*$,\1,p' $filename | sed -e 's/ *$//'`
+    (do_header $headtext ;
+     do_href_toc ;
+     do_href_prev "$prevsect" "$prevtext" ;
+     sed -n -e '/<A NAME="'$sect'"/,/<A NAME=/p' $filename |
+     do_href_next |
+     change_hrefs ;
+     do_footer) > $htmldir/$sect.html
+     prevsect="$sect"
+     prevtext="$headtext"
+done
+
+
+		# Produce table of contents
+echo "toc"
+(do_header Table of Contents ;
+ sed -n -e '/<A NAME="toc">/,$p' $filename | \
+ change_hrefs ;
+) > $htmldir/toc.html
+
+
+		# Finally, make sure the symlinks index.html and
+		# sect1.html -> sect0.html are in place, and if not,
+		# create them.  If --tocfirst is not specified, then
+		# link index.html to section 0, otherwise link it to
+		# index.html
+
+echo "Doing symlinks..."
+cd $htmldir
+
+rm -f index.html
+if [ -z "$tocfirst" ]
+then
+    echo "Linking index.html -> sect0.html"
+    ln -s sect0.html index.html
+else
+    echo "Linking index.html -> toc.html"
+    ln -s toc.html index.html
+fi
+
+for sect in $mergesects
+do
+  if [ "$sect" != "sect0" ]
+  then
+    echo "Linking $sect.html -> sect0.html"
+    rm -f $sect.html
+    ln -fs sect0.html $sect.html
+  fi
+done
+
+echo "Done."
diff --git a/xc/extras/rman/contrib/rman_html_split.1 b/xc/extras/rman/contrib/rman_html_split.1
new file mode 100644
index 000000000..5d3cb4d59
--- /dev/null
+++ b/xc/extras/rman/contrib/rman_html_split.1
@@ -0,0 +1,71 @@
+.TH rman_html_split 1 "4 Aug 1998"
+.SH NAME
+rman_html_split \-  split rman html output into separate sections
+.SH SYNOPSIS
+.B rman_html_split
+[
+.I options
+] [
+.I file
+]
+.SH DESCRIPTION
+The \fBrman_html_split\fP script reads a file containing HTML produced
+from a manual page by the \fBrman\fP(1) program, and splits it into a
+set of files, each containing one section or subsection of the manual.
+Each section file has added to it an optional header identifying the
+program described by the manual page, a link to the table of contents,
+and links to the previous section and the next section.  The separate
+section files are placed in a subdirectory.  A symbolic link named
+\fIindex.\|html\fP is created in this subdirectory, linked to either
+\fIsect0.\|html\fP (the first section) or \fItoc.\|html\fP (the table
+of contents section).
+.PP
+Use this script to split a large manual page into smaller, more
+manageable parts for use as online web-based documentation.
+.SH OPTIONS
+.TP
+.BR  "\-d \fIdir\fP" " | "  "\-\-html\-dir \fIdir\fP"
+Place output into directory \fIdir\fP.  The default directory name is
+\fIhtml\fP.
+.TP
+.BR "\-m \fInum\fP" " | "  "\-\-merge \fInum\fP"
+Merge sections 0 thru \fInum\fP into one file.  The default for
+\fInum\fP is 1, since usually section 0 is the NAME section and
+section 1 is SYNOPSIS, which are usually both short and go well
+together.  Set
+.I num
+to 0 to prevent merging, or set it to a larger number to merge more
+initial sections.  Maximum value of \fInum\fP is 9. There is no
+provision for merging sections other than consecutive numbers from
+section 0.
+.TP
+.BR \-t " | "  \-\-toc\-first
+Link \fIindex.\|html\fP to the table of contents.  By default, it is
+linked to \fIsect0.\|html\fP.  Use this if you prefer the initial page
+to be the table of contents rather than the first section of the
+manual page.
+.TP
+.BR "\-T \fItitle\fP" " | "  "\-\-title \fItitle\fP"
+Use \fItitle\fP as a header on top of each page.  This header is
+displayed using the H2 HTML tag.  By default, \fBrman_html_split\fP uses
+the title generated by \fBrman\fP.  By making \fItitle\fP the empty
+string (""), the header can be suppressed entirely.
+.TP
+.BR \-h " | "  \-\-help
+Print synopsis and exit.
+.TP
+.BR \-V " | "  \-\-version
+Print version information and exit.
+.SH SEE ALSO
+.BR man (1),
+.BR rman (1).
+.SH BUGS
+You tell me.
+.SH AUTHOR
+Robert K. Moniot <moniot@fordham.edu>
+.br
+Fordham University
+.br
+Department of Computer and Information Sciences
+.br
+New York, NY
diff --git a/xc/extras/rman/contrib/sco-wrapper.sh b/xc/extras/rman/contrib/sco-wrapper.sh
new file mode 100755
index 000000000..95d7610ed
--- /dev/null
+++ b/xc/extras/rman/contrib/sco-wrapper.sh
@@ -0,0 +1,6 @@
+#
+REALRMAN=/usr/local/bin/rman.real
+VOLLISTFILE=/etc/default/man
+
+VOLLIST=`grep '^ORDER' $VOLLISTFILE | cut -f2 -d=`
+exec $REALRMAN -V "=$VOLLIST" $*
diff --git a/xc/extras/rman/contrib/sutter.txt b/xc/extras/rman/contrib/sutter.txt
new file mode 100644
index 000000000..b467bb348
--- /dev/null
+++ b/xc/extras/rman/contrib/sutter.txt
@@ -0,0 +1,155 @@
+"Edward L. Sutter" <els@sage.sage.att.com>
+
+Perhaps there is already a better way to do this, but since I couldn'd find
+one, I'd like to make a suggestion that has helped me quite a bit for getting
+the proper hypertext links.  Keep in mind that this is under the context of
+me converting "man" directories with their typical sub-directories of
+man1, man2, etc... to an equivalently structured "html" directory with
+the same sub-directory heirarchy.
+
+I added an option to rman that allows it to search for the files over
+a specified set of directories.  This allows (for example) manpages under
+man1 that reference something under man3 to be properly linked.
+
+rman.c v2.4 ...
+
+...
+
+/* ELS: added to support a smarter external reference generator. */
+char	*SearchDirs=0;
+int 	HrefSearch();
+
+...
+
+/* ELS: added the -S option: */
+	while ((c=getopt(argc,argv,"Kh?f:l:r:bckmTpvn:t:s:yS:"))!=-1)
+		switch (c) {
+		   case 'k': fHeadfoot=1; break;
+		   case 'b': fSubsections=1; break;
+		   case 'c': fChangeleft=1; break;
+		   case 'S': SearchDirs=optarg; break;
+
+
+...
+
+
+void
+HTML(enum command cmd) {
+
+...
+
+	
+	   case BEGINMANREF:
+		for (p=hitxt; *p && *p!='('; p++) /* empty */;
+		*p++='\0'; p0=p;
+		for (; *p && *p!=')'; p++) /* empty */;
+		*p='\0';
+/* ELS: added a call to HrefSearch() if the -S option is set.. */
+		if (SearchDirs)
+			HrefSearch(hitxt,p0);
+		else {
+			printf("<A HREF=\"");
+			printf(manRef, hitxt, p0);
+			printf("\">");
+		}
+		break;
+
+
+...
+
+
+/* ELS... 
+   HrefSearch():
+	Active only with command line option -S...
+	Called when rman -fHTML has determined that it is going to add a
+	hypertext link.  The user tells rman where to search for the hypertext
+	links (local machine search only) and if HrefSearch() finds the file
+
+			SRCHDIR/manname.section
+
+	where
+		SRCHDIR is one of the colon-delimited paths specified with
+			the -S option;
+		manname is the text that rman found preceding a "manname(##)"
+			detection;
+		section is the string within the parens of the manname spec;
+
+
+	then it will use that path to build the HREF line.  If not found,
+	then <A> is all that is inserted.
+	This is generally only helpful when you are simply attempting to
+	turn a man directory into an html directory.
+
+	Note that if the first char of SearchDirs is a colon, then if
+	HrefSearch does not find the reference, it defaults to what rman
+	used to do (use manRef, -r option); otherwise, it will not add
+	a hypertext link at all.
+*/
+HrefSearch(manname,section)
+char	*manname, *section;
+{
+	char	*dir, *colon, tmp;
+	int	DefaultToManRef;
+	FILE	*fp;
+	static char	path[256];
+
+	tmp = 0;
+
+again:
+	if (SearchDirs[0] == ':') {
+		dir = &SearchDirs[1];
+		DefaultToManRef = 1;
+	}
+	else {
+		dir = SearchDirs;
+		DefaultToManRef = 0;
+	}
+
+	/* Make 2 passes on all search directories... */
+	/* First pass is with the path dir/manname.section */
+	/* Second pass is with the path dir/manname.section[0] */
+	/* This allows the spec manname(3x) to be found as manname.3 */
+	/* just in cast manname.3x doesn't exist. */
+	/* Note that the second pass is only necessary if the section */
+	/* string is more than one character in length. */
+	while(1) {
+		colon =  strchr(dir,':');
+		if (colon) *colon = 0;
+		sprintf(path,"%s/%s.%s.html",dir,manname,section); 
+		if ((fp = fopen(path,"rw")) != NULL) {
+			printf("<A HREF=\"");
+			printf("%s",path);
+			printf("\">");
+			fclose(fp);
+			if (colon) *colon = ':';
+			fprintf(stderr,"HREF @ %s\n",path);
+			return(1);
+		}
+		if (colon) {
+			*colon = ':';
+			dir = colon+1;
+		}
+		else
+			break;
+	}
+	if (section[1]) {
+		tmp = section[1];
+		section[1] = 0;
+		dir = SearchDirs;
+		goto again;
+	}
+	if (tmp)
+		section[1] = tmp;
+
+	if (DefaultToManRef) {
+		printf("<A HREF=\"");
+		printf(manRef, manname, section);
+		printf("\">");
+	}
+	else
+		printf("<A>");
+	return(1);
+}
+
+/* End ELS additions. */
+
diff --git a/xc/extras/rman/contrib/youki.pl b/xc/extras/rman/contrib/youki.pl
new file mode 100755
index 000000000..8af5865d6
--- /dev/null
+++ b/xc/extras/rman/contrib/youki.pl
@@ -0,0 +1,195 @@
+#!/usr/local/bin/perl5
+#
+# CGI script for translating manpage into html on the fly.
+# Front-end for PolyglotMan (formerly called RosettaMan)
+#
+# Author: Youki Kadobayashi  <youki@center.osaka-u.ac.jp>
+#
+# NOTE: Replace 'CGI::Apache' with just 'CGI' for systems without mod_perl.
+#
+# ALSO: You may want to recompile rman like this:
+#	MANREFPRINTF = "/mod-bin/cgi-rman.pl?keyword=%s&section=%s"
+#
+
+use CGI::Apache font;			# for people with mod_perl and apache
+## use CGI font;			# for people without mod_perl
+$par = "<P>\n";
+$brk = "<BR>\n";
+$bg = '#c0ffff';
+
+$query = new CGI::Apache;		# for people with mod_perl and apache
+## $query = new CGI;			# for people without mod_perl
+
+%mandatory = ('keyword' => 'Name or keyword',
+	      'section' => 'Manual page section');
+
+@given = $query->param;
+if ($#given < 0) {
+	&request_page;
+	exit 0;
+}
+
+foreach $field (keys %mandatory) {
+	if ($query->param($field) eq '') {
+		push(@missing, $mandatory{$field});
+	}
+}
+if ($#missing >= 0) {
+    &info_missing_page(@missing);
+} else {
+    if ($query->param('type') eq 'apropos') {
+	&apropos_page;
+    } else {
+	&manual_page;
+    }
+}
+exit 0;
+
+sub standout {
+    my ($level, $color, $string) = @_;
+
+    # As per CGI.pm documentation "Generating new HTML tags"
+    return $query->font({color => "$color"}, "<$level>$string</$level>");
+}
+
+sub error_page {
+    my ($message) = @_;
+
+#    print $query->header,
+#		$query->start_html(-title=>$message, -BGCOLOR=>$bg);
+
+    print &standout("H2", "brown", $message),
+		"The above error occured during the manual page generation",
+		" process.  Please check keyword and section number,",
+		" then try again.", $par;
+
+#    print $query->end_html;
+}
+
+sub info_missing_page {
+    my (@missing) = @_;
+
+    print $query->header,
+		$query->start_html(-title=>"Information is missing",
+				   -BGCOLOR=>$bg);
+
+    print &standout("H2", "brown", "Information is missing"),
+		"Sorry but your request was not fulfilled because",
+		" the following information is missing in your entry:", $par,
+		join(' ', @missing), $par,
+		"Please go back and make sure to enter data on the missing field.";
+
+    print $query->end_html;
+}
+
+sub request_page {
+    print $query->header,
+		$query->start_html(-title=>'Hypertext Manual Page',
+				   -author=>'Youki Kadobayashi',
+				   -BGCOLOR=>$bg);
+
+    print &standout("H2", "green", "Hypertext Manual Page");
+
+    print $query->start_form,
+	"Type of Search: ",
+		$query->radio_group(-name=>'type',
+				    -values=>['man', 'apropos'],
+				    -default=>'man'), $brk,
+	"Name or Keyword: ",
+		$query->textfield(-name=>'keyword'), $brk,
+	"Manpage Section: ",
+		$query->popup_menu(-name=>'section',
+				   -labels=>{
+				       0 => 'All Sections',
+				       1 => '1 - General Commands',
+				       2 => '2 - System Calls',
+				       3 => '3 - Library Functions',
+				       4 => '4 - Device Special Files',
+				       5 => '5 - File Formats',
+				       6 => '6 - Games',
+				       7 => '7 - Macros and Conventions',
+				       8 => '8 - System Administration',
+				       'pgsql' => 'PostgreSQL',
+				       'tcl' => 'Tcl/Tk',
+				       'mh' => 'Message Handler',
+				       'isode' => 'ISODE',
+				       'X11' => 'X Window System'},
+				   -values=>[0, 1, 2, 3, 4, 5, 6, 7, 8,
+					     'pgsql', 'tcl', 'mh',
+					     'isode', 'X11'],
+				   -default=>0), $brk;
+
+    print $query->submit(-name=>'Submit'), $query->reset,
+		$query->end_form, $par;
+
+    print $query->end_html;
+}
+
+sub manual_page {
+    my $keyword = $query->param('keyword');
+    my $section = $query->param('section');
+    my $man = "man";
+    my $ok = "-a-zA-Z0-9._";
+
+    # sanitize for security
+    $section =~ s/[^$ok]+/_/go;
+    $keyword =~ s/[^$ok]+/_/go;
+
+    if ($section =~ /^[A-Za-z]/) {
+	$man .= " -M$section";
+    }
+    elsif ($section =~ /^\d/) {
+	$section =~ s/^(\d)\w*/\1/go;
+	$man .= " -s $section";
+    }
+    $man .= ' ' . $keyword;
+    open (MAN, "$man | /usr/local/bin/rman -f html -n $keyword -s $section |");
+
+    print $query->header;
+	# start_html and end_html not needed here, since rman tacks them.
+    print "<!-- text generated with '$man' by cgi-rman.pl -->\n";
+    while (<MAN>) {
+	print $_;
+    }
+    if ($? != 0 || $. < 15) {
+	&error_page("Your request for manual page '$keyword' failed.");
+#	print "return code $?  line $.\n";
+	close(MAN);
+	return;
+    }
+    close(MAN);
+}
+
+sub apropos_page {
+    my $keyword = $query->param('keyword');
+    my $section = $query->param('section');	# igored
+    my $man = "man -k";
+    my $matches = 0;
+
+    $man .= ' ' . $keyword;
+    open (MAN, "$man |");
+
+    $url = $query->url;
+    print $query->header, $query->start_html(-title=>$man);
+
+    while (<MAN>) {
+	if (/^([a-zA-Z0-9_.-]+)[, ][a-zA-Z0-9_., -]*\((\d)\w*\)/) {
+	    print $`, qq{ <A HREF="$url?keyword=$1&section=$2&type=man"> },
+		$&, "</A>", $';
+	    ++$matches;
+	} else {
+	    print $_;
+	}
+	print $brk;
+    }
+    if ($? != 0 || $matches == 0) {
+	&error_page("Your search request with keyword '$keyword' failed.");
+#	print "return code $?  matches $matches\n";
+	close(MAN);
+	return;
+    }
+    close(MAN);
+
+    print $query->end_html;
+}
+
diff --git a/xc/extras/rman/getopt.c b/xc/extras/rman/getopt.c
new file mode 100644
index 000000000..b5e8403ef
--- /dev/null
+++ b/xc/extras/rman/getopt.c
@@ -0,0 +1,79 @@
+/*
+Newsgroups: mod.std.unix
+Subject: public domain AT&T getopt source
+Date: 3 Nov 85 19:34:15 GMT
+
+Here's something you've all been waiting for:  the AT&T public domain
+source for getopt(3).  It is the code which was given out at the 1985
+UNIFORUM conference in Dallas.  I obtained it by electronic mail
+directly from AT&T.  The people there assure me that it is indeed
+in the public domain.
+*/
+
+/*LINTLIBRARY*/
+
+extern int strlen();
+extern int strcmp();
+extern char *strchr();
+extern int write();
+
+#define NULL	0
+#define EOF	(-1)
+#define ERR(s, c)	if(opterr){\
+	char errbuf[2];\
+	errbuf[0] = c; errbuf[1] = '\n';\
+	(void) write(2, argv[0], (unsigned)strlen(argv[0]));\
+	(void) write(2, s, (unsigned)strlen(s));\
+	(void) write(2, errbuf, 2);}
+
+
+int	opterr = 1;
+int	optind = 1;
+int	optopt;
+char	*optarg;
+
+int
+getopt(argc, argv, opts)
+int	argc;
+char	**argv, *opts;
+{
+	static int sp = 1;
+	register int c;
+	register char *cp;
+
+	if(sp == 1)
+		if(optind >= argc ||
+		   argv[optind][0] != '-' || argv[optind][1] == '\0')
+			return(EOF);
+		else if(strcmp(argv[optind], "--") == NULL) {
+			optind++;
+			return(EOF);
+		}
+	optopt = c = argv[optind][sp];
+	if(c == ':' || (cp=strchr(opts, c)) == NULL) {
+		ERR(": illegal option -- ", c);
+		if(argv[optind][++sp] == '\0') {
+			optind++;
+			sp = 1;
+		}
+		return('?');
+	}
+	if(*++cp == ':') {
+		if(argv[optind][sp+1] != '\0')
+			optarg = &argv[optind++][sp+1];
+		else if(++optind >= argc) {
+			ERR(": option requires an argument -- ", c);
+			sp = 1;
+			return('?');
+		} else
+			optarg = argv[optind++];
+		sp = 1;
+	} else {
+		if(argv[optind][++sp] == '\0') {
+			sp = 1;
+			optind++;
+		}
+		optarg = NULL;
+	}
+	return(c);
+}
diff --git a/xc/extras/rman/macish.c b/xc/extras/rman/macish.c
new file mode 100644
index 000000000..57dd8de8b
--- /dev/null
+++ b/xc/extras/rman/macish.c
@@ -0,0 +1,6 @@
+#include <QuickDraw.h>
+
+void InitToolbox()
+{
+	InitGraf(&qd.thePort);
+}
diff --git a/xc/extras/rman/rman.1 b/xc/extras/rman/rman.1
new file mode 100644
index 000000000..7fa1cd200
--- /dev/null
+++ b/xc/extras/rman/rman.1
@@ -0,0 +1,269 @@
+.TH PolyglotMan 1
+.SH "NAME "
+PolyglotMan, rman - reverse compile man pages from formatted 
+form to a number of source formats 
+.SH "SYNOPSIS "
+rman [ \fIoptions \fR] [ \fIfile \fR] 
+.SH "DESCRIPTION "
+\fIPolyglotMan \fR takes man pages from most of the popular flavors 
+of UNIX and transforms them into any of a number of text source 
+formats. PolyglotMan was formerly known as RosettaMan. The name 
+of the binary is still called \fIrman \fR, for scripts that depend 
+on that name; mnemonically, just think "reverse man". Previously \fI
+PolyglotMan \fR required pages to be formatted by nroff prior 
+to its processing. With version 3.0, it \fIprefers [tn]roff source \fR
+and usually produces results that are better yet. And source 
+processing is the only way to translate tables. Source format 
+translation is not as mature as formatted, however, so try formatted 
+translation as a backup. 
+.PP
+In parsing [tn]roff source, one could implement an arbitrarily 
+large subset of [tn]roff, which I did not and will not do, so 
+the results can be off. I did implement a significant subset 
+of those use in man pages, however, including tbl (but not eqn), 
+if tests, and general macro definitions, so usually the results 
+look great. If they don't, format the page with nroff before 
+sending it to PolyglotMan. If PolyglotMan doesn't recognize a 
+key macro used by a large class of pages, however, e-mail me 
+the source and a uuencoded nroff-formatted page and I'll see 
+what I can do. When running PolyglotMan with man page source 
+that includes or redirects to other [tn]roff source using the .so (source 
+or inclusion) macro, you should be in the parent directory of 
+the page, since pages are written with this assumption. For example, 
+if you are translating /usr/man/man1/ls.1, first cd into /usr/man. 
+.PP
+\fIPolyglotMan \fR accepts man pages from: SunOS, Sun Solaris, 
+Hewlett-Packard HP-UX, AT&T System V, OSF/1 aka Digital UNIX, 
+DEC Ultrix, SGI IRIX, Linux, FreeBSD, SCO. Source processing 
+works for: SunOS, Sun Solaris, Hewlett-Packard HP-UX, AT&T System 
+V, OSF/1 aka Digital UNIX, DEC Ultrix. It can produce printable 
+ASCII-only (control characters stripped), section headers-only, 
+Tk, TkMan, [tn]roff (traditional man page source), SGML, HTML, 
+MIME, LaTeX, LaTeX2e, RTF, Perl 5 POD. A modular architecture 
+permits easy addition of additional output formats. 
+.PP
+The latest version of PolyglotMan is always available from \fI
+ftp://ftp.cs.berkeley.edu/ucb/people/phelps/tcltk/rman.tar.Z \fR. 
+.SH "OPTIONS "
+The following options should not be used with any others and 
+exit PolyglotMan without processing any input. 
+.TP 15
+-h|--help 
+Show list of command line options and exit. 
+.TP 15
+-v|--version 
+Show version number and exit. 
+.PP
+\fIYou should specify the filter first, as this sets a number 
+of parameters, and then specify other options. 
+.TP 15
+-f|--filter <ASCII|roff|TkMan|Tk|Sections|HTML|SGML|MIME|LaTeX|LaTeX2e|RTF|POD> 
+Set the output filter. Defaults to ASCII. 
+.TP 15
+-S|--source 
+PolyglotMan tries to automatically determine whether its input 
+is source or formatted; use this option to declare source input. 
+.TP 15
+-F|--format|--formatted 
+PolyglotMan tries to automatically determine whether its input 
+is source or formatted; use this option to declare formatted 
+input. 
+.TP 15
+-l|--title \fIprintf-string \fR
+In HTML mode this sets the <TITLE> of the man pages, given the 
+same parameters as \fI-r \fR. 
+.TP 15
+-r|--reference|--manref \fIprintf-string \fR
+In HTML and SGML modes this sets the URL form by which to retrieve 
+other man pages. The string can use two supplied parameters: 
+the man page name and its section. (See the Examples section.) 
+If the string is null (as if set from a shell by "-r ''"), `-' 
+or `off', then man page references will not be HREFs, just set 
+in italics. If your printf supports XPG3 positions specifier, 
+this can be quite flexible. 
+.TP 15
+-V|--volumes \fI<colon-separated list> \fR
+Set the list of valid volumes to check against when looking for 
+cross-references to other man pages. Defaults to \fI1:2:3:4:5:6:7:8:9:o:l:n:p \fR(volume 
+names can be multicharacter). If an non-whitespace string in 
+the page is immediately followed by a left parenthesis, then 
+one of the valid volumes, and ends with optional other characters 
+and then a right parenthesis--then that string is reported as 
+a reference to another manual page. If this -V string starts 
+with an equals sign, then no optional characters are allowed 
+between the match to the list of valids and the right parenthesis. (This 
+option is needed for SCO UNIX.) 
+.PP
+The following options apply only when formatted pages are given 
+as input. They do not apply or are always handled correctly with 
+the source. 
+.TP 15
+-b|--subsections 
+Try to recognize subsection titles in addition to section titles. 
+This can cause problems on some UNIX flavors. 
+.TP 15
+-K|--nobreak 
+Indicate manual pages don't have page breaks, so don't look for 
+footers and headers around them. (Older nroff -man macros always 
+put in page breaks, but lately some vendors have realized that 
+printout are made through troff, whereas nroff -man is used to 
+format pages for reading on screen, and so have eliminated page 
+breaks.) \fIPolyglotMan \fR usually gets this right even without 
+this flag. 
+.TP 15
+-k|--keep 
+Keep headers and footers, as a canonical report at the end of 
+the page. changeleft 
+Move changebars, such as those found in the Tcl/Tk manual pages, 
+to the left. --> notaggressive 
+\fIDisable \fR aggressive man page parsing. Aggressive manual, 
+which is on by default, page parsing elides headers and footers, 
+identifies sections and more. --> 
+.TP 15
+-n|--name \fIname \fR
+Set name of man page (used in roff format). If the filename is 
+given in the form " \fIname \fR. \fIsection \fR", the name and 
+section are automatically determined. If the page is being parsed 
+from [tn]roff source and it has a .TH line, this information 
+is extracted from that line. 
+.TP 15
+-p|--paragraph 
+paragraph mode toggle. The filter determines whether lines should 
+be linebroken as they were by nroff, or whether lines should 
+be flowed together into paragraphs. Mainly for internal use. 
+.TP 15
+-s|section \fI# \fR
+Set volume (aka section) number of man page (used in roff format). 
+tables 
+Turn on aggressive table parsing. --> 
+.TP 15
+-t|--tabstops \fI# \fR
+For those macros sets that use tabs in place of spaces where 
+possible in order to reduce the number of characters used, set 
+tabstops every \fI# \fR columns. Defaults to 8. 
+.SH "NOTES ON FILTER TYPES "
+.SS "ROFF "
+Some flavors of UNIX ship man page without [tn]roff source, making 
+one's laser printer little more than a laser-powered daisy wheel. 
+This filer tries to intuit the original [tn]roff directives, 
+which can then be recompiled by [tn]roff. 
+.SS "TkMan "
+TkMan, a hypertext man page browser, uses \fIPolyglotMan \fR 
+to show man pages without the (usually) useless headers and footers 
+on each pages. It also collects section and (optionally) subsection 
+heads for direct access from a pulldown menu. TkMan and Tcl/Tk, 
+the toolkit in which it's written, are available via anonymous 
+ftp from \fIftp://ftp.smli.com/pub/tcl/ \fR
+.SS "Tk "
+This option outputs the text in a series of Tcl lists consisting 
+of text-tags pairs, where tag names roughly correspond to HTML. 
+This output can be inserted into a Tk text widget by doing an \fI
+eval <textwidget> insert end <text> \fR. This format should be 
+relatively easily parsible by other programs that want both the 
+text and the tags. Also see ASCII. 
+.SS "ASCII "
+When printed on a line printer, man pages try to produce special 
+text effects by overstriking characters with themselves (to produce 
+bold) and underscores (underlining). Other text processing software, 
+such as text editors, searchers, and indexers, must counteract 
+this. The ASCII filter strips away this formatting. Piping nroff 
+output through \fIcol -b \fR also strips away this formatting, 
+but it leaves behind unsightly page headers and footers. Also 
+see Tk. 
+.SS "Sections "
+Dumps section and (optionally) subsection titles. This might 
+be useful for another program that processes man pages. 
+.SS "HTML "
+With a simple extention to an HTTP server for Mosaic or other 
+World Wide Web browser, \fIPolyglotMan \fR can produce high quality 
+HTML on the fly. Several such extensions and pointers to several 
+others are included in \fIPolyglotMan \fR's \fIcontrib \fR directory. 
+.SS "SGML "
+This is appoaching the Docbook DTD, but I'm hoping that someone 
+that someone with a real interest in this will polish the tags 
+generated. Try it to see how close the tags are now. 
+.SS "MIME "
+MIME (Multipurpose Internet Mail Extensions) as defined by RFC 1563, 
+good for consumption by MIME-aware e-mailers or as Emacs (>=19.29) 
+enriched documents. 
+.SS "LaTeX and LaTeX2e "
+Why not? 
+.SS "RTF "
+Use output on Mac or NeXT or whatever. Maybe take random man 
+pages and integrate with NeXT's documentation system better. 
+Maybe NeXT has own man page macros that do this. 
+.SS "PostScript and FrameMaker "
+To produce PostScript, use \fIgroff \fR or \fIpsroff \fR. To 
+produce FrameMaker MIF, use FrameMaker's builtin filter. In both 
+cases you need \fI[tn]roff \fR source, so if you only have a 
+formatted version of the manual page, use \fIPolyglotMan \fR's 
+roff filter first. 
+.SH "EXAMPLES "
+To convert the \fIformatted \fR man page named \fIls.1 \fR back 
+into [tn]roff source form: 
+.PP
+\fIrman -f roff /usr/local/man/cat1/ls.1 > /usr/local/man/man1/ls.1 \fR
+.br
+.PP
+Long man pages are often compressed to conserve space (compression 
+is especially effective on formatted man pages as many of the 
+characters are spaces). As it is a long man page, it probably 
+has subsections, which we try to separate out (some macro sets 
+don't distinguish subsections well enough for \fIPolyglotMan \fR
+to detect them). Let's convert this to LaTeX format: 
+.br
+.PP
+\fIpcat /usr/catman/a_man/cat1/automount.z | rman -b -n automount -s 1 -f 
+latex > automount.man \fR
+.br
+.PP
+Alternatively, \fIman 1 automount | rman -b -n automount -s 1 -f 
+latex > automount.man \fR
+.br
+.PP
+For HTML/Mosaic users, \fIPolyglotMan \fR can, without modification 
+of the source code, produce HTML links that point to other HTML 
+man pages either pregenerated or generated on the fly. First 
+let's assume pregenerated HTML versions of man pages stored in \fI/usr/man/html \fR. 
+Generate these one-by-one with the following form: 
+.br
+\fIrman -f html -r 'http:/usr/man/html/%s.%s.html' /usr/man/cat1/ls.1 > /usr/man/html/ls.1.html \fR
+.br
+.PP
+If you've extended your HTML client to generate HTML on the fly 
+you should use something like: 
+.br
+\fIrman -f html -r 'http:~/bin/man2html?%s:%s' /usr/man/cat1/ls.1 \fR
+.br
+when generating HTML. 
+.SH "BUGS/INCOMPATIBILITIES "
+\fIPolyglotMan \fR is not perfect in all cases, but it usually 
+does a good job, and in any case reduces the problem of converting 
+man pages to light editing. 
+.PP
+Tables in formatted pages, especially H-P's, aren't handled very 
+well. Be sure to pass in source for the page to recognize tables. 
+.PP
+The man pager \fIwoman \fR applies its own idea of formatting 
+for man pages, which can confuse \fIPolyglotMan \fR. Bypass \fI
+woman \fR by passing the formatted manual page text directly 
+into \fIPolyglotMan \fR. 
+.PP
+The [tn]roff output format uses fB to turn on boldface. If your 
+macro set requires .B, you'll have to a postprocess the \fIPolyglotMan \fR
+output. 
+.SH "SEE ALSO "
+\fItkman(1) \fR, \fIxman(1) \fR, \fIman(1) \fR, \fIman(7) \fR
+or \fIman(5) \fR depending on your flavor of UNIX 
+.SH "AUTHOR "
+PolyglotMan 
+.br
+by Thomas A. Phelps ( \fIphelps@ACM.org \fR) 
+.br
+developed at the 
+.br
+University of California, Berkeley 
+.br
+Computer Science Division 
+.PP
+Manual page last updated on $Date: 2000/07/11 19:36:05 $ 
diff --git a/xc/extras/rman/rman.c b/xc/extras/rman/rman.c
new file mode 100644
index 000000000..1e20cc254
--- /dev/null
+++ b/xc/extras/rman/rman.c
@@ -0,0 +1,5460 @@
+static char rcsid[] = "Header: /home/cs/phelps/spine/rman/RCS/rman.c,v 1.144 1999/08/10 00:41:55 phelps Exp phelps $";
+
+/*
+   PolyglotMan by Thomas A. Phelps (phelps@ACM.org)
+
+  accept man pages as formatted by (10)
+     Hewlett-Packard HP-UX, AT&T System V, SunOS, Sun Solaris, OSF/1, DEC Ultrix,
+	SGI IRIX, Linux, FreeBSD, SCO
+
+  output as (9)
+     printable ASCII, section headers only, TkMan, [tn]roff, HTML,
+	LaTeX, LaTeX2e, RTF, Perl pod, MIME, and soon SGML
+
+     written March 24, 1993
+	bs2tk transformed into RosettaMan November 4-5, 1993
+     source interpretation added September 24, 1996
+	renamed PolyglotMan due to lawsuit by Rosetta, Inc. August 8, 1997
+*/
+/* $XFree86: xc/extras/rman/rman.c,v 1.9 2000/06/13 23:15:47 dawes Exp $ */
+
+
+/* TO DO ****
+
+   clean up relationship between source and formatted filtering support routines
+
+   output to SGML with Davenport DTD
+
+   don't give SHORTLINE if just finished bullet of bultxt, ended section head, ... other cases?
+   make sure text following bullet is aligned correctly
+
+   output to WinHelp?  don't have specs (anybody interested?)
+   collect header and footer until hit blank line?
+   what to do about tables?   count second gap of spaces & average gap? ==>
+      good idea but tables too variable for this to work
+   internal, outline-like header section for HTML documents?  how to put this *first*?  (can't in single pass)
+   one line look ahead to enable better parsing (item lists, et cetera)
+   alluc (==nonlc) flag, copy curline to last line vector (works well with lookahead cache)
+   ??      collect sundry globals into vectors (i.e., arrays and enum indexes)
+              (if compiler has good constant propagation, then shouldn't slow access)
+   collect scattered globals into vectors (e.g., curline[ispcnt]): array + enum
+      curline, lastline, flags, pending, bufs+lens
+*/
+
+
+#include <unistd.h>
+#include <stdio.h>
+#include <string.h>
+#include <ctype.h>
+#include <stdlib.h>
+#include <assert.h>
+/* I'm sure I'll need some #ifdef's here to include the right headers */
+/* ... but compiles swell on Solaris, DEC Alpha, HP, SunOS */
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <fcntl.h>
+
+#ifdef Lynx
+extern int optind;
+extern char *optarg;
+#endif
+
+/*** make #define's into consts? => can't because compilers not smart enough ***/
+/* maximum number of tags per line */
+#define MAXTAGS 50*100
+/* BUFSIZ = 1024 on Solaris */
+#define MAXBUF 2*5000
+#define MAXLINES 20000
+#define MAXTOC 500
+#define xputchar(c)		(fcharout? putchar(c): (c))
+#define sputchar(c)		(fcharout? plain[sI++]=(char)c: (char)(c))
+#define stagadd(tag)	tagadd(tag,sI,0)
+enum { c_rsquote='\''/*\x27*/, c_lsquote='`'/*\x60*/, c_dagger='\xa7', c_bullet='\xb7', c_plusminus='\xb1' };
+
+
+/*** tag management ***/
+
+enum tagtype { NOTAG, TITLE, ITALICS, BOLD, SYMBOL, SMALLCAPS, BOLDITALICS, MONO, MANREF };	/* MANREF last */
+struct { enum tagtype type; int first; int last; } tags[MAXTAGS];
+int tagc=0;
+struct { char *text; int type; int line; } toc[MAXTOC];
+int tocc=0;
+
+
+/* characters in this list automatically prefixed by a backslash (set in output format function */
+char *escchars="";
+char *vollist = VOLLIST;
+const char *manvalid = "._-+:";	/* in addition to alphanumerics, valid characters to find in a man page name */
+char *manrefname;
+char *manrefsect;
+
+enum command {
+
+	/*BEGINCHARTAGS,*/
+	CHARTAB='\t',
+	CHARPERIOD='.', CHARLSQUOTE='`', CHARRSQUOTE='\'', CHARGT='>', CHARLT='<',
+	CHARAMP='&', CHARBACKSLASH='\\', CHARDASH='-', CHARHAT='^', CHARVBAR='|',
+ 	CHARNBSP=0xa0,	CHARCENT=0xa2, CHARSECT=0xa7, CHARCOPYR=0xa9, CHARNOT=0xac,
+ 	CHARDAGGER=0xad, CHARREGTM=0xae, CHARDEG=0xb0, CHARPLUSMINUS=0xb1,
+ 	CHARACUTE=0xb4, CHARBULLET=0xb7, CHAR14=0xbc, CHAR12=0xbd, CHAR34=0xbe,
+ 	CHARMUL=0xd7, CHARDIV=0xf7,
+	CHANGEBAR=0x100, CHARLQUOTE, CHARRQUOTE, HR,
+	/*ENDCHARTAGS,*/
+
+	/*BEGINFONTTAGS,*/
+	BEGINBOLD, ENDBOLD, BEGINITALICS, ENDITALICS, BEGINBOLDITALICS, ENDBOLDITALICS,
+	BEGINSC, ENDSC, BEGINY, ENDY, BEGINCODE, ENDCODE, BEGINMANREF, ENDMANREF,
+	FONTSIZE,
+	/*ENDFONTTAGS*/
+
+	/*BEGINLAYOUTTAGS,*/
+	ITAB, BEGINCENTER, ENDCENTER,
+	/*ENDLAYOUTTAGS,*/
+
+	/*BEGINSTRUCTTAGS,*/
+	BEGINDOC, ENDDOC, BEGINCOMMENT, ENDCOMMENT, COMMENTLINE, BEGINBODY, ENDBODY, 
+	BEGINHEADER, ENDHEADER, BEGINFOOTER, ENDFOOTER, BEGINLINE, ENDLINE, SHORTLINE,
+	BEGINSECTION, ENDSECTION, BEGINSUBSECTION, ENDSUBSECTION,
+	BEGINSECTHEAD, ENDSECTHEAD, BEGINSUBSECTHEAD, ENDSUBSECTHEAD,
+	BEGINBULPAIR, ENDBULPAIR, BEGINBULLET, ENDBULLET, BEGINBULTXT, ENDBULTXT,
+	BEGINTABLE, ENDTABLE, BEGINTABLELINE, ENDTABLELINE, BEGINTABLEENTRY, ENDTABLEENTRY,
+	BEGININDENT, ENDINDENT, BEGINCODEBLOCK, ENDCODEBLOCK,
+
+	BEGINDIFFA, ENDDIFFA, BEGINDIFFD, ENDDIFFD
+	/*,*//*ENDSTRUCTTAGS,*/
+};
+
+const char *tcltkOP[] = { "Command-Line Name", "Database Name", "Database Class" };
+
+
+/* characters that need special handling in any output format, *more than just a backslash* */
+/* characters in this list need a corresponding case statement in each output format */
+/*char *trouble="\t.`'><&\\^|-\xa7\xb7\xb1";*/
+const unsigned char trouble[]= { CHARTAB, CHARPERIOD, CHARLSQUOTE, CHARRSQUOTE,
+	CHARGT, CHARLT, CHARAMP, CHARBACKSLASH, CHARDASH, CHARHAT, CHARVBAR, CHARCENT,
+	CHARSECT, CHARCOPYR, CHARNOT, CHARDAGGER, CHARREGTM, CHARDEG, CHARPLUSMINUS,
+	CHARACUTE, CHARBULLET, CHAR14, CHAR12, CHAR34, CHARMUL, CHARDIV,
+	0 };
+
+
+enum command tagbeginend[][2] = {	/* parallel to enum tagtype */
+	{ -1,-1 },
+	{ -1,-1 },
+	{ BEGINITALICS, ENDITALICS },
+	{ BEGINBOLD, ENDBOLD },
+	{ BEGINY, ENDY },
+	{ BEGINSC, ENDSC },
+	{ BEGINBOLDITALICS, ENDBOLDITALICS },
+	{ -1,-1 },
+	{ BEGINMANREF, ENDMANREF }
+};
+
+void (*fn)(enum command) = NULL;
+enum command prevcmd = BEGINDOC;
+
+
+/*** globals ***/
+/* move all flags into an array?
+enum { fSubsX, fLast };
+int flags[fLast];
+*/
+
+int fSource=-1;	/* -1 => not determined yet */
+int fDiff=0;
+FILE *difffd;
+char diffline[MAXBUF];
+char diffline2[MAXBUF];
+char *message = NULL;
+int fontdelta=0;
+int intArg;
+
+int fPara=0;			/* line or paragraph groupings of text */
+int fSubsections=0;	/* extract subsection titles too? */
+int fChangeleft=0;	/* move change bars to left? (-1 => delete them) */
+int fReflow=0;
+int fURL=0;		/* scan for URLs too? */
+#if 0
+int fMan=1;		/* invoke agressive man page filtering? */
+#endif
+int fQS=0;		/* squeeze out spaces (scnt and interword)? */
+int fIQS=0;		/* squeeze out initial spaces (controlled separately from fQS) */
+int fILQS=0;		/* squeeze out spaces for usual indent */
+int fHeadfoot=0;	/* show canonical header and footer at bottom? */
+int falluc=0;
+int itabcnt=0;
+int fQuiet=0;
+int fTclTk=0;
+
+/* patterns observed in section heads that don't conform to first-letter-uppercase-rest-lowercase pattern (stay all uc, or go all lc, or have subsequent uc) */
+int lcexceptionslen = -1;	/* computed by system */
+char *lcexceptions[] = {
+/* new rule: double/all consonants == UC? */
+	/* articles, verbs, conjunctions, prepositions, pronouns */
+	"a", "an", "the", 
+	"am", "are", "is", "were",
+	"and", "or",
+	"by", "for", "from", "in", "into", "it", "of", "on", "to", "with",
+	"that", "this",
+
+	/* terms */
+	"API", "CD", "GUI", "UI", /*I/O=>I/O already*/ "ID", "IDs", "OO",
+	"IOCTLS", "IPC", "RPC",
+
+	/* system names */
+	"AWK", "cvs", "rcs", "GL", "vi", "PGP", "QuickTime", "DDD", "XPG/3",
+	"NFS", "NIS", "NIS+", "AFS", 
+	"UNIX", "SysV", 
+	"XFree86", "ICCCM",
+	"MH", "MIME",
+	"TeX", "LaTeX", "PicTeX",
+	"PostScript", "EPS", "EPSF", "EPSI", 
+	"HTML", "URL", "WWW",
+
+	/* institution names */
+	"ANSI", "CERN", "GNU", "ISO", "NCSA",
+
+	/* Sun-specific */	
+	"MT-Level", "SPARC",
+
+	NULL
+};
+/* what exceptions have you encountered? */
+
+int TabStops=8;
+int hanging=0;		/* location of hanging indent (if ==0, none) */
+enum { NAME, SYNOPSIS, DESCRIPTION, SEEALSO, FILES, AUTHOR, RANDOM };	/* RANDOM last */
+char *sectheadname[] = {
+  "NAME", "SYNOPSIS", "DESCRIPTION:INTRODUCTION", "SEE ALSO:RELATED INFORMATION", "FILES", "AUTHOR:AUTHORS", "RANDOM"
+};
+int sectheadid = RANDOM;
+int oldsectheadid = RANDOM;
+
+int fCodeline=0;
+int fNOHY=0;		/* re-linebreak so no words are hyphenated; not used by TkMan, but gotta keep for people converting formatted text */
+int fNORM=0;		/* normalize?  initial space => tabs, no changebars, exactly one blank line between sections */
+const char TABLEOFCONTENTS[] = "Table of Contents";
+const char HEADERANDFOOTER[] = "Header and Footer";
+char manName[80]="man page";
+char manSect[10]="1";
+const char provenance[] =
+	"manual page source format generated by PolyglotMan v" POLYGLOTMANVERSION;
+const char anonftp[] =
+	"available via anonymous ftp from ftp.cs.berkeley.edu:/ucb/people/phelps/tcltk/rman.tar.Z";
+const char horizontalrule[] = "------------------------------------------------------------";
+
+const int LINEBREAK=70;
+int linelen=0;			/* length of result in plain[] */
+int spcsqz;			/* number of spaces squeezed out */
+int ccnt=0;			/* # of changebars */
+int scnt,scnt2;		/* counts of initial spaces in line */
+int s_sum,s_cnt;
+int bs_sum, bs_cnt;
+int ncnt=0,oncnt=0;		/* count of interline newlines */
+int CurLine=1;
+int AbsLine=1-1;		/* absolute line number */
+int indent=0;			/* global indentation */
+int lindent=0;			/* usual local indent */
+int auxindent=0;		/* aux indent */
+int I;			/* index into line/paragraph */
+int fcharout=1;		/* show text or not */
+char lookahead;
+/*int tabgram[MAXBUF];*/	/* histogram of first character positions */
+char buf[MAXBUF];
+char plain[MAXBUF];		/* current text line with control characters stripped out */
+char hitxt[MAXBUF];		/* highlighted text (available at time of BEGIN<highlight> signal */
+char header[MAXBUF]/*=""*/;		/* complete line */
+char header2[MAXBUF]/*=""*/;		/* SGIs have two lines of headers and footers */
+char header3[MAXBUF]/*=""*/;		/* GNU and some others have a third! */
+char footer[MAXBUF]/*=""*/;
+char footer2[MAXBUF]/*=""*/;
+#define CRUFTS 5
+char *cruft[CRUFTS] = { header, header2, header3, footer, footer2 };
+char *File, *in;		/* File = pointer to full file contents, in = current file pointer */
+char *argv0;
+int finTable=0;
+char tableSep='\0';	/*\t';*/
+/*int fTable=0;
+int fotable=0;*/
+char *tblcellformat;
+int tblcellspan;
+/*int tblspanmax;*/
+int listtype=-1;	/* current list type bogus to begin with */
+enum listtypes { DL, OL, UL };
+
+int fIP=0;
+
+
+
+/*** utility functions ***/
+
+
+/* case insensitive versions of strcmp and strncmp */
+
+static int
+stricmp(const char *s1, const char *s2)
+{
+	assert(s1!=NULL && s2!=NULL);
+	/*strincmp(s1, s2, strlen(s1)+1);*/
+
+	while (tolower(*s1)==tolower(*s2)) {
+		if (*s1=='\0' /*&& *s2=='\0'*/) return 0;
+		s1++; s2++;
+	}
+
+	if (tolower(*s1)<tolower(*s2)) return -1;
+	else return 1;
+}
+
+static int
+lcexceptionscmp(const void *a, const void *b)
+{
+	return stricmp(*(char **)a, *(char **)b);
+}
+
+static int
+strincmp(const char *s1, const char *s2, size_t n)
+{
+	assert(s1!=NULL && s2!=NULL && n>0);
+
+	while (n>0 && tolower(*s1)==tolower(*s2)) {
+		n--; s1++; s2++;
+	}
+	if (n==0) return 0;
+	else if (tolower(*s1)<tolower(*s2)) return -1;
+	else return 1;
+}
+
+/* compare string and a colon-separated list of strings */
+static int
+strcoloncmp2(char *candidate, int end, const char *list, int sen)
+{
+	const char *l = list;
+	char *c,c2;
+
+	assert(candidate!=NULL && list!=NULL);
+	assert(end>=-1 && end<=255);
+	assert(sen==0 || sen==1);
+
+	if (*l==':') l++;	/* tolerate a leading colon */
+
+	/* invariant: c and v point to start of strings to compare */
+	while (*l) {
+		assert(l==list || l[-1]==':');
+		for (c=candidate; *c && *l; c++,l++)
+			if ((sen && *c!=*l) || (!sen && tolower(*c)!=tolower(*l)))
+				break;
+
+		/* if candidate matches a valid one as far as valid goes, it's a keeper */
+		if ((*l=='\0' || *l==':') && (*c==end || end==-1)) {
+		  if (*c=='\b') {
+		    c2 = c[-1];
+		    while (*c=='\b' && c[1]==c2) c+=2;
+		  }
+		  /* no volume qualifiers with digits */
+		  if (!isdigit(*c)) return 1;
+		}
+
+		/* bump to start of next valid */
+		while (*l && *l++!=':') /* nada */;
+	}
+
+	return 0;
+}
+
+static int
+strcoloncmp(char *candidate, int end, const char *list)
+{
+	int sen=1;
+	const char *l = list;
+
+	assert(candidate!=NULL && list!=NULL);
+	assert(end>=-1 && end<=255);
+
+	if (*l=='=') l++; else end=-1;
+	if (*l=='i') { sen=0; l++; }
+
+	return strcoloncmp2(candidate, end, l, sen);
+}
+
+/* strdup not universally available */
+static char *
+mystrdup(const char *p)
+{
+  char *q;
+
+  if (p==NULL) return NULL;
+
+  q = malloc(strlen(p)+1);	/* +1 gives space for \0 that is not reported by strlen */
+  if (q!=NULL) strcpy(q,p);
+  return q;
+}
+
+
+/* given line of text, return "casified" version in place:
+   if word in exceptions list, return exception conversion
+   else uc first letter, lc rest
+*/
+static void
+casify(char *p)
+{
+	char tmpch, *q, **exp;
+	int fuc;
+
+	for (fuc=1; *p; p++) {
+		if (isspace(*p) || strchr("&/",*p)!=NULL) fuc=1;
+		else if (fuc) {
+			/* usually */
+			if (p[1] && isupper(p[1]) /*&& p[2] && isupper(p[2])*/) fuc=0;
+			/* check for exceptions */
+			for (q=p; *q && !isspace(*q); q++) /*nada*/;
+			tmpch = *q; *q='\0';
+			exp = (char **)bsearch(&p, lcexceptions, lcexceptionslen, sizeof(char *), lcexceptionscmp);
+			*q = tmpch;
+			if (exp!=NULL) {
+				for (q=*exp; *q; q++) *p++=*q;
+				fuc = 1;
+			}
+		} else *p=tolower(*p);
+	}
+}
+
+
+/* add an attribute tag to a range of characters */
+
+static void
+tagadd(int /*enum tagtype--abused in source parsing*/ type, int first, int last)
+{
+	assert(type!=NOTAG);
+
+	if (tagc<MAXTAGS) {
+		tags[tagc].type = type;
+		tags[tagc].first = first;
+		tags[tagc].last = last;
+		tagc++;
+	}
+}
+
+
+/*
+   collect all saves to string table one one place, so that
+   if decide to go with string table instead of multiple malloc, it's easy
+   (probably few enough malloc's that more sophistication is unnecessary)
+*/
+
+static void
+tocadd(char *text, enum command type, int line)
+{
+	char *r;
+
+	assert(text!=NULL && strlen(text)>0);
+	assert(type==BEGINSECTION || type==BEGINSUBSECTION);
+
+	if (tocc<MAXTOC) {
+		r = malloc(strlen(text)+1); if (r==NULL) return;
+		strcpy(r,text);
+		toc[tocc].text = r;
+		toc[tocc].type = type;
+		toc[tocc].line = line;
+		tocc++;
+	}
+}
+
+
+
+char *manTitle = MANTITLEPRINTF;
+char *manRef = MANREFPRINTF;
+char *href;
+int fmanRef=1;	/* make 'em links or just show 'em? */
+
+static void
+manrefextract(char *p)
+{
+  char *p0;
+  static char *nonhref = "\">'";
+
+  while (*p==' ') p++;
+  if (strincmp(p,"http",4)==0) {
+	href="%s"; manrefname = p;
+	p+=4;
+	while (*p && !isspace(*p) && !strchr(nonhref,*p)) p++;
+  } else {
+	href = manRef;
+
+	manrefname = p;
+	while (*p && *p!=' ' && *p!='(') p++; *p++='\0';
+	while (*p==' ' || *p=='(') p++; p0=p;
+#ifdef XFree86
+	/* Don't allow a letter after the man section number */
+	p++;
+#else
+	while (*p && *p!=')') p++;
+#endif
+	manrefsect = p0;
+  }
+  *p='\0';
+}
+
+
+
+
+/*
+ * OUTPUT FORMATS
+ *    *** break these out so can selectively include them in the binary ***
+ *    *** does this save significant space? ***
+ */
+
+static void
+formattedonly(void)
+{
+	fprintf(stderr, "The output formats for Tk and TkMan require nroff-formatted input\n");
+	exit(1);
+}
+
+
+/*
+ * DefaultFormat -- in weak OO inheritance, top of hierarchy for everybody
+ */
+static void
+DefaultFormat(enum command cmd)
+{
+  int i;
+  
+  switch (cmd) {
+  case ITAB:
+    for (i=0; i<itabcnt; i++) putchar('\t');
+    break;
+  default:
+    /* nada */
+    break;
+  }
+}
+
+
+/*
+ * DefaultLine -- in weak OO inheritance, top of hierarchy for line-based formats
+ * for output format to "inherit", have "default: DefaultLine(cmd)" and override case statement "methods"
+ */
+
+static void
+DefaultLine(enum command cmd) {
+	switch (cmd) {
+	default:
+	  DefaultFormat(cmd);
+	}
+}
+
+
+/*
+ * DefaultPara -- top of hierarchy for output formats that are formatted by their viewers
+ */
+
+static void
+DefaultPara(enum command cmd)
+{
+	switch (cmd) {
+	default:
+	  DefaultFormat(cmd);
+	}
+}
+
+
+
+/*
+ * Tk -- just emit list of text-tags pairs
+ */
+
+static void
+Tk(enum command cmd)
+{
+	static int skip=0;	/* skip==1 when line has no text */
+	int i;
+
+	if (fSource) formattedonly();
+
+	/* invariant: always ready to insert text */
+
+	switch (cmd) {
+	   case BEGINDOC:
+		I=0; CurLine=1;
+		escchars = "\"[]$";
+		printf(/*$t insert end */ "\"");
+		break;
+	   case ENDDOC:
+		if (fHeadfoot) {
+/*	grr, should have +mark syntax for Tk text widget! -- maybe just just +sect#, +subsect#
+		printf("\\n\\n\" {} \"%s\\n\" {+headfoot h2}\n",HEADERANDFOOTER);
+*/
+			printf("\\n\\n\" {} \"%s\\n\" h2\n",HEADERANDFOOTER);
+			/*printf("$t mark set headfoot %d.0\n",CurLine);*/
+			CurLine++;
+
+			for (i=0; i<CRUFTS; i++) {
+				if (*cruft[i]) {
+					printf(/*$t insert end */"{%s} sc \\n\n",cruft[i]);
+					CurLine++;
+				}
+			}
+		} else printf("\"\n");
+		break;
+
+	   case COMMENTLINE: printf("# "); break;
+
+	   case BEGINLINE:
+		/*I=0; -- need to do this at end of line so set for filterline() */
+		/* nothing to do at start of line except catch up on newlines */
+		for (i=0; i<ncnt; i++) printf("\\n");
+		CurLine+=ncnt;
+		/*if (fSource) for (i=0; i<indent; i++) putchar('\t');*/
+		break;
+	   case ENDLINE:
+		/*if (!fSource) {*/
+		  if (!skip) /*if (ncnt)*/ printf("\\n"); /*else xputchar(' ');*/
+		  skip=0;
+		  CurLine++; I=0;
+		/*
+		} else {
+		  putchar(' '); I++;
+		}
+		*/
+		break;
+
+	   case ENDSECTHEAD:
+		printf("\\n\" h2 \"");
+		tagc=0;
+		skip=1;
+		break;
+	   case ENDSUBSECTHEAD:
+		printf("\\n\" h3 \"");	/* add h3? */
+		tagc=0;
+		skip=1;
+		break;
+	   case HR:			/*printf("\\n%s\\n", horizontalrule); CurLine+=2; I=0;*/ break;
+	   case BEGINTABLEENTRY:
+		/*if (fSource) putchar('\t');*/
+		break;
+	   case BEGINTABLELINE:
+	   case ENDTABLEENTRY:
+		break;
+	   case ENDTABLELINE:
+		printf("\" tt \"");
+		/*tagadd(MONO, 0, I);*/
+		break;
+
+	   case CHANGEBAR:		putchar('|'); I++; break;
+	   case CHARLQUOTE:
+	   case CHARRQUOTE:
+		putchar('\\'); putchar('"'); I++;
+		break;
+	   case CHARLSQUOTE:
+	   case CHARRSQUOTE:
+	   case CHARPERIOD:
+	   case CHARTAB:
+	   case CHARDASH:
+	   case CHARLT:
+	   case CHARGT:
+	   case CHARHAT:
+	   case CHARVBAR:
+	   case CHARAMP:
+	   case CHARPLUSMINUS:
+	   case CHARNBSP:
+ 	   case CHARCENT:
+ 	   case CHARSECT:
+ 	   case CHARCOPYR:
+ 	   case CHARNOT:
+ 	   case CHARREGTM:
+ 	   case CHARDEG:
+ 	   case CHARACUTE:
+ 	   case CHAR14:
+ 	   case CHAR12:
+ 	   case CHAR34:
+ 	   case CHARMUL:
+ 	   case CHARDIV:
+  		putchar(cmd); I++; break;
+ 	   case CHARDAGGER:
+ 		putchar('+'); I++; break;
+	   case CHARBACKSLASH:	printf("\\\\"); I++; break;
+	   case CHARBULLET:		printf("\" {} %c symbol \"",c_bullet); I++; break;
+
+
+	   case BEGINSECTHEAD:
+	   case BEGINSUBSECTHEAD:
+		/*if (fSource && sectheadid!=NAME) { printf("\\n\\n"); CurLine+=2; I=0; }*/
+		tagc=0;	/* section and subsection formatting controlled descriptively */
+		/* no break;*/
+
+	   case BEGINBOLD:
+	   case BEGINITALICS:
+	   case BEGINBOLDITALICS:
+	   case BEGINCODE:
+	   case BEGINY:
+	   case BEGINSC:
+	   case BEGINMANREF:
+		/* end text, begin attributed text */
+		printf("\" {} \"");
+		break;
+
+	   /* rely on the fact that no more than one tag per range of text */
+	   case ENDBOLD:		printf("\" b \""); break;
+	   case ENDITALICS:		printf("\" i \""); break;
+	   case ENDBOLDITALICS:	printf("\" bi \""); break;
+	   case ENDCODE:		printf("\" tt \""); break;
+	   case ENDY:			printf("\" symbol \""); break;
+	   case ENDSC:			printf("\" sc \""); break;
+	   case ENDMANREF:		printf("\" manref \""); break;
+		/* presentation attributes dealt with at end of line */
+
+	   case BEGINBODY:
+		/*if (fSource) { printf("\\n\\n"); CurLine+=2; I=0; }*/
+		break;
+	   case SHORTLINE:
+		/*if (fSource) { printf("\\n"); CurLine++; I=0; }*/
+		break;
+	   case ENDBODY:
+	   case BEGINBULPAIR: case ENDBULPAIR:
+		/*if (fSource) { printf("\\n"); CurLine++; I=0; }*/
+		break;
+	   case BEGINBULTXT:
+		/*if (fSource) putchar('\t');*/
+		break;
+	   case BEGINBULLET: case ENDBULLET:
+	   case ENDBULTXT:
+	   case BEGINSECTION: case ENDSECTION:
+	   case BEGINSUBSECTION: case ENDSUBSECTION:
+	   case BEGINHEADER: case ENDHEADER:
+	   case BEGINFOOTER: case ENDFOOTER:
+	   case BEGINTABLE: case ENDTABLE:
+	   case FONTSIZE:
+	   case BEGININDENT: case ENDINDENT:
+		/* no action */
+		break;
+	   default:
+		DefaultLine(cmd);
+	}
+}
+
+
+/*
+ * TkMan -- Tk format wrapped with commands
+ */
+
+int linetabcnt[MAXLINES];	/* don't want to bother with realloc */
+int clocnt=0, clo[MAXLINES];
+int paracnt=0, para[MAXLINES];
+int rebuscnt=0, rebus[MAXLINES];
+int rebuspatcnt=0, rebuspatlen[25];
+char *rebuspat[25];
+
+static void
+TkMan(enum command cmd)
+{
+	static int lastscnt=-1;
+	static int lastlinelen=-1;
+	static int lastsect=0;
+	/*static int coalese=0;*/
+	static int finflow=0;
+	int i;
+	char c,*p;
+
+	/* invariant: always ready to insert text */
+
+	switch (cmd) {
+	   case BEGINDOC:
+		printf("$t insert end ");	/* opening quote supplied in Tk() below */
+		Tk(cmd);
+		break;
+	   case ENDDOC:
+		Tk(ENDLINE);
+
+		if (fHeadfoot) {
+/*	grr, should have +mark syntax for Tk text widget!
+		printf("\\n\\n\" {} \"%s\\n\" {+headfoot h2}\n", HEADERANDFOOTER);
+*/
+			printf("\\n\\n\" {} \"%s\\n\" h2\n", HEADERANDFOOTER);
+/*			printf("$t mark set headfoot end-2l\n");*/
+			CurLine++;
+
+			for (i=0; i<CRUFTS; i++) {
+				if (*cruft[i]) {
+					printf("$t insert end {%s} sc \\n\n",cruft[i]);
+					CurLine++;
+				}
+			}
+		} else printf("\"\n");
+
+/*
+		printf("$t insert 1.0 {");
+		for (i=0; i<MAXBUF; i++) if (tabgram[i]) printf("%d=%d, ", i, tabgram[i]);
+		printf("\\n\\n}\n");
+*/
+
+		printf("set manx(tabcnts) {"); for (i=1; i<CurLine; i++) printf("%d ", linetabcnt[i]); printf("}\n");
+		printf("set manx(clo) {"); for (i=0; i<clocnt; i++) printf("%d ", clo[i]); printf("}\n");
+		printf("set manx(para) {"); for (i=0; i<paracnt; i++) printf("%d ", para[i]); printf("}\n");
+		printf("set manx(reb) {"); for (i=0; i<rebuscnt; i++) printf("%d ", rebus[i]); printf("}\n");
+
+		break;
+
+	   case BEGINCOMMENT: fcharout=0; break;
+	   case ENDCOMMENT: fcharout=1; break;
+	   case COMMENTLINE: break;
+
+	   case ENDSECTHEAD:
+	   case ENDSUBSECTHEAD:
+		lastsect=1;
+		Tk(cmd);
+		break;
+
+	   case BEGINLINE:
+	     Tk(cmd);
+		linetabcnt[CurLine] = itabcnt;
+		/* old pattern for command line options "^\\|*\[ \t\]+-\[^-\].*\[^ \t\]" */
+		c = plain[0];
+		if (linelen>=2 && ((c=='-' || c=='%' || c=='\\' || c=='$' /**/ /* not much talk of money in man pages so reasonable */) && (isalnum(plain[1]) /*<= plain[1]!='-'*//*no dash*/ || ncnt/*GNU long option*/) && plain[1]!=' ') ) clo[clocnt++] = CurLine;
+		/*
+		   would like to require second letter to be a capital letter to cut down on number of matches,
+		   but command names usually start with lowercase letter
+		   maybe use a uppercase requirement as secondary strategy, but probably not
+		*/
+		if ((ncnt || lastsect) && linelen>0 && scnt>0 && scnt<=5) para[paracnt++] = CurLine;
+		lastsect=0;
+
+
+		/* rebus too, instead of search through whole Tk widget */
+		if (rebuspatcnt && scnt>=5 /* not sect or subsect heads */) {
+			for (p=plain; *p && *p!=' '; p++) /*empty*/;	/* never first word */
+			while (*p) {
+				for (i=0; i<rebuspatcnt; i++) {
+					if (tolower(*p) == tolower(*rebuspat[i]) && strincmp(p, rebuspat[i], rebuspatlen[i])==0) {
+						/* don't interfere with man page refs */
+						for (; *p && !isspace(*p); p++) if (*p=='(') continue;
+						rebus[rebuscnt++] = CurLine;
+						p="";	/* break for outer */
+						break;	/* just locating any line with any rebus, not exact positions */
+					}
+				}
+				/* just check start of words, though doesn't have to be full word (if did, could use strlen rather than strnlen) */
+				while (*p && *p!=' ') p++;
+				while (*p && *p==' ') p++;
+			}
+		}
+
+
+		if (fReflow && !ncnt && (finflow || lastlinelen>50) && (abs(scnt-lastscnt)<=1 || abs(scnt-hanging)<=1)) {
+		  finflow=1;
+		  putchar(' ');
+		} else {
+		  Tk(ENDLINE);
+		  /*if ((CurLine&0x3f)==0x3f) printf("\"\nupdate idletasks\n$t insert end \""); blows up some Tk text buffer, apparently, on long lines*/
+		  if ((CurLine&0x1f)==0x1f) printf("\"\nupdate idletasks\n$t insert end \"");
+		  finflow=0;
+
+		  /*if (fCodeline) printf("CODE");*/
+		}
+		lastlinelen=linelen; lastscnt=scnt;
+		break;
+
+	   case ENDLINE:
+		/* don't call Tk(ENDLINE) */
+		break;
+
+	   default:	/* if not caught above, it's the same as Tk */
+		Tk(cmd);
+	}
+}
+
+
+
+
+/*
+ * ASCII
+ */
+
+static void
+ASCII(enum command cmd)
+{
+	int i;
+
+	switch (cmd) {
+	   case ENDDOC:
+		if (fHeadfoot)	{
+			printf("\n%s\n", HEADERANDFOOTER);
+			for (i=0; i<CRUFTS; i++) if (*cruft[i]) printf("%s\n",cruft[i]);
+		}
+		break;
+	   case CHARRQUOTE:
+	   case CHARLQUOTE:
+		putchar('"');
+		break;
+	   case CHARLSQUOTE:
+		putchar('`');
+		break;
+	   case CHARRSQUOTE:
+	   case CHARACUTE:
+		putchar('\'');
+		break;
+	   case CHARPERIOD:
+	   case CHARTAB:
+	   case CHARDASH:
+	   case CHARLT:
+	   case CHARAMP:
+	   case CHARBACKSLASH:
+	   case CHARGT:
+	   case CHARHAT:
+	   case CHARVBAR:
+	   case CHARNBSP:
+		putchar(cmd); break;
+	   case CHARDAGGER:	putchar('+'); break;
+	   case CHARBULLET:	putchar('*'); break;
+	   case CHARPLUSMINUS: printf("+-"); break;
+	   case CHANGEBAR:	putchar('|'); break;
+ 	   case CHARCENT:	putchar('c'); break;
+ 	   case CHARSECT:	putchar('S'); break;
+ 	   case CHARCOPYR:	printf("(C)"); break;
+ 	   case CHARNOT:	putchar('~'); break;
+ 	   case CHARREGTM:	printf("(R)"); break;
+ 	   case CHARDEG:	putchar('o'); break;
+ 	   case CHAR14:		printf("1/4"); break;
+ 	   case CHAR12:		printf("1/2"); break;
+ 	   case CHAR34:		printf("3/4"); break;
+ 	   case CHARMUL:	putchar('X'); break;
+ 	   case CHARDIV:	putchar('/'); break;
+	   case HR:		printf("\n%s\n", horizontalrule); break;
+
+	   case BEGINLINE:
+		for (i=0; i<ncnt; i++) putchar('\n');
+		break;
+	   case BEGINBODY:
+	   case SHORTLINE:
+		if (!fSource) break;
+	   case ENDLINE:
+		putchar('\n');
+		CurLine++;
+		break;
+
+	   case BEGINDOC:
+	   case ENDBODY:
+	   case BEGINHEADER: case ENDHEADER:
+	   case BEGINFOOTER: case ENDFOOTER:
+	   case BEGINSECTION: case ENDSECTION:
+	   case BEGINSECTHEAD: case ENDSECTHEAD:
+	   case BEGINSUBSECTHEAD: case ENDSUBSECTHEAD:
+	   case BEGINBULPAIR: case ENDBULPAIR:
+	   case BEGINBULLET: case ENDBULLET:
+	   case BEGINBULTXT: case ENDBULTXT:
+	   case BEGINSUBSECTION: case ENDSUBSECTION:
+
+	   case BEGINTABLE: case ENDTABLE:
+	   case BEGINTABLELINE: case ENDTABLELINE: case BEGINTABLEENTRY: case ENDTABLEENTRY:
+	   case BEGININDENT: case ENDINDENT:
+	   case FONTSIZE:
+	   case BEGINBOLD: case ENDBOLD:
+	   case BEGINCODE: case ENDCODE:
+	   case BEGINITALICS: case ENDITALICS:
+	   case BEGINMANREF: case ENDMANREF:
+	   case BEGINBOLDITALICS: case ENDBOLDITALICS:
+	   case BEGINY: case ENDY:
+	   case BEGINSC: case ENDSC:
+		/* nothing */
+		break;
+	   default:
+		DefaultLine(cmd);
+	}
+}
+
+
+
+/*
+ * Perl 5 pod ("plain old documentation")
+ */
+
+static void
+pod(enum command cmd)
+{
+	static int curindent=0;
+	int i;
+
+	if (hanging==-1) {
+		if (curindent) hanging=curindent; else hanging=5;
+	}
+
+
+	if (cmd==BEGINBULPAIR) {
+		/* want to have multiply indented text */
+		if (curindent && hanging!=curindent) printf("\n=back\n\n");
+		if (hanging!=curindent) printf("\n=over %d\n\n",hanging);
+		curindent=hanging;
+	} else if (cmd==ENDBULPAIR) {
+		/* nothing--wait until next command */
+	} else if (cmd==BEGINLINE && !scnt) {
+		if (curindent) printf("\n=back\n\n");
+		curindent=0;
+	} else if (cmd==BEGINBODY) {
+		if (curindent) {
+			printf("\n=back\n\n");
+			curindent=0;
+			auxindent=0;
+		}
+	}
+/*
+	   case BEGINBULPAIR:
+		printf("=over %d\n\n", hanging);
+		break;
+	   case ENDBULPAIR:
+		printf("\n=back\n\n");
+		break;
+*/
+	switch (cmd) {
+	   case BEGINDOC: I=0; break;
+
+	   case BEGINCOMMENT: fcharout=0; break;
+	   case ENDCOMMENT: fcharout=1; break;
+	   case COMMENTLINE: break;
+
+	   case CHARRQUOTE:
+	   case CHARLQUOTE:
+		putchar('"');
+		break;
+	   case CHARLSQUOTE:
+		putchar('`');
+		break;
+	   case CHARRSQUOTE:
+	   case CHARACUTE:
+		putchar('\'');
+		break;
+	   case CHARPERIOD:
+	   case CHARTAB:
+	   case CHARDASH:
+	   case CHARLT:
+	   case CHARAMP:
+	   case CHARBACKSLASH:
+	   case CHARGT:
+	   case CHARHAT:
+	   case CHARVBAR:
+	   case CHARNBSP:
+		putchar(cmd); break;
+	   case CHARDAGGER:	putchar('+'); break;
+	   case CHARPLUSMINUS: printf("+-"); break;
+	   case CHANGEBAR:	putchar('|'); break;
+ 	   case CHARCENT:	putchar('c'); break;
+ 	   case CHARSECT:	putchar('S'); break;
+ 	   case CHARCOPYR:	printf("(C)"); break;
+ 	   case CHARNOT:	putchar('~'); break;
+	   case CHARREGTM:	printf("(R)"); break;
+ 	   case CHARDEG:	putchar('o'); break;
+ 	   case CHAR14:	printf("1/4"); break;
+ 	   case CHAR12:	printf("1/2"); break;
+ 	   case CHAR34:	printf("3/4"); break;
+ 	   case CHARMUL:	putchar('X'); break;
+ 	   case CHARDIV:	putchar('/'); break;
+	   case HR:		printf("\n%s\n", horizontalrule); break;
+	   case CHARBULLET:	putchar('*'); break;
+
+	   case BEGINLINE:
+		for (i=0; i<ncnt; i++) putchar('\n');
+		CurLine+=ncnt;
+		break;
+	   case ENDLINE:
+		putchar('\n');
+		CurLine++;
+		I=0;
+		break;
+
+	   case BEGINSECTHEAD:	printf("=head1 "); break;
+	   case BEGINSUBSECTHEAD:	printf("=head2 "); break;
+
+	   case ENDSECTHEAD:
+	   case ENDSUBSECTHEAD:
+		printf("\n");
+		break;
+
+	   case BEGINCODE:
+	   case BEGINBOLD:		printf("B<"); break;
+	   case BEGINITALICS:	printf("I<"); break;
+	   case BEGINMANREF:	printf("L<"); break;
+
+	   case ENDBOLD:
+	   case ENDCODE:
+	   case ENDITALICS:
+	   case ENDMANREF:
+		printf(">");
+		break;
+
+	   case BEGINBULLET:
+		printf("\n=item ");
+		break;
+	   case ENDBULLET:
+		printf("\n\n");
+		fcharout=0;
+		break;
+	   case BEGINBULTXT:
+		fcharout=1;
+		auxindent=hanging;
+		break;
+	   case ENDBULTXT:
+		auxindent=0;
+		break;
+
+
+	   case ENDDOC:
+	   case BEGINBODY: case ENDBODY:
+	   case BEGINHEADER: case ENDHEADER:
+	   case BEGINFOOTER: case ENDFOOTER:
+	   case BEGINSECTION: case ENDSECTION:
+	   case BEGINSUBSECTION: case ENDSUBSECTION:
+	   case BEGINBULPAIR: case ENDBULPAIR:
+
+	   case SHORTLINE:
+	   case BEGINTABLE: case ENDTABLE:
+	   case BEGINTABLELINE: case ENDTABLELINE: case BEGINTABLEENTRY: case ENDTABLEENTRY:
+	   case BEGININDENT: case ENDINDENT:
+	   case FONTSIZE:
+	   case BEGINBOLDITALICS: case ENDBOLDITALICS:
+	   case BEGINY: case ENDY:
+	   case BEGINSC: case ENDSC:
+		/* nothing */
+		break;
+	   default:
+		DefaultLine(cmd);
+	}
+}
+
+
+
+static void
+Sections(enum command cmd)
+{
+
+	switch (cmd) {
+	   case ENDSECTHEAD:
+	   case ENDSUBSECTHEAD:
+		putchar('\n');
+	   case BEGINDOC:
+		fcharout=0;
+		break;
+
+	   case BEGINCOMMENT: fcharout=0; break;
+	   case ENDCOMMENT: fcharout=1; break;
+	   case COMMENTLINE: break;
+
+	   case BEGINSUBSECTHEAD:
+		printf("  ");
+		/* no break */
+	   case BEGINSECTHEAD:
+		fcharout=1;
+		break;
+	   case CHARRQUOTE:
+	   case CHARLQUOTE:
+		xputchar('"');
+		break;
+	   case CHARLSQUOTE:
+		xputchar('`');
+		break;
+	   case CHARRSQUOTE:
+	   case CHARACUTE:
+		xputchar('\'');
+		break;
+	   case BEGINTABLE: case ENDTABLE:
+	   case BEGINTABLELINE: case ENDTABLELINE: case BEGINTABLEENTRY: case ENDTABLEENTRY:
+	   case BEGININDENT: case ENDINDENT:
+	   case FONTSIZE:
+		break;
+	   case CHARPERIOD:
+	   case CHARTAB:
+	   case CHARDASH:
+	   case CHARBACKSLASH:
+	   case CHARLT:
+	   case CHARGT:
+	   case CHARHAT:
+	   case CHARVBAR:
+	   case CHARAMP:
+	   case CHARNBSP:
+		xputchar(cmd); break;
+	   case CHARDAGGER:	xputchar('+'); break;
+	   case CHARBULLET:	xputchar('*'); break;
+	   case CHARPLUSMINUS: xputchar('+'); xputchar('-'); break;
+ 	   case CHARCENT:	xputchar('c'); break;
+ 	   case CHARSECT:	xputchar('S'); break;
+	   case CHARCOPYR:	xputchar('('); xputchar('C'); xputchar(')'); break;
+ 	   case CHARNOT:	xputchar('~'); break;
+ 	   case CHARREGTM:	xputchar('('); xputchar('R'); xputchar(')'); break;
+ 	   case CHARDEG:	xputchar('o'); break;
+ 	   case CHAR14:	xputchar('1'); xputchar('/'); xputchar('4'); break;
+ 	   case CHAR12:	xputchar('1'); xputchar('/'); xputchar('2'); break;
+ 	   case CHAR34:	xputchar('3'); xputchar('/'); xputchar('4'); break;
+ 	   case CHARMUL:	xputchar('X'); break;
+ 	   case CHARDIV:	xputchar('/'); break;
+	   case ITAB:		DefaultLine(cmd); break;
+
+
+	   default:
+		/* nothing */
+		break;
+	}
+}
+
+
+
+static void
+Roff(enum command cmd)
+{
+	switch (cmd) {
+	   case BEGINDOC:
+		I=1;
+		printf(".TH %s %s \"generated by PolyglotMan\" UCB\n",manName,manSect);
+		printf(".\\\"  %s,\n",provenance);
+		printf(".\\\"  %s\n",anonftp);
+		CurLine=1;
+		break;
+	   case BEGINBODY:		printf(".LP\n"); break;
+
+	   case BEGINCOMMENT: 
+	   case ENDCOMMENT:
+		break;
+	   case COMMENTLINE: printf("'\\\" "); break;
+
+	   case BEGINSECTHEAD:	printf(".SH "); break;
+	   case BEGINSUBSECTHEAD:printf(".SS "); break;
+	   case BEGINBULPAIR:	printf(".IP "); break;
+	   case SHORTLINE:		printf("\n.br"); break;
+	   case BEGINBOLD:		printf("\\fB"); break;	/* \n.B -- grr! */
+	   case ENDCODE:
+	   case ENDBOLD:		printf("\\fR"); break;	/* putchar('\n'); */
+	   case BEGINITALICS:	printf("\\fI"); break;
+	   case ENDITALICS:		printf("\\fR"); break;
+	   case BEGINCODE:
+	   case BEGINBOLDITALICS:printf("\\f4"); break;
+	   case ENDBOLDITALICS:	printf("\\fR"); break;
+
+	   case CHARLQUOTE:		printf("\\*(rq"); break;
+	   case CHARRQUOTE:		printf("\\*(lq"); break;
+	   case CHARNBSP:		printf("\\|"); break;
+	   case CHARLSQUOTE:	putchar('`');	break;
+	   case CHARRSQUOTE:	putchar('\'');	break;
+	   case CHARPERIOD:		if (I==1) printf("\\&"); putchar('.'); I++; break;
+	   case CHARDASH:		printf("\\-"); break;
+	   case CHARTAB:
+	   case CHARLT:
+	   case CHARGT:
+	   case CHARHAT:
+	   case CHARVBAR:
+	   case CHARAMP:
+		putchar(cmd); break;
+	   case CHARBULLET:		printf("\\(bu"); break;
+	   case CHARDAGGER:		printf("\\(dg"); break;
+	   case CHARPLUSMINUS:	printf("\\(+-"); break;
+	   case CHANGEBAR:		putchar('|'); break;
+ 	   case CHARCENT:		printf("\\(ct"); break;
+ 	   case CHARSECT:		printf("\\(sc"); break;
+ 	   case CHARCOPYR:		printf("\\(co"); break;
+ 	   case CHARNOT:		printf("\\(no"); break;
+ 	   case CHARREGTM:		printf("\\(rg"); break;
+ 	   case CHARDEG:		printf("\\(de"); break;
+ 	   case CHARACUTE:		printf("\\(aa"); break;
+ 	   case CHAR14:		printf("\\(14"); break;
+ 	   case CHAR12:		printf("\\(12"); break;
+ 	   case CHAR34:		printf("\\(34"); break;
+ 	   case CHARMUL:		printf("\\(mu"); break;
+ 	   case CHARDIV:		printf("\\(di"); break;
+	   case HR:			/*printf("\n%s\n", horizontalrule);*/ break;
+	   case CHARBACKSLASH:	printf("\\\\"); break;  /* correct? */
+
+	   case BEGINLINE:
+		/*for (i=0; i<ncnt; i++) putchar('\n');*/
+		break;
+
+	   case BEGINBULLET:	putchar('"'); break;
+	   case ENDBULLET:		printf("\"\n"); break;
+
+	   case ENDLINE:
+		CurLine++;
+		I=1;
+		/* no break */
+	   case ENDSUBSECTHEAD:
+	   case ENDSECTHEAD:
+	   case ENDDOC:
+		putchar('\n');
+		break;
+
+	   case BEGINCODEBLOCK:	printf(".nf\n");
+	   case ENDCODEBLOCK:	printf(".fi\n");
+
+	   case ENDBODY:
+	   case ENDBULPAIR:
+	   case BEGINBULTXT: case ENDBULTXT:
+	   case BEGINSECTION: case ENDSECTION:
+	   case BEGINSUBSECTION: case ENDSUBSECTION:
+	   case BEGINY: case ENDY:
+	   case BEGINSC: case ENDSC:
+	   case BEGINTABLE: case ENDTABLE:
+	   case BEGINTABLELINE: case ENDTABLELINE: case BEGINTABLEENTRY: case ENDTABLEENTRY:
+	   case BEGININDENT: case ENDINDENT:
+	   case FONTSIZE:
+	   case BEGINHEADER: case ENDHEADER:
+	   case BEGINFOOTER: case ENDFOOTER:
+	   case BEGINMANREF: case ENDMANREF:
+		/* nothing */
+		break;
+	   default:
+		DefaultPara(cmd);
+	}
+}
+
+
+/* can't test Ensemble, probably not used, so why take up space? */
+#if 0
+
+/*
+ * Ensemble
+ */
+
+static void
+EnsembleDumpTags(void)
+{
+	int i,tag;
+	int fI=0, fB=0;
+
+	if (!tagc) return;
+
+	/* { */
+	printf("}{}{");		/* } header */
+
+	/* italics */
+	for (i=0; i<tagc; i++) {
+		tag = tags[i].type;
+		if (tag==ITALICS||tag==BOLDITALICS) {
+			if (!fI) {printf("ITALIC=("); fI=1;}
+			printf("(%d,%d,[T])", tags[i].first, tags[i].last);
+		}
+	}
+	if (fI) printf(")");
+
+	/* bold */
+	for (i=0; i<tagc; i++) {
+		tag = tags[i].type;
+		if (tag==BOLD||tag==BOLDITALICS) {
+			if (!fB) {printf(",BOLD=("); fB=1;}
+			printf("(%d,%d,[T])", tags[i].first, tags[i].last);
+		}
+	}
+	if (fB) printf(")");
+
+	/* man ref? */
+/*
+	for (i=0; i<tagc; i++) {
+		tag = tags[i].type;
+		if (tag==MANREF) {
+			if (!fH) {printf(",HYPER=("); fH=1;}
+			printf("(%d,%d,[???])", tags[i].first, tags[i].last);
+		}
+	}
+	if (fH) printf(")");
+*/
+
+	/* { */
+	/* don't put  printf("}");  here as trailer -- controlling function expects to close it */
+
+	tagc=0;
+}
+
+static void
+Ensemble(enum command cmd)
+{
+
+	switch (cmd) {
+	   case BEGINDOC:
+		I=0;
+		printf("DOCUMENT MANPAGE\n<MANPAGE>\n");
+		escchars = "{}\\";
+		break;
+	   case ENDDOC:	printf("</MANPAGE>\n"); break;
+
+	   case BEGINCOMMENT: printf("\n<!--\n"); break;
+	   case ENDCOMMENT: printf("\n-->\n"); break;
+	   case COMMENTLINE: break;
+
+	   case BEGINBODY:
+		printf("<SUBSECTIONBODY><BODY>{");
+		break;
+	   case ENDBODY:
+		CurLine++;
+		EnsembleDumpTags(); printf("}</BODY></SUBSECTIONBODY>\n");
+		tagc=0;
+		break;
+	   case BEGINSECTION:		printf("<SECTION>"); break;
+	   case ENDSECTION:			printf("</SECTION>\n"); break;
+	   case BEGINSECTHEAD:		printf("<SECTHEAD>{"); break;
+	   case ENDSECTHEAD:		tagc=0; I=0; printf("}</SECTHEAD>\n"); break;
+	   case BEGINSUBSECTHEAD:	printf("<SUBSECTHEAD>{"); break;
+	   case ENDSUBSECTHEAD:		tagc=0; I=0; printf("}</SUBSECTHEAD>\n"); break;
+	   case BEGINBULPAIR:
+		printf("<SUBSECTIONBODY><LISTELEMENT>");
+		break;
+	   case ENDBULPAIR:
+		printf("</LISTELEMENT></SUBSECTIONBODY>\n");
+		break;
+	   case BEGINBULLET:		printf("<BULLET>{"); break;
+	   case ENDBULLET:			tagc=0; I=0; printf("}</BULLET>"); break;
+	   case BEGINBULTXT:		printf("<BULLETTEXT>{"); break;
+	   case ENDBULTXT:
+		EnsembleDumpTags();
+		CurLine++;
+		printf("}</BULLETTEXT>");
+		break;
+	   case BEGINSUBSECTION:	printf("<SUBSECTIONBODY><SUBSECTION>\n"); break;
+	   case ENDSUBSECTION:	printf("</SUBSECTION></SUBSECTIONBODY>\n"); break;
+	   case SHORTLINE:		/*poppush(prevcmd);*/ break;
+
+
+	   case CHARRQUOTE:
+	   case CHARLQUOTE:
+		putchar('"'); I++;
+		break;
+	   case CHARLSQUOTE:
+		putchar('\'');
+		break;
+	   case CHARRSQUOTE:
+		putchar('\'');
+		break;
+	   case CHARPERIOD:
+	   case CHARTAB:
+	   case CHARDASH:
+	   case CHARBACKSLASH:
+	   case CHARLT:
+	   case CHARGT:
+	   case CHARHAT:
+	   case CHARVBAR:
+	   case CHARAMP:
+	   case CHARBULLET:
+	   case CHARDAGGER:
+	   case CHARPLUSMINUS:
+	   case CHARNBSP:
+ 	   case CHARCENT:
+ 	   case CHARSECT:
+ 	   case CHARCOPYR:
+ 	   case CHARNOT:
+ 	   case CHARREGTM:
+ 	   case CHARDEG:
+ 	   case CHARACUTE:
+ 	   case CHAR14:
+ 	   case CHAR12:
+ 	   case CHAR34:
+ 	   case CHARMUL:
+ 	   case CHARDIV:
+		putchar(cmd); I++; break;
+
+	   case ENDLINE:		putchar(' '); I++; break;
+	   case HR:			/*printf("\n%s\n", horizontalrule);*/ break;
+	   case CHANGEBAR:
+		/* maybe something later */
+	   case BEGINLINE: 
+	   case BEGINY: case ENDY:
+	   case BEGINHEADER: case ENDHEADER:
+	   case BEGINFOOTER: case ENDFOOTER:
+	   case BEGINBOLD: case ENDBOLD:
+	   case BEGINCODE: case ENDCODE:
+	   case BEGINITALICS: case ENDITALICS:
+	   case BEGINBOLDITALICS: case ENDBOLDITALICS:
+	   case BEGINSC: case ENDSC:
+	   case BEGINTABLE: case ENDTABLE:
+	   case BEGINTABLELINE: case ENDTABLELINE: case BEGINTABLEENTRY: case ENDTABLEENTRY:
+	   case BEGININDENT: case ENDINDENT:
+	   case FONTSIZE:
+
+	   case BEGINMANREF:
+	   case ENDMANREF:
+		/* easy strike for hypertext--want to dynamically generate, though */
+		/* nothing */
+		break;
+	   default:
+		DefaultPara(cmd);
+	}
+}
+#endif
+
+
+
+/*
+ * HTML
+ */
+
+static void
+HTML(enum command cmd)
+{
+	static int pre=0;
+	int i;
+	int lasttoc;
+
+	/* always respond to these signals */
+	switch (cmd) {
+	   case CHARNBSP:		printf("&nbsp;"); I++; break;
+	   case CHARTAB:		printf("<tt>&#32;</tt>&nbsp;<tt>&#32;</tt>&nbsp;");		break;
+	  /* old browsers--who uses these?--don't understand symbolic codes */
+	  /*
+ 	   case CHARNBSP:		printf("&#160;"); I++; break;
+ 	   case CHARLQUOTE:		printf("&#171;"); break;
+ 	   case CHARRQUOTE:		printf("&#187;"); break;
+ 	   case CHARTAB:		printf("<tt> </tt>&#160;<tt> </tt>&#160;"); break;
+		*/
+	   case CHARLQUOTE:
+	   case CHARRQUOTE:
+	   case CHARLSQUOTE:
+	   case CHARRSQUOTE:
+	   case CHARPERIOD:
+	   case CHARDASH:
+	   case CHARBACKSLASH:
+	   case CHARVBAR:		/*printf("&brvbar;"); -- broken bar no good */
+	   case CHARHAT:
+		putchar(cmd); break;
+	   case CHARDAGGER:		printf("*"); break;
+	   case CHARBULLET:		printf("&#183;"/*"&middot;"*//*&sect;--middot hardly visible*/); break;
+	   case CHARPLUSMINUS:	printf("&#177;"/*"&plusmn;"*/); break;
+	   case CHARGT:		printf("&gt;"); break;
+	   case CHARLT:		printf("&lt;"); break;
+	   case CHARAMP:		printf("&amp;"); break;
+	   case CHARCENT:		printf("&#162;"); break;	/* translate these to symbolic forms, sometime */
+ 	   case CHARSECT:		printf("&#167;"); break;
+ 	   case CHARCOPYR:		printf("&#169;"); break;
+ 	   case CHARNOT:		printf("&#172;"); break;
+ 	   case CHARREGTM:		printf("&#174;"); break;
+ 	   case CHARDEG:		printf("&#176;"); break;
+ 	   case CHARACUTE:		printf("&#180;"); break;
+ 	   case CHAR14:		printf("&#188;"); break;
+ 	   case CHAR12:		printf("&#189;"); break;
+ 	   case CHAR34:		printf("&#190;"); break;
+ 	   case CHARMUL:		printf("&#215;"); break;
+ 	   case CHARDIV:		printf("&#247;"); break;
+	   default:
+		break;
+	}
+
+	/* while in pre mode... */
+	if (pre) {
+		switch (cmd) {
+		   case ENDLINE:	I=0; CurLine++; if (!fPara && scnt) printf("<BR>"); printf("\n"); break;
+		   case ENDTABLE:
+			if (fSource) {
+			  printf("</TABLE>\n");
+			} else {
+			  printf("</PRE><BR>\n"); pre=0; fQS=fIQS=fPara=1;
+			}
+			break;
+		   case ENDCODEBLOCK: printf("</PRE>"); pre=0; break;
+		   case SHORTLINE:
+		   case ENDBODY:
+			printf("\n");
+			break;
+	   	   case BEGINBOLD:		printf("<B>"); break;
+	   	   case ENDBOLD:		printf("</B>"); break;
+	   	   case BEGINITALICS:	printf("<I>"); break;
+	   	   case ENDITALICS:		printf("</I>"); break;
+		   default:
+			/* nothing */
+			break;
+		}
+		return;
+	}
+
+	/* usual operation */
+	switch (cmd) {
+	   case BEGINDOC:
+		/* escchars = ...  => HTML doesn't backslash-quote metacharacters */
+		printf("<!-- %s, -->\n",provenance);
+		printf("<!-- %s -->\n\n",anonftp);
+		printf("<HTML>\n<HEAD>\n");
+/*		printf("<ISINDEX>\n");*/
+		/* better title possible? */
+		printf("<TITLE>"); printf(manTitle, manName, manSect); printf("</TITLE>\n");
+#ifdef XFree86
+		printf("</HEAD>\n<BODY BGCOLOR=\"#efefef\" TEXT=\"black\" "
+			"LINK=\"blue\" VLINK=\"#551A8B\" ALINK=\"red\">\n");
+#else
+		printf("</HEAD>\n<BODY bgcolor=white>\n");
+#endif
+		printf("<A HREF=\"#toc\">%s</A><P>\n", TABLEOFCONTENTS);
+		I=0;
+		break;
+	   case ENDDOC:
+		/* header and footer wanted? */
+		printf("<P>\n");
+		if (fHeadfoot) {
+			printf("<HR><H2>%s</H2>\n", HEADERANDFOOTER);
+			for (i=0; i<CRUFTS; i++) if (*cruft[i]) printf("%s<BR>\n",cruft[i]);
+		}
+
+		if (!tocc) {
+			/*printf("\n<H1>ERROR: Empty man page</H1>\n");*/
+		} else {
+			printf("\n<HR><P>\n");
+			printf("<A NAME=\"toc\"><B>%s</B></A><P>\n", TABLEOFCONTENTS);
+			printf("<UL>\n");
+			for (i=0, lasttoc=BEGINSECTION; i<tocc; lasttoc=toc[i].type, i++) {
+				if (lasttoc!=toc[i].type) {
+					if (toc[i].type==BEGINSUBSECTION) printf("<UL>\n");
+					else printf("</UL>\n");
+				}
+				printf("<LI><A NAME=\"toc%d\" HREF=\"#sect%d\">%s</A></LI>\n", i, i, toc[i].text);
+			}
+			if (lasttoc==BEGINSUBSECTION) printf("</UL>");
+			printf("</UL>\n");
+		}
+		printf("</BODY></HTML>\n");
+		break;
+	   case BEGINBODY:
+		printf("<P>\n");
+		break;
+	   case ENDBODY:		break;
+
+	   case BEGINCOMMENT: printf("\n<!--\n"); break;
+	   case ENDCOMMENT: printf("\n-->\n"); break;
+	   case COMMENTLINE: break;
+
+	   case BEGINSECTHEAD:
+		printf("\n<H2><A NAME=\"sect%d\" HREF=\"#toc%d\">", tocc, tocc);
+		break;
+	   case ENDSECTHEAD:
+		printf("</A></H2>\n");
+		/* useful extraction from FILES, ENVIRONMENT? */
+		break;
+	   case BEGINSUBSECTHEAD:
+		printf("\n<H3><A NAME=\"sect%d\" HREF=\"#toc%d\">", tocc, tocc);
+		break;
+	   case ENDSUBSECTHEAD:
+		printf("</A></H3>\n");
+		break;
+	   case BEGINSECTION:	break;
+	   case ENDSECTION:
+		if (sectheadid==NAME && message!=NULL) printf(message);
+		break;
+	   case BEGINSUBSECTION:	break;
+	   case ENDSUBSECTION:	break;
+
+	   case BEGINBULPAIR:
+		if (listtype==OL) printf("\n<OL>\n");
+		else if (listtype==UL) printf("\n<UL>\n");
+		else printf("\n<DL>\n");
+		break;
+	   case ENDBULPAIR:
+		if (listtype==OL) printf("\n</OL>\n");
+		else if (listtype==UL) printf("\n</UL>\n");
+		else printf("</DL>\n");
+		break;
+	   case BEGINBULLET:
+		if (listtype==OL || listtype==UL) fcharout=0;
+		else printf("\n<DT>");
+		break;
+	   case ENDBULLET:
+		if (listtype==OL || listtype==UL) fcharout=1;
+		else printf("</DT>");
+		break;
+	   case BEGINBULTXT:
+		if (listtype==OL || listtype==UL) printf("<LI>");
+		else printf("\n<DD>");
+		break;
+	   case ENDBULTXT:
+		if (listtype==OL || listtype==UL) printf("</LI>");
+		else printf("</DD>\n");
+		break;
+
+	   case BEGINLINE:
+		/*		if (ncnt) printf("<P>\n"); -- if haven't already generated structural tag */
+		if (ncnt) printf("<P>\n");
+
+		/* trailing spaces already trimmed off, so look for eol now */
+		if (fCodeline) {
+		  printf("<CODE>");
+		  for (i=0; i<scnt-indent; i++) printf("&nbsp;"/*&#160;*/);		/* ? */
+		  tagc=0;
+
+		  /* already have .tag=BOLDITALICS, .first=0 */
+		  /* would be more elegant, but can't print initial spaces before first tag
+		  tags[0].last = linelen;
+		  tagc=1;
+		  fIQS=0;
+		  */
+		}
+
+		break;
+
+	   case ENDLINE:
+		/*if (fCodeline) { fIQS=1; fCodeline=0; }*/
+		if (fCodeline) { printf("</CODE><BR>"); fCodeline=0; }
+		I=0; CurLine++; if (!fPara && scnt) printf("<BR>"); printf("\n");
+		break;
+
+	   case SHORTLINE:
+		if (fCodeline) { printf("</CODE>"); fCodeline=0; }
+		if (!fIP) printf("<BR>\n");
+		break;
+
+
+	   case BEGINTABLE:
+		if (fSource) {
+		  /*printf("<CENTER><TABLE BORDER>\n");*/
+		  printf("<TABLE BORDER=0>\n");
+		} else {
+		  printf("<BR><PRE>\n"); pre=1; fQS=fIQS=fPara=0;
+		}
+		break;
+	   case ENDTABLE:
+		if (fSource) {
+		  printf("</TABLE>\n");
+		} else {
+		  printf("</PRE><BR>\n"); pre=0; fQS=fIQS=fPara=1;
+		}
+		break;
+	   case BEGINTABLELINE:		printf("<TR>"); break;
+	   case ENDTABLELINE:		printf("</TR>\n"); break;
+	   case BEGINTABLEENTRY:
+		printf("<TD ALIGN=");
+		switch (tblcellformat[0]) {
+		case 'c':	printf("CENTER"); break;
+		case 'n':	/*printf("DECIMAL"); break;  -- fall through to right for now */
+		case 'r':	printf("RIGHT"); break;
+		case 'l':
+		default:
+		  printf("LEFT");
+		}
+		if (tblcellspan>1) printf(" COLSPAN=%d", tblcellspan);
+		printf(">");
+		break;
+	   case ENDTABLEENTRY:
+		printf("</TD>");
+		break;
+
+	   /* something better with CSS */
+	   case BEGININDENT:		printf("<blockquote>"); break;
+	   case ENDINDENT:			printf("</blockquote>\n"); break;
+
+	   case FONTSIZE:
+		/* HTML font step sizes are bigger than troff's */
+		if ((fontdelta+=intArg)!=0) printf("<FONT SIZE=%c1>", (intArg>0)?'+':'-'); else printf("</FONT>\n");
+		break;
+
+	   case BEGINBOLD:		printf("<B>"); break;
+	   case ENDBOLD:		printf("</B>"); break;
+	   case BEGINITALICS:	printf("<I>"); break;
+	   case ENDITALICS:		printf("</I>"); break;
+	   case BEGINBOLDITALICS:
+	   case BEGINCODE:		printf("<CODE>"); break;
+	   case ENDBOLDITALICS:
+	   case ENDCODE:		printf("</CODE>"); break;
+	   case BEGINCODEBLOCK:	printf("<PRE>"); pre=1; break;	/* wrong for two-column lists in kermit.1, pine.1, perl4.1 */
+	   case ENDCODEBLOCK:	printf("</PRE>"); pre=0; break;
+	   case BEGINCENTER:		printf("<CENTER>"); break;
+	   case ENDCENTER:		printf("</CENTER>"); break;
+	   case BEGINMANREF:
+		manrefextract(hitxt);
+		if (fmanRef) { printf("<A HREF=\""); printf(href, manrefname, manrefsect); printf("\">"); }
+		else printf("<I>");
+		break;
+	   case ENDMANREF:
+		if (fmanRef) printf("</A>\n"); else printf("</I>");
+		break;
+	   case HR:		printf("\n<HR>\n"); break;
+
+		/* U (was B, I), strike -- all temporary until HTML 4.0's INS and DEL widespread */
+	   case BEGINDIFFA: printf("<INS><U>"); break;
+	   case ENDDIFFA: printf("</U></INS>"); break;
+	   case BEGINDIFFD: printf("<DEL><strike>"); break;
+	   case ENDDIFFD: printf("</strike></DEL>"); break;
+
+	   case BEGINSC: case ENDSC:
+	   case BEGINY: case ENDY:
+	   case BEGINHEADER: case ENDHEADER:
+	   case BEGINFOOTER: case ENDFOOTER:
+	   case CHANGEBAR:
+		/* nothing */
+		break;
+	   default:
+		DefaultPara(cmd);
+	}
+}
+
+
+
+/*
+ * SGML
+ */
+
+/* same as HTML but just has man page-specific DTD */
+/* follows the Davenport DocBook DTD v2.3, availble from ftp.ora.com */
+
+/*char *docbookpath = "docbook.dtd";*/
+
+static void
+SGML(enum command cmd)
+{
+	static int pre=0;
+	int i;
+	char *p;
+	static int fRefEntry=0;
+	static int fRefPurpose=0;
+	/*static char *bads => SGML doesn't backslash-quote metacharacters */
+
+/*
+	fprintf(stderr,
+		   "The framework for SGML is in place but not done.  If you\n"
+		   "are familiar with the DocBook DTD, however, it shouldn't be\n"
+ 		   "too difficult to finish it.  If you do so, please send your\n"
+		   "code to me so that I may share the wealth in the next release.\n"
+		   );
+	exit(1);
+*/
+
+	/* always respond to these signals */
+	switch (cmd) {
+	   case CHARLQUOTE:
+	   case CHARRQUOTE:
+		printf("&quot;");
+		break;
+	   case CHARLSQUOTE:
+	   case CHARRSQUOTE:
+	   case CHARPERIOD:
+	   case CHARTAB:
+	   case CHARHAT:
+	   case CHARVBAR:
+	   case CHARBULLET:
+	   case CHARDAGGER:
+	   case CHARPLUSMINUS:
+	   case CHARNBSP:
+ 	   case CHARCENT:
+ 	   case CHARSECT:
+ 	   case CHARCOPYR:
+ 	   case CHARNOT:
+ 	   case CHARREGTM:
+ 	   case CHARDEG:
+ 	   case CHARACUTE:
+ 	   case CHAR14:
+ 	   case CHAR12:
+ 	   case CHAR34:
+ 	   case CHARMUL:
+ 	   case CHARDIV:
+		putchar(cmd); break;
+	   case CHARAMP:		printf("&amp;"); break;
+	   case CHARDASH:
+		if (sectheadid==NAME && !fRefPurpose) {
+			printf("</RefEntry><RefPurpose>");
+			fRefPurpose=1;
+		} else putchar('-');
+		break;
+	   case CHARBACKSLASH:	putchar('\\'); break;
+	   case CHARGT:		printf("&gt;"); break;
+	   case CHARLT:		printf("&lt;"); break;
+	   default:
+		break;
+	}
+
+	/* while in pre mode... */
+	if (pre) {
+		switch (cmd) {
+		   case ENDLINE:	I=0; CurLine++; if (!fPara && scnt) printf("<BR>"); printf("\n"); break;
+		   case ENDTABLE:
+			if (fSource) {
+			  printf("</TABLE>\n");
+			} else {
+			  printf("</PRE><BR>\n"); pre=0; fQS=fIQS=fPara=1;
+			}
+			break;
+		   default:
+			/* nothing */
+			break;
+		}
+		return;
+	}
+
+	/* usual operation */
+	switch (cmd) {
+	   case BEGINDOC:
+		/*printf("<!DOCTYPE chapter SYSTEM \"%s\">\n", docbookpath);*/
+		printf("<!--\n\n\tI am looking for help to finish SGML.\n\n-->\n");
+
+		printf("<!-- %s\n",provenance);
+		printf("     %s -->\n\n",anonftp);
+		/* better title possible? */
+		printf("<RefEntry ID=%s.%s>\n", manName, manSect);
+		printf("<RefMeta><RefEntryTitle>%s</RefEntryTitle>", manName);
+		printf("<ManVolNum>%s</ManVolNum></RefMeta>\n\n", manSect);
+
+		I=0;
+		break;
+
+	   case ENDDOC:
+		/* header and footer wanted? */
+		if (fHeadfoot) {
+			printf("<RefSect1><Title>%s</Title>\n", HEADERANDFOOTER);
+			for (i=0; i<CRUFTS; i++) if (*cruft[i]) printf("<Para>%s\n",cruft[i]);
+			printf("</RefSect1>");
+		}
+
+		/* table of contents, such as found in HTML, can be generated automatically by SGML software */
+
+		printf("</RefEntry>\n");
+		break;
+	   case BEGINBODY:		printf("\n\n<Para>"); break;
+	   case ENDBODY:		break;
+
+	   case BEGINCOMMENT: printf("\n<!--\n"); break;
+	   case ENDCOMMENT: printf("\n-->\n"); break;
+	   case COMMENTLINE: break;
+
+	   case BEGINSECTHEAD:
+	   case BEGINSUBSECTHEAD:
+		printf("<Title>");
+		break;
+	   case ENDSECTHEAD:
+	   case ENDSUBSECTHEAD:
+		printf("</Title>");
+		break;
+
+	   case BEGINSECTION:
+		if (sectheadid==NAME) printf("<RefNameDiv>");
+			/*printf("<RefEntry>");  -- do lotsa parsing here for RefName, RefPurpose*/
+		else if (sectheadid==SYNOPSIS) printf("<RefSynopsisDiv>");
+		else printf("<RefSect1>");
+		break;
+	   case ENDSECTION:
+		if (oldsectheadid==NAME) printf("</RefNameDiv>\n\n");
+		else if (oldsectheadid==SYNOPSIS) printf("</RefSynopsisDiv>\n\n");
+		else printf("</RefSect1>\n\n");
+		break;
+
+	   case BEGINSUBSECTION:	printf("<RefSect2>"); break;
+	   case ENDSUBSECTION:	printf("</RefSect2>"); break;
+
+	   /* need to update this for enumerated and plain lists */
+	   case BEGINBULPAIR:	printf("<ItemizedList MARK=Bullet>\n"); break;
+	   case ENDBULPAIR:		printf("</ItemizedList>\n"); break;
+	   case BEGINBULLET:	printf("<Term>"); break;
+	   case ENDBULLET:		printf("</Term>"); break;
+	   case BEGINBULTXT:	printf("<ListItem><Para>"); break;
+	   case ENDBULTXT:		printf("</Para></ListItem>\n"); break;
+
+	   case BEGINLINE:
+		/* remember, get BEGINBODY call at start of paragraph */
+		if (fRefEntry) {
+			if (fRefPurpose) {
+				for (p=plain; *p!='-'; p++) {
+					/* nothing?! */
+				}
+			}
+		}
+
+		break;
+
+	   case ENDLINE:
+		/*if (fCodeline) { fIQS=1; fCodeline=0; }*/
+		if (fCodeline) { printf("</CODE><BR>"); fCodeline=0; }
+		I=0; CurLine++; if (!fPara && scnt) printf("<BR>"); printf("\n");
+		break;
+
+	   case SHORTLINE:
+		if (fCodeline) { printf("</CODE>"); fCodeline=0; }
+		if (!fIP) printf("<BR>\n");
+		break;
+
+	   case BEGINTABLE:
+		if (fSource) {
+		  printf("<TABLE>\n");
+		} else {
+		  printf("<BR><PRE>\n"); pre=1; fQS=fIQS=fPara=0;
+		}
+		break;
+	   case ENDTABLE:
+		if (fSource) {
+		  printf("</TABLE>\n");
+		} else {
+		  printf("</PRE><BR>\n"); pre=0; fQS=fIQS=fPara=1;
+		}
+		break;
+	   case BEGINTABLELINE:		printf("<TR>"); break;
+	   case ENDTABLELINE:		printf("</TR>\n"); break;
+	   case BEGINTABLEENTRY:		printf("<TD>"); break;
+	   case ENDTABLEENTRY:		printf("</TD>"); break;
+
+	   case BEGININDENT: case ENDINDENT:
+	   case FONTSIZE:
+		break;
+
+	   /* have to make some guess about bold and italics */
+	   case BEGINBOLD:		printf("<B>"); break;
+	   case ENDBOLD:		printf("</B>"); break;
+	   case BEGINITALICS:	printf("<I>"); break;
+	   case ENDITALICS:		printf("</I>"); break;
+	   case BEGINBOLDITALICS:
+	   case BEGINCODE:		printf("<CODE>"); break;
+	   case ENDBOLDITALICS:	
+	   case ENDCODE:		printf("</CODE>"); break;
+	   case BEGINMANREF:
+/*
+		manrefextract(hitxt);
+		if (fmanRef) { printf("<LINK LINKEND=\""); printf(href, manrefname, manrefsect); printf("\">\n"); }
+		else printf("<I>");
+		break;
+*/
+		printf("<Command>");
+		break;
+	   case ENDMANREF:
+/*		if (fmanRef) printf("</LINK>"); else printf("</I>");*/
+		printf("</Command>");
+		break;
+
+	   case HR:		/*printf("\n<HR>\n", horizontalrule);*/ break;
+
+	   case BEGINSC: case ENDSC:
+	   case BEGINY: case ENDY:
+	   case BEGINHEADER: case ENDHEADER:
+	   case BEGINFOOTER: case ENDFOOTER:
+	   case CHANGEBAR:
+		/* nothing */
+		break;
+	   default:
+		DefaultPara(cmd);
+	}
+}
+
+#if 0
+/*
+ * GNU Texinfo -- somebody should finish this up
+ */
+
+static void
+Texinfo(enum command cmd)
+{
+	static int pre=0;
+	int i;
+
+	/* always respond to these signals */
+	switch (cmd) {
+	   case CHARNBSP:		printf("&nbsp;"); I++; break;
+	   case CHARLQUOTE:		printf("&laquo;");		break;
+	   case CHARRQUOTE:		printf("&raquo;");		break;
+	   case CHARTAB:		printf("<tt> </tt>&nbsp;<tt> </tt>&nbsp;");		break;
+	  /* old browsers--who uses these?--don't understand symbolic codes */
+	  /*
+ 	   case CHARNBSP:		printf("&#160;"); I++; break;
+ 	   case CHARLQUOTE:		printf("&#171;"); break;
+ 	   case CHARRQUOTE:		printf("&#187;"); break;
+ 	   case CHARTAB:		printf("<tt> </tt>&#160;<tt> </tt>&#160;"); break;
+		*/
+	   case CHARLSQUOTE:
+	   case CHARRSQUOTE:
+	   case CHARPERIOD:
+	   case CHARDASH:
+	   case CHARBACKSLASH:
+	   case CHARVBAR:		/*printf("&brvbar;"); -- broken bar no good */
+	   case CHARHAT:
+		putchar(cmd); break;
+	   case CHARDAGGER:		printf("*"); break;
+	   case CHARBULLET:		printf("&#183;"/*"&middot;"*//*&sect;--middot hardly visible*/); break;
+	   case CHARPLUSMINUS:	printf("&#177;"/*"&plusmn;"*/); break;
+	   case CHARGT:		printf("&gt;"); break;
+	   case CHARLT:		printf("&lt;"); break;
+	   case CHARAMP:		printf("&amp;"); break;
+	   case CHARCENT:		printf("&#162;"); break;	/* translate these to symbolic forms, sometime */
+ 	   case CHARSECT:		printf("&#167;"); break;
+ 	   case CHARCOPYR:		printf("&#169;"); break;
+ 	   case CHARNOT:		printf("&#172;"); break;
+ 	   case CHARREGTM:		printf("&#174;"); break;
+ 	   case CHARDEG:		printf("&#176;"); break;
+ 	   case CHARACUTE:		printf("&#180;"); break;
+ 	   case CHAR14:		printf("&#188;"); break;
+ 	   case CHAR12:		printf("&#189;"); break;
+ 	   case CHAR34:		printf("&#190;"); break;
+ 	   case CHARMUL:		printf("&#215;"); break;
+ 	   case CHARDIV:		printf("&#247;"); break;
+	   default:
+		break;
+	}
+
+	/* while in pre mode... */
+	if (pre) {
+		switch (cmd) {
+		   case ENDLINE:	I=0; CurLine++; if (!fPara && scnt) printf("<BR>"); printf("\n"); break;
+		   case ENDTABLE:
+			if (fSource) {
+			  printf("</TABLE>\n");
+			} else {
+			  printf("</PRE><BR>\n"); pre=0; fQS=fIQS=fPara=1;
+			}
+			break;
+		   default:
+			/* nothing */
+			break;
+		}
+		return;
+	}
+
+	/* usual operation */
+	switch (cmd) {
+	   case BEGINDOC:
+		/* escchars = ...  => HTML doesn't backslash-quote metacharacters */
+		printf("<!-- %s, -->\n",provenance);
+		printf("<!-- %s -->\n\n",anonftp);
+		printf("<HTML>\n<HEAD>\n");
+/*		printf("<ISINDEX>\n");*/
+		/* better title possible? */
+		printf("<TITLE>"); printf(manTitle, manName, manSect); printf("</TITLE>\n");
+		printf("</HEAD>\n<BODY>\n");
+		printf("<A HREF=\"#toc\">%s</A><P>\n", TABLEOFCONTENTS);
+		I=0;
+		break;
+	   case ENDDOC:
+		/* header and footer wanted? */
+		printf("<P>\n");
+		if (fHeadfoot) {
+			printf("<HR><H2>%s</H2>\n", HEADERANDFOOTER);
+			for (i=0; i<CRUFTS; i++) if (*cruft[i]) printf("%s<BR>\n",cruft[i]);
+		}
+
+		if (!tocc) {
+			/*printf("\n<H1>ERROR: Empty man page</H1>\n");*/
+		} else {
+			printf("\n<HR><P>\n");
+			printf("<A NAME=\"toc\"><B>%s</B></A><P>\n", TABLEOFCONTENTS);
+			printf("<UL>\n");
+			for (i=0, lasttoc=BEGINSECTION; i<tocc; lasttoc=toc[i].type, i++) {
+				if (lasttoc!=toc[i].type) {
+					if (toc[i].type==BEGINSUBSECTION) printf("<UL>\n");
+					else printf("</UL>\n");
+				}
+				printf("<LI><A NAME=\"toc%d\" HREF=\"#sect%d\">%s</A></LI>\n", i, i, toc[i].text);
+			}
+			if (lasttoc==BEGINSUBSECTION) printf("</UL>");
+			printf("</UL>\n");
+		}
+		printf("</BODY></HTML>\n");
+		break;
+	   case BEGINBODY:
+		printf("<P>\n");
+		break;
+	   case ENDBODY:		break;
+
+	   case BEGINCOMMENT:
+	   case ENDCOMMENT:
+		break;
+	   case COMMENTLINE: printf("@c ");	break;
+
+	   case BEGINSECTHEAD:
+		break;
+	   case ENDSECTHEAD:
+		printf("\n@node %s\n", toc[tocc].text);
+		printf("\n@section %s\n\n", toc[tocc].text);
+		/* useful extraction from FILES, ENVIRONMENT? */
+		break;
+	   case BEGINSUBSECTHEAD:
+		break;
+	   case ENDSUBSECTHEAD:
+		printf("\n@node %s\n", toc[tocc].text);
+		printf("\n@subsection %s\n\n", toc[tocc].text);
+		break;
+	   case BEGINSECTION:	break;
+	   case ENDSECTION:		break;
+	   case BEGINSUBSECTION:	break;
+	   case ENDSUBSECTION:	break;
+
+	   case BEGINBULPAIR:
+		if (listtype==OL) printf("\n<OL>\n");
+		else if (listtype==UL) printf("\n<UL>\n");
+		else printf("\n<DL>\n");
+		break;
+	   case ENDBULPAIR:
+		if (listtype==OL) printf("\n</OL>\n");
+		else if (listtype==UL) printf("\n</UL>\n");
+		else printf("</DL>\n");
+		break;
+	   case BEGINBULLET:
+		if (listtype==OL || listtype==UL) fcharout=0;
+		else printf("\n<DT>");
+		break;
+	   case ENDBULLET:
+		if (listtype==OL || listtype==UL) fcharout=1;
+		else printf("</DT>");
+		break;
+	   case BEGINBULTXT:
+		if (listtype==OL || listtype==UL) printf("<LI>");
+		else printf("\n<DD>");
+		break;
+	   case ENDBULTXT:
+		if (listtype==OL || listtype==UL) printf("</LI>");
+		else printf("</DD>\n");
+		break;
+
+	   case BEGINLINE:
+		/*		if (ncnt) printf("<P>\n");*/
+
+		/* trailing spaces already trimmed off, so look for eol now */
+		if (fCodeline) {
+		  printf("<CODE>");
+		  for (i=0; i<scnt-indent; i++) printf("&nbsp;"/*&#160;*/);		/* ? */
+		  tagc=0;
+
+		  /* already have .tag=BOLDITALICS, .first=0 */
+		  /* would be more elegant, but can't print initial spaces before first tag
+		  tags[0].last = linelen;
+		  tagc=1;
+		  fIQS=0;
+		  */
+		}
+
+		break;
+
+	   case ENDLINE:
+		/*if (fCodeline) { fIQS=1; fCodeline=0; }*/
+		if (fCodeline) { printf("</CODE><BR>"); fCodeline=0; }
+		I=0; CurLine++; if (!fPara && scnt) printf("<BR>"); printf("\n");
+		break;
+
+	   case SHORTLINE:
+		if (fCodeline) { printf("</CODE>"); fCodeline=0; }
+		if (!fIP) printf("<BR>\n");
+		break;
+
+
+	   case BEGINTABLE:
+		if (fSource) {
+		  /*printf("<CENTER><TABLE BORDER>\n");*/
+		  printf("<TABLE BORDER=0>\n");
+		} else {
+		  printf("<BR><PRE>\n"); pre=1; fQS=fIQS=fPara=0;
+		}
+		break;
+	   case ENDTABLE:
+		if (fSource) {
+		  printf("</TABLE>\n");
+		} else {
+		  printf("</PRE><BR>\n"); pre=0; fQS=fIQS=fPara=1;
+		}
+		break;
+	   case BEGINTABLELINE:		printf("<TR>"); break;
+	   case ENDTABLELINE:		printf("</TR>\n"); break;
+	   case BEGINTABLEENTRY:
+		printf("<TD ALIGN=");
+		switch (tblcellformat[0]) {
+		case 'c':	printf("CENTER"); break;
+		case 'n':	/*printf("DECIMAL"); break;  -- fall through to right for now */
+		case 'r':	printf("RIGHT"); break;
+		case 'l':
+		default:
+		  printf("LEFT");
+		}
+		if (tblcellspan>1) printf(" COLSPAN=%d", tblcellspan);
+		printf(">");
+		break;
+	   case ENDTABLEENTRY:
+		printf("</TD>");
+		break;
+
+	   /* something better with CSS */
+	   case BEGININDENT:		printf("<blockquote>"); break;
+	   case ENDINDENT:			printf("</blockquote>\n"); break;
+
+	   case FONTSIZE:
+		/* HTML font step sizes are bigger than troff's */
+		if ((fontdelta+=intArg)!=0) printf("<FONT SIZE=%c1>", (intArg>0)?'+':'-'); else printf("</FONT>\n");
+		break;
+
+	   case BEGINBOLD:		printf("@b{"); break; /* } */
+	   case BEGINITALICS:	printf("@i{"); break;
+	   case BEGINSC:		printf("@sc{"); break; /* } */
+	   case ENDITALICS:
+	   case ENDBOLD:
+	   case ENDSC: /* { */
+		printf("}");
+		break;
+	   case BEGINBOLDITALICS:
+	   case BEGINCODE:		printf("<CODE>"); break;
+	   case ENDBOLDITALICS:
+	   case ENDCODE:		printf("</CODE>"); break;
+	   case BEGINMANREF:
+		manrefextract(hitxt);
+		if (fmanRef) { printf("@xref{}"); }
+/*<A HREF=\""); printf(href, manrefname, manrefsect); printf("\">"); }*/
+		else printf("<I>");
+		break;
+	   case ENDMANREF:
+		if (fmanRef) printf("</A>\n"); else printf("</I>");
+		break;
+	   case HR:		printf("\n<HR>\n"); break;
+
+	   case BEGINY: case ENDY:
+	   case BEGINHEADER: case ENDHEADER:
+	   case BEGINFOOTER: case ENDFOOTER:
+	   case CHANGEBAR:
+		/* nothing */
+		break;
+	   default:
+		DefaultPara(cmd);
+	}
+}
+#endif
+
+
+
+/* generates MIME compliant to RFC 1563 */
+
+static void
+MIME(enum command cmd)
+{
+	static int pre=0;
+	int i;
+
+	/* always respond to these signals */
+	switch (cmd) {
+	   case CHARDASH:
+	   case CHARAMP:
+	   case CHARPERIOD:
+	   case CHARTAB:
+		putchar(cmd); break;
+	   case CHARLSQUOTE:	putchar('`'); break;
+	   case CHARACUTE:
+	   case CHARRSQUOTE:	putchar('\''); break;
+	   case CHARBULLET:		putchar('*'); break;
+	   case CHARDAGGER:		putchar('|'); break;
+	   case CHARPLUSMINUS:	printf("+-"); break;
+	   case CHARNBSP:		putchar(' '); break;
+ 	   case CHARCENT:	putchar('c'); break;
+ 	   case CHARSECT:	putchar('S'); break;
+ 	   case CHARCOPYR:	printf("(C)"); break;
+ 	   case CHARNOT:	putchar('~'); break;
+ 	   case CHARREGTM:	printf("(R)"); break;
+ 	   case CHARDEG:	putchar('o'); break;
+ 	   case CHAR14:		printf("1/4"); break;
+ 	   case CHAR12:		printf("1/2"); break;
+ 	   case CHAR34:		printf("3/4"); break;
+ 	   case CHARMUL:	putchar('X'); break;
+ 	   case CHARDIV:	putchar('/'); break;
+	   case CHARLQUOTE:
+	   case CHARRQUOTE:
+		putchar('"');
+		break;
+	   case CHARBACKSLASH:	/* these should be caught as escaped chars */
+	   case CHARGT:
+	   case CHARLT:
+#if 0
+		assert(1);
+#endif
+		break;
+	   default:
+		break;
+	}
+
+	/* while in pre mode... */
+	if (pre) {
+		switch (cmd) {
+		   case ENDLINE:	I=0; CurLine++; if (!fPara && scnt) printf("\n\n"); break;
+		   case ENDTABLE:	printf("</fixed>\n\n"); pre=0; fQS=fIQS=fPara=1; break;
+		   default:
+			/* nothing */
+			break;
+		}
+		return;
+	}
+
+	/* usual operation */
+	switch (cmd) {
+	   case BEGINDOC:
+		printf("Content-Type: text/enriched\n");
+		printf("Text-Width: 60\n");
+		escchars = "<>\\";
+
+		I=0;
+		break;
+	   case ENDDOC:
+		/* header and footer wanted? */
+		printf("\n\n");
+		if (fHeadfoot) {
+			printf("\n");
+			MIME(BEGINSECTHEAD); printf("%s",HEADERANDFOOTER); MIME(ENDSECTHEAD);
+			for (i=0; i<CRUFTS; i++) if (*cruft[i]) printf("\n%s\n",cruft[i]);
+		}
+
+/*
+		printf("\n<comment>\n");
+		printf("%s\n%s\n", provenance, anonftp);
+		printf("</comment>\n\n");
+*/
+
+/*
+		printf("\n<HR><P>\n");
+		printf("<A NAME=\"toc\"><B>%s</B></A><P>\n", TABLEOFCONTENTS);
+		printf("<UL>\n");
+		for (i=0, lasttoc=BEGINSECTION; i<tocc; lasttoc=toc[i].type, i++) {
+		  if (lasttoc!=toc[i].type) {
+		    if (toc[i].type==BEGINSUBSECTION) printf("<UL>\n");
+		    else printf("</UL>\n");
+		  }
+		  printf("<LI><A NAME=\"toc%d\" HREF=\"#sect%d\">%s</A></LI>\n", i, i, toc[i].text);
+		}
+		if (lasttoc==BEGINSUBSECTION) printf("</UL>");
+		printf("</UL>\n");
+		printf("</BODY></HTML>\n");
+*/
+		break;
+	   case BEGINBODY:
+		printf("\n\n");
+		break;
+	   case ENDBODY:		break;
+
+	   case BEGINCOMMENT: fcharout=0; break;
+	   case ENDCOMMENT: fcharout=1; break;
+	   case COMMENTLINE: break;
+
+	   case BEGINSECTHEAD:
+		printf("\n<bigger><bigger><underline>");
+		/*A NAME=\"sect%d\" HREF=\"#toc%d\"><H2>", tocc, tocc);*/
+		break;
+	   case ENDSECTHEAD:
+		printf("</underline></bigger></bigger>\n\n<indent>");
+		/* useful extraction from files, environment? */
+		break;
+	   case BEGINSUBSECTHEAD:
+		printf("<bigger>");
+		/*\n<A NAME=\"sect%d\" HREF=\"#toc%d\"><H3>", tocc, tocc);*/
+		break;
+	   case ENDSUBSECTHEAD:
+		printf("</bigger>\n\n</indent>");
+		break;
+	   case BEGINSECTION:
+	   case BEGINSUBSECTION:
+		break;
+	   case ENDSECTION:
+	   case ENDSUBSECTION:
+		printf("</indent>\n");
+		break;
+
+	   case BEGINBULPAIR:	break;
+	   case ENDBULPAIR:		break;
+	   case BEGINBULLET:	printf("<bold>"); break;
+	   case ENDBULLET:		printf("</bold>\t"); break;
+	   case BEGINBULTXT:
+	   case BEGININDENT:
+		printf("<indent>");
+		break;
+	   case ENDBULTXT:
+	   case ENDINDENT:
+		printf("</indent>\n");
+		break;
+
+	   case FONTSIZE:
+		if ((fontdelta+=intArg)==0) {
+		  if (intArg>0) printf("</smaller>"); else printf("</bigger>");
+		} else {
+		  if (intArg>0) printf("<bigger>"); else printf("<smaller>");
+		}
+		break;
+
+	   case BEGINLINE:		/*if (ncnt) printf("\n\n");*/ break;
+	   case ENDLINE:		I=0; CurLine++; printf("\n"); break;
+	   case SHORTLINE:		if (!fIP) printf("\n\n"); break;
+	   case BEGINTABLE:		printf("<nl><fixed>\n"); pre=1; fQS=fIQS=fPara=0; break;
+	   case ENDTABLE:		printf("</fixed><nl>\n"); pre=0; fQS=fIQS=fPara=1; break;
+	   case BEGINTABLELINE: case ENDTABLELINE: case BEGINTABLEENTRY: case ENDTABLEENTRY:
+		break;
+		/* could use a new list type */
+
+	   case BEGINBOLD:		printf("<bold>"); break;
+	   case ENDBOLD:		printf("</bold>"); break;
+	   case BEGINITALICS:	printf("<italics>"); break;
+	   case ENDITALICS:		printf("</italics>"); break;
+	   case BEGINCODE:
+	   case BEGINBOLDITALICS:printf("<bold><italics>"); break;
+	   case ENDCODE:
+	   case ENDBOLDITALICS:	printf("</bold></italics>"); break;
+	   case BEGINMANREF:
+		printf("<x-color><param>blue</param>");
+/* how to make this hypertext?
+   		manrefextract(hitxt);
+		if (fmanRef) { printf("<A HREF=\""); printf(href, manrefname, manrefsect); printf("\">\n"); }
+		else printf("<I>");
+		break;
+*/
+		break;
+	   case ENDMANREF:
+		printf("</x-color>");
+		break;
+
+	   case HR:		printf("\n\n%s\n\n", horizontalrule); break;
+
+	   case BEGINSC: case ENDSC:
+	   case BEGINY: case ENDY:
+	   case BEGINHEADER: case ENDHEADER:
+	   case BEGINFOOTER: case ENDFOOTER:
+	   case CHANGEBAR:
+		/* nothing */
+		break;
+	   default:
+		DefaultPara(cmd);
+	}
+}
+
+
+
+/*
+ * LaTeX
+ */
+
+static void
+LaTeX(enum command cmd)
+{
+
+	switch (cmd) {
+	   case BEGINDOC:
+		escchars = "$&%#_{}"; /* and more to come? */
+		printf("%% %s,\n", provenance);
+		printf("%% %s\n\n", anonftp);
+		/* definitions */
+		printf(
+		  "\\documentstyle{article}\n"
+		  "\\def\\thefootnote{\\fnsymbol{footnote}}\n"
+		  "\\setlength{\\parindent}{0pt}\n"
+		  "\\setlength{\\parskip}{0.5\\baselineskip plus 2pt minus 1pt}\n"
+		  "\\begin{document}\n"
+			  );
+		I=0;
+		break;
+	   case ENDDOC:
+		/* header and footer wanted? */
+		printf("\n\\end{document}\n");
+
+		break;
+	   case BEGINBODY:
+		printf("\n\n");
+		break;
+	   case ENDBODY:		break;
+
+	   case BEGINCOMMENT:
+	   case ENDCOMMENT:
+		break;
+	   case COMMENTLINE: printf("%% "); break;
+
+
+	   case BEGINSECTION:	break;
+	   case ENDSECTION:		break;
+	   case BEGINSECTHEAD:	printf("\n\\section{"); tagc=0; break;
+	   case ENDSECTHEAD:
+		printf("}");
+/*
+		if (CurLine==1) printf("\\footnote{"
+		  "\\it conversion to \\LaTeX\ format by PolyglotMan "
+		  "available via anonymous ftp from {\\tt ftp.berkeley.edu:/ucb/people/phelps/tcltk}}"
+			  );
+*/
+		/* useful extraction from files, environment? */
+		printf("\n");
+		break;
+	   case BEGINSUBSECTHEAD:printf("\n\\subsection{"); break;
+	   case ENDSUBSECTHEAD:
+		printf("}");
+		break;
+	   case BEGINSUBSECTION:	break;
+	   case ENDSUBSECTION:	break;
+	   case BEGINBULPAIR:	printf("\\begin{itemize}\n"); break;
+	   case ENDBULPAIR:		printf("\\end{itemize}\n"); break;
+	   case BEGINBULLET:	printf("\\item ["); break;
+	   case ENDBULLET:		printf("] "); break;
+	   case BEGINLINE:		/*if (ncnt) printf("\n\n");*/ break;
+	   case ENDLINE:		I=0; putchar('\n'); CurLine++; break;
+	   case BEGINTABLE:		printf("\\begin{verbatim}\n"); break;
+	   case ENDTABLE:		printf("\\end{verbatim}\n"); break;
+	   case BEGINTABLELINE: case ENDTABLELINE: case BEGINTABLEENTRY: case ENDTABLEENTRY:
+		break;
+	   case BEGININDENT: case ENDINDENT:
+	   case FONTSIZE:
+		break;
+	   case SHORTLINE:		if (!fIP) printf("\n\n"); break;
+	   case BEGINBULTXT:	break;
+	   case ENDBULTXT:		putchar('\n'); break;
+
+	   case CHARLQUOTE:		printf("``"); break;
+	   case CHARRQUOTE:		printf("''"); break;
+	   case CHARLSQUOTE:
+	   case CHARRSQUOTE:
+	   case CHARPERIOD:
+	   case CHARTAB:
+	   case CHARDASH:
+	   case CHARNBSP:
+		putchar(cmd); break;
+	   case CHARBACKSLASH:	printf("$\\backslash$"); break;
+	   case CHARGT:		printf("$>$"); break;
+	   case CHARLT:		printf("$<$"); break;
+	   case CHARHAT:		printf("$\\char94{}$"); break;
+	   case CHARVBAR:		printf("$|$"); break;
+	   case CHARAMP:		printf("\\&"); break;
+	   case CHARBULLET:		printf("$\\bullet$ "); break;
+	   case CHARDAGGER:		printf("\\dag "); break;
+	   case CHARPLUSMINUS:	printf("\\pm "); break;
+ 	   case CHARCENT:		printf("\\hbox{\\rm\\rlap/c}"); break;
+ 	   case CHARSECT:		printf("\\S "); break;
+ 	   case CHARCOPYR:		printf("\\copyright "); break;
+ 	   case CHARNOT:		printf("$\\neg$"); break;
+ 	   case CHARREGTM:		printf("(R)"); break;
+ 	   case CHARDEG:		printf("$^\\circ$"); break;
+ 	   case CHARACUTE:		putchar('\''); break;
+ 	   case CHAR14:		printf("$\\frac{1}{4}$"); break;
+ 	   case CHAR12:		printf("$\\frac{1}{2}$"); break;
+ 	   case CHAR34:		printf("$\\frac{3}{4}$"); break;
+ 	   case CHARMUL:		printf("\\times "); break;
+ 	   case CHARDIV:		printf("\\div "); break;
+
+	   case BEGINCODE:
+	   case BEGINBOLD:		printf("{\\bf "); break; /* } */
+	   case BEGINSC:		printf("{\\sc "); break; /* } */
+	   case BEGINITALICS:	printf("{\\it "); break; /* } */
+	   case BEGINBOLDITALICS:printf("{\\bf\\it "); break; /* } */
+	   case BEGINMANREF:	printf("{\\sf "); break; /* } */
+	   case ENDCODE:
+	   case ENDBOLD:
+	   case ENDSC:
+	   case ENDITALICS:
+	   case ENDBOLDITALICS:
+	   case ENDMANREF: /* { */
+		putchar('}');
+		break;
+	   case HR:		/*printf("\n%s\n", horizontalrule);*/ break;
+
+	   case BEGINY: case ENDY:
+	   case BEGINHEADER: case ENDHEADER:
+	   case BEGINFOOTER: case ENDFOOTER:
+	   case CHANGEBAR:
+		/* nothing */
+		break;
+	   default:
+		DefaultPara(cmd);
+	}
+}
+
+
+static void
+LaTeX2e(enum command cmd)
+{
+	switch (cmd) {
+		/* replace selected commands ... */
+	   case BEGINDOC:
+		escchars = "$&%#_{}";
+		printf("%% %s,\n", provenance);
+		printf("%% %s\n\n", anonftp);
+		/* definitions */
+		printf(
+		  "\\documentclass{article}\n"
+		  "\\def\\thefootnote{\\fnsymbol{footnote}}\n"
+		  "\\setlength{\\parindent}{0pt}\n"
+		  "\\setlength{\\parskip}{0.5\\baselineskip plus 2pt minus 1pt}\n"
+		  "\\begin{document}\n"
+			  );
+		I=0;
+		break;
+	   case BEGINCODE:
+	   case BEGINBOLD:		printf("\\textbf{"); break; /* } */
+	   case BEGINSC:		printf("\\textsc{"); break; /* } */
+	   case BEGINITALICS:	printf("\\textit{"); break; /* } */
+	   case BEGINBOLDITALICS:printf("\\textbf{\\textit{"); break; /* }} */
+	   case BEGINMANREF:	printf("\\textsf{"); break; /* } {{ */
+	   case ENDBOLDITALICS:	printf("}}"); break;
+
+		/* ... rest same as old LaTeX */
+	   default:
+		LaTeX(cmd);
+	}
+}
+
+
+
+/*
+ * Rich Text Format (RTF)
+ */
+
+/* RTF could use more work */
+
+static void
+RTF(enum command cmd)
+{
+
+	switch (cmd) {
+	   case BEGINDOC:
+		escchars = "{}";
+		/* definitions */
+		printf(
+		  /* fonts */
+		  "{\\rtf1\\deff2 {\\fonttbl"
+		  "{\\f20\\froman Times;}{\\f150\\fnil I Times Italic;}"
+		  "{\\f151\\fnil B Times Bold;}{\\f152\\fnil BI Times BoldItalic;}"
+		  "{\\f22\\fmodern Courier;}{\\f23\\ftech Symbol;}"
+		  "{\\f135\\fnil I Courier Oblique;}{\\f136\\fnil B Courier Bold;}{\\f137\\fnil BI Courier BoldOblique;}"
+		  "{\\f138\\fnil I Helvetica Oblique;}{\\f139\\fnil B Helvetica Bold;}}"
+		  "\n"
+
+		  /* style sheets */
+		  "{\\stylesheet{\\li720\\sa120 \\f20 \\sbasedon222\\snext0 Normal;}"
+		  "{\\s2\\sb200\\sa120 \\b\\f3\\fs20 \\sbasedon0\\snext2 section head;}"
+		  "{\\s3\\li180\\sa120 \\b\\f20 \\sbasedon0\\snext3 subsection head;}"
+		  "{\\s4\\fi-1440\\li2160\\sa240\\tx2160 \\f20 \\sbasedon0\\snext4 detailed list;}}"
+		  "\n"
+
+/* more header to come--do undefined values default to nice values? */
+			   );
+		I=0;
+		break;
+	   case ENDDOC:
+		/* header and footer wanted? */
+		printf("\\par{\\f150 %s,\n%s}", provenance, anonftp);
+		printf("}\n");
+		break;
+	   case BEGINBODY:
+		printf("\n\n");
+		break;
+	   case ENDBODY:
+		CurLine++;
+		printf("\\par\n");
+		tagc=0;
+		break;
+
+	   case BEGINCOMMENT: fcharout=0; break;
+	   case ENDCOMMENT: fcharout=1; break;
+	   case COMMENTLINE: break;
+
+	   case BEGINSECTION:	break;
+	   case ENDSECTION:		printf("\n\\par\n"); break;
+	   case BEGINSECTHEAD:	printf("{\\s2 "); tagc=0; break;
+	   case ENDSECTHEAD:
+		printf("}\\par");
+		/* useful extraction from files, environment? */
+		printf("\n");
+		break;
+	   case BEGINSUBSECTHEAD:printf("{\\s3 "); break;
+	   case ENDSUBSECTHEAD:
+		printf("}\\par\n");
+		break;
+	   case BEGINSUBSECTION:	break;
+	   case ENDSUBSECTION:	break;
+	   case BEGINLINE:		/*if (ncnt) printf("\n\n");*/ break;
+	   case ENDLINE:		I=0; putchar(' '); /*putchar('\n'); CurLine++;*/ break;
+	   case SHORTLINE:		if (!fIP) printf("\\line\n"); break;
+	   case BEGINBULPAIR:	printf("{\\s4 "); break;
+	   case ENDBULPAIR:		printf("}\\par\n"); break;
+	   case BEGINBULLET:	break;
+	   case ENDBULLET:		printf("\\tab "); fcharout=0; break;
+	   case BEGINBULTXT:	fcharout=1; break;
+	   case ENDBULTXT:		break;
+
+	   case CHARLQUOTE:		printf("``"); break;
+	   case CHARRQUOTE:		printf("''"); break;
+	   case CHARLSQUOTE:
+	   case CHARRSQUOTE:
+	   case CHARPERIOD:
+	   case CHARTAB:
+	   case CHARDASH:
+	   case CHARBACKSLASH:
+	   case CHARGT:
+	   case CHARLT:
+	   case CHARHAT:
+	   case CHARVBAR:
+	   case CHARAMP:
+	   case CHARNBSP:
+ 	   case CHARCENT:
+ 	   case CHARSECT:
+ 	   case CHARCOPYR:
+ 	   case CHARNOT:
+ 	   case CHARREGTM:
+ 	   case CHARDEG:
+ 	   case CHARACUTE:
+ 	   case CHAR14:
+ 	   case CHAR12:
+ 	   case CHAR34:
+ 	   case CHARMUL:
+ 	   case CHARDIV:
+		putchar(cmd); break;
+	   case CHARBULLET:		printf("\\bullet "); break;
+	   case CHARDAGGER:		printf("\\dag "); break;
+	   case CHARPLUSMINUS:	printf("\\pm "); break;
+
+	   case BEGINCODE:
+	   case BEGINBOLD:		printf("{\\b "); break; /* } */
+	   case BEGINSC:		printf("{\\fs20 "); break; /* } */
+	   case BEGINITALICS:	printf("{\\i "); break; /* } */
+	   case BEGINBOLDITALICS:printf("{\\b \\i "); break; /* } */
+	   case BEGINMANREF:	printf("{\\f22 "); break; /* } */
+	   case ENDBOLD:
+	   case ENDCODE:
+	   case ENDSC:
+	   case ENDITALICS:
+	   case ENDBOLDITALICS:
+	   case ENDMANREF: /* { */
+		putchar('}');
+		break;
+	   case HR:		printf("\n%s\n", horizontalrule); break;
+
+	   case BEGINY: case ENDY:
+	   case BEGINHEADER: case ENDHEADER:
+	   case BEGINFOOTER: case ENDFOOTER:
+	   case BEGINTABLE: case ENDTABLE:
+	   case BEGINTABLELINE: case ENDTABLELINE: case BEGINTABLEENTRY: case ENDTABLEENTRY:
+	   case BEGININDENT: case ENDINDENT:
+	   case FONTSIZE:
+	   case CHANGEBAR:
+		/* nothing */
+		break;
+	   default:
+		DefaultPara(cmd);
+	}
+}
+
+
+
+/*
+ * pointers to existing tools
+ */
+
+static void
+PostScript(enum command cmd)
+{
+	fprintf(stderr, "Use groff or psroff to generate PostScript.\n");
+	exit(1);
+}
+
+
+static void
+FrameMaker(enum command cmd)
+{
+	fprintf(stderr, "FrameMaker comes with filters that convert from roff to MIF.\n");
+	exit(1);
+}
+
+
+
+
+/*
+ * Utilities common to both parses
+ */
+
+
+/*
+ level 0: DOC - need match
+ level 1: SECTION - need match
+ level 2: SUBSECTION | BODY | BULLETPAIR
+ level 3: BODY (within SUB) | BULLETPAIR (within SUB) | BULTXT (within BULLETPAIR)
+ level 4: BULTXT (within BULLETPAIR within SUBSECTION)
+
+ never see: SECTHEAD, SUBSECTHEAD, BULLET
+*/
+
+int Psect=0, Psub=0, Pbp=0, Pbt=0, Pb=0, Pbul=0;
+
+static void
+pop(enum command cmd)
+{
+	assert(cmd==ENDINDENT || cmd==BEGINBULLET || cmd==BEGINBULTXT || cmd==BEGINBULPAIR || cmd==BEGINBODY || cmd==BEGINSECTION || cmd==BEGINSUBSECTION || cmd==ENDDOC);
+/*
+	int i;
+	int p;
+	int match;
+
+	p=cmdp-1;
+	for (i=cmdp-1;i>=0; i--)
+		if (cmd==cmdstack[i]) { match=i; break; }
+*/
+
+	/* if match, pop off all up to and including match */
+	/* otherwise, pop off one level*/
+
+	if (Pbul) {
+	  (*fn)(ENDBULLET); Pbul=0;
+	  if (cmd==BEGINBULLET) return;
+	} /* else close off ENDBULTXT */
+
+	if (Pbt) { (*fn)(ENDBULTXT); Pbt=0; }
+	if (cmd==BEGINBULTXT || cmd==BEGINBULLET) return;
+
+	if (Pb && cmd==BEGINBULPAIR) { (*fn)(ENDBODY); Pb=0; }	/* special */
+	if (Pbp) { (*fn)(ENDBULPAIR); Pbp=0; }
+	if (cmd==BEGINBULPAIR || cmd==ENDINDENT) return;
+
+	if (Pb) { (*fn)(ENDBODY); Pb=0; }
+	if (cmd==BEGINBODY) return;
+
+	if (Psub) { (*fn)(ENDSUBSECTION); Psub=0; }
+	if (cmd==BEGINSUBSECTION) return;
+
+	if (Psect) { (*fn)(ENDSECTION); Psect=0; }
+	if (cmd==BEGINSECTION) return;
+}
+
+
+static void
+poppush(enum command cmd)
+{
+	assert(cmd==ENDINDENT || cmd==BEGINBULLET || cmd==BEGINBULTXT || cmd==BEGINBULPAIR || cmd==BEGINBODY || cmd==BEGINSECTION || cmd==BEGINSUBSECTION);
+
+	pop(cmd);
+
+	switch (cmd) {
+	   case BEGINBULLET: Pbul=1; break;
+	   case BEGINBULTXT: Pbt=1; break;
+	   case BEGINBULPAIR: Pbp=1; break;
+	   case BEGINBODY: Pb=1; break;
+	   case BEGINSECTION: Psect=1; break;
+	   case BEGINSUBSECTION: Psub=1; break;
+	   default:
+		if (!fQuiet) fprintf(stderr, "poppush: unrecognized code %d\n", cmd);
+	}
+
+	(*fn)(cmd);
+	prevcmd = cmd;
+}
+
+
+
+/*
+ * PREFORMATTED PAGES PARSING
+ */
+
+/* wrapper for getchar() that expands tabs, and sends maximum of n=40 consecutive spaces */
+
+static int
+getchartab(void)
+{
+	static int tabexp = 0;
+	static int charinline = 0;
+	static int cspccnt = 0;
+	char c;
+
+	c = lookahead;
+	if (tabexp) tabexp--;
+	else if (c=='\n') {
+		charinline=0;
+		cspccnt=0;
+	} else if (c=='\t') {
+		tabexp = TabStops-(charinline%TabStops); if (tabexp==TabStops) tabexp=0;
+		lookahead = c = ' ';
+	} else if (cspccnt>=40) {
+		if (*in==' ') {
+			while (*in==' '||*in=='\t') in++;
+			in--;
+		}
+		cspccnt=0;
+	}
+
+	if (!tabexp && lookahead) lookahead = *in++;
+	if (c=='\b') charinline--; else charinline++;
+	if (c==' ') cspccnt++;
+	return c;
+}
+
+
+/* replace gets.  handles hyphenation too */
+static char *
+la_gets(char *buf)
+{
+	static char la_buf[MAXBUF];	/* can lookahead a full line, but nobody does now */
+	static int fla=0, hy=0;
+	char *ret,*p;
+	int c,i;
+
+	assert(buf!=NULL);
+
+	if (fla) {
+		/* could avoid copying if callers used return value */
+		strcpy(buf,la_buf); fla=0;
+		ret=buf;	/* correct? */
+	} else {
+		/*ret=gets(buf); -- gets is deprecated (since it can read too much?) */
+		/* could do this...
+		ret=fgets(buf, MAXBUF, stdin);
+		buf[strlen(buf)-1]='\0';
+		... but don't want to have to rescan line with strlen, so... */
+
+		i=0; p=buf;
+
+		/* recover spaces if re-linebreaking */
+		for (; hy; hy--) { *p++=' '; i++; }
+
+		while (lookahead && (c=getchartab())!='\n' && i<MAXBUF) { *p++=c; i++; }
+		assert(i<MAXBUF);
+
+#if 0
+		lookahead=ungetc(getchar(), stdin);	/* only looking ahead one character for now */
+#endif
+
+		/* very special case: if in SEE ALSO section, re-linebreak so references aren't linebroken
+		   (also do this if fNOHY flag is set) -- doesn't affect lookahead */
+		/* 0xad is an en dash on Linux? */
+		if ((fPara || sectheadid==SEEALSO || fNOHY) && p>buf && p[-1]=='-' && isspace(lookahead)) {
+			p--;	/* zap hyphen */
+			/* zap boldfaced hyphens, gr! */
+			while (p[-1]=='\b' && p[-2]=='-') p-=2;
+
+			/* start getting next line, spaces first ... */
+			while (lookahead && isspace(lookahead) && lookahead!='\n') { getchartab(); hy++; }
+
+			/* ... append next nonspace string to previous ... */
+			while (lookahead && !isspace(lookahead) && i++<MAXBUF) *p++=getchartab();
+
+			/* gobble following spaces (until, perhaps including, end of line) */
+			while (lookahead && isspace(lookahead) && lookahead!='\n') getchartab();
+			if (lookahead=='\n') { getchartab(); hy=0; }
+		}
+
+		*p='\0';
+		ret=(lookahead)?buf:NULL;
+	}
+
+	AbsLine++;
+	return ret;	/* change this to line length? (-1==old NULL) */
+}
+
+
+/*** Kong ***/
+/*
+  I hope the compiler has good common subexpression elimination
+     for all the pointer arithmetic.
+*/
+
+
+char phrase[MAXBUF];	/* first "phrase" (space of >=3 spaces) */
+int phraselen;
+
+static void
+filterline(char *buf, char *plain)
+{
+	char *p,*q,*r;
+	char *ph;
+	int iq;
+	int i,j;
+	int hl=-1, hl2=-1;
+	int iscnt=0;	/* interword space count */
+	int tagci;
+	int I0;
+	int etype;
+	int efirst;
+	enum tagtype tag = NOTAG;
+
+	assert(buf!=NULL && plain!=NULL);
+
+	etype=NOTAG;
+	efirst=-1;
+	tagci=tagc;
+	ph=phrase; phraselen=0;
+	scnt=scnt2=0;
+	s_sum=s_cnt=0;
+	bs_sum=bs_cnt=0;
+	ccnt=0;
+	spcsqz=0;
+
+	/* strip only certain \x1b's and only at very beginning of line */
+	for (p=buf; *p=='\x1b' && (p[1]=='8'||p[1]=='9'); p+=2)
+		/* nop */;
+
+	strcpy(plain,p);
+	q=&plain[strlen(p)];
+
+	/*** spaces and change bars ***/
+	for (scnt=0,p=plain; *p==' '; p++) scnt++;	/* initial space count */
+	if (scnt>200) scnt=130-(q-p);
+
+	assert(*q=='\0');
+	q--;
+	if (fChangeleft)
+		for (; q-40>plain && *q=='|'; q--)	{	/* change bars */
+			if (fChangeleft!=-1) ccnt++;
+			while (q-2>=plain && q[-1]=='\b' && q[-2]=='|') q-=2;	/* boldface changebars! */
+		}
+
+	/*if (q!=&plain[scnt-1])*/			/* zap trailing spaces */
+		for (; *q==' ' && q>plain; q--) /* nop */;
+
+	/* second changebar way out east! HACK HACK HACK */
+	if (q-plain>100 && *q=='|') {
+	  while (*q=='|' && q>plain) { q--; if (fChangeleft!=-1) ccnt++; }
+	  while ((*q==' ' || *q=='_' || *q=='-') && q>plain) q--;
+	}
+
+	for (r=q; (*r&0xff)==CHARDAGGER; r--) *r='-';	/* convert daggers at end of line to hyphens */
+
+	if (q-plain < scnt) scnt = q-plain+1;
+	q[1]='\0';
+
+	/* set I for tags below */
+	if (indent>=0 && scnt>=indent) scnt-=indent;
+	if (!fPara && !fIQS) {
+		if (fChangeleft) I+=(scnt>ccnt)?scnt:ccnt;
+		else I+=scnt;
+	}
+	I0=I;
+
+	/*** tags and filler spaces ***/
+
+	iq=0; falluc=1;
+	for (q=plain; *p; p++) {
+
+		iscnt=0;
+		if (*p==' ') {
+			for (r=p; *r==' '; r++) { iscnt++; spcsqz++; }
+			s_sum+=iscnt; s_cnt++;
+			if (iscnt>1 && !scnt2 && *p==' ') scnt2=iscnt;
+			if (iscnt>2) { bs_cnt++; bs_sum+=iscnt; }	/* keep track of large gaps */
+			iscnt--;		/* leave last space for tail portion of loop */
+
+			/* write out spaces */
+			if (fQS && iscnt<3) { p=r-1;	iscnt=0; } /* reduce strings of <3 spaces to 1 */
+			/* else if (fQS && iscnt>=3) { replace with tab? } */
+			else {
+				for (i=0; i<iscnt; i++) { p++; *q++=' '; }
+			}
+		} /* need to go through if chain for closing off annotations */
+
+		/** backspace-related filtering **/
+
+		if (*p=='\b' && p[1]=='_' && q>plain && q[-1]=='+') {
+			/* bold plus/minus(!) */
+			q[-1]=c_plusminus;
+			while (*p=='\b' && p[1]=='_') p+=2;
+			continue;
+		} else if ((*p=='_' && p[1]=='\b' && p[2]!='_' && p[3]!='\b')
+			|| (*p=='\b' && p[1]=='_')) {
+			/* italics */
+			if (tag!=ITALICS && hl>=0) { tagadd(tag, hl, I+iq); hl=-1; }
+			if (hl==-1) hl=I+iq;
+			tag=ITALICS;
+			p+=2;
+		} else if (*p=='_' && p[2]==p[4] && p[1]=='\b' && p[3]=='\b' && p[2]!='_') {
+			/* bold italics (for Solaris) */
+			for (p+=2; *p==p[2] && p[1]=='\b';) p+=2;
+			if (tag!=BOLDITALICS && hl>=0) { tagadd(tag, hl, I+iq); hl=-1; }
+			if (hl==-1) hl=I+iq;
+			tag=BOLDITALICS;
+		} else if (*p==p[2] && p[1]=='\b') {
+			/* boldface */
+			while (*p==p[2] && p[1]=='\b') p+=2;
+			if (tag!=BOLD && hl>=0) { tagadd(tag, hl, I+iq); hl=-1; }
+			if (hl==-1) hl=I+iq;
+			tag=BOLD;
+		} else if (p[1]=='\b' &&
+				 ((*p=='o' && p[2]=='+') ||
+				  (*p=='+' && p[2]=='o')) ) {
+			/* bullets */
+			p+=2;
+			while (p[1]=='\b' && (*p=='o' || p[2]=='+') ) p+=2;		/* bold bullets(!) */
+			*q++=c_bullet; iq++;
+			continue;
+		} else if (*p=='\b' && p>plain && p[-1]=='o' && p[1]=='+') {
+			/* OSF bullets */
+			while (*p=='\b' && p[1]=='+') p+=2;	/* bold bullets(!) */
+			q[-1]=c_bullet; p--;
+			continue;
+		} else if (p[1]=='\b' && *p=='+' && p[2]=='_') {
+			/* plus/minus */
+			p+=2;
+			*q++=c_plusminus; iq++;
+			continue;
+		} else if (p[1]=='\b' && *p=='|' && p[2]=='-') {
+			/* dagger */
+			*q++=c_dagger; iq++;
+			p+=2; continue;
+		} else if (*p=='\b') {
+			/* supress unattended backspaces */
+			continue;
+		} else if (*p=='\x1b' /*&& (p[1]=='9'||p[1]=='8')*/) {
+			p++;
+			if (*p=='[') {
+				p++;
+				if (*p=='1' && hl==-1) {
+					/* stash attributes in "invalid" array element */
+					efirst=I+iq; etype=BOLD;
+					/*hl=I+iq; tag=BOLD; -- faces immediate end of range */
+				} else if (*p=='0' /*&& hl>=0 && hl2==-1 && tags[MAXTAGS].first<I+iq*/) {
+					/* doesn't catch tag if spans line -- just make tag and hl static? */
+					/*tagadd(tags[MAXTAGS].type, tags[MAXTAGS].first, I+iq);*/
+					if (hl==-1 && hl2==-1 && efirst!=-1/*<I+iq*/)
+						tagadd(etype, efirst, I+iq);
+					efirst=-1;
+				} /* else skip unrecognized escape codes like 8/9 */
+			}
+			p+=2;	/* gobble 0/1 and mysterious following 'm' */
+			/* following 'm' (why?) gobbled in overarching for */
+			/*continue;*/
+		} else if ((isupper(*p) /*|| *p=='_' || *p=='&'*/) &&
+				 (hl>=0 || isupper(p[1]) || (p[1]=='_' && p[2]!='\b') || p[1]=='&')) {
+			if (hl==-1 && efirst==-1) { hl=I+iq; tag=SMALLCAPS; }
+		} else {
+			/* end of tag, one way or another */
+			/* collect tags in this pass, interspersed later if need be */
+			/* can't handle overlapping tags */
+			if (hl>=0) {
+				if (hl2==-1) tagadd(tag, hl, I+iq);
+				hl=-1;
+			}
+		}
+
+		/** non-backspace related filtering **/
+		/* case statement here in place of if chain? */
+/* Tk 3.x's text widget tabs too crazy
+		if (*p==' ' && strncmp("     ",p,5)==0) {
+			xputchar('\t'); i+=5-1; ci++; continue;
+		} else
+*/
+/* copyright symbol: too much work for so little
+		if (p[i]=='o' && (strncmp("opyright (C) 19",&p[i],15)==0
+				    || strncmp("opyright (c) 19",&p[i],15)==0)) {
+			printf("opyright \xd3 19");
+			tagadd(SYMBOL, ci+9, ci+10);
+			i+=15-1; ci+=13; continue;
+		} else
+*/
+		if (*p=='(' && q>plain && (isalnum(q[-1])||strchr(manvalid/*"._-+"*/,q[-1])!=NULL)
+		    && strcoloncmp(&p[1],')',vollist)
+		    /* && p[1]!='s' && p[-1]!='`' && p[-1]!='\'' && p[-1]!='"'*/ ) {
+			hl2=I+iq;
+			for (r=q-1; r>=plain && (isalnum(*r)||strchr(manvalid/*"._-+:"*/,*r)!=NULL); r--)
+				hl2--;
+			/* else ref to a function? */
+			/* maybe save position of opening paren so don't highlight it later */
+		} else if (*p==')' && hl2!=-1) {
+			/* don't overlap tags on man page referenes */
+			while (tagc>0 && tags[tagc-1].last>hl2) tagc--;
+			tagadd(MANREF, hl2, I+iq+1);
+			hl2=hl=-1;
+		} else if (hl2!=-1) {
+			/* section names are alphanumic or '+' for C++ */
+			if (!isalnum(*p) && *p!='+') hl2=-1;
+		}
+
+
+		/*assert(*p!='\0');*/
+		if (!*p) break;	/* not just safety check -- check out sgmls.1 */
+
+		*q++=*p;
+/*		falluc = falluc && (isupper(*p) || isspace(*p) || isdigit(*p) || strchr("-+&_'/()?!.,;",*p)!=NULL);*/
+		falluc = falluc && !islower(*p);
+		if (!scnt2) { *ph++=*p; phraselen++; }
+		iq+=iscnt+1;
+	}
+	if (hl>=0) tagadd(tag, hl, I+iq);
+	else if (efirst>=0) tagadd(etype, efirst, I+iq);
+	*q=*ph='\0';
+	linelen=iq+ccnt;
+
+
+	/* special case for Solaris:
+	   if line has ONLY <CODE> tags AND they SPAN line, convert to one tag */
+	fCodeline=0;
+	if (tagc && tags[0].first==0 && tags[tagc-1].last==linelen) {
+		fCodeline=1;
+		j=0;
+		/* invariant: at start of a tag */
+		for (i=0; fCodeline && i<tagc; i++) {
+			if (tags[i].type!=BOLDITALICS /*&& tags[i].type!=BOLD*/) fCodeline=0;
+			else if ((j=tags[i].last)<linelen) {
+			  for (; j < tags[i+1].first ; j++)
+			  	if (!isspace(plain[j])) { fCodeline=0; break; }
+			}
+		}
+	}
+
+
+	/* verify tag lists -- in production, compiler should kill with dead code elimination */
+	for (i=tagci; i<tagc; i++) {
+		/* verify valid ranges */
+		assert(tags[i].type>NOTAG && tags[i].type<=MANREF);
+		assert(tags[i].first>=I0 && tags[i].last<=linelen+I0);
+		assert(tags[i].first<=tags[i].last);
+
+		/* verify for no overlap with other tags */
+		for (j=i+1; j<tagc; j++) {
+			assert(tags[i].last<=tags[j].first /*|| tags[i].first>=tags[j].last*/);
+		}
+	}
+}
+
+
+/*
+  buf[] == input text (read only)
+  plain[] == output (initial, trailing spaces stripped; tabs=>spaces;
+     underlines, overstrikes => tag array; spaces squeezed, if requested)
+  ccnt = count of changebars
+  scnt = count of initial spaces
+  linelen = length result in plain[]
+*/
+
+/*#define MAXINDENT 15*/
+/*#define HEADFOOTMATCH 20*/
+
+int fHead=0;
+int fFoot=0;
+
+static void
+preformatted_filter(void)
+{
+	const int MINRM=50;		/* minimum column for right margin */
+	const int MINMID=20;
+	const int HEADFOOTSKIP=20;
+	const int HEADFOOTMAX=25;
+	int curtag;
+	char *p,*r;
+	char head[MAXBUF]="";		/* first "word" */
+	char foot[MAXBUF]="";
+	int header_m=0, footer_m=0;
+	int headlen=0, footlen=0;
+/*	int line=1-1; */
+	int i,j,k,l,off;
+	int sect=0,subsect=0,bulpair=0,osubsect=0;
+	int title=1;
+	int oscnt=-1;
+	int empty=0,oempty;
+	int fcont=0;
+	int Pnew=0,I0;
+	float s_avg=0.0;
+	int spaceout;
+	int skiplines=0;
+	int c;
+
+	/* try to keep tabeginend[][] in parallel with enum tagtype */
+	assert(tagbeginend[ITALICS][0]==BEGINITALICS);
+	assert(tagbeginend[MANREF][1]==ENDMANREF);
+	in++;	/* lookahead = current character, in points to following */
+
+	/*	for (i=0; i<MAXBUF; i++) tabgram[i]=0;*/
+
+	/*if (fMan) */indent=-1;
+	I=1;
+	CurLine=1;
+	(*fn)(BEGINDOC); I0=I;
+
+	/* run through each line */
+	while (la_gets(buf)!=NULL) {
+		if (title) I=I0;
+		/* strip out Ousterhout box: it's confusing the section line counts in TkMan outlining */
+		if (fNORM && *buf=='_'
+		    && strncmp(buf,"_________________________________________________________________",65)==0) {
+			fTclTk = 1;
+			if (fChangeleft==0) fChangeleft=1;
+			skiplines = 2;
+		}
+		if (skiplines) { skiplines--; AbsLine++; continue; }
+		filterline(buf,plain);	/* ALL LINES ARE FILTERED */
+
+#if 0
+		/* dealing with tables in formatted pages is hopeless */
+		finTable = fTable &&
+			((!ncnt && fotable) ||
+			(ncnt && bs_cnt>=2 && bs_cnt<=5 && ((float) bs_sum / (float) bs_cnt)>3.0));
+		if (finTable) {
+			if (!fotable) (*fn)(BEGINTABLE);
+		} else if (fotable) {
+			(*fn)(ENDTABLE);
+			I=I0; tagc=0; filterline(buf,plain);	/* rescan first line out of table */
+		}
+#endif
+
+		s_avg=(float) s_sum;
+		if (s_cnt>=2) {
+			/* don't count large second space gap */
+			if (scnt2) s_avg= (float) (s_sum - scnt2) / (float) (s_cnt-1);
+			else s_avg= (float) (s_sum) / (float) (s_cnt);
+		}
+
+		p=plain;	/* points to current character in plain */
+
+		/*** determine header and global indentation ***/
+		if (/*fMan && (*/!fHead || indent==-1/*)*/) {
+			if (!linelen) continue;
+			if (!*header) {
+				/* check for missing first header--but this doesn't catch subsequent pages */
+				if (stricmp(p,"NAME")==0 || stricmp(p,"NOMBRE")==0) {	/* works because line already filtered */
+					indent=scnt; /*filterline(buf,plain);*/ scnt=0; I=I0; fHead=1;
+				} else {
+					fHead=1;
+					(*fn)(BEGINHEADER);
+					/* grab header and its first word */
+					strcpy(header,p);
+					if ((header_m=HEADFOOTSKIP)>linelen) header_m=0;
+					strcpy(head,phrase); headlen=phraselen;
+					la_gets(buf); filterline(buf,plain);
+					if (linelen) {
+						strcpy(header2,plain);
+						if (strincmp(plain,"Digital",7)==0 || strincmp(plain,"OSF",3)==0) {
+							fFoot=1;
+							fSubsections=0;
+						}
+					}
+					(*fn)(ENDHEADER); tagc=0;
+					continue;
+				}
+			} else {
+				/* some idiot pages have a *third* header line, possibly after a null line */
+				if (*header && scnt>MINMID) { strcpy(header3,p); ncnt=0; continue; }
+				/* indent of first line ("NAME") after header sets global indent */
+				/* check '<' for Plan 9(?) */
+				if (*p!='<') {
+					indent=scnt; I=I0; scnt=0;
+				} else continue;
+			}
+/*			if (indent==-1) continue;*/
+		}
+		if (!lindent && scnt) lindent=scnt;
+/*printf("lindent = %d, scnt=%d\n",lindent,scnt);*/
+
+
+		/**** for each ordinary line... *****/
+
+		/*** skip over global indentation */
+		oempty=empty; empty=(linelen==0);
+		if (empty) {ncnt++; continue;}
+
+		/*** strip out per-page titles ***/
+
+		if (/*fMan && (*/scnt==0 || scnt>MINMID/*)*/) {
+/*printf("***ncnt = %d, fFoot = %d, line = %d***", ncnt,fFoot,AbsLine);*/
+			if (!fFoot && !isspace(*p) && (scnt>5 || (*p!='-' && *p!='_')) &&
+			    /* don't add ncnt -- AbsLine gets absolute line number */
+			    (((ncnt>=2 && AbsLine/*+ncnt*/>=61/*was 58*/ && AbsLine/*+ncnt*/<70)
+				 || (ncnt>=4 && AbsLine/*+ncnt*/>=59 && AbsLine/*+ncnt*/<74)
+				 || (ncnt && AbsLine/*+ncnt*/>=61 && AbsLine/*+ncnt*/<=66))
+				 && (/*lookahead!=' ' ||*/ (s_cnt>=1 && s_avg>1.1) || !falluc) )
+			    ) {
+				(*fn)(BEGINFOOTER);
+				/* grab footer and its first word */
+				strcpy(footer,p);
+/*				if ((footer_m=linelen-HEADFOOTSKIP)<0) footer_m=0;*/
+				if ((footer_m=HEADFOOTSKIP)>linelen) footer_m=0;
+				/*grabphrase(p);*/ strcpy(foot,phrase); footlen=phraselen;
+				/* permit variations at end, as for SGI "Page N", but keep minimum length */
+				if (footlen>3) footlen--;
+				la_gets(buf); filterline(buf,plain); if (linelen) strcpy(footer2,plain);
+				title=1;
+				(*fn)(ENDFOOTER); tagc=0;
+
+				/* if no header on first page, try again after first footer */
+				if (!fFoot && *header=='\0') fHead=0;	/* this is dangerous */
+				fFoot=1;
+				continue;
+			} else
+				/* a lot of work, but only for a few lines (about 4%) */
+				if (fFoot && (scnt==0 || scnt+indent>MINMID) &&
+					 (   (headlen && strncmp(head,p,headlen)==0)
+					  || strcmp(header2,p)==0 || strcmp(header3,p)==0
+					  || (footlen && strncmp(foot,p,footlen)==0)
+					  || strcmp(footer2,p)==0
+					  /* try to recognize lines with dates and page numbers */
+					  /* skip into line */
+					  || (header_m && header_m<linelen &&
+						 strncmp(&header[header_m],&p[header_m],HEADFOOTMAX)==0)
+					  || (footer_m && footer_m<linelen &&
+						 strncmp(&footer[footer_m],&p[footer_m],HEADFOOTMAX)==0)
+					  /* skip into line allowing for off-by-one */
+					  || (header_m && header_m<linelen &&
+						 strncmp(&header[header_m],&p[header_m+1],HEADFOOTMAX)==0)
+					  || (footer_m && footer_m<linelen &&
+						 strncmp(&footer[footer_m],&p[footer_m+1],HEADFOOTMAX)==0)
+					  /* or two */
+					  || (header_m && header_m<linelen &&
+						 strncmp(&header[header_m],&p[header_m+2],HEADFOOTMAX)==0)
+					  || (footer_m && footer_m<linelen &&
+						 strncmp(&footer[footer_m],&p[footer_m+2],HEADFOOTMAX)==0)
+					  /* or with reflected odd and even pages */
+					  || (headlen && headlen<linelen &&
+						 strncmp(head,&p[linelen-headlen],headlen)==0)
+					  || (footlen && footlen<linelen &&
+						 strncmp(foot,&p[linelen-footlen],footlen)==0)
+					  )) {
+				tagc=0; title=1; continue;
+			}
+
+			/* page numbers at end of line */
+			for(i=0; p[i] && isdigit(p[i]); i++)
+				/* empty */;
+			if (&p[i]!=plain && !p[i]) {title=1; fFoot=1; continue;}
+		}
+
+		/*** interline spacing ***/
+		/* multiple \n: paragraph mode=>new paragraph, line mode=>blank lines */
+		/* need to chop up lines for Roff */
+
+		/*tabgram[scnt]++;*/
+		if (title) ncnt=(scnt!=oscnt || (/*scnt<4 &&*/ isupper(*p)));
+		itabcnt = scnt/5;
+		if (CurLine==1) {ncnt=0; tagc=0;} /* gobble all newlines before first text line */
+		sect = (scnt==0 && isupper(*p));
+		subsect = (fSubsections && (scnt==2||scnt==3));
+		if ((sect || subsect) && ncnt>1) ncnt=1; /* single blank line between sections */
+		(*fn)(BEGINLINE);
+		if (/*fPara &&*/ ncnt) Pnew=1;
+		title=0; /*ncnt=0;--moved down*/
+		/*if (finTable) (*fn)(BEGINTABLELINE);*/
+		oscnt=scnt; /*fotable=finTable;*/
+
+/* let output modules decide what to do at the start of a paragraph
+		if (fPara && !Pnew && (prevcmd==BEGINBODY || prevcmd==BEGINBULTXT)) {
+			putchar(' '); I++;
+		}
+*/
+
+		/*** identify structural sections and notify fn */
+
+		/*if (fMan) {*/
+/*			bulpair = (scnt<7 && (*p==c_bullet || *p=='-'));*/
+			/* decode the below */
+			bulpair = ((!auxindent || scnt!=lindent+auxindent) /*!bulpair*/
+					 && ((scnt>=2 && scnt2>5) || scnt>=5 || (tagc>0 && tags[0].first==scnt) ) /* scnt>=2?? */
+					 && (((*p==c_bullet || strchr("-+.",*p)!=NULL || falluc) && (ncnt || scnt2>4)) || 
+					  (scnt2-s_avg>=2 && phrase[phraselen-1]!='.') ||
+					  (scnt2>3 && s_cnt==1)
+					  ));
+			if (bulpair) {
+				if (tagc>0 && tags[0].first==scnt) {
+					k=tags[0].last;
+					for (l=1; l<tagc; l++) {
+						if (tags[l].first - k <=3)
+							k=tags[l].last;
+						else break;
+					}
+					phraselen=k-scnt;
+					for (k=phraselen; plain[k]==' ' && k<linelen; k++) /* nothing */;
+					if (k>=5 && k<linelen) hanging=k; else hanging=-1;
+				} else if (scnt2) hanging=phraselen+scnt2;
+				else hanging=5;
+			} else hanging=0;
+
+/*			hanging = bulpair? phraselen+scnt2 : 0;*/
+/*if (bulpair) printf("hanging = %d\n",hanging);*/
+			/* maybe, bulpair=0 would be best */
+			/*end fMan}*/
+
+		/* certain sections (subsections too?) special, like SEE ALSO */
+		/* to make canonical name as plain, all lowercase */
+		if (sect /*||subsect -- no widespread subsection names*/) {
+		  for (j=0; (sectheadid=j)<RANDOM; j++) if (strcoloncmp2(plain,'\0',sectheadname[j],0)) break;
+		}
+
+		/* normalized section headers are put into mixed case */
+		if (/*fNORM &&*/falluc && (sect || subsect)) casify(plain);
+
+		if (sect) {
+			poppush(BEGINSECTION); (*fn)(BEGINSECTHEAD);
+			tocadd(plain, BEGINSECTION, CurLine);
+		} else if (subsect && !osubsect) {
+			poppush(BEGINSUBSECTION); (*fn)(BEGINSUBSECTHEAD);
+			tocadd(plain, BEGINSUBSECTION, CurLine);
+		} else if (bulpair) {
+			/* used to be just poppush(BEGINBULPAIR); */
+			if (!Pbp) poppush(BEGINBULPAIR);
+			(poppush)(BEGINBULLET);
+			fIP=1; /*grabphrase(plain);*/
+		} else if (Pnew) {
+			poppush(BEGINBODY);
+		}
+		Pnew=0;
+		oldsectheadid = sectheadid;
+
+
+		/* move change bars to left */
+		if (fChangeleft && !fNORM) {
+			if (fPara) (*fn)(CHANGEBAR);
+			/* replace initial spaces with changebars */
+			else for (i=0; i<ccnt; i++) { /*xputchar('|'); */ (*fn)(CHANGEBAR); }
+		}
+
+		/* show initial spaces */
+		if (!fIQS && fcharout) {
+			spaceout = (scnt>ccnt)?(scnt-ccnt):0;
+			if (fILQS) { if (spaceout>=lindent) spaceout-=lindent; else spaceout=0; }
+			if (auxindent) { if (spaceout>=auxindent) spaceout-=auxindent; else spaceout=0; }
+			if (fNORM) {
+			  	if (itabcnt>0) (*fn)(ITAB);
+				for (i=0; i<(scnt%5); i++) putchar(' ');
+			} else printf("%*s",spaceout,"");
+		}
+
+
+		/*** iterate over each character in line, ***/
+		/*** handling underlining, tabbing, copyrights ***/
+
+		off=(!fIQS&&!fPara)?scnt:0;
+		for (i=0, p=plain, curtag=0, fcont=0; *p; p++,i++,fcont=0) {
+			/* interspersed presentation signals */
+			/* start tags in reverse order of addition (so structural first) */
+			if (curtag<tagc && i+I0+off==tags[curtag].first) {
+				for (r=hitxt, j=tags[curtag].last-tags[curtag].first, hitxt[j]='\0'; j; j--)
+					hitxt[j-1]=p[j-1];
+				(*fn)(tagbeginend[tags[curtag].type][0]);
+			}
+
+			/* special characters */
+			switch(*p) {
+			   case '"':
+				if (p==plain || isspace(p[-1])) { (*fn)(CHARLQUOTE); fcont=1; }
+				else if (isspace(p[1])) { (*fn)(CHARRQUOTE); fcont=1; }
+				break;
+			   case '\'':
+				if (p==plain || isspace(p[-1])) { (*fn)(CHARLSQUOTE); fcont=1; }
+				else if (isspace(p[1])) { (*fn)(CHARRSQUOTE); fcont=1; }
+				break;
+			   case '-':
+				/* check for -opt => \-opt */
+				if (p==plain || (isspace(p[-1]) && !isspace(p[1]))) {
+				  (*fn)(CHARDASH); fcont=1;
+				}
+				break;
+			}
+
+			/* troublemaker characters */
+			c = (*p)&0xff;
+			if (!fcont && fcharout) {
+				if (strchr(escchars,c)!=NULL) {
+					putchar('\\'); putchar(c); I++;
+				} else if (strchr((char *)trouble,c)!=NULL) {
+					(*fn)(c); fcont=1;
+				} else {
+					putchar(c); I++;
+				}
+			}
+
+/*default:*/
+			if (curtag<tagc && i+I0+off+1==tags[curtag].last) {
+				(*fn)(tagbeginend[tags[curtag].type][1]);
+				curtag++;
+			}
+
+			if (fIP && ((*p==' ' && i==phraselen) || *p=='\0')) {
+				p++;  /* needed but why? */
+				(*fn)(ENDBULLET); fIP=0;
+				if (*p!='\0') {
+					/*oscnt+=phraselen;*/
+					oscnt+=i;
+					for (r=p; *r==' '; r++) {
+						oscnt++;
+/*
+						i++;
+						if (fQS || !fcharout) p++;
+*/
+					}
+				}
+				p--;	/* to counteract increment in loop */
+
+				poppush(BEGINBULTXT);
+			}
+		}
+
+
+		/*** end of line in buf[] ***/
+		/*** deal with section titles, hyperlinks ***/
+
+		if (sect) { (*fn)(ENDSECTHEAD); Pnew=1; }
+		else if (subsect) { (*fn)(ENDSUBSECTHEAD); Pnew=1; }
+		else if (fIP) { (*fn)(ENDBULLET); fIP=0; poppush(BEGINBULTXT); }
+/* oscnt not right here */
+		else if (scnt+linelen+spcsqz<MINRM /*&& ncnt*/ && lookahead!='\n'
+			    && prevcmd!=BEGINBULTXT && prevcmd!=ENDSUBSECTHEAD && prevcmd!=ENDSUBSECTHEAD)
+			(*fn)(SHORTLINE);
+		osubsect=subsect;
+
+		/*if (finTable) (*fn)(ENDTABLELINE);*/
+		/*if (!fPara)*/ (*fn)(ENDLINE); tagc=0;
+		ncnt=0;
+		I0=I;	/* save I here in case skipping lines screws it up */
+	}
+
+	/* wrap up at end */
+	pop(ENDDOC);	/* clear up all tags on stack */
+	(*fn)(ENDDOC);
+}
+
+
+
+/*
+ * SOURCE CODE PARSING
+ *    for better transcription short of full nroff interpreter
+ *
+ *    Macros derived empirically, except for weird register ones that were looked up in groff
+ *
+ * to do
+ *   X fix up output modules to be aware of Source: they need to do own formatting (Tk, ASCII, ...)
+ *	X profile if's and reorder, or centralized id (fast enough already?),
+ *   X how to format tables for Tk?
+ *   + coalese lists, distinguish DL vs OL vs UL,
+ *   + recognize more table ops (differences across flavors)
+ *   make lines with tabs into tables (tabs make no sense to HTML)
+ *
+ * buffer usage
+ *    buf = incoming text from man page file
+ *    plain = "second pass" buffer used to identify man page references
+ *
+ * test pages
+ *    Solaris: fdisk.1m, fcntl.2, curs_getwch.3x, locale.5 (numbered lists),
+ *		getservbyname.3n (font size changes)
+ */
+
+/* macros */
+/* put as much in here, as opposed to in code, as possible.
+   less expensive and here they can be overridden by other macros */
+/*const int macromax=100; -- dumb compiler */
+#define MACROMAX 1000
+struct { char *key; char *subst; } macro[MACROMAX] = {
+  /* Solaris */
+  {"NA", ".SH NAME"},
+  {"SB", "\\s-2\\fB\\$1\\fR\\s0"},
+  /* HP-UX */
+#if 0
+  {"SM", "\\s-2\\$1\\s0"},
+  {"C", "\\f3\\$1\\fR"},
+  {"CR", "\\f3\\$1\\fR\\$2"},
+  {"CI", "\\f3\\$1\\fI\\$2\\fR"},
+  {"RC", "\\fR\\$1\\f3\\$2\\fR"},
+#endif
+  /* SGI -- doesn't ship man page source */
+
+  /* 4.4BSD  -  http://intergate.sonyinteractive.com/cgi-bin/manlink/7/mdoc */
+  /* scads more, but for them definition in -mandoc is sufficient */
+#if 0
+  {"Dt", ".TH \\$1 \\$2 \\$3"},
+  {"Sh", ".SH \\$1 \\$2 \\$3 \\$4 \\$5 \\$6 \\$7 \\$8 \\$9"},
+  {"Ss", ".SS \\$1 \\$2 \\$3 \\$4 \\$5 \\$6 \\$7 \\$8 \\$9"},
+  {"Pp", ".P"},
+  {"Nm", ".BR \\$1 \\$2 \\$3 \\$4 \\$5 \\$6 \\$7 \\$8 \\$9"},	/* name */
+  {"Ar", ".IR \\$1 \\$2 \\$3 \\$4 \\$5 \\$6 \\$7 \\$8 \\$9"},
+#endif
+  { NULL, NULL }
+};
+/* what all can be represented as a macro? */
+int macrocnt=-1;	/* length of table counted at start */
+
+char *macnotfound[MACROMAX];
+int macnotcnt=0;
+
+#define SUBSTMAX 1000
+/* "*" or "*(" prefixed */
+struct { char *key; char *subst; } subst[SUBSTMAX] = {
+  {"rq", "'"}, {"lq", "`"}, {"L\"", "``"}, {"R\"", "''"},
+  {"L", "\\f3"}, {"E", "\\f2"}, {"V", "\\f4"}, {"O", "\\f1"}
+};
+int substcnt=8;
+
+#define REGISTERMAX 1000
+struct { char *name; char *value; } reg[REGISTERMAX];
+int regcnt=0;
+
+/* special characters */
+const struct { char key[4]; unsigned char subst[4]; } spec[] = {
+  { "**", "*" },
+  { "+-", { CHARPLUSMINUS, '\0' }},
+  { "12", { CHAR12, '\0' }},
+  { "14", { CHAR14, '\0' }},
+  { "34", { CHAR34, '\0' }},
+  { "aa", { CHARACUTE, '\0' }},
+  { "ap", "~" },
+  { "br", "|" },
+  { "bu", { CHARBULLET, '\0' }},
+  { "bv", "|" },
+  { "ci", "O" },
+  { "co", { CHARCOPYR, '\0' }},
+  { "ct", { CHARCENT, '\0' }},
+  { "de", { CHARDEG, '\0' }},
+  { "dg", { CHARDAGGER, '\0' }},
+  { "di", { CHARDIV, '\0' }},
+  { "em", "--" },
+  { "eq", "=" },
+  { "hy", "-" },
+  { "mi", "-" },
+  { "mu", { CHARMUL, '\0' }},
+  { "no", { CHARNOT, '\0' }},
+  { "or", "|" },
+  { "pl", "+" },
+  { "rg", { CHARREGTM, '\0' }},
+  { "ru", "_" },
+  { "sc", { CHARSECT, '\0' }},
+  { "sl", "/" },
+  { "ua", "^" },
+  { "ul", "_" }
+};
+#define speccnt (sizeof spec / sizeof spec[0])
+ 
+/* tbl line descriptions */
+char *tbl[20][20];	/* space enough for twenty description lines, twenty parts each */
+int tblc=0;
+int tbli;
+
+int fsourceTab = 0, fosourceTab=0;
+int supresseol=0;
+int finitDoc=0;
+int finlist=0;
+int sublevel=0;
+
+static char *
+source_gets(void)
+{
+  char *p,*q;
+  char *ret = (*in)?buf:NULL;
+  int i;
+  char *begin;
+  char tmpbuf[MAXBUF];
+  char name[3];
+
+  if (!sublevel) AbsLine++;
+
+  p = tmpbuf;
+  falluc = 1;
+  while (1) {
+    /* collect characters in line */
+    while (*in && *in!='\n') {
+	 if (p[0]=='\\' && p[1]=='\n') p+=2;	/* \<newline> */
+	 falluc = falluc && !islower(*in);
+	 *p++ = *in++;
+    }
+    if (*in) in++;
+    *p='\0';
+
+    /* normalize commands */
+    p=tmpbuf; q=buf;	/* copy from tmpbuf to buf */
+    /* no spaces between command-initiating period and command letters */
+    if (*p=='\'') { *p='.'; }	/* what's the difference? */
+    if (*p=='.') { *q++ = *p++; while (isspace(*p)) p++; }
+
+
+    /* convert lines with tabs to tables? */
+    fsourceTab=0;
+
+    /* if comment at start of line, OK */
+    /* dynamically determine iff Tcl/Tk page by scanning comments */
+    begin = p;
+    if (*p=='\\' && *(p+1)=='"') {
+	if (!fTclTk && strstr(p+1,"supplemental macros used in Tcl/Tk")!=NULL) fTclTk=1;
+	if (fTclTk)
+	    p+=2;
+    }
+
+    while (*p) {
+	 if (*p=='\t') fsourceTab++;
+	 if (*p=='\\') {
+	   p++;
+	   if (*p=='n') {
+		p++;
+		if (*p=='(') {
+		  p++; name[0]=*p++; name[1]=*p++; name[2]='\0';
+		} else {
+		  name[0]=*p++; name[1]='\0';
+		}
+		*q='0'; *(q+1)='\0';	/* defaults to 0, in case doesn't exist */
+		for (i=0; i<regcnt; i++) {
+		  if (strcmp(reg[i].name,name)==0) {
+		    strcpy(q, reg[i].value);
+		    break;
+		  }
+		}
+		q+=strlen(q);
+	   } else if (*p=='"' && (p != begin+1)) {	/* comment in Digital UNIX, OK elsewhere? */
+		*p='\0';
+		q--; while (q>buf && isspace(*q)) q--;	/* trim tailing whitespace */
+		q++; *q='\0';
+	   } else {
+		/* verbatim character (often a backslash) */
+		*q++ = '\\';	/* postpone interpretation (not the right thing but...) */
+		*q++ = *p++;
+	   }
+	 } else *q++ = *p++;
+    }
+
+    /* dumb Digital--later */
+    /*if (q-3>plain && q[-1]=='{' && q[-2]=='\\' && q[-3]==' ') q[-3]='\n'; }*/
+
+    /* close off buf */
+    *q='\0';
+
+#if 0
+    if (q>buf && q[-1]=='\\' && *in=='.') /* append next line * / else break;*/
+#endif
+    break;
+  }
+
+  /*printf("*ret = |%s|\n", ret!=NULL?ret : "NULL");*/
+  return ret;
+}
+
+
+/* dump characters from buffer, signalling right tags along the way */
+/* all this work to introduce an internal second pass to recognize man page references */
+/* now for HTTP references too */
+
+int sI=0;
+/* use int linelen from up top */
+int fFlush=1;
+
+static void
+source_flush(void)
+{
+  int i,j;
+  char *p,*q,*r;
+  int c;
+  int manoff,posn;
+
+  if (!sI) return;
+  plain[sI] = '\0';
+
+  /* flush called often enough that all man page references are at end of text to be flushed */
+  /* find man page ref */
+  if (sI>=4/*+1*/ && (plain[sI-(manoff=1)-1]==')' || plain[sI-(manoff=0)-1]==')')) {
+    for (q=&plain[sI-manoff-1-1]; q>plain && isalnum(*q) && *q!='('; q--) /* nada */;
+    if (*q=='(' && strcoloncmp(&q[1],')',vollist)) {
+	 r=q-1;
+	 if (*r==' ' && (sectheadid==SEEALSO || /*single letter volume */ *(q+2)==')' || *(q+3)==')')) r--;	/* permitted single intervening space */
+	 for ( ; r>=plain && (isalnum(*r) || strchr(manvalid,*r)!=NULL); r--) /* nada */;
+	 r++;
+	 if (isalpha(*r) && r<q) {
+	   /* got one: clear out tags and spaces to make normalized form */
+	   posn = r-plain;
+	   /*while (tagc && tags[tagc-1].first >= posn) tagc--;*/
+
+	   /* add MANREF tags */
+	   strcpy(hitxt,r);
+	   tagadd(BEGINMANREF, posn, 0);
+	   tagadd(ENDMANREF, sI-manoff-1+1, 0);
+	 }
+    }
+
+  /* HTML hyperlinks */
+  } else if (fURL && sI>=4 && (p=strstr(plain,"http"))!=NULL) {
+    i = p-plain;
+    tagadd(BEGINMANREF, i, 0);
+    for (j=0; i<sI && !isspace(*p) && *p!='"' && *p!='>'; i++,j++) hitxt[j] = *p++;
+    hitxt[j]='\0';
+    tagadd(ENDMANREF, i, 0);
+  }
+
+  if (!fFlush) return;
+
+  /* output text */
+  for (i=0,j=0,p=plain; i<sI && *p; i++,p++) {
+    if (!linelen) (*fn)(BEGINLINE);	/* issue BEGINLINE when know will be chars on line */
+
+    /* dump tags */
+    /*for ( ; j<tagc && tags[j].first == i; j++) (*fn)(tags[j].type);*/
+    for (j=0 ; j<tagc; j++) if (tags[j].first == i) (*fn)((enum command)tags[j].type);
+
+    /* dump text */
+    c = (*p)&0xff;	/* just make c unsigned? */
+    if (strchr(escchars,c)!=NULL) {
+	 xputchar('\\'); xputchar(c);
+	 if (fcharout) linelen++;
+    } else if (strchr((char *)trouble,c)!=NULL) {
+	 (*fn)(c);
+    } else if (linelen>=LINEBREAK && c==' ') { (*fn)(ENDLINE); linelen=0;
+    } else {		/* normal character */
+	 xputchar(c);
+	 if (fcharout) linelen++;
+    }
+
+    /*if (linelen>=LINEBREAK && c==' ') { (*fn)(ENDLINE); linelen=0; } -- leaves space at end of line*/
+  }
+  /* dump tags at end */
+  /*for ( ; j<tagc && tags[j].first == sI; j++) (*fn)(tags[j].type);*/
+    for (j=0 ; j<tagc; j++) if (tags[j].first==sI) (*fn)((enum command)tags[j].type);
+
+  sI=0; tagc=0;
+}
+
+
+/* source_out stuffs characters in a buffer */
+static char *
+source_out0(const char *pin, char end)
+{
+  /* stack of character formattings */
+  static enum tagtype styles[20];
+  static int style=-1;
+  int funwind=0;
+  int i, j;
+  int len;
+  int sign;
+  char *p, *pstart;
+
+  pstart = p = mystrdup(pin);
+  while (*p && *p!=end) {
+    if (*p=='\\') {	/* escape character */
+	 switch (*++p) {
+ 
+ 	 case '&':	/* no space.  used as a no-op sometimes */
+ 	 case '^':	/* 1/12 em space */
+ 	 case '|':      /* 1/6 em space */
+ 	 case '%':      /* hyphenation indicator */
+ 	   /* just ignore it */
+ 	   p++;
+ 	   break;
+ 	 case '0':	/* digit width space */
+ 	   p++;
+ 	   sputchar(' ');
+ 	   break;
+ 	 case ' ':	/* unpaddable space */
+ 	   stagadd(CHARNBSP);	/* nonbreaking space */
+ 	   /*sputchar(' ');*/
+ 	   p++;
+ 	   break;
+ 	 case 's':	/* font size change */
+	   p++;
+	   sign=1;
+	   if (*p=='-' || *p=='+') if (*p++=='-') sign=-1;
+	   intArg = sign * ((*p++)-'0');
+	   if (intArg==0) intArg = -fontdelta;	/* s0 returns to normal height */
+	   if (fontdelta || intArg) { source_flush(); (*fn)(FONTSIZE); }
+	   break;
+ 	 case 'v':	/* vertical motion */
+ 	 case 'h':	/* horizontal motion */
+ 	 case 'L':	/* vertical line */
+ 	 case 'l':	/* horizontal line */
+	   /* ignore */
+	   p++;
+	   if (*p=='\'') { p++; while (*p++!='\''); }
+	   break;
+	 case '"':	/* comment */
+	   *p='\0';	/* rest of line is comment */
+	   break;
+	 case 'f':
+	   p++;
+	   switch (*p++) {
+	   case '3': case 'B':		/* boldface */
+		styles[++style] = BOLD;
+		stagadd(BEGINBOLD);
+		break;
+	   case '2': case 'I':		/* italics */
+		styles[++style] = ITALICS;
+		stagadd(BEGINITALICS);
+		break;
+	   case '4':	 /* bolditalics mode => program code */
+		styles[++style] = BOLDITALICS;
+		stagadd(BEGINBOLDITALICS);
+		break;
+	   case '1': case '0': case 'R': case 'P':	/* back to Roman */
+		/*sputchar(' '); -- taken out; not needed, I hope */
+		funwind=1;
+		break;
+	   case '-':
+		p++;
+		break;
+	   }
+	   break;
+	 case '(':	/* multicharacter macros */
+	   p++;
+	   for (i=0; (unsigned)i<speccnt; i++) {
+		if (p[0]==spec[i].key[0] && p[1]==spec[i].key[1]) {
+		  p+=2;
+		  for (j=0; spec[i].subst[j]; j++) sputchar(spec[i].subst[j]);
+		  break;
+		}
+	   }
+	   break;
+	 case '*':	/* strings */
+	   p++;
+ 	   if (*p!='(') {	/* single character */
+ 	     for (i=0; i<substcnt; i++) {
+		  if (*p==subst[i].key[0] && subst[i].key[1]=='\0') {
+		    source_out0(subst[i].subst,'\0');
+		    break;
+		  }
+ 	     }
+ 	     p++;
+	   } else {	/* multicharacter macros */
+		p++;
+		for (i=0; i<substcnt; i++) {
+		  len = strlen(subst[i].key);
+		  if (strncmp(p,subst[i].key,len)==0) {
+		    source_out0(subst[i].subst,'\0');
+		    p+=len;
+		    break;
+		  }
+		}
+	   }
+	   break;
+#if 0
+	 case '|':
+	   stagadd(CHARNBSP);	/* nonbreaking space */
+#if 0
+	   sputchar(' ');
+#endif
+	   p++;
+#endif
+	 case 'e':		/* escape */
+	   sputchar('\\');
+	   p++;
+	   break;
+	 case 'c':
+		/* supress following carriage return-induced space */
+	   /* handled in source_gets(); ignore within line => can't because next line might start with a command */
+	   supresseol=1;
+	   p++;
+	   break;
+	 case '-':		/* minus sign */
+	   sputchar(CHARDASH);
+	   p++;
+	   break;
+#if 0
+	 case '^':
+	   /* end stylings? (found in Solaris) */
+	   p++;
+#endif
+	 case 'N':
+	   p++;
+	   if (*p == '\'') {
+	     char *tmp;
+	     p++;
+	     if ((tmp = strchr(p, '\'')) != NULL) {
+	       sputchar(atoi(p));
+	       p = tmp + 1;
+	     }
+	   }
+	   break;
+	 default:		/* unknown escaped character */
+	   sputchar(*p++);
+	 }
+
+    } else {		/* normal character */
+	 if (*p) sputchar(*p++);
+    }
+
+
+    /* unwind character formatting stack */
+    if (funwind) {
+	 for ( ; style>=0; style--) {
+	   if (styles[style]==BOLD) {
+		stagadd(ENDBOLD);
+	   } else if (styles[style]==ITALICS) {
+		stagadd(ENDITALICS);
+	   } else {
+		stagadd(ENDBOLDITALICS);
+	   }
+	 } /* else error */
+	 assert(style==-1);
+
+	 funwind=0;
+    }
+
+    /* check for man page reference and flush buffer if safe */
+    /* postpone check until after following character so catch closing tags */
+    if ((sI>=4+1 && plain[sI-1-1]==')') ||
+	   /*  (plain[sI-1]==' ' && (q=strchr(plain,' '))!=NULL && q<&plain[sI-1])) */
+	 (plain[sI-1]==' ' && !isalnum(plain[sI-1-1]))) {
+	 /* regardless, flush buffer */
+	 source_flush();
+    }
+  }
+
+  if (*p && *p!=' ') p++;		/* skip over end character */
+  free(pstart);
+  return (char *)pin + (p - pstart);
+
+}
+
+/* oh, for function overloading.  inlined by compiler, probably */
+static char *
+source_out(const char *p)
+{
+  return source_out0(p,'\0');
+}
+
+
+static char *
+source_out_word(const char *p)
+{
+  char end = ' ';
+
+  while (*p && isspace(*p)) p++;
+  if (*p=='"' /* || *p=='`' ? */) {
+    end = *p;
+    p++;
+  }
+  p = source_out0(p,end);
+  /*while (*p && isspace(*p)) p++;*/
+  return (char *)p;
+}
+
+
+static void
+source_struct(enum command cmd)
+{
+  source_out("\\fR\\s0");	/* don't let run-on stylings run past structural units */
+  source_flush();
+  if (cmd==SHORTLINE) linelen=0;
+  (*fn)(cmd);
+}
+
+#define checkcmd(str)	strcmp(cmd,str)==0
+
+int finnf=0;
+
+static void source_line(char *p);
+
+static void
+source_subfile(char *newin)
+{
+  char *p;
+  char *oldin = in;
+
+  sublevel++;
+
+  in = newin;
+  while ((p=source_gets())!=NULL) {
+    source_line(p);
+  }
+  in = oldin;
+
+  sublevel--;
+}
+
+/* have to delay acquisition of list tag */
+static void
+source_list(void)
+{
+  static int oldlisttype;	/* OK to have just one because nested lists done with RS/RE */
+  char *q;
+  int i;
+
+  /* guard against empty bullet */
+  for (i=0, q=plain; i<sI && isspace(*q); q++,i++) /* empty */;
+  if (i==sI) return;
+
+  assert(finlist);
+
+  fFlush=1;
+
+  /* try to determine type of list: DL, OL, UL */
+  q=plain; plain[sI]='\0';
+  if (((int)*q&0xff)==CHARBULLET /*||*/) {
+    listtype=UL;
+    q+=4;
+  } else {
+    if (strchr("[(",*q)) q++;
+    while (isdigit(*q)) { listtype=OL; q++; }	/* I hope this gives the right number */
+    if (*q=='.') q++;
+    if (strchr(")]",*q)) q++;
+    if (*q=='.') q++;
+    while (isspace(*q)) q++;
+    if (*q) listtype=DL;
+  }
+  oldlisttype = listtype;
+
+  /* interpretation left to output formats based on listtype (HTML: DL, OL, UL) */
+  i = sI; sI=0;
+  if (!Pbp || listtype!=oldlisttype) poppush(BEGINBULPAIR);
+  poppush(BEGINBULLET);
+  /*if (tphp) source_line(p); else source_out_word(p);*/
+  /*if (listtype!=OL && listtype!=UL)*/ sI=i;
+  source_struct(ENDBULLET); Pbul=0; /* handled immediately below */
+  poppush(BEGINBULTXT);
+
+  finlist=0;
+}
+
+static int inComment=0;
+static int isComment=0;
+
+static void
+source_command(char *p)
+{
+  static int lastif=1;
+  int mylastif;
+  char *cmd=p;
+  char *q;
+  int i,j,endch;
+  int fid;
+  struct stat fileinfo;
+  char *sobuf;
+  char *macroArg[9];
+  char endig[10];
+  int err=0;
+  char ch;
+  int tphp=0;
+  int ie=0;
+  int cond,invcond=0;
+  char delim,op;
+  char if0[80], if1[80];
+  float nif0, nif1;
+  int insertat;
+  char macrobuf[MAXBUF];		/* local so can have nested macros */
+  static char ft='\0';
+  static int fTableCenter=0;
+
+  /* should centralize command matching (binary search?), pointer bumping here
+     if for no other reason than to catch conflicts -- and allow overrides? */
+  /* parse out command */
+  while (*p && !isspace(*p)) p++;
+  if (*p) { *p='\0'; p++; }
+  /* should set up argv, argc for command arguments--it's regular enough that everyone doesn't have to do it itself */
+
+  isComment = (/*checkcmd("") ||*/ !strncmp(cmd, "\\\"", 2) || /*DEC triple dot*/checkcmd(".."));
+  if (inComment && !isComment) { source_struct(ENDCOMMENT); inComment=0; }	/* special case to handle transition */
+
+  if (isComment) {
+    /* maybe have BEGINCOMMENT, ENDCOMMENT, COMMENTLINE */
+    supresseol=0;
+    if (!inComment) { source_flush(); source_struct(BEGINCOMMENT); inComment=1; }
+    source_struct(COMMENTLINE);
+    printf("%s\n", p);	/* could contain --> or other comment closer, but unlikely */
+
+  /* structural commands */
+  } else if (checkcmd("TH")) {
+    /* sample: .TH CC 1 "Dec 1990" */
+    /* overrides command line -- should fix this */
+    if (!finitDoc) {
+	 while (isspace(*p)) p++;
+	 if (*p) {
+	   q=strchr(p, ' ');
+	   if (q!=NULL)
+	   {
+	     *q++='\0';
+	     strcpy(manName, p);
+	     for (p=q; isspace(*p); p++) /*nada*/;
+	     if (*p) {
+		   q=strchr(p,' ');
+		   if (q!=NULL) *q++='\0';
+		   strcpy(manSect, p);
+	     }
+	   }
+	 }
+	 sI=0;
+	 finitDoc=1;
+	 (*fn)(BEGINDOC);
+	 /* emit information in .TH line? */
+    } /* else complain about multiple definitions? */
+
+  } else if (checkcmd("SH") || checkcmd("Sh")) {	/* section title */
+    while (indent) { source_command("RE"); }
+    source_flush();
+
+    pop(BEGINSECTION);	/* before reset sectheadid */
+
+    if (*p) {
+	 if (*p=='"') { p++; q=p; while (*q && *q!='"') q++; *q='\0'; }
+	 finnf=0;
+	 for (j=0; (sectheadid=j)<RANDOM; j++) if (strcoloncmp2(p,'\0',sectheadname[j],0)) break;
+	 if (!finitDoc) {
+	   /* handle missing .TH */
+	   /* if secthead!=NAME -- insist on this?
+	   fprintf(stderr, "Bogus man page: no .TH or \".SH NAME\" lines\n");
+	   exit(1);
+	   */
+	   (*fn)(BEGINDOC);
+	   finitDoc=1;
+	 }
+	 poppush(BEGINSECTION); source_struct(BEGINSECTHEAD);
+	 fFlush=0;
+	 if (falluc) casify(p);
+	 source_out(p);		/* people forget the quotes around multiple words */
+	 while (isspace(plain[--sI])) /*nada*/;
+	 plain[++sI]='\0'; tocadd(plain, BEGINSECTION, CurLine);	/* flushed with source_struct above */
+	 fFlush=1;
+	 source_struct(ENDSECTHEAD);
+    }
+  } else if (checkcmd("SS")) {	/* subsection title */
+    while (indent) { source_command("RE"); }
+    source_flush();
+
+    if (*p) {
+	 if (*p=='"') { p++; q=p; while (*q && *q!='"') q++; *q='\0'; }
+	 finnf=0;
+	 source_flush();
+	 poppush(BEGINSUBSECTION); source_struct(BEGINSUBSECTHEAD);
+	 fFlush=0;
+
+	 if (falluc) casify(p);
+	 source_out(p);		/* people forget the quotes around multiple words */
+	 while (isspace(plain[--sI])) /*nada*/;
+	 plain[++sI]='\0'; tocadd(plain, BEGINSUBSECTION, CurLine);
+	 fFlush=1;
+	 source_struct(ENDSUBSECTHEAD);
+    }
+
+  } else if (checkcmd("P") || checkcmd("PP") || checkcmd("LP")) {	 /* new paragraph */
+    source_flush();
+    poppush(BEGINBODY);
+
+  } else if ((tphp=checkcmd("TP")) || (tphp=checkcmd("HP")) || checkcmd("IP") || checkcmd("LI")) {
+    /* TP, HP: indented paragraph, tag on next line (DL list) */
+    /* IP, LI: tag as argument */
+    source_flush();
+    fFlush=0;
+    finlist=1;
+    if (!tphp) { source_out_word(p); source_list(); }
+    /* lists terminated only at start of non-lists */
+  } else if (checkcmd("RS")) {	/* set indent */
+    source_struct(BEGININDENT);
+    indent++;
+  } else if (checkcmd("RE")) {
+    if (indent) indent--;
+    pop(ENDINDENT);
+    source_struct(ENDINDENT);
+#if 0
+  } else if (checkcmd("Xr")) {
+    /* 4.4BSD man ref */
+    supresseol=0;
+    p=source_out_word(p);
+    source_out("(");
+    p=source_out_word(p);
+    source_out(")");
+#endif
+
+  } else if (checkcmd("nf")) {
+    source_struct(SHORTLINE); 
+    finnf=1;
+    source_struct(BEGINCODEBLOCK);
+  } else if (checkcmd("fi")) {
+    source_struct(ENDCODEBLOCK);
+    finnf=0;
+  } else if (checkcmd("br")) {
+    source_struct(SHORTLINE);
+  } else if (checkcmd("sp") || checkcmd("SP")) {	/* blank lines */
+    /*if (!finTable) {*/
+	 if (finnf) source_struct(SHORTLINE); else source_struct(BEGINBODY);
+	 /*}*/
+  } else if (checkcmd("ta")) {	/* set tab stop(s?) */
+    /* argument is tab stop -- handle these as tables => leave to output format */
+    /* HTML handles tables but not tabs, Tk's text tabs but not tables */
+    /* does cause a linebreak */
+    stagadd(BEGINBODY);
+  } else if (checkcmd("ce")) {
+	  /* get line count, recursively filter for that many lines */
+	  if (sscanf(p, "%d", &i)) {
+		  source_struct(BEGINCENTER);
+		  for (; i>0 && (p=source_gets())!=NULL; i--) source_line(p);
+		  source_struct(ENDCENTER);
+	  }
+
+  /* limited selection of control structures */
+  } else if (checkcmd("if") || (checkcmd("ie"))) {	/* if <test> cmd, if <test> command and else on next line */
+    supresseol=1;
+    ie = checkcmd("ie");
+    mylastif=lastif;
+
+    if (*p=='!') { invcond=1; p++; }
+
+    if (*p=='n') { cond=1; p++; }		/* masquerading as nroff the right thing to do? */
+    else if (*p=='t') { cond=0; p++; }
+    else if (*p=='(' || *p=='-' || *p=='+' || isdigit(*p)) {
+	 if (*p=='(') p++;
+	 nif0=atof(p);
+	 if (*p=='-' || *p=='+') p++; while (isdigit(*p)) p++;
+	 op = *p++;	/* operator: =, >, < */
+	 if (op==' ') {
+	   cond = (nif0!=0);
+	 } else {
+	   nif1=atoi(p);
+	   while (isdigit(*p)) p++;
+	   if (*p==')') p++;
+	   if (op=='=') cond = (nif0==nif1);
+	   else if (op=='<') cond = (nif0<nif1);
+	   else /* op=='>' -- ignore >=, <= */ cond = (nif0>nif1);
+	 }
+    } else if (!isalpha(*p)) {	/* usually quote, ^G in Digital UNIX */
+	 /* gobble up comparators between delimiters */
+	 delim = *p++;
+	 q = if0; while (*p!=delim) { *q++=*p++; } *q='\0'; p++;
+	 q = if1; while (*p!=delim) { *q++=*p++; } *q='\0'; p++;
+	 cond = (strcmp(if0,if1)==0);
+    } else cond=0;	/* a guess, seems to be right bettern than half the time */
+    if (invcond) cond=1-cond;
+    while (isspace(*p)) p++;
+
+    lastif = cond;
+    if (strncmp(p,"\\{",2)==0) {	/* rather than handle groups here, have turn on/off output flag? */
+	    p+=2; while (isspace(*p)) p++; /* {{ */
+	    while (strncmp(p,".\\}",3)!=0 || strncmp(p,"\\}",2)!=0 /*Solaris*/) {
+		    if (cond) source_line(p);
+		    if ((p=source_gets())==NULL) break;
+	    }
+    } else if (cond) source_line(p);
+
+    if (ie) source_line(source_gets());		/* do else part with prevailing lastif */
+
+    lastif=mylastif;
+
+  } else if (checkcmd("el")) {
+    mylastif=lastif;
+
+    /* should centralize gobbling of groups */
+    cond = lastif = !lastif;
+    if (strncmp(p,"\\{",2)==0) {
+	 p+=2; while (isspace(*p)) p++;
+	 while (strncmp(p,".\\}",3)!=0 || strncmp(p,"\\}",2)!=0 /*Solaris*/) {
+	   if (cond) source_line(p);
+	   if ((p=source_gets())==NULL) break;
+	 }
+    } else if (cond) source_line(p);
+
+    lastif=mylastif;
+
+  } else if (checkcmd("ig")) {	/* "ignore group" */
+    strcpy(endig,".."); if (*p) { endig[0]='.'; strcpy(&endig[1],p); }
+    while ((p=source_gets())!=NULL) {
+	 if (strcmp(p,endig)==0) break;
+	 if (!lastif) source_line(p);		/* usually ignore line, except in one weird case */
+    }
+
+
+  /* macros and substitutions */
+  } else if (checkcmd("de")) {
+    /* grab key */
+    q=p; while (*q && !isspace(*q)) q++; *q='\0';
+
+    /* if already have a macro of that name, override it */
+    /* could use a good dictionary class */
+    for (insertat=0; insertat<macrocnt; insertat++) {
+	 if (strcmp(p,macro[insertat].key)==0) break;
+    }
+    if (insertat==macrocnt) macrocnt++;
+
+    /* should replace one with same name, if one exists */
+    macro[insertat].key = mystrdup(p);
+
+    /* build up macro in plain[] ... */
+    /* everything until ".." line part of macro */
+    q=plain; i=0;
+    while ((p=source_gets())!=NULL) {
+	 if (strcmp(p,"..")==0) break;
+	 while (*p) {	/* append string, interpreting quotes along the way--just double backslash to single now */
+	   if (*p=='\\' && p[1]=='\\') p++;
+	   *q++=*p++;
+	 }
+	 *q++='\n';
+    }
+    *q='\0';
+
+    /* ... then copy once have whole thing */
+    macro[insertat].subst = mystrdup(plain);
+    /*fprintf(stderr, "defining macro %s as %s\n", macro[insertat].key, macro[insertat].subst);*/
+    sI=0;
+
+  } else if (checkcmd("rm")) {	/* remove macro definition, can have multiple arguments */
+    for (i=0; i<macrocnt; i++) {	/* moot as new definitions replace old when conflicts */
+	 if (strcmp(p,macro[i].key)) {
+	   macro[i] = macro[--macrocnt];
+	   break;
+	 }
+    }
+
+  } else if (checkcmd("ds")) {	/* text substitutions (like macros) */
+    /* /usr/sww/man/man1/CC.1 a good test of this */
+    q = strchr(p,' ');
+    if (q!=NULL) {
+	 *q='\0'; q++;
+#if 0
+	 while (*++q) /*nada*/;
+#endif
+	 if (*q=='"') q++;
+	 if (substcnt<SUBSTMAX) {
+	   subst[substcnt].key = mystrdup(p); subst[substcnt].subst = mystrdup(q); substcnt++;
+	 }
+	 /*fprintf(stderr, "defining substitution: name=%s, body=%s\n", p, q);*/
+    }
+
+  } else if (checkcmd("so")) {
+    /* assuming in .../man like nroff, source in file and execute it as nested file, */
+    /* so nested .so's OK */
+
+	err = 1;	/* assume error unless successful */
+	if (fTclTk) {
+		err=0;
+	} else if (stat(p, &fileinfo)==0) {
+		sobuf = malloc(fileinfo.st_size + 1);
+		if (sobuf!=NULL) {
+			/* suck in entire file */
+			fid = open(p, O_RDONLY);
+			if (fid!=-1) {
+				if (read(fid, sobuf, fileinfo.st_size) == fileinfo.st_size) {
+				  sobuf[fileinfo.st_size]='\0';
+				  /* dumb Digital puts \\} closers on same line */
+				  /* {{ */
+				  for (q=sobuf; (q=strstr(q," \\}"))!=NULL; q+=3) *q='\n';
+				  source_subfile(sobuf);
+				  err = 0;
+				}
+				close(fid);
+			}
+			free(sobuf);
+		}
+	}
+
+	if (err) {
+		fprintf(stderr, "%s: couldn't read in referenced file %s.\n", argv0, p);
+		if (strchr(p,'/')==NULL) {
+		  fprintf(stderr, "\tTry cd'ing into same directory of man page first.\n");
+		} else if (strchr(p,'/')==strrchr(p,'.')) {
+		  fprintf(stderr, "\tTry cd'ing into parent directory of man page first.\n");
+		} else {
+		  fprintf(stderr, "\tTry cd'ing into ancestor directory that makes relative path valid first.\n");
+		}
+		exit(1);
+	}
+
+
+  /* character formatting */
+  /* reencode m/any as macro definitions? would like to but the below don't have "words" */
+  } else if (checkcmd("ft")) {	/* change font, next char is R,I,B.  P=previous not supported */
+    if (ft=='B') stagadd(ENDBOLD); else if (ft=='I') stagadd(ENDITALICS);
+    ft = *p++;
+    if (ft=='B') stagadd(BEGINBOLD); else if (ft=='I') stagadd(BEGINITALICS);
+  } else if (checkcmd("B")) {
+    supresseol=0;
+    stagadd(BEGINBOLD); p = source_out_word(p); source_out(p); stagadd(ENDBOLD);
+    if (finnf) source_struct(SHORTLINE);
+  } else if (checkcmd("I")) {
+    supresseol=0;
+    stagadd(BEGINITALICS); p = source_out_word(p); stagadd(ENDITALICS);
+    source_out(p);
+    if (finnf) source_struct(SHORTLINE);
+  } else if (checkcmd("BI")) {
+    supresseol=0;
+    while (*p) {
+	 stagadd(BEGINBOLD); p = source_out_word(p); stagadd(ENDBOLD);
+	 if (*p) { stagadd(BEGINITALICS); p = source_out_word(p); stagadd(ENDITALICS); }
+    }
+    if (finnf) source_struct(SHORTLINE);
+  } else if (checkcmd("IB")) {
+    supresseol=0;
+    while (*p) {
+	 stagadd(BEGINITALICS); p = source_out_word(p); stagadd(ENDITALICS);
+	 if (*p) { stagadd(BEGINBOLD); p = source_out_word(p); stagadd(ENDBOLD); }
+    }
+    if (finnf) source_struct(SHORTLINE);
+  } else if (checkcmd("RB")) {
+    supresseol=0;
+    while (*p) {
+	 p = source_out_word(p);
+	 if (*p) { stagadd(BEGINBOLD); p = source_out_word(p); stagadd(ENDBOLD); }
+    }
+    if (finnf) source_struct(SHORTLINE);
+  } else if (checkcmd("BR")) {
+    supresseol=0;
+    while (*p) {
+	 stagadd(BEGINBOLD); p = source_out_word(p); stagadd(ENDBOLD);
+	 p = source_out_word(p);
+    }
+    if (finnf) source_struct(SHORTLINE);
+  } else if (checkcmd("IR")) {
+    supresseol=0;
+    while (*p) {
+	 stagadd(BEGINITALICS); p=source_out_word(p); stagadd(ENDITALICS);
+	 p=source_out_word(p);
+    }
+    if (finnf) source_struct(SHORTLINE);
+  } else if (checkcmd("RI")) {
+    supresseol=0;
+    while (*p) {
+	 p=source_out_word(p);
+	 stagadd(BEGINITALICS); p=source_out_word(p); stagadd(ENDITALICS);
+    }
+    if (finnf) source_struct(SHORTLINE);
+
+
+  /* HP-UX */
+  } else if (checkcmd("SM")) {
+    supresseol=0; source_out("\\s-1"); while (*p) p=source_out(p); source_out("\\s0");
+  } else if (checkcmd("C")) {
+    supresseol=0;
+    stagadd(BEGINCODE); while (*p) p=source_out_word(p); stagadd(ENDCODE);
+  } else if (checkcmd("CR")) {
+    supresseol=0;
+    while (*p) {
+	 stagadd(BEGINCODE); p=source_out_word(p); stagadd(ENDCODE);
+	 if (*p) p=source_out_word(p);
+    }
+  } else if (checkcmd("RC")) {
+    supresseol=0;
+    while (*p) {
+	 p=source_out_word(p);
+	 if (*p) { stagadd(BEGINCODE); p=source_out_word(p); stagadd(ENDCODE); }
+    }
+  } else if (checkcmd("CI")) {
+    supresseol=0;
+    while (*p) {
+	 stagadd(BEGINCODE); p=source_out_word(p); stagadd(ENDCODE);
+	 if (*p) { stagadd(BEGINITALICS); p=source_out_word(p); stagadd(ENDITALICS); }
+    }
+
+
+  /* tables */
+  } else if (checkcmd("TS")) {
+    tblc=0; /*tblspanmax=0;*/ tableSep='\0';		/* need to reset each time because tabbed lines (.ta) made into tables too */
+    while ((p = source_gets())!=NULL) {
+	 if ((q=strstr(p,"tab"))!=NULL) {	/* "tab(" or "tab (".  table entry separator */
+	   p=(q+3); while (isspace(*p)) p++;
+	   p++;	/* jump over '(' */
+	   tableSep=*p;
+	   continue;
+	 }
+	 if (strincmp(p,"center",strlen("center"))==0) {	/* center entire table; should look for "left" and "right", probably */
+	   fTableCenter=1; source_struct(BEGINCENTER);
+	   p+=strlen("center"); while (isspace(*p)) p++;
+	   continue;
+	 }
+	 if (p[strlen(p)-1]==';') { tblc=0; continue; }	/* HP has a prequel terminated by ';' */
+
+	 for (i=0; *p; i++,p=q) {
+	   if (*p=='.') break;
+	   if (*p=='f') p+=2;	/* DEC sets font here */
+	   q=p+1;
+	   if (strchr("lrcn",*q)==NULL) {	/* dumb DEC doesn't put spaces between them */
+		while (*q && *q!='.' && !isspace(*q)) q++;
+	   }
+	   ch=*q; *q='\0';
+	   tbl[tblc][i] = mystrdup(p);
+	   tbl[tblc][i+1] = "";	/* mark end */
+	   *q=ch;
+	   while (*q && isspace(*q)) q++;
+	 }
+	 /*if (i>tblspanmax) tblspanmax=i;*/
+	 tbl[tblc++][i]="";		/* mark end */
+	 if (*p=='.') break;
+    }
+    tbli=0;
+    source_struct(BEGINTABLE);
+
+    while ((p=source_gets())!=NULL) {
+	 if (strncmp(p,".TE",3)==0) break;
+	 if (*p=='.') { source_line(p); continue; }
+
+	 /* count number of entries on line.  if >1, can use to set tableSep */
+	 insertat=0; for (j=0; *tbl[tbli][j]; j++) if (*tbl[tbli][j]!='s') insertat++;
+	 if (!tableSep && insertat>1) { if (fsourceTab) tableSep='\t'; else tableSep='@'; }
+	 source_struct(BEGINTABLELINE);
+	 if (strcmp(p,"_")==0 || /* double line */ strcmp(p,"=")==0) {
+	   source_out(" ");
+	   /*stagadd(HR);*/	/* empty row -- need ROWSPAN for HTML */
+	   continue;
+	 }
+
+	 for (i=0; *tbl[tbli][i] && *p; i++) {
+	   tblcellspan=1;
+	   tblcellformat = tbl[tbli][i];
+	   if (*tblcellformat=='^') {	/* vertical span => blank entry */
+		tblcellformat="l";
+	   } else if (*tblcellformat=='|') {
+		/* stagadd(VBAR); */
+		continue;
+	   } else if (strchr("lrcn", *tblcellformat)==NULL) {
+		tblcellformat="l";
+		/*continue;*/
+	   }
+
+	   while (strncmp(tbl[tbli][i+1],"s",1)==0) { tblcellspan++; i++; }
+
+	   source_struct(BEGINTABLEENTRY);
+	   if (toupper(tblcellformat[1])=='B') stagadd(BEGINBOLD);
+	   else if (toupper(tblcellformat[1])=='I') stagadd(BEGINITALICS);
+	   /* not supporting DEC's w(<num><unit>) */
+
+	   if (strcmp(p,"T{")==0) {	/* DEC, HP */
+		while (strncmp(p=source_gets(),"T}",2)!=0) source_line(p);
+		p+=2; if (*p) p++;
+	   } else {
+		p = source_out0(p, tableSep);
+	   }
+	   if (toupper(tblcellformat[1])=='B') stagadd(ENDBOLD);
+	   else if (toupper(tblcellformat[1])=='I') stagadd(ENDITALICS);
+	   source_struct(ENDTABLEENTRY);
+	 }
+	 if (tbli+1<tblc) tbli++;
+	 source_struct(ENDTABLELINE);
+    }
+    source_struct(ENDTABLE);
+    if (fTableCenter) { source_struct(ENDCENTER); fTableCenter=0; }
+
+
+  } else if (checkcmd("nr")) {
+    q=p; while (*q && !isspace(*q)) q++; *q='\0'; q++;
+
+    for (insertat=0; insertat<regcnt; insertat++) {
+	 if (strcmp(reg[insertat].name,p)==0) break;
+    }
+    if (insertat==regcnt) { regcnt++; reg[insertat].name = mystrdup(p); } /* else use same name */
+    p=q;
+    if (*q=='+' || *q=='-') q++;	/* accept signed, floating point numbers */
+    if (*q=='.') q++;
+    if (!*q || isspace(*q)) { *q++='0'; *q++='\0'; }
+    while (isdigit(*q)) { q++; } *q='\0'; /* ignore units */
+    reg[insertat].value = mystrdup(p);
+
+  } else if (checkcmd("EQ")) {	/* eqn not supported */
+    source_out("\\s-1\\fBeqn not supported\\fR\\s0");
+    while ((p=source_gets())!=NULL) {
+	 if (strncmp(p,".EN",3)==0) break;
+    }
+
+
+
+  /* Tcl/Tk macros */
+  } else if (fTclTk && (checkcmd("VS") || checkcmd("VE"))) {
+	  /* nothing for sidebars */
+  } else if (fTclTk && checkcmd("OP")) {
+    source_struct(BEGINBODY);
+    for (i=0; i<3; i++) {
+	 if (fcharout) { source_out(tcltkOP[i]); source_out(": "); }
+	 stagadd(BEGINBOLD); p=source_out_word(p); stagadd(ENDBOLD); 
+	 source_struct(SHORTLINE);
+    }
+    source_struct(BEGINBODY);
+
+  } else if (fTclTk && checkcmd("BS")) {	/* box */
+    /*source_struct(HR); -- ugh, no Ouster box */
+  } else if (fTclTk && checkcmd("BE")) {
+    /*source_struct(HR);*/
+
+  } else if (fTclTk && (checkcmd("CS")||checkcmd("DS"))) {	/* code excerpt */
+    /* respect line ends, set in teletype */
+    /* source_struct(SHORTLINE); -- done as part of CS's ENDLINE */
+    finnf=1;
+    source_struct(SHORTLINE);
+    if (checkcmd("DS")) source_line(".P");
+    stagadd(BEGINCODE);
+  } else if (fTclTk && (checkcmd("CE")||checkcmd("DE"))) {
+    stagadd(ENDCODE);
+    finnf=0;
+
+  } else if (fTclTk && checkcmd("SO")) {
+    source_struct(BEGINSECTION);
+    source_struct(BEGINSECTHEAD); source_out("STANDARD OPTIONS"); source_struct(ENDSECTHEAD);
+    tblc=1; tbl[0][0]=tbl[0][1]=tbl[0][2]="l"; tbl[0][3]="";
+    source_struct(BEGINTABLE);
+    while (1) {
+	 p = source_gets();
+	 if ((strncmp(p,".SE",3))==0) break;
+	 tblcellformat = "l";
+	 source_struct(BEGINTABLELINE);
+	 if (*p=='.') {
+	   source_command(p);
+	 } else {
+	   while (*p) {
+		source_struct(BEGINTABLEENTRY);
+		p = source_out0(p, '\t');
+		source_struct(ENDTABLEENTRY);
+	   }
+	 }
+	 source_struct(ENDTABLELINE);
+    }
+    source_struct(ENDTABLE);
+    source_struct(ENDSECTION);
+
+  } else if (fTclTk && checkcmd("AP")) {	/* arguments */
+    source_struct(BEGINBODY);
+    p = source_out_word(p); source_out("   ");
+    stagadd(BEGINITALICS); p = source_out_word(p); stagadd(ENDITALICS); source_out("\t");
+    source_out("("); p = source_out_word(p); source_out(")");
+    source_struct(SHORTLINE);
+    source_out("\t");
+  } else if (fTclTk && checkcmd("AS")) {	/* arguments */
+
+#if 0
+/* let these be defined as macros.  if they're not, they're just caught as unrecognized macros */
+  } else if (checkcmd("ll") || checkcmd("IX") ||
+		   checkcmd("nh")||checkcmd("hy")||checkcmd("hc")||checkcmd("hw")	/* hyphenation */
+	) {	/* unsupported macros -- usually roff specific */
+
+    fprintf(stderr, "macro \"%s\" not supported -- ignoring\n", cmd);
+#endif
+
+  } else {		/* could be a macro definition */
+    supresseol=0;
+
+    for (i=0; i<macrocnt; i++) {
+	 if (macro[i].key == NULL) continue;	/* !!! how does this happen? */
+	 if (checkcmd(macro[i].key)) {
+
+	   /* it is, collect arguments */
+	   for (j=0; j<9; j++) macroArg[j]="";
+	   for (j=0; p!=NULL && *p && j<9; j++, p=q) {
+		endch=' '; if (*p=='"') { endch='"'; p++; }
+		q = strchr(p,endch);
+		if (q!=NULL) {
+		  *q++='\0';
+		  if (*q && endch!=' ') q++;
+		}
+		macroArg[j] = p;
+	   }
+
+	   /* instantiate that text, substituting \\[1-9]'s */
+	   p = macro[i].subst;
+	   q = macrobuf;	/* allocated on stack */
+	   while (*p) {
+		if (*p=='\\') {
+		  p++;
+		  if (*p=='t') {
+		    *q++ = '\t';
+		    p++;
+		  } else if (*p=='$' && isdigit(p[1])) {
+		    j = p[1]-'0'-1;	/* convert to ASCII and align with macroArg array */
+		    p+=2;
+		    /* *q++='"'; -- no */
+		    strcpy(q, macroArg[j]); q += strlen(q);
+		    /* *q++='"'; -- no */
+
+		  } else {
+		    *q++ = '\\';
+		  }
+		} else {
+		  *q++ = *p++;
+		}
+	   }
+	   *q='\0';
+
+	   /* execute that text */
+	   /*fprintf(stderr, "for macro %s, substituted text is \n%s\n", macro[i].key, macrobuf);*/
+	   source_subfile(macrobuf);
+
+	   break;
+	 }
+    }
+
+    /* macro not found */
+    if (i==macrocnt) {
+	   /* report missing macros only once */
+	   for (j=0; j<macnotcnt; j++) if (strcmp(macnotfound[j],cmd)==0) break;
+	   if (j==macnotcnt) {
+		  if (!fQuiet) fprintf(stderr, "macro \"%s\" not recognized -- ignoring\n", cmd);
+		  q = malloc(strlen(cmd)+1); strcpy(q,cmd);
+		  macnotfound[macnotcnt++] = q;
+	   }
+    }
+  } /* else command is unrecognized -- ignore it: we're not a complete [tn]roff implementation */
+
+  /* popular but meaningless commands: .ne (need <n> lines--on infinite scroll */
+}
+
+
+static void
+source_line(char *p)
+{
+  /*stagadd(BEGINLINE);*/
+  char *cmd=p;
+  if (p==NULL) return;	/* bug somewhere else, but where? */
+
+#if 0
+  if (*p!='.' && *p!='\'' && !finlist) {
+    if (fsourceTab && !fosourceTab) {
+	 tblc=1; tbli=0; tableSep='\t';
+	 tbl[0][0]=tbl[0][1]=tbl[0][2]=tbl[0][3]=tbl[0][4]=tbl[0][5]=tbl[0][6]=tbl[0][7]=tbl[0][8]="l";
+	 source_struct(BEGINTABLE); finTable=1;
+    } else if (!fsourceTab && fosourceTab) {
+	 source_struct(ENDTABLE); finTable=0;
+    }
+    fosourceTab=fsourceTab;
+  }
+#endif
+
+  if (*p=='.' /*|| *p=='\''  -- normalized */) {	/* command == starts with "." */
+    p++;
+    supresseol=1;
+    source_command(p);
+
+  } else if (!*p) {		/* blank line */
+    /*source_command("P");*/
+    ncnt=1; source_struct(BEGINLINE); ncnt=0;	/* empty line => paragraph break */
+
+#if 0
+  } else if (fsourceTab && !finlist /* && pmode */) {	/* can't handle tabs, so try tables */
+    source_struct(BEGINTABLE);
+    tblcellformat = "l";
+    do {
+	 source_struct(BEGINTABLELINE);
+	 while (*p) {
+	   source_struct(BEGINTABLEENTRY);
+	   p = source_out0(p, '\t');
+	   source_struct(ENDTABLEENTRY);
+	 }
+	 source_struct(ENDTABLELINE);
+    } while ((p=source_gets())!=NULL && fsourceTab);
+    source_struct(ENDTABLE);
+    source_line(p);
+#endif
+
+  } else {			/* otherwise normal text */
+    source_out(p);
+    if (finnf || isspace(*cmd)) source_struct(SHORTLINE);
+  }
+
+  if (!supresseol && !finnf) { source_out(" "); if (finlist) source_list(); }
+  supresseol=0;
+  /*stagadd(ENDLINE);*/
+}
+
+
+static void
+source_filter(void)
+{
+  char *p = in, *q;
+  char *oldv,*newv,*shiftp,*shiftq,*endq;
+  int lenp,lenq;
+  int i,on1,on2,nn1,nn2,first;
+  int insertcnt=0, deletecnt=0, insertcnt0;
+  int nextDiffLine=-1;
+  char diffcmd, tmpc, tmpendq;
+
+  AbsLine=0;
+
+  /* just count length of macro table! */
+  for (i=0; macro[i].key!=NULL; i++) /*empty*/;
+  macrocnt = i;
+
+  /* {{ dumb Digital puts \\} closers on same line */
+  for (p=in; (p=strstr(p," \\}"))!=NULL; p+=3) *p='\n';
+
+  sI=0;
+  /* (*fn)(BEGINDOC); -- done at .TH or first .SH */
+
+
+  /* was: source_subfile(in); */
+  while (fDiff && fgets(diffline, MAXBUF, difffd)!=NULL) {
+	  /* requirements: no context lines, no errors in files, ...
+		 change-command: 8a12,15 or 5,7c8,10 or 5,7d3
+		 < from-file-line
+		 < from-file-line...
+		 --
+		 > to-file-line
+		 > to-file-line...
+	  */
+	  for (q=diffline; ; q++) { diffcmd=*q; if (diffcmd=='a'||diffcmd=='c'||diffcmd=='d') break; }
+	  if (sscanf(diffline, "%d,%d", &on1,&on2)==1) on2=on1-1+(diffcmd=='d'||diffcmd=='c');
+	  if (sscanf(++q, "%d,%d", &nn1,&nn2)==1) nn2=nn1-1+(diffcmd=='a'||diffcmd=='c');
+
+	  deletecnt = on2-on1+1;
+	  insertcnt = nn2-nn1+1;
+
+	  nextDiffLine = nn1;
+	  /*assert(nextDiffLine>=AbsLine); -- can happen if inside a macro? */
+	  if (nextDiffLine<AbsLine) continue;
+
+	  while (AbsLine<nextDiffLine && (p=source_gets())!=NULL) {
+		  source_line(p);
+	  }
+
+	  insertcnt0=insertcnt+1;	/* eat duplicate insert lines and '---' too */
+	  diffline2[0] = '\0';
+	  while (insertcnt && deletecnt) {
+		  if (ungetc(fgetc(difffd),difffd)=='<') { fgetc(difffd); fgetc(difffd); }	/* skip '<' */		  
+		  /* fill buffer with old line -- but replace if command */
+		  /* stay away from commands -- too careful if .B <word> */
+		  do {
+			  p = oldv = fgets(diffline, MAXBUF, difffd);
+			  p[strlen(p)-1]='\0';	/* fgets's \n ending => \0 */
+			  deletecnt--;
+		  } while (deletecnt && *p=='.');	/* throw out commands in old version */
+
+		  q = newv = source_gets();
+		  insertcnt--;
+		  while (insertcnt && *q=='.') {
+			  source_line(q);
+			  insertcnt--;
+		  }
+
+		  if (*p=='.' || *q=='.') break;
+
+
+		  /* make larger chunk for better diff -- but still keep away from commands */
+		  lenp=strlen(p); lenq=strlen(q);
+		  while (deletecnt && MAXBUF-lenq>80*2) {
+			  fgetc(difffd); fgetc(difffd);	/* skip '<' */
+			  if (ungetc(fgetc(difffd),difffd)=='.') break;
+			  p=&diffline[lenp]; *p++=' '; lenp++;
+			  fgets(p, MAXBUF-lenp, difffd); p[strlen(p)-1]='\0'; lenp+=strlen(p);
+			  deletecnt--;
+		  }
+
+		  while (insertcnt && *in!='.' && MAXBUF-lenq>80*2) {
+			  if (newv!=diffline2) { strcpy(diffline2,q); newv=diffline2; }
+			  q=source_gets(); diffline2[lenq]=' '; lenq++;
+			  strcpy(&diffline2[lenq],q); lenq+=strlen(q);
+			  insertcnt--;
+		  }
+
+		  /* common endings */
+		  p = &p[strlen(oldv)]; q=&q[strlen(newv)];
+		  while (p>oldv && q>newv && p[-1]==q[-1]) { p--; q--; }
+		  if ((p>oldv && p[-1]=='\\') || (q>newv && q[-1]=='\\'))
+			  while (*p && *q && !isspace(*p)) { p++; q++; }	/* steer clear of escapes */
+		  tmpendq=*q; *p=*q='\0'; endq=q;
+
+		  p=oldv; q=newv;
+		  while (*p && *q) {
+			  /* common starts */
+			  newv=q; while (*p && *q && *p==*q) { p++; q++; }
+			  if (q>newv) {
+				  tmpc=*q; *q='\0'; source_line(newv); *q=tmpc;
+			  }
+
+			  /* too hard to read */
+			  /* difference: try to find hunk of p in remainder of q */
+			  if (strlen(p)<15 || (shiftp=strchr(&p[15],' ') /*|| shiftp-p>30*/)==NULL) break;
+			  shiftp++; /* include the space */
+			  tmpc=*shiftp; *shiftp='\0'; shiftq=strstr(q,p); *shiftp=tmpc;	/* includes space */
+			  if (shiftq!=NULL) {
+			  	/* call that part of q inserted */
+				tmpc=*shiftq; *shiftq='\0';
+				stagadd(BEGINDIFFA); source_line(q); stagadd(ENDDIFFA); source_line(" ");
+				*shiftq=tmpc; q=shiftq;
+			  } else {
+				/* call that part of p deleted */
+				shiftp--;	*shiftp='\0';	/* squash the trailing space */
+				stagadd(BEGINDIFFD); source_line(p); stagadd(ENDDIFFD); source_line(" ");
+				p=shiftp+1;
+			  }
+/*#endif*/
+		  }
+
+		  if (*p) { stagadd(BEGINDIFFD); source_line(p); stagadd(ENDDIFFD); }
+		  if (*q) { stagadd(BEGINDIFFA); source_line(q); stagadd(ENDDIFFA); }
+		  if (tmpendq!='\0') { *endq=tmpendq; source_line(endq); }
+		  source_line(" ");
+	  }
+
+	  /* even if diffcmd=='c', could still have remaining old version lines */
+	  first=1;
+	  while (deletecnt--) {
+		  fgets(diffline, MAXBUF, difffd);
+		  if (diffline[2]!='.') {
+			  if (first) { stagadd(BEGINDIFFD); first=0; }
+			  source_line(&diffline[2]);	/* don't do commands; skip initial '<' */
+		  }
+	  }
+	  if (!first) { stagadd(ENDDIFFD); source_line(" "); }
+
+	  /* skip over duplicated from old */
+	  if (diffcmd=='c') while (insertcnt0--) fgets(diffline, MAXBUF, difffd);
+
+	  /* even if diffcmd=='c', could still have remaining new version lines */
+	  first=1;
+	  nextDiffLine = AbsLine + insertcnt;
+	  while (insertcnt--) fgets(diffline, MAXBUF, difffd);	/* eat duplicate text of above */
+	  while (/*insertcnt--*/AbsLine<nextDiffLine && (p=source_gets())!=NULL) {
+		  if (first && *p!='.') { stagadd(BEGINDIFFA); first=0; }
+		  source_line(p);
+	  }
+	  if (!first) { stagadd(ENDDIFFA); source_line(" "); }
+  }
+  /* finish up remainder (identical to both) */
+  while ((p=source_gets())!=NULL) {
+	  source_line(p);
+  }
+
+  source_flush();
+  pop(ENDDOC); (*fn)(ENDDOC);
+}
+
+
+
+/*
+ * STARTUP
+ */
+
+static int
+setFilterDefaults(char *outputformat)
+{
+	static struct {
+		void (*fn)(enum command);
+		int fPara; int fQS; int fIQS; int fNOHY; int fChangeleft; int fURL; char *names;
+		/* (could set parameters at BEGINDOC call...) */
+	} format[] = {
+		/*{default	0par 0qs	0iqs	0hy	0cl	0url	"name"},*/
+		{ ASCII,		0,	0,	0,	0,	0,	0,	"ASCII" },
+		{ TkMan,		0,	0,	0,	0,	0,	0,	"TkMan" },
+		{ Tk,		0,	0,	0,	0,	0,	0,	"Tk:Tcl" },
+		{ Sections,	0,	0,	0,	0,	0,	0,	"sections" },
+		{ Roff,		0,	1,	1,	1,	-1,	0,	"roff:troff:nroff" },
+		{ HTML,		1,	1,	1,	1,	1,	1,	"HTML:WWW:htm" },
+		{ SGML,		1,	1,	1,	1,	1,	1,	"SGML:XML" },
+/*		{ Texinfo,	1,	1,	1,	1,	1,	1,	"Texinfo:info:GNU" },*/
+		{ MIME,		1,	1,	1,	1,	1,	0,	"MIME:Emacs:enriched" },
+		{ LaTeX,		1,	1,	1,	1,	1,	0,	"LaTeX:LaTeX209:209:TeX" },
+		{ LaTeX2e,	1,	1,	1,	1,	1,	0,	"LaTeX2e:2e" },
+		{ RTF,		1,	1,	1,	1,	1,	0,	"RTF" },
+		{ pod,		0,	0,	1,	0,	-1,	0,	"pod:Perl" },
+		/*{ Ensemble,	1,	1,	1,	1,	1,	0,	"Ensemble" },*/
+
+		{ PostScript,	0,	0,	0,	0,	0,	0,	"PostScript:ps" },
+		{ FrameMaker,	0,	0,	0,	0,	0,	0,	"FrameMaker:Frame:Maker:MIF" },
+
+		{ NULL,		0,	0,	0,	0,	0,	0,	NULL }
+	};
+
+	int i, found=0;
+	for (i=0; format[i].fn!=NULL; i++) {
+		if (strcoloncmp2(outputformat,'\0',format[i].names,0)) {
+			fn = format[i].fn;
+			fPara = format[i].fPara;
+			fQS = format[i].fQS;
+			fIQS = format[i].fIQS;
+			fNOHY = format[i].fNOHY;
+			fChangeleft = format[i].fChangeleft;
+			fURL = format[i].fURL;
+			found=1;
+			break;
+		}
+	}
+
+	return (found==0);
+}
+
+
+/* read in whole file.  caller responsible for freeing memory */
+static char *
+filesuck(FILE *in)
+{
+  const int inc=1024*100;	/* what's 100K these days? */
+  int len=0,cnt;
+  char *file = malloc(1);  /*NULL -- relloc on NULL not reducing to malloc on some machines? */
+
+  do {
+    file = realloc(file, len+inc+1);	/* what's the ANSI way to find the file size? */
+    cnt = fread(&file[len], 1, inc, in);
+    len+=cnt;
+  } while (cnt==inc);
+  file[len]='\0';
+
+  return file;
+}
+
+int
+main(int argc, char *argv[])
+{
+	int c;
+	int i,j;
+	char *p,*oldp;
+	int fname=0;
+	const int helpbreak=75;
+	const int helpispace=4;
+	int helplen=0;
+	int desclen;
+	char **argvch;	/* remapped argv */
+	char *argvbuf;
+	char *processing = "stdin";
+/*	FILE *macros; -- interpret -man macros? */
+
+	char strgetopt[80];
+	/* options with an arg must have a '<' in the description */
+	static struct { char letter; int arg; char *longnames; char *desc; } option[] = {
+		{ 'f', 1, "filter", " <ASCII|roff|TkMan|Tk|Sections|HTML|SGML|MIME|LaTeX|LaTeX2e|RTF|pod>" },
+		{ 'S', 0, "source", "(ource of man page passed in)" },	/* autodetected */
+		{ 'F', 0, "formatted:format", "(ormatted man page passed in)" },	/* autodetected */
+
+		{ 'r', 1, "reference:manref:ref", " <man reference printf string>" },
+		{ 'l', 1, "title", " <title printf string>" },
+		{ 'V', 1, "volumes:vol", "(olume) <colon-separated list>" },
+		{ 'U', 0, "url:urls", "(RLs as hyperlinks)" },
+
+		/* following options apply to formatted pages only */
+		{ 'b', 0, "subsections:sub", " (show subsections)" },
+		{ 'k', 0, "keep:head:foot:header:footer", "(eep head/foot)" },
+		{ 'n', 1, "name", "(ame of man page) <string>" },
+		{ 's', 1, "section:sect", "(ection) <string>" },
+		{ 'p', 0, "paragraph:para", "(aragraph mode toggle)" },
+		{ 't', 1, "tabstop:tabstops", "(abstops spacing) <number>" },
+		{ 'N', 0, "normalize:normal", "(ormalize spacing, changebars)" },
+		{ 'y', 0, "zap:nohyphens", " (zap hyphens toggle)" },
+		{ 'K', 0, "nobreak", " (declare that page has no breaks)" },	/* autodetected */
+		{ 'd', 1, "diff", "(iff) <file> (diff of old page source to incorporate)" },
+		{ 'M', 1, "message", "(essage) <text> (included verbatim at end of Name section)" },
+		/*{ 'l', 0, "number lines", "... can number lines in a pipe" } */
+		/*{ 'T', 0, "tables", "(able agressive parsing ON)" },*/
+/*		{ 'c', 0, "changeleft:changebar", "(hangebarstoleft toggle)" }, -- default is perfect */
+		/*{ 'R', 0, "reflow", "(eflow text lines)" },*/
+		{ 'R', 1, "rebus", "(ebus words for TkMan)" },
+		{ 'C', 0, "TclTk", " (enable Tcl/Tk formatting)" },	/* autodetected */
+
+		/*{ 'D', 0, "debug", "(ebugging mode)" }, -- dump unrecognized macros, e.g.*/
+		{ 'o', 0, "noop", " (no op)" },
+		{ 'O', 0, "noop", " <arg> (no op with arg)" },
+		{ 'q', 0, "quiet", "(uiet--don't report warnings)" },
+		{ 'h', 0, "help", "(elp)" },
+		/*{ '?', 0, "help", " (help)" },	-- getopt returns '?' as error flag */
+		{ 'v', 0, "version", "(ersion)" },
+		{ '\0', 0, "", NULL }
+	};
+
+	/* calculate strgetopt from options list */
+	for (i=0,p=strgetopt; option[i].letter!='\0'; i++) {
+		*p++ = option[i].letter;
+		/* check for duplicate option letters */
+		assert(strchr(strgetopt,option[i].letter)==&p[-1]);
+		if (option[i].arg) *p++=':';
+	}
+	*p='\0';
+
+	/* spot check construction of strgetopt */
+	assert(p<strgetopt+80);
+	assert(strlen(strgetopt)>10);
+	assert(strchr(strgetopt,'f')!=NULL);
+	assert(strchr(strgetopt,'v')!=NULL);
+	assert(strchr(strgetopt,':')!=NULL);
+
+#ifdef macintosh
+	extern void InitToolbox();
+	InitToolbox();
+#endif
+
+	/* count, sort exception strings */
+	for (lcexceptionslen=0; (p=lcexceptions[lcexceptionslen])!=NULL; lcexceptionslen++) /*empty*/;
+	qsort(lcexceptions, lcexceptionslen, sizeof(char*), lcexceptionscmp);
+
+	/* map long option names to single letters for switching */
+	/* (GNU probably has a reusable function to do this...) */
+	/* deep six getopt in favor of integrated long names + letters? */
+	argvch = malloc(argc * sizeof(char*));
+	p = argvbuf = malloc(argc*3 * sizeof(char));	/* either -<char>'\0' or no space used */
+	for (i=0; i<argc; i++) argvch[i]=argv[i];	/* need argvch[0] for getopt? */
+	argv0 = mystrdup(argv[0]);
+	for (i=1; i<argc; i++) {
+		if (argv[i][0]=='-' && argv[i][1]=='-') {
+			if (argv[i][2]=='\0') break;	/* end of options */
+			for (j=0; option[j].letter!='\0'; j++) {
+				if (strcoloncmp2(&argv[i][2],'\0',option[j].longnames,0)) {
+					argvch[i] = p;
+					*p++ = '-'; *p++ = option[j].letter; *p++ = '\0';
+					if (option[j].arg) i++;	/* skip arguments of options */
+					break;
+				}
+			}
+			if (option[j].letter=='\0') fprintf(stderr, "%s: unknown option %s\n", argv[0], argv[i]);
+		}
+	}
+
+
+
+	/* pass through options to set defaults for chosen format */
+	setFilterDefaults("ASCII");		/* default to ASCII (used by TkMan's Glimpse indexing */
+
+	/* initialize header/footer buffers (save room in binary) */
+	for (i=0; i<CRUFTS; i++) { *cruft[i] = '\0'; }	/* automatically done, guaranteed? */
+	/*for (i=0; i<MAXLINES; i++) { linetabcnt[i] = 0; } */
+
+	while ((c=getopt(argc,argvch,strgetopt))!=-1) {
+
+		switch (c) {
+		   case 'k': fHeadfoot=1; break;
+		   case 'b': fSubsections=1; break;
+/*		   case 'c': fChangeleft=1; break; -- obsolete */
+			/* case 'R': fReflow=1; break;*/
+		   case 'n': strcpy(manName,optarg); fname=1; break;	/* name & section for when using stdin */
+		   case 's': strcpy(manSect,optarg); break;
+		   /*case 'D': docbookpath = optarg; break;*/
+		   case 'V': vollist = optarg; break;
+		   case 'l': manTitle = optarg; break;
+		   case 'r': manRef = optarg;
+			if (strlen(manRef)==0 || strcmp(manRef,"-")==0 || strcmp(manRef,"off")==0) fmanRef=0;
+			break;
+		   case 't': TabStops=atoi(optarg); break;
+		   /*case 'T': fTable=1; break;	-- if preformatted doesn't work, if source automatic */
+		   case 'p': fPara=!fPara; break;
+		   case 'K': fFoot=1; break;
+		   case 'y': fNOHY=1; break;
+		   case 'N': fNORM=1; break;
+
+		   case 'f': /* set format */
+			if (setFilterDefaults(optarg)) {
+				fprintf(stderr, "%s: unknown format: %s\n", argv0, optarg);
+				exit(1);
+			}
+			break;
+		   case 'F': fSource=0; break;
+		   case 'S': fSource=1; break;
+
+		   case 'd':
+			difffd = fopen(optarg, "r");
+			if (difffd==NULL) { fprintf(stderr, "%s: can't open %s\n", argv0, optarg); exit(1); }
+/* read in a line at a time
+			diff = filesuck(fd);
+			fclose(fd);
+*/
+			fDiff=1;
+			break;
+
+		   case 'M': message = optarg; break;
+
+		   case 'C': fTclTk=1; break;
+		   case 'R':
+			p = malloc(strlen(optarg)+1);
+			strcpy(p, optarg);	/* string may not be in writable address space */
+			oldp = "";
+			for (; *p; oldp=p, p++) {
+				if (*oldp=='\0') rebuspat[rebuspatcnt++] = p;
+				if (*p=='|') *p='\0';
+			}
+			for (i=0; i<rebuspatcnt; i++) rebuspatlen[i] = strlen(rebuspat[i]);	/* for strnlen() */
+			break;
+
+		   case 'q': fQuiet=1; break;
+		   case 'o': /*no op*/ break;
+		   case 'O': /* no op with arg */ break;
+		   case 'h':
+			printf("rman"); helplen=strlen("rman");
+
+			/* linebreak options */
+			assert(helplen>0);
+			for (i=0; option[i].letter!='\0'; i++) {
+			  desclen = strlen(option[i].desc);
+			  if (helplen+desclen+5 > helpbreak) { printf("\n%*s",helpispace,""); helplen=helpispace; }
+			  printf(" [-%c%s]", option[i].letter, option[i].desc);
+			  helplen += desclen+5;
+			}
+			if (helplen>helpispace) printf("\n");
+			printf("%*s [<filename>]\n",helpispace,"");
+			exit(0);
+
+		   case 'v': /*case '?':*/
+			printf("PolyglotMan v" POLYGLOTMANVERSION "\n");
+			exit(0);
+
+		   default:
+			fprintf(stderr, "%s: unidentified option -%c (-h for help)\n",argvch[0],c);
+			exit(2);
+		}
+	}
+
+
+
+	/* read from given file name(s) */
+	if (optind<argc) {
+		processing = argvch[optind];
+
+		if (!fname) {	/* if no name given, create from file name */
+			/* take name from tail of path */
+			if ((p=strrchr(argvch[optind],'/'))!=NULL) p++; else p=argvch[optind];
+			strcpy(manName,p);
+
+			/* search backward from end for final dot. split there */
+			if ((p=strrchr(manName,'.'))!=NULL) {
+				strcpy(manSect,p+1);
+				*p='\0';
+			}
+		}
+
+		strcpy(plain,argvch[optind]);
+
+		if (freopen(argvch[optind], "r", stdin)==NULL) {
+			fprintf(stderr, "%s: can't open %s\n", argvch[0],argvch[optind]);
+			exit(1);
+		}
+	}
+
+	/* need to read macros, ok if fail; from /usr/lib/tmac/an => needs to be set in Makefile, maybe a searchpath */
+	/*
+	if ((macros=fopen("/usr/lib/tmac/an", "r"))!=NULL) {
+		in = File = filesuck(macros);
+		lookahead = File[0];
+		source_filter();
+		free(File);
+	}
+	*/
+
+	/* suck in whole file and just operate on pointers */
+	in = File = filesuck(stdin);
+
+
+	/* minimal check for roff source: first character dot command or apostrophe comment */
+	/* MUST initialize lookahead here, BEFORE first call to la_gets */
+	if (fSource==-1) {
+		lookahead = File[0];
+		fSource = (lookahead=='.' || lookahead=='\'' || /*dumb HP*/lookahead=='/'
+		    /* HP needs this too but causes problems || isalpha(lookahead)--use --source flag*/);
+	}
+
+	if (fDiff && (!fSource || fn!=HTML)) {
+		fprintf(stderr, "diff incorporation supported for man page source, generating HTML\n");
+		exit(1);
+	}
+
+	if (fSource) source_filter(); else preformatted_filter();
+	if (fDiff) fclose(difffd);
+	/*free(File);	-- let system clean up, perhaps more efficiently */
+
+	return 0;
+}
diff --git a/xc/extras/rman/rman.html b/xc/extras/rman/rman.html
new file mode 100644
index 000000000..b0e22587c
--- /dev/null
+++ b/xc/extras/rman/rman.html
@@ -0,0 +1,322 @@
+<HTML><HEAD><TITLE>PolyglotMan Manual Page</TITLE></HEAD>
+<BODY>
+<H1>Name</H1>
+
+PolyglotMan, rman - reverse compile man pages from formatted form to a number of source formats
+
+<H1>Synopsis</H1>
+
+rman [<I>options</I>] [<I>file</I>]
+
+<H1>Description</H1>
+
+<P><I>PolyglotMan</I> takes man pages from most of the
+popular flavors of UNIX and transforms them into any of a number of
+text source formats.  PolyglotMan was formerly known as RosettaMan. 
+The name of the binary is still called <tt>rman</tt>, for scripts
+that depend on that name; mnemonically, just think "reverse man".
+Previously <I>PolyglotMan</I> required pages to
+be formatted by nroff prior to its processing.  With version 3.0, it <i>prefers
+[tn]roff source</i> and usually produces results that are better yet.
+And source processing is the only way to translate tables.
+Source format translation is not as mature as formatted, however, so 
+try formatted translation as a backup.
+
+<P>In parsing [tn]roff source, one could implement an arbitrarily
+large subset of [tn]roff, which I did not and will not do, so the
+results can be off.  I did implement a significant subset of those use
+in man pages, however, including tbl (but not eqn), if tests, and
+general macro definitions, so usually the results look great.  If they
+don't, format the page with nroff before sending it to PolyglotMan.  If
+PolyglotMan doesn't recognize a key macro used by a large class of
+pages, however, e-mail me the source and a uuencoded nroff-formatted
+page and I'll see what I can do.  When running PolyglotMan with man
+page source that includes or redirects to other [tn]roff source using
+the .so (source or inclusion) macro, you should be in the parent
+directory of the page, since pages are written with this assumption.
+For example, if you are translating /usr/man/man1/ls.1, first cd into
+/usr/man.
+
+<P><I>PolyglotMan</I> accepts man pages from: SunOS, Sun Solaris, Hewlett-Packard HP-UX, 
+AT&amp;T System V, OSF/1 aka Digital UNIX, DEC Ultrix, SGI IRIX, Linux, 
+FreeBSD, SCO.    Source processing works for: SunOS, Sun Solaris, Hewlett-Packard HP-UX, 
+AT&amp;T System V, OSF/1 aka Digital UNIX, DEC Ultrix.
+It can produce printable ASCII-only (control characters
+stripped), section headers-only,
+Tk, TkMan, [tn]roff (traditional man page source), <!--Ensemble,--> SGML, HTML, MIME, 
+LaTeX, LaTeX2e, RTF, Perl 5 POD.  
+A modular architecture permits easy addition of additional output
+formats.</P>
+
+<P>The latest version of PolyglotMan is always available from 
+<tt>ftp://ftp.cs.berkeley.edu/ucb/people/phelps/tcltk/rman.tar.Z</tt>.
+
+
+<H1>Options</H1>
+
+<P>The following options should not be used with any others and exit PolyglotMan 
+without processing any input.
+
+<DL>
+<DT>-h|--help</DT>
+<DD>Show list of command line options and exit.</DD>
+
+<DT>-v|--version</DT>
+<DD>Show version number and exit.</DD>
+</DL>
+
+
+<P><em>You should specify the filter first, as this sets a number of parameters,
+and then specify other options.</em>
+
+<DL>
+<DT>-f|--filter &lt;ASCII|roff|TkMan|Tk|<!--Ensemble|-->Sections|HTML|SGML|MIME|LaTeX|LaTeX2e|RTF|POD&gt;</DT>
+
+<DD>Set the output filter.  Defaults to ASCII.
+<!--  If you are converting
+from formatted roff source, it is recommended that you prevent hyphenation by using
+groff, making file with the contents ".hpm 20", can reading this in
+before the roff source, e.g., groff -Tascii -man <hpm-file> <roff-source>.
+-->
+</DD>
+
+<DT>-S|--source</DT>
+<DD>PolyglotMan tries to automatically determine whether its input is source or formatted;
+use this option to declare source input.</DD>
+
+<DT>-F|--format|--formatted</DT>
+<DD>PolyglotMan tries to automatically determine whether its input is source or formatted;
+use this option to declare formatted input.</DD>
+
+<DT>-l|--title <I>printf-string</I></DT>
+<DD>In HTML mode this sets the &lt;TITLE&gt; of the man pages, given the same
+parameters as <tt>-r</tt>.</DD>
+
+<DT>-r|--reference|--manref <I>printf-string</I></DT>
+<DD>In HTML and SGML modes this sets the URL form by which to retrieve other man pages.
+The string can use two supplied parameters: the man page name and its section.
+(See the Examples section.)  If the string is null (as if set from a shell
+by "-r ''"), `-' or `off', then man page references will not be HREFs, just set in italics.
+If your printf supports XPG3 positions specifier, this can be quite flexible.</DD>
+
+<DT>-V|--volumes <I>&lt;colon-separated list&gt;</I></DT>
+<DD>Set the list of valid volumes to check against when looking for 
+cross-references to other man pages.  Defaults to <tt>1:2:3:4:5:6:7:8:9:o:l:n:p</tt>
+(volume names can be multicharacter).
+If an non-whitespace string in the page is immediately followed by a left
+parenthesis, then one of the valid volumes, and ends with optional other
+characters and then a right parenthesis--then that string is reported as
+a reference to another manual page.  If this -V string starts with an equals
+sign, then no optional characters are allowed between the match to the list of
+valids and the right parenthesis.  (This option is needed for SCO UNIX.)
+</DD>
+
+</DL>
+
+
+<P>The following options apply only when formatted pages are given as input.
+They do not apply or are always handled correctly with the source.
+
+<DL>
+<DT>-b|--subsections</DT>
+<DD>Try to recognize subsection titles in addition to section titles.
+This can cause problems on some UNIX flavors.</DD>
+
+<DT>-K|--nobreak</DT>
+<DD>Indicate manual pages don't have page breaks, so don't look for footers and headers
+around them.  (Older nroff -man macros always put in page breaks, but lately
+some vendors have realized that printout are made through troff, whereas
+nroff -man is used to format pages for reading on screen, and so have eliminated
+page breaks.)  <I>PolyglotMan</I> usually gets this right even without this flag.</DD>
+
+<DT>-k|--keep</DT>
+<DD>Keep headers and footers, as a canonical report at the end of the page.</DD>
+
+<!-- this done automatically for Tcl/Tk pages; doesn't apply for others
+<DT>-c|--changeleft</DT>
+<DD>Move changebars, such as those found in the Tcl/Tk manual pages,
+to the left.</DD>
+-->
+
+<!-- agressive parsing works so well that this option has been removed
+<DT>-m|--notaggressive</DT>
+<DD><I>Disable</I> aggressive man page parsing.  Aggressive manual,
+which is on by default, page parsing elides headers and footers,
+identifies sections and more.</DD>
+-->
+
+<DT>-n|--name <I>name</I></DT>
+<DD>Set name of man page (used in roff format).
+If the filename is given in the form "<I>name</I>.<I>section</I>", the name
+and section are automatically determined.  If the page is being parsed from
+[tn]roff source and it has a .TH line, this information is extracted from that line.</DD>
+
+<DT>-p|--paragraph</DT>
+<DD>paragraph mode toggle.  The filter determines whether lines should be linebroken
+as they were by nroff, or whether lines should be flowed together into paragraphs.
+Mainly for internal use.</DD>
+
+<DT>-s|section <I>#</I></DT>
+<DD>Set volume (aka section) number of man page (used in roff format).</DD>
+
+<!-- if in source automatic, if in preformatted really doesn't work
+<DT>-T|--tables</DT>
+<DD>Turn on aggressive table parsing.</DD>
+-->
+
+<DT>-t|--tabstops <I>#</I></DT>
+<DD>For those macros sets that use tabs in place of spaces where
+possible in order to reduce the number of characters used, set
+tabstops every <I>#</I> columns.  Defaults to 8.</DD>
+
+
+</DL>
+
+
+<H1>Notes on Filter Types</H1>
+
+<H2>ROFF</H2>
+<P>Some flavors of UNIX ship man page without [tn]roff source, making one's laser printer
+little more than a laser-powered daisy wheel.  This filer tries to intuit
+the original [tn]roff directives, which can then be recompiled by [tn]roff.</P>
+
+<H2>TkMan</H2>
+<P>TkMan, a hypertext man page browser, uses <I>PolyglotMan</I> to show
+man pages without the (usually) useless headers and footers on each
+pages.  It also collects section and (optionally) subsection heads for
+direct access from a pulldown menu.  TkMan and Tcl/Tk, the toolkit in
+which it's written, are available via anonymous ftp from
+<tt>ftp://ftp.smli.com/pub/tcl/</tt></P>
+
+<H2>Tk</H2>
+
+<P>This option outputs the text in a series of Tcl lists consisting of
+text-tags pairs, where tag names roughly correspond to HTML.  This
+output can be inserted into a Tk text widget by doing an <tt>eval
+&lt;textwidget&gt; insert end &lt;text&gt;</tt>.  This format should be relatively
+easily parsible by other programs that want both the text and the
+tags.  Also see ASCII.</P>
+
+<!-- hard to track what Ensemble's input format is
+<H2>Ensemble</H2>
+<P>Ensemble, a multimedia editor of structured documents, is currently
+being developed by the research groups of Professors Michael A. Harrison and
+Susan L. Graham at the University of California, Berkeley.  With proper
+structure and presentation specifications (schemas), the appearance of
+a manual page can be radically transformed by Ensemble.</P>
+-->
+
+<H2>ASCII</H2>
+<P>When printed on a line printer, man pages try to produce special text effects
+by overstriking characters with themselves (to produce bold) and underscores
+(underlining).  Other text processing software, such as text editors, searchers,
+and indexers, must counteract this.  The ASCII filter strips away this formatting.
+Piping nroff output through <tt>col -b</tt> also strips away this formatting,
+but it leaves behind unsightly page headers and footers.  Also see Tk.</P>
+
+<H2>Sections</H2>
+<P>Dumps section and (optionally) subsection titles.  This might be useful for
+another program that processes man pages.</P>
+
+<H2>HTML</H2>
+<P>With a simple extention to an HTTP server for Mosaic or other World Wide Web
+browser, <I>PolyglotMan</I> can produce high quality HTML on the fly.
+Several such extensions and pointers to several others are included in <I>PolyglotMan</I>'s
+<tt>contrib</tt> directory.</P>
+
+<H2>SGML</H2>
+<P>This is appoaching the Docbook DTD, but I'm hoping that someone that someone
+with a real interest in this will polish the tags generated.  Try it to see
+how close the tags are now.</P>
+
+<H2>MIME</H2>
+<P>MIME (Multipurpose Internet Mail Extensions) as defined by RFC 1563,
+good for consumption by MIME-aware e-mailers or as Emacs (>=19.29)
+enriched documents.</P>
+
+<H2>LaTeX and LaTeX2e</H2>
+Why not?
+
+<H2>RTF</H2>
+<P>Use output on Mac or NeXT or whatever.  Maybe take random man pages
+and integrate with NeXT's documentation system better.  Maybe NeXT has
+own man page macros that do this.</P>
+
+<H2>PostScript and FrameMaker</H2>
+<P>To produce PostScript, use <tt>groff</tt> or <tt>psroff</tt>.  To produce FrameMaker MIF,
+use FrameMaker's builtin filter.  In both cases you need <tt>[tn]roff</tt> source,
+so if you only have a formatted version of the manual page, use <I>PolyglotMan</I>'s
+roff filter first.</P>
+
+
+<H1>Examples</H1>
+
+<P>To convert the <I>formatted</I> man page named <tt>ls.1</tt> back into
+[tn]roff source form:</P>
+
+<P>
+    <tt>rman -f roff /usr/local/man/cat1/ls.1 > /usr/local/man/man1/ls.1</tt><BR>
+
+<P>Long man pages are often compressed to conserve space (compression is
+especially effective on formatted man pages as many of the characters
+are spaces).  As it is a long man page, it probably has subsections,
+which we try to separate out (some macro sets don't distinguish
+subsections well enough for <I>PolyglotMan</I> to detect them).  Let's convert
+this to LaTeX format:<BR>
+
+<P>
+    <tt>pcat /usr/catman/a_man/cat1/automount.z | rman -b -n automount -s 1 -f latex > automount.man</tt><BR>
+
+<P>Alternatively,
+
+    <tt>man 1 automount | rman -b -n automount -s 1 -f latex > automount.man</tt><BR>
+
+<P>For HTML/Mosaic users, <I>PolyglotMan</I> can, without modification of the
+source code, produce HTML links that point to other HTML man pages
+either pregenerated or generated on the fly.  First let's assume
+pregenerated HTML versions of man pages stored in <I>/usr/man/html</I>.
+Generate these one-by-one with the following form:<BR>
+
+    <tt>rman -f html -r 'http:/usr/man/html/%s.%s.html' /usr/man/cat1/ls.1 > /usr/man/html/ls.1.html</tt><BR>
+
+<P>If you've extended your HTML client to generate HTML on the fly you should use
+something like:<BR>
+
+    <tt>rman -f html -r 'http:~/bin/man2html?%s:%s' /usr/man/cat1/ls.1</tt><BR>
+
+when generating HTML.</P>
+
+
+<H1>Bugs/Incompatibilities</H1>
+
+<P><I>PolyglotMan</I> is not perfect in all cases, but it usually does a
+good job, and in any case reduces the problem of converting man pages
+to light editing.</P>
+
+<P>Tables in formatted pages, especially H-P's, aren't handled very well.
+Be sure to pass in source for the page to recognize tables.</P>
+
+<P>The man pager <I>woman</I> applies its own idea of formatting for
+man pages, which can confuse <I>PolyglotMan</I>.  Bypass <I>woman</I>
+by passing the formatted manual page text directly into
+<I>PolyglotMan</I>.</P>
+
+<P>The [tn]roff output format uses fB to turn on boldface.  If your macro set
+requires .B, you'll have to a postprocess the <I>PolyglotMan</I> output.</P>
+
+
+<H1>See Also</H1>
+
+<tt>tkman(1)</tt>, <tt>xman(1)</tt>, <tt>man(1)</tt>, <tt>man(7)</tt> or <tt>man(5)</tt> depending on your flavor of UNIX
+
+<H1>Author</H1>
+
+<P>PolyglotMan<BR>
+by Thomas A. Phelps  (<tt>phelps@ACM.org</tt>)<BR>
+developed at the<BR>
+University of California, Berkeley<BR>
+Computer Science Division
+
+<P>Manual page last updated on $Date: 2000/07/11 19:36:05 $
+
+</BODY></HTML>