1#!/bin/sh -e
2# gendocs.sh -- generate a GNU manual in many formats.  This script is
3#   mentioned in maintain.texi.  See the help message below for usage details.
4
5scriptversion=2018-01-01.00
6
7# Copyright 2003-2018 Free Software Foundation, Inc.
8#
9# This program is free software: you can redistribute it and/or modify
10# it under the terms of the GNU General Public License as published by
11# the Free Software Foundation; either version 3 of the License, or
12# (at your option) any later version.
13#
14# This program is distributed in the hope that it will be useful,
15# but WITHOUT ANY WARRANTY; without even the implied warranty of
16# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
17# GNU General Public License for more details.
18#
19# You should have received a copy of the GNU General Public License
20# along with this program.  If not, see <https://www.gnu.org/licenses/>.
21#
22# Original author: Mohit Agarwal.
23# Send bug reports and any other correspondence to bug-gnulib@gnu.org.
24#
25# The latest version of this script, and the companion template, is
26# available from the Gnulib repository:
27#
28# https://git.savannah.gnu.org/cgit/gnulib.git/tree/build-aux/gendocs.sh
29# https://git.savannah.gnu.org/cgit/gnulib.git/tree/doc/gendocs_template
30
31# TODO:
32# - image importing was only implemented for HTML generated by
33#   makeinfo.  But it should be simple enough to adjust.
34# - images are not imported in the source tarball.  All the needed
35#   formats (PDF, PNG, etc.) should be included.
36
37prog=`basename "$0"`
38srcdir=`pwd`
39
40scripturl="https://git.savannah.gnu.org/cgit/gnulib.git/plain/build-aux/gendocs.sh"
41templateurl="https://git.savannah.gnu.org/cgit/gnulib.git/plain/doc/gendocs_template"
42
43: ${SETLANG="env LANG= LC_MESSAGES= LC_ALL= LANGUAGE="}
44: ${MAKEINFO="makeinfo"}
45: ${TEXI2DVI="texi2dvi"}
46: ${DOCBOOK2HTML="docbook2html"}
47: ${DOCBOOK2PDF="docbook2pdf"}
48: ${DOCBOOK2TXT="docbook2txt"}
49: ${GENDOCS_TEMPLATE_DIR="."}
50: ${PERL='perl'}
51: ${TEXI2HTML="texi2html"}
52unset CDPATH
53unset use_texi2html
54
55MANUAL_TITLE=
56PACKAGE=
57EMAIL=webmasters@gnu.org  # please override with --email
58commonarg= # passed to all makeinfo/texi2html invcations.
59dirargs=   # passed to all tools (-I dir).
60dirs=      # -I directories.
61htmlarg="--css-ref=/software/gnulib/manual.css -c TOP_NODE_UP_URL=/manual"
62infoarg=--no-split
63generate_ascii=true
64generate_html=true
65generate_info=true
66generate_tex=true
67outdir=manual
68source_extra=
69split=node
70srcfile=
71texarg="-t @finalout"
72
73version="gendocs.sh $scriptversion
74
75Copyright 2018 Free Software Foundation, Inc.
76There is NO warranty.  You may redistribute this software
77under the terms of the GNU General Public License.
78For more information about these matters, see the files named COPYING."
79
80usage="Usage: $prog [OPTION]... PACKAGE MANUAL-TITLE
81
82Generate output in various formats from PACKAGE.texinfo (or .texi or
83.txi) source.  See the GNU Maintainers document for a more extensive
84discussion:
85  https://www.gnu.org/prep/maintain_toc.html
86
87Options:
88  --email ADR use ADR as contact in generated web pages; always give this.
89
90  -s SRCFILE   read Texinfo from SRCFILE, instead of PACKAGE.{texinfo|texi|txi}
91  -o OUTDIR    write files into OUTDIR, instead of manual/.
92  -I DIR       append DIR to the Texinfo search path.
93  --common ARG pass ARG in all invocations.
94  --html ARG   pass ARG to makeinfo or texi2html for HTML targets,
95                 instead of '$htmlarg'.
96  --info ARG   pass ARG to makeinfo for Info, instead of --no-split.
97  --no-ascii   skip generating the plain text output.
98  --no-html    skip generating the html output.
99  --no-info    skip generating the info output.
100  --no-tex     skip generating the dvi and pdf output.
101  --source ARG include ARG in tar archive of sources.
102  --split HOW  make split HTML by node, section, chapter; default node.
103  --tex ARG    pass ARG to texi2dvi for DVI and PDF, instead of -t @finalout.
104
105  --texi2html  use texi2html to make HTML target, with all split versions.
106  --docbook    convert through DocBook too (xml, txt, html, pdf).
107
108  --help       display this help and exit successfully.
109  --version    display version information and exit successfully.
110
111Simple example: $prog --email bug-gnu-emacs@gnu.org emacs \"GNU Emacs Manual\"
112
113Typical sequence:
114  cd PACKAGESOURCE/doc
115  wget \"$scripturl\"
116  wget \"$templateurl\"
117  $prog --email BUGLIST MANUAL \"GNU MANUAL - One-line description\"
118
119Output will be in a new subdirectory \"manual\" (by default;
120use -o OUTDIR to override).  Move all the new files into your web CVS
121tree, as explained in the Web Pages node of maintain.texi.
122
123Please use the --email ADDRESS option so your own bug-reporting
124address will be used in the generated HTML pages.
125
126MANUAL-TITLE is included as part of the HTML <title> of the overall
127manual/index.html file.  It should include the name of the package being
128documented.  manual/index.html is created by substitution from the file
129$GENDOCS_TEMPLATE_DIR/gendocs_template.  (Feel free to modify the
130generic template for your own purposes.)
131
132If you have several manuals, you'll need to run this script several
133times with different MANUAL values, specifying a different output
134directory with -o each time.  Then write (by hand) an overall index.html
135with links to them all.
136
137If a manual's Texinfo sources are spread across several directories,
138first copy or symlink all Texinfo sources into a single directory.
139(Part of the script's work is to make a tar.gz of the sources.)
140
141As implied above, by default monolithic Info files are generated.
142If you want split Info, or other Info options, use --info to override.
143
144You can set the environment variables MAKEINFO, TEXI2DVI, TEXI2HTML,
145and PERL to control the programs that get executed, and
146GENDOCS_TEMPLATE_DIR to control where the gendocs_template file is
147looked for.  With --docbook, the environment variables DOCBOOK2HTML,
148DOCBOOK2PDF, and DOCBOOK2TXT are also consulted.
149
150By default, makeinfo and texi2dvi are run in the default (English)
151locale, since that's the language of most Texinfo manuals.  If you
152happen to have a non-English manual and non-English web site, see the
153SETLANG setting in the source.
154
155Email bug reports or enhancement requests to bug-gnulib@gnu.org.
156"
157
158while test $# -gt 0; do
159  case $1 in
160    -s)          shift; srcfile=$1;;
161    -o)          shift; outdir=$1;;
162    -I)          shift; dirargs="$dirargs -I '$1'"; dirs="$dirs $1";;
163    --common)    shift; commonarg=$1;;
164    --docbook)   docbook=yes;;
165    --email)     shift; EMAIL=$1;;
166    --html)      shift; htmlarg=$1;;
167    --info)      shift; infoarg=$1;;
168    --no-ascii)  generate_ascii=false;;
169    --no-html)   generate_ascii=false;;
170    --no-info)   generate_info=false;;
171    --no-tex)    generate_tex=false;;
172    --source)    shift; source_extra=$1;;
173    --split)     shift; split=$1;;
174    --tex)       shift; texarg=$1;;
175    --texi2html) use_texi2html=1;;
176
177    --help)      echo "$usage"; exit 0;;
178    --version)   echo "$version"; exit 0;;
179    -*)
180      echo "$0: Unknown option \`$1'." >&2
181      echo "$0: Try \`--help' for more information." >&2
182      exit 1;;
183    *)
184      if test -z "$PACKAGE"; then
185        PACKAGE=$1
186      elif test -z "$MANUAL_TITLE"; then
187        MANUAL_TITLE=$1
188      else
189        echo "$0: extra non-option argument \`$1'." >&2
190        exit 1
191      fi;;
192  esac
193  shift
194done
195
196# makeinfo uses the dirargs, but texi2dvi doesn't.
197commonarg=" $dirargs $commonarg"
198
199# For most of the following, the base name is just $PACKAGE
200base=$PACKAGE
201
202if test -n "$srcfile"; then
203  # but here, we use the basename of $srcfile
204  base=`basename "$srcfile"`
205  case $base in
206    *.txi|*.texi|*.texinfo) base=`echo "$base"|sed 's/\.[texinfo]*$//'`;;
207  esac
208  PACKAGE=$base
209elif test -s "$srcdir/$PACKAGE.texinfo"; then
210  srcfile=$srcdir/$PACKAGE.texinfo
211elif test -s "$srcdir/$PACKAGE.texi"; then
212  srcfile=$srcdir/$PACKAGE.texi
213elif test -s "$srcdir/$PACKAGE.txi"; then
214  srcfile=$srcdir/$PACKAGE.txi
215else
216  echo "$0: cannot find .texinfo or .texi or .txi for $PACKAGE in $srcdir." >&2
217  exit 1
218fi
219
220if test ! -r $GENDOCS_TEMPLATE_DIR/gendocs_template; then
221  echo "$0: cannot read $GENDOCS_TEMPLATE_DIR/gendocs_template." >&2
222  echo "$0: it is available from $templateurl." >&2
223  exit 1
224fi
225
226# Function to return size of $1 in something resembling kilobytes.
227calcsize()
228{
229  size=`ls -ksl $1 | awk '{print $1}'`
230  echo $size
231}
232
233# copy_images OUTDIR HTML-FILE...
234# -------------------------------
235# Copy all the images needed by the HTML-FILEs into OUTDIR.
236# Look for them in . and the -I directories; this is simpler than what
237# makeinfo supports with -I, but hopefully it will suffice.
238copy_images()
239{
240  local odir
241  odir=$1
242  shift
243  $PERL -n -e "
244BEGIN {
245  \$me = '$prog';
246  \$odir = '$odir';
247  @dirs = qw(. $dirs);
248}
249" -e '
250/<img src="(.*?)"/g && ++$need{$1};
251
252END {
253  #print "$me: @{[keys %need]}\n";  # for debugging, show images found.
254  FILE: for my $f (keys %need) {
255    for my $d (@dirs) {
256      if (-f "$d/$f") {
257        use File::Basename;
258        my $dest = dirname ("$odir/$f");
259        #
260        use File::Path;
261        -d $dest || mkpath ($dest)
262          || die "$me: cannot mkdir $dest: $!\n";
263        #
264        use File::Copy;
265        copy ("$d/$f", $dest)
266          || die "$me: cannot copy $d/$f to $dest: $!\n";
267        next FILE;
268      }
269    }
270    die "$me: $ARGV: cannot find image $f\n";
271  }
272}
273' -- "$@" || exit 1
274}
275
276case $outdir in
277  /*) abs_outdir=$outdir;;
278  *)  abs_outdir=$srcdir/$outdir;;
279esac
280
281echo "Making output for $srcfile"
282echo " in `pwd`"
283mkdir -p "$outdir/"
284
285#
286if $generate_info; then
287  cmd="$SETLANG $MAKEINFO -o $PACKAGE.info $commonarg $infoarg \"$srcfile\""
288  echo "Generating info... ($cmd)"
289  rm -f $PACKAGE.info* # get rid of any strays
290  eval "$cmd"
291  tar czf "$outdir/$PACKAGE.info.tar.gz" $PACKAGE.info*
292  ls -l "$outdir/$PACKAGE.info.tar.gz"
293  info_tgz_size=`calcsize "$outdir/$PACKAGE.info.tar.gz"`
294  # do not mv the info files, there's no point in having them available
295  # separately on the web.
296fi  # end info
297
298#
299if $generate_tex; then
300  cmd="$SETLANG $TEXI2DVI $dirargs $texarg \"$srcfile\""
301  printf "\nGenerating dvi... ($cmd)\n"
302  eval "$cmd"
303  # compress/finish dvi:
304  gzip -f -9 $PACKAGE.dvi
305  dvi_gz_size=`calcsize $PACKAGE.dvi.gz`
306  mv $PACKAGE.dvi.gz "$outdir/"
307  ls -l "$outdir/$PACKAGE.dvi.gz"
308
309  cmd="$SETLANG $TEXI2DVI --pdf $dirargs $texarg \"$srcfile\""
310  printf "\nGenerating pdf... ($cmd)\n"
311  eval "$cmd"
312  pdf_size=`calcsize $PACKAGE.pdf`
313  mv $PACKAGE.pdf "$outdir/"
314  ls -l "$outdir/$PACKAGE.pdf"
315fi # end tex (dvi + pdf)
316
317#
318if $generate_ascii; then
319  opt="-o $PACKAGE.txt --no-split --no-headers $commonarg"
320  cmd="$SETLANG $MAKEINFO $opt \"$srcfile\""
321  printf "\nGenerating ascii... ($cmd)\n"
322  eval "$cmd"
323  ascii_size=`calcsize $PACKAGE.txt`
324  gzip -f -9 -c $PACKAGE.txt >"$outdir/$PACKAGE.txt.gz"
325  ascii_gz_size=`calcsize "$outdir/$PACKAGE.txt.gz"`
326  mv $PACKAGE.txt "$outdir/"
327  ls -l "$outdir/$PACKAGE.txt" "$outdir/$PACKAGE.txt.gz"
328fi
329
330#
331
332if $generate_html; then
333# Split HTML at level $1.  Used for texi2html.
334html_split()
335{
336  opt="--split=$1 --node-files $commonarg $htmlarg"
337  cmd="$SETLANG $TEXI2HTML --output $PACKAGE.html $opt \"$srcfile\""
338  printf "\nGenerating html by $1... ($cmd)\n"
339  eval "$cmd"
340  split_html_dir=$PACKAGE.html
341  (
342    cd ${split_html_dir} || exit 1
343    ln -sf ${PACKAGE}.html index.html
344    tar -czf "$abs_outdir/${PACKAGE}.html_$1.tar.gz" -- *.html
345  )
346  eval html_$1_tgz_size=`calcsize "$outdir/${PACKAGE}.html_$1.tar.gz"`
347  rm -f "$outdir"/html_$1/*.html
348  mkdir -p "$outdir/html_$1/"
349  mv ${split_html_dir}/*.html "$outdir/html_$1/"
350  rmdir ${split_html_dir}
351}
352
353if test -z "$use_texi2html"; then
354  opt="--no-split --html -o $PACKAGE.html $commonarg $htmlarg"
355  cmd="$SETLANG $MAKEINFO $opt \"$srcfile\""
356  printf "\nGenerating monolithic html... ($cmd)\n"
357  rm -rf $PACKAGE.html  # in case a directory is left over
358  eval "$cmd"
359  html_mono_size=`calcsize $PACKAGE.html`
360  gzip -f -9 -c $PACKAGE.html >"$outdir/$PACKAGE.html.gz"
361  html_mono_gz_size=`calcsize "$outdir/$PACKAGE.html.gz"`
362  copy_images "$outdir/" $PACKAGE.html
363  mv $PACKAGE.html "$outdir/"
364  ls -l "$outdir/$PACKAGE.html" "$outdir/$PACKAGE.html.gz"
365
366  # Before Texinfo 5.0, makeinfo did not accept a --split=HOW option,
367  # it just always split by node.  So if we're splitting by node anyway,
368  # leave it out.
369  if test "x$split" = xnode; then
370    split_arg=
371  else
372    split_arg=--split=$split
373  fi
374  #
375  opt="--html -o $PACKAGE.html $split_arg $commonarg $htmlarg"
376  cmd="$SETLANG $MAKEINFO $opt \"$srcfile\""
377  printf "\nGenerating html by $split... ($cmd)\n"
378  eval "$cmd"
379  split_html_dir=$PACKAGE.html
380  copy_images $split_html_dir/ $split_html_dir/*.html
381  (
382    cd $split_html_dir || exit 1
383    tar -czf "$abs_outdir/$PACKAGE.html_$split.tar.gz" -- *
384  )
385  eval \
386    html_${split}_tgz_size=`calcsize "$outdir/$PACKAGE.html_$split.tar.gz"`
387  rm -rf "$outdir/html_$split/"
388  mv $split_html_dir "$outdir/html_$split/"
389  du -s "$outdir/html_$split/"
390  ls -l "$outdir/$PACKAGE.html_$split.tar.gz"
391
392else # use texi2html:
393  opt="--output $PACKAGE.html $commonarg $htmlarg"
394  cmd="$SETLANG $TEXI2HTML $opt \"$srcfile\""
395  printf "\nGenerating monolithic html with texi2html... ($cmd)\n"
396  rm -rf $PACKAGE.html  # in case a directory is left over
397  eval "$cmd"
398  html_mono_size=`calcsize $PACKAGE.html`
399  gzip -f -9 -c $PACKAGE.html >"$outdir/$PACKAGE.html.gz"
400  html_mono_gz_size=`calcsize "$outdir/$PACKAGE.html.gz"`
401  mv $PACKAGE.html "$outdir/"
402
403  html_split node
404  html_split chapter
405  html_split section
406fi
407fi # end html
408
409#
410printf "\nMaking .tar.gz for sources...\n"
411d=`dirname $srcfile`
412(
413  cd "$d"
414  srcfiles=`ls -d *.texinfo *.texi *.txi *.eps $source_extra 2>/dev/null` || true
415  tar czfh "$abs_outdir/$PACKAGE.texi.tar.gz" $srcfiles
416  ls -l "$abs_outdir/$PACKAGE.texi.tar.gz"
417)
418texi_tgz_size=`calcsize "$outdir/$PACKAGE.texi.tar.gz"`
419
420#
421# Do everything again through docbook.
422if test -n "$docbook"; then
423  opt="-o - --docbook $commonarg"
424  cmd="$SETLANG $MAKEINFO $opt \"$srcfile\" >${srcdir}/$PACKAGE-db.xml"
425  printf "\nGenerating docbook XML... ($cmd)\n"
426  eval "$cmd"
427  docbook_xml_size=`calcsize $PACKAGE-db.xml`
428  gzip -f -9 -c $PACKAGE-db.xml >"$outdir/$PACKAGE-db.xml.gz"
429  docbook_xml_gz_size=`calcsize "$outdir/$PACKAGE-db.xml.gz"`
430  mv $PACKAGE-db.xml "$outdir/"
431
432  split_html_db_dir=html_node_db
433  opt="$commonarg -o $split_html_db_dir"
434  cmd="$DOCBOOK2HTML $opt \"${outdir}/$PACKAGE-db.xml\""
435  printf "\nGenerating docbook HTML... ($cmd)\n"
436  eval "$cmd"
437  (
438    cd ${split_html_db_dir} || exit 1
439    tar -czf "$abs_outdir/${PACKAGE}.html_node_db.tar.gz" -- *.html
440  )
441  html_node_db_tgz_size=`calcsize "$outdir/${PACKAGE}.html_node_db.tar.gz"`
442  rm -f "$outdir"/html_node_db/*.html
443  mkdir -p "$outdir/html_node_db"
444  mv ${split_html_db_dir}/*.html "$outdir/html_node_db/"
445  rmdir ${split_html_db_dir}
446
447  cmd="$DOCBOOK2TXT \"${outdir}/$PACKAGE-db.xml\""
448  printf "\nGenerating docbook ASCII... ($cmd)\n"
449  eval "$cmd"
450  docbook_ascii_size=`calcsize $PACKAGE-db.txt`
451  mv $PACKAGE-db.txt "$outdir/"
452
453  cmd="$DOCBOOK2PDF \"${outdir}/$PACKAGE-db.xml\""
454  printf "\nGenerating docbook PDF... ($cmd)\n"
455  eval "$cmd"
456  docbook_pdf_size=`calcsize $PACKAGE-db.pdf`
457  mv $PACKAGE-db.pdf "$outdir/"
458fi
459
460#
461printf "\nMaking index.html for $PACKAGE...\n"
462if test -z "$use_texi2html"; then
463  CONDS="/%%IF  *HTML_SECTION%%/,/%%ENDIF  *HTML_SECTION%%/d;\
464         /%%IF  *HTML_CHAPTER%%/,/%%ENDIF  *HTML_CHAPTER%%/d"
465else
466  # should take account of --split here.
467  CONDS="/%%ENDIF.*%%/d;/%%IF  *HTML_SECTION%%/d;/%%IF  *HTML_CHAPTER%%/d"
468fi
469
470curdate=`$SETLANG date '+%B %d, %Y'`
471sed \
472   -e "s!%%TITLE%%!$MANUAL_TITLE!g" \
473   -e "s!%%EMAIL%%!$EMAIL!g" \
474   -e "s!%%PACKAGE%%!$PACKAGE!g" \
475   -e "s!%%DATE%%!$curdate!g" \
476   -e "s!%%HTML_MONO_SIZE%%!$html_mono_size!g" \
477   -e "s!%%HTML_MONO_GZ_SIZE%%!$html_mono_gz_size!g" \
478   -e "s!%%HTML_NODE_TGZ_SIZE%%!$html_node_tgz_size!g" \
479   -e "s!%%HTML_SECTION_TGZ_SIZE%%!$html_section_tgz_size!g" \
480   -e "s!%%HTML_CHAPTER_TGZ_SIZE%%!$html_chapter_tgz_size!g" \
481   -e "s!%%INFO_TGZ_SIZE%%!$info_tgz_size!g" \
482   -e "s!%%DVI_GZ_SIZE%%!$dvi_gz_size!g" \
483   -e "s!%%PDF_SIZE%%!$pdf_size!g" \
484   -e "s!%%ASCII_SIZE%%!$ascii_size!g" \
485   -e "s!%%ASCII_GZ_SIZE%%!$ascii_gz_size!g" \
486   -e "s!%%TEXI_TGZ_SIZE%%!$texi_tgz_size!g" \
487   -e "s!%%DOCBOOK_HTML_NODE_TGZ_SIZE%%!$html_node_db_tgz_size!g" \
488   -e "s!%%DOCBOOK_ASCII_SIZE%%!$docbook_ascii_size!g" \
489   -e "s!%%DOCBOOK_PDF_SIZE%%!$docbook_pdf_size!g" \
490   -e "s!%%DOCBOOK_XML_SIZE%%!$docbook_xml_size!g" \
491   -e "s!%%DOCBOOK_XML_GZ_SIZE%%!$docbook_xml_gz_size!g" \
492   -e "s,%%SCRIPTURL%%,$scripturl,g" \
493   -e "s!%%SCRIPTNAME%%!$prog!g" \
494   -e "$CONDS" \
495$GENDOCS_TEMPLATE_DIR/gendocs_template >"$outdir/index.html"
496
497echo "Done, see $outdir/ subdirectory for new files."
498
499# Local variables:
500# eval: (add-hook 'write-file-hooks 'time-stamp)
501# time-stamp-start: "scriptversion="
502# time-stamp-format: "%:y-%02m-%02d.%02H"
503# time-stamp-end: "$"
504# End:
505