PROBLEMS revision 75584
174467SrwatsonThis file describes various problems that have been encountered in
274467Srwatsoncompiling, installing and running groff.  Suggestions for additions or
374467Srwatsonother improvements to this file are welcome.
474467Srwatson
574467Srwatson----------------------------------------------------------------------
6
7* My document says that the current year is 19100, not 2000.
8
9In groff, as in traditional troff, the yr number register yields the
10year minus 1900.  Unfortunately, there is a longstanding bug in the
11Troff User's Manual <http://cm.bell-labs.com/cm/cs/cstr/54.ps.gz>,
12which incorrectly claims that yr is the last two digits of the year.
13This claim was never true of either Unix troff or of groff.
14
15If your text looks like this:
16
17	.\" Wrong:
18	This document was formatted in 19\n(yr.
19
20you can correct it as follows:
21
22	This document was formatted in \n[year].
23
24or, if you want to be portable to older troff versions, as follows:
25
26	.nr y4 1900+\n(yr
27	This document was formatted in \n(y4.
28
29----------------------------------------------------------------------
30
31* I get lots of `numeric overflow' error messages whenever I run
32  groff; I compiled groff with AT&T C++ 2.0 with an ANSI C compiler.
33
34Make sure -DCFRONT_ANSI_BUG is included in DEFINES in the top-level
35Makefile.  If that doesn't solve the problem, define INT_MIN as
36-INT_MAX in libgroff/lib.h.
37
38----------------------------------------------------------------------
39
40* I get errors when I try to compile groff with Sun C++ version 5.0
41  or 5.1.
42
43This is a known problem; see Sun bug #4301919.  As of this writing, no
44patch is available.  Use GCC 2.95.2 or later instead.
45
46----------------------------------------------------------------------
47
48* I get errors when I try to compile groff with Sun C++ version 3 or
49  earlier.
50
51Groff requires header files that are moderately compatible with AT&T
52C++ and ANSI C.  With some versions of Sun C++, the supplied header
53files need some of the following changes to meet this requirement:
54<string.h> must declare the mem* functions, (just add `#include
55<memory.h>' to <string.h>); the first argument to fopen and freopen
56should be declared as `const char *'; the first argument to fread
57should be declared as `void *'; the first argument to fwrite should be
58declared as `const void *'; malloc should be declared to return
59`void *'; in <alloca.h>, the declaration `extern "C" { void
60*__builtin_alloca(int); }' should be added; in <sys/signal.h> the
61return type and the second argument type of signal() should be changed
62to be `void (*)(int)'.
63
64You can either change them in place, or copy them to some other
65directory and include that directory with a -I option.
66
67----------------------------------------------------------------------
68
69* The configure script fails on OS/390 Unix.
70
71There is a bug in the Language Environment (LE) whereby the test
72program for static destructors fails.  You will see the message
73'configure: error: a working C++ compiler is required'
74
75Applying PTF UQ42006 is supposed to fix this, but the test program is
76still returning the wrong value (1).  To work around this problem, you
77can comment out the following in the configure script (near line 956).
78This will effectively bypass the test (static constructors and
79destructors do actually work properly):
80
81#if { (eval echo configure:957: \"$ac_link\") 1>&5; (eval $ac_link) 2>&5; } && test -s conftest${ac_exeext} && (./conftest; exit) 2>/dev/null
82#then
83  echo "$ac_t""yes" 1>&6
84#else
85#  echo "configure: failed program was:" >&5
86#  cat conftest.$ac_ext >&5
87#  rm -fr conftest*
88#  echo "$ac_t""no" 1>&6;{ echo "configure: error: a working C++ compiler is required" 1>&2; exit 1; }
89#fi
90
91----------------------------------------------------------------------
92
93* I get errors when I try to compile groff with DEC C++.
94
95Fix the declaration of write() in <unistd.h> so that the second
96argument is a const char *.  Fix the declaration of open() in
97<sys/file.h> so that the first argument is a const char *.
98
99----------------------------------------------------------------------
100
101* On a host using Unix make (e.g. Solaris), if you are compiling for
102  multiple architectures by building in a subdirectory, the make stops
103  with a message like this:
104
105    make: Fatal error: Don't know how to make target `assert.o'
106
107  or like this:
108
109    make: Fatal error: Can't find /u/src/groff/src/include/Makefile.sub': No such file or directory
110
111This occurs because GNU make and Unix make handle VPATH differently,
112and the groff build relies on GNU make's VPATH handling.
113
114Use GNU make to work around the problem.  In Solaris 8, GNU make is
115on the Software Companion CD and is installed as /opt/sfw/bin/gmake.
116
117----------------------------------------------------------------------
118
119* On Ultrix, the make stops with the message
120
121    *** Error code 1
122
123    Stop.
124
125  for no apparent reason.
126
127Use GNU make.
128
129----------------------------------------------------------------------
130
131* I'm having problems compiling groff on 386BSD 0.1.
132
133If you're using ash as /bin/sh, you'll need the following patch.
134
135*** gendef.sh.org	Sun Jun 30 13:30:36 1991
136--- gendef.sh	Sun Feb 28 10:23:49 1993
137***************
138*** 3,9 ****
139  file=$1
140  shift
141  
142! defs="#define $1"
143  shift
144  for def
145  do
146--- 3,10 ----
147  file=$1
148  shift
149  
150! x=$1
151! defs="#define $x"
152  shift
153  for def
154  do
155
156You'll also need to change dirnamemax.c so that it doesn't use
157pathconf().
158
159----------------------------------------------------------------------
160
161* While compiling on Xenix, ranlib libgroff.a fails.
162
163The system ranlib can't handle externals longer than 40 characters.
164Use the ranlib included in demon.co.uk:/pub/xenix/g++-1.40.3a.v1
165instead.
166
167----------------------------------------------------------------------
168
169* There are many empty `Makefile.dep' files.  Is this a bug?
170
171No.  Real dependency files are created with a `make depend' call.
172
173----------------------------------------------------------------------
174
175* Groff can't handle my troff document.  It works fine with AT&T troff.
176
177Read the section on incompatibilities in gtroff(1).  Try using the -C
178option.  Alternatively there's the sed script `tmac/fixmacros.sed'
179which will attempt to edit a file of macros so that it can be used
180with groff without the -C flag.
181
182----------------------------------------------------------------------
183
184* groff -Tdvi produces dvi files that use fonts at weird
185  magnifications.
186
187Yes, it does.  You may need to compile fonts with Metafont at these
188magnifications.  The CompileFonts script in the devdvi/generate
189directory may help you to do this.  (It will take a *long* time on
190slow computers.)
191
192----------------------------------------------------------------------
193
194* pic output is not centered horizontally; pictures sometimes run off
195  the bottom of the page.
196
197The macro package you are using is not supplying appropriate
198definitions of PS and PE.  Give groff a -mpic option.
199
200----------------------------------------------------------------------
201
202* I'm having problems including PostScript illustrations using the
203  PSPIC macro.
204
205A PostScript document must meet three requirements in order to be
206included with the PSPIC macro: it must comply with the Adobe Document
207Structuring Conventions; it must contain a BoundingBox line; it must
208be ``well-behaved''.  The BoundingBox line should be of the form:
209
210  %%BoundingBox: llx lly urx ury
211
212where llx, lly, urx, ury are the coordinates of the lower left x,
213lower left y, upper right x, upper right y of the bounding box of
214marks on the page expressed as integers in the default PostScript
215coordinate system (72 units per inch, origin at bottom left corner).
216A useful tactic is to print out the illustration by itself (you may
217need to add a `showpage' at the end), and physically measure the
218bounding box.  For more detail on these requirements, read the
219specification of Encapsulated PostScript format.  (This is available
220from the Adobe file server; send a message with a body of `help' to
221ps-file-server@adobe.com.)
222
223----------------------------------------------------------------------
224
225* I've configured groff for A4 paper, but gtroff still seems to think
226  that the length of a page (as returned by `\n(.p') is 11 inches.
227
228This is intentional.  The PAGE option is used only by grops.  For
229compatibility with ditroff, the default page length in gtroff is
230always 11 inches.  The page length can be changed with the `pl'
231request.
232
233----------------------------------------------------------------------
234
235* groff produces wrapper macros for `ms' and friends which call the
236  system's original macros.  Then, to get groff's ms macro package I
237  have to use `-mgs' instead `-ms'.  Can I avoid this?
238
239Yes.  Configure and compile groff as usual, but install it with
240
241  make install tmac_wrap=""
242
243Then no wrapper files are produced, and `-ms' will use groff's `ms'
244macros.
245
246----------------------------------------------------------------------
247
248* Groff doesn't use the font names I'm used to.
249
250Use the `ftr' request.  See (g)troff(1).
251
252----------------------------------------------------------------------
253
254* I get errors using the Unix -ms macros with groff -e -C.
255
256Apply this change:
257
258*** /usr/lib/ms/ms.eqn	Tue Apr 25 02:14:28 1989
259--- ms.eqn	Sun Nov 11 10:33:59 1990
260***************
261*** 22,29 ****
262  ..
263  .	\" EN - end of a displayed equation
264  .de EN
265! .if !\\*(10 .br
266  .di
267  .rm EZ
268  .nr ZN \\n(dn
269  .if \\n(ZN>0 .if \\n(YE=0 .LP
270--- 22,30 ----
271  ..
272  .	\" EN - end of a displayed equation
273  .de EN
274! .if \\n(.k>0 .br
275  .di
276+ .ds 10 \\*(EZ\\
277  .rm EZ
278  .nr ZN \\n(dn
279  .if \\n(ZN>0 .if \\n(YE=0 .LP
280
281----------------------------------------------------------------------
282
283* gpic doesn't accept the syntax `chop N M' for chopping both ends of
284  a line.
285
286The correct syntax is `chop N chop M'.
287
288----------------------------------------------------------------------
289
290* With gpic -t, when I print `line ->; box' using a dvi to ps program,
291  the arrow head sticks through into the inside of the box.
292
293The dvi to ps program should be modified to set the line cap and line
294join parameters to 1 while printing tpic specials.
295
296----------------------------------------------------------------------
297
298* When I print the output groff -Tps, the output is always shifted up
299  by about 0.7 inches; I'm using 8.5x11 inch paper.
300
301Make sure that PAGE is defined to be `letter' in the top-level
302Makefile.
303
304----------------------------------------------------------------------
305
306* When I try to print the output of groff -Tps, I get no output at all
307  from the printer, and the log file shows the error
308  %%[ error: undefined; offendingcommand: BP ]%%
309  I'm using TranScript spooling software.
310
311This is a bug in the page reversal filter in early versions of
312TranScript.  Change the `broken' parameter in
313/usr/local/lib/groff/font/devps/DESC to 7.
314
315----------------------------------------------------------------------
316
317* When I preview groff -Tps output using the Sun OpenWindows 2.0
318  pageview program, all the pages are displayed on top of each other.
319
320This is a defect in pageview.  Change the `broken' parameter in
321/usr/local/lib/groff/font/devps/DESC to 2.
322
323----------------------------------------------------------------------
324
325* With groff -TX75, -TX100 or -X, I can only view the first page.
326
327The left mouse button brings up a menu that allows you to view other
328pages.
329
330----------------------------------------------------------------------
331
332* When I print the output of groff -Tdvi, I just get a black dot in
333  upper left corner.
334
335Some dvi drivers (notably early versions of xtex) do not correctly
336handle dvi files that use a resolution different from that used by dvi
337files produced by TeX.  Try getting a more up to date driver.
338
339----------------------------------------------------------------------
340
341* I get lots of errors when I use groff with the AT&T -mm macros.
342
343Use the groff -mm macros.
344
345----------------------------------------------------------------------
346
347* gtroff doesn't understand lines like `.ce99' with no space between
348  the name of the request or macro and the arguments.
349
350gtroff requires a space between macro or request and its arguments
351because it allows the use of long names for macros and requests.  You
352can use the -C option or the `cp' request to put gtroff into a
353compatibility mode in which it is not possible to use long names for
354macros but in which no space is required between macros and their
355arguments.  The use of compatibility mode is strongly discouraged.
356
357----------------------------------------------------------------------
358
359* gtroff gives warnings about lines like
360  .ev	\" a comment
361  (with a tab after the .ev).
362
363A tab character cannot be used as a substitute for a space character
364(except in one case: between a control character at the beginning of a
365line and the name of a macro or request).  For example, in Unix troff
366
367  .ps	\" restore the previous point size
368
369(with a tab after the .ps) will NOT restore the previous point-size;
370instead it will be silently ignored.  Since this is very likely to be
371an error, gtroff can give a warning about it.  If you want to align
372comments, you can do it like this:
373
374  .ev\"				\" a comment
375
376----------------------------------------------------------------------
377
378* I don't like the page headers and footers produced by groff -man.
379
380There seem to be many different styles of page header and footer
381produced by different versions of the -man macros.  You will need to
382modify tmac/an-old.tmac to suit your personal taste.  For example, if
383you want the center of the page header to say
384
385  System Programmer's Manual
386
387you will need to change the line
388
389  .el .ds an-extra3 \"System Programmer's Manual
390
391to
392
393  .el .ds an-extra3 System Programmer's Manual
394
395----------------------------------------------------------------------
396
397* While formatting a manual page, groff complains about not being able
398  to break lines.  The problem seems to be caused by a line like:
399  .TP \w'label'+2
400
401The -man documentation says that the default scale indicator for TP
402macro is `n'.  The groff -man macros implement this correctly, so that
403the argument will be evaluated as if it were
404
405  \w'label'n+2n
406
407The Unix -man macros don't implement this correctly (probably because
408it's hard to do in Unix troff); they just append `n' to the entire
409argument, so that it will be evaluated as if it were
410
411  \w'label'u+2n
412
413The solution is to fix the manual page:
414
415  .TP \w'label'u+2
416
417----------------------------------------------------------------------
418
419* I'm having problems formatting Ultrix man pages with groff -man.
420
421The Ultrix man pages use a number of non-standard extensions to the
422Unix man macros.  One solution is to use the Ultrix -man macros with
423groff.  Copy /usr/lib/tmac/tmac.an to
424/usr/local/share/groff/site-tmac/an.tmac and apply the following patch
425(from Frank Wortner):
426
427*** /usr/local/lib/groff/tmac/tmac.an     Wed Sep  9 12:29:28 1992
428--- /usr/lib/tmac/tmac.an       Fri Jul 24 19:58:19 1992
429***************
430*** 489,495 ****
431  .     \" make special case of shift out of italic
432  .de }S
433  .ds ]F
434! .if \\$12 .if !\\$5 .ds ]F \^
435  .ie !\\$4 .}S \\$2 \\$1 "\\$3\f\\$1\\$4\\*(]F" "\\$5" "\\$6" "\\$7" "\\$8" "\\$9"
436  .el \\$3
437  .}f
438--- 489,495 ----
439  .     \" make special case of shift out of italic
440  .de }S
441  .ds ]F
442! .if \\$12 .if !\\$5 .ds ]F\^
443  .ie !\\$4 .}S \\$2 \\$1 "\\$3\f\\$1\\$4\\*(]F" "\\$5" "\\$6" "\\$7" "\\$8" "\\$9"
444  .el \\$3
445  .}f
446
447Another possible solution is to install tmac/man.ultrix as
448/usr/local/share/groff/site-tmac/man.local.
449
450----------------------------------------------------------------------
451
452* I'm having problems formatting HP-UX 9.0 man pages with groff -man.
453
454Copy HP's tmac.an into /usr/local/share/groff/site-tmac/an.tmac, and
455either put `.cp 1' at the beginning or filter it (and any files it .so's)
456through tmac/fixmacros.sed.
457
458----------------------------------------------------------------------
459
460* On HP-UX, the compiler complains about missing symbol `alloca'.
461
462Say
463
464  export LDFLAGS=-lPW
465
466before starting the configure script.
467
468----------------------------------------------------------------------
469
470* I'm having problems formatting man pages produced by the perl
471  wrapman script.
472
473Some versions of wrapman have a superfluous blank line before the .TH
474line.  This must be deleted.  Then either use groff -C, or apply the
475following patch:
476
477*** wrapman.~2~	Sun Jan 19 12:10:24 1992
478--- wrapman	Tue Aug 10 02:06:41 1993
479***************
480*** 35,41 ****
481      $line1 .= <IN> if $line1 =~ /eval/;
482      $line1 .= <IN> if $line1 =~ /argv/;
483      $line2 = <IN>;
484!     next if $line2 eq "'di';\n";
485  
486      # Pull the old switcheroo.
487  
488--- 35,41 ----
489      $line1 .= <IN> if $line1 =~ /eval/;
490      $line1 .= <IN> if $line1 =~ /argv/;
491      $line2 = <IN>;
492!     next if $line2 eq "'di ';\n" || $line2 eq "'di';\n";
493  
494      # Pull the old switcheroo.
495  
496***************
497*** 49,56 ****
498  
499      print OUT $line1;
500      print OUT <<EOF;
501! 'di';
502! 'ig00';
503  #
504  # $header
505  #
506--- 49,58 ----
507  
508      print OUT $line1;
509      print OUT <<EOF;
510! 'di ';
511! 'ds 00 \\"';
512! 'eo ';
513! 'ig 00 ';
514  #
515  # $header
516  #
517***************
518*** 72,85 ****
519  
520      # These next few lines are legal in both Perl and nroff.
521  
522! $null.00;                       # finish .ig
523   
524  'di           \\" finish diversion--previous line must be blank
525  .nr nl 0-1    \\" fake up transition to first page again
526  .nr % 0         \\" start at page 1
527! '; __END__ ##### From here on it's a standard manual page #####
528  .TH $PROG 1 "$month $mday, 19$year"
529- .AT 3
530  .SH NAME
531  $prog \\- whatever
532  .SH SYNOPSIS
533--- 74,87 ----
534  
535      # These next few lines are legal in both Perl and nroff.
536  
537! $null.00 ;                      # finish .ig
538! 'ec \\';
539   
540  'di           \\" finish diversion--previous line must be blank
541  .nr nl 0-1    \\" fake up transition to first page again
542  .nr % 0         \\" start at page 1
543! .\\"'; __END__ ##### From here on it's a standard manual page #####
544  .TH $PROG 1 "$month $mday, 19$year"
545  .SH NAME
546  $prog \\- whatever
547  .SH SYNOPSIS
548
549
550----------------------------------------------------------------------
551
552* When I preview documents using -TX75 or -TX100, the layout is not
553  the same as when I print the document with -Tps: the line and page
554  breaks come in different places.
555
556Use groff -X -Tps.
557
558----------------------------------------------------------------------
559
560* When I try to run gxditview, I get the error:
561  Error: Widget viewport has zero width and/or height
562
563This error means you haven't correctly installed the application
564defaults file, GXditview.ad; ``make install'' does this for you
565automatically, so either you didn't do ``make install'', or you don't
566have imake configured correctly.
567
568----------------------------------------------------------------------
569
570* groff uses up an enormous amount of memory processing large files.
571  I'm using 386BSD 0.1.
572
573386BSD includes an old version of g++, 1.39, which has a bug that
574causes a major memory leak in gtroff.  Apply the following fix to g++
575and recompile groff:
576
577*** cplus-decl.c.~1~	Mon Aug  6 05:28:59 1990
578--- cplus-decl.c	Wed Jun  5 08:55:04 1991
579***************
580*** 7951,7961 ****
581  
582        /* At the end, call delete if that's what's requested.  */
583        if (TREE_GETS_DELETE (current_class_type))
584  	exprstmt = build_method_call (build1 (NOP_EXPR, TYPE_POINTER_TO (current_class_type), error_mark_node),
585  				      get_identifier (OPERATOR_DELETE_FORMAT),
586! 				      build_tree_list (NULL_TREE, integer_zero_node),
587  				      NULL_TREE, LOOKUP_NORMAL);
588        else if (TYPE_USES_VIRTUAL_BASECLASSES (current_class_type))
589  	exprstmt = build_x_delete (ptr_type_node, current_class_decl, 0);
590        else
591  	exprstmt = 0;
592--- 7951,7961 ----
593  
594        /* At the end, call delete if that's what's requested.  */
595        if (TREE_GETS_DELETE (current_class_type))
596  	exprstmt = build_method_call (build1 (NOP_EXPR, TYPE_POINTER_TO (current_class_type), error_mark_node),
597  				      get_identifier (OPERATOR_DELETE_FORMAT),
598! 				      build_tree_list (NULL_TREE, current_class_decl),
599  				      NULL_TREE, LOOKUP_NORMAL);
600        else if (TYPE_USES_VIRTUAL_BASECLASSES (current_class_type))
601  	exprstmt = build_x_delete (ptr_type_node, current_class_decl, 0);
602        else
603  	exprstmt = 0;
604
605----------------------------------------------------------------------
606
607* Where can I get grap?
608
609Ted Faber <faber@lunabase.org> has written a freely available grap:
610
611  http://www.lunabase.org/~faber/Vault/software/grap/
612
613----------------------------------------------------------------------
614
615* How can I use groff with an old LaserJet printer that doesn't work
616  with groff -Tlj4?
617
618You have at least 3 options:
619
620- use groff -Tps with GNU Ghostscript;
621
622- use groff -Tdvi with a TeX .dvi to Laserjet driver;
623
624- use groff with the LaserJet driver in Chris Lewis' psroff package
625  (available for ftp from:
626  ftp.uunet.ca:/distrib/chris_lewis/psroff3.0pl17).
627
628----------------------------------------------------------------------
629
630* Groff seems to generate level 3 Postscript, but my printer is only a
631  level 1 or 2 PostScript printer.
632
633In fact groff generates only level 1 PostScript.  The `%!PS-Adobe-3.0'
634comment at the beginning of PostScript output generated by groff
635indicates that the file conforms to version 3.0 of the Adobe Document
636Structuring Conventions.  The output generated by groff should be
637printable on any PostScript printer.  Problems with groff output's not
638printing are most often caused by the spooling system.
639
640----------------------------------------------------------------------
641
642* The \n(st and \n(sb registers don't seem to work.  I thought \w set
643  them to the height and depth of its argument, but the registers
644  always seem to be 0.
645
646\n(st and \n(sb aren't supposed to give the height and depth of the
647string rather they give the minimum and maximum vertical displacement
648of the baseline.  For example for \v'2u'\v'-3u', \n(st will be 1 and
649\n(sb will be -2.  The height and depth of the string is available in
650the \n[rst] and \n[rsb] registers: these are groff extensions.
651
652----------------------------------------------------------------------
653
654* On an SGI system, how can I make the man command use groff?
655
656From David Hinds <dhinds@allegro.stanford.edu> (some of these steps
657are unnecessary if you install with the `g' Makefile variable defined
658as empty):
659
660Create a script called 'eqn':
661
662 > #!/bin/sh
663 > if [ ${1:-""} = /usr/pub/eqnchar ] ; then shift ; fi
664 > geqn $*
665
666and a script called 'neqn':
667
668 > #!/bin/sh
669 > if [ ${1:-""} = /usr/pub/eqnchar ] ; then shift ; fi
670 > geqn -Tascii $*
671
672and do:
673
674 > ln -s gnroff nroff
675
676and edit the end of the gnroff script to be:
677
678 > rest=`echo ${1+"$@"} | sed -e 's+/usr/lib/tmac+/usr/local/lib/groff/tmac+'`
679 > exec groff -Wall -mtty-char $T $opts $rest
680
681To get PostScript output from 'man -t', you also need to create a
682'psroff' script similar to 'nroff'.  Here are the context diffs:
683
684*** /usr/local/bin/nroff        Sat Feb 13 15:51:09 1993
685--- /usr/local/bin/psroff       Sat Feb 13 17:45:46 1993
686***************
687*** 1,8 ****
688  #!/bin/sh
689! # Emulate nroff with groff.
690
691  prog="$0"
692! T=-Tascii
693  opts=
694
695  for i
696--- 1,8 ----
697  #!/bin/sh
698! # Emulate psroff with groff.
699
700  prog="$0"
701! T=-Tps
702  opts=
703
704  for i
705***************
706*** 25,30 ****
707--- 25,33 ----
708        -Tascii|-Tlatin1)
709                T=$1
710                ;;
711+       -t)
712+               # ignore -- default is send to stdout
713+               ;;
714        -T*)
715                # ignore other devices
716                ;;
717***************
718*** 49,53 ****
719  rest=`echo ${1+"$@"} | sed -e 's+/usr/lib/tmac+/usr/local/lib/groff/tmac+'`
720
721  # This shell script is intended for use with man, so warnings are
722  # probably not wanted.  Also load nroff-style character definitions.
723! exec groff -Wall -mtty-char $T $opts $rest
724--- 52,56 ----
725  rest=`echo ${1+"$@"} | sed -e 's+/usr/lib/tmac+/usr/local/lib/groff/tmac+'`
726
727  # This shell script is intended for use with man, so warnings are
728! # probably not wanted.
729! exec groff -Wall $T $opts $rest
730