#!/bin/sh -e
# gendocs.sh -- generate a GNU manual in many formats.  This script is
#   mentioned in maintain.texi.  See the help message below for usage details.

setvar scriptversion = '2016-01-01.00'

# Copyright 2003-2016 Free Software Foundation, Inc.
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program.  If not, see <http://www.gnu.org/licenses/>.
#
# Original author: Mohit Agarwal.
# Send bug reports and any other correspondence to bug-gnulib@gnu.org.
#
# The latest version of this script, and the companion template, is
# available from the Gnulib repository:
#
# http://git.savannah.gnu.org/cgit/gnulib.git/tree/build-aux/gendocs.sh
# http://git.savannah.gnu.org/cgit/gnulib.git/tree/doc/gendocs_template

# TODO:
# - image importing was only implemented for HTML generated by
#   makeinfo.  But it should be simple enough to adjust.
# - images are not imported in the source tarball.  All the needed
#   formats (PDF, PNG, etc.) should be included.

setvar prog = $(basename $0)
setvar srcdir = $(pwd)

setvar scripturl = ""http://git.savannah.gnu.org/cgit/gnulib.git/plain/build-aux/gendocs.sh""
setvar templateurl = ""http://git.savannah.gnu.org/cgit/gnulib.git/plain/doc/gendocs_template""

: ${SETLANG="env LANG= LC_MESSAGES= LC_ALL= LANGUAGE="}
: ${MAKEINFO="makeinfo"}
: ${TEXI2DVI="texi2dvi"}
: ${DOCBOOK2HTML="docbook2html"}
: ${DOCBOOK2PDF="docbook2pdf"}
: ${DOCBOOK2TXT="docbook2txt"}
: ${GENDOCS_TEMPLATE_DIR="."}
: ${PERL='perl'}
: ${TEXI2HTML="texi2html"}
unset CDPATH
unset use_texi2html

setvar version = ""gendocs.sh $scriptversion

Copyright 2016 Free Software Foundation, Inc.
There is NO warranty.  You may redistribute this software
under the terms of the GNU General Public License.
For more information about these matters, see the files named COPYING.""

setvar usage = ""Usage: $prog [OPTION]... PACKAGE MANUAL-TITLE

Generate output in various formats from PACKAGE.texinfo (or .texi or
.txi) source.  See the GNU Maintainers document for a more extensive
discussion:
  http://www.gnu.org/prep/maintain_toc.html

Options:
  --email ADR use ADR as contact in generated web pages; always give this.

  -s SRCFILE   read Texinfo from SRCFILE, instead of PACKAGE.{texinfo|texi|txi}
  -o OUTDIR    write files into OUTDIR, instead of manual/.
  -I DIR       append DIR to the Texinfo search path.
  --common ARG pass ARG in all invocations.
  --html ARG   pass ARG to makeinfo or texi2html for HTML targets,
                 instead of --css-ref=/software/gnulib/manual.css.
  --info ARG   pass ARG to makeinfo for Info, instead of --no-split.
  --no-ascii   skip generating the plain text output.
  --no-html    skip generating the html output.
  --no-info    skip generating the info output.
  --no-tex     skip generating the dvi and pdf output.
  --source ARG include ARG in tar archive of sources.
  --split HOW  make split HTML by node, section, chapter; default node.
  --tex ARG    pass ARG to texi2dvi for DVI and PDF, instead of -t @finalout.

  --texi2html  use texi2html to make HTML target, with all split versions.
  --docbook    convert through DocBook too (xml, txt, html, pdf).

  --help       display this help and exit successfully.
  --version    display version information and exit successfully.

Simple example: $prog --email bug-gnu-emacs@gnu.org emacs \"GNU Emacs Manual\"

Typical sequence:
  cd PACKAGESOURCE/doc
  wget \"$scripturl\"
  wget \"$templateurl\"
  $prog --email BUGLIST MANUAL \"GNU MANUAL - One-line description\"

Output will be in a new subdirectory \"manual\" (by default;
use -o OUTDIR to override).  Move all the new files into your web CVS
tree, as explained in the Web Pages node of maintain.texi.

Please use the --email ADDRESS option so your own bug-reporting
address will be used in the generated HTML pages.

MANUAL-TITLE is included as part of the HTML <title> of the overall
manual/index.html file.  It should include the name of the package being
documented.  manual/index.html is created by substitution from the file
$GENDOCS_TEMPLATE_DIR/gendocs_template.  (Feel free to modify the
generic template for your own purposes.)

If you have several manuals, you'll need to run this script several
times with different MANUAL values, specifying a different output
directory with -o each time.  Then write (by hand) an overall index.html
with links to them all.

If a manual's Texinfo sources are spread across several directories,
first copy or symlink all Texinfo sources into a single directory.
(Part of the script's work is to make a tar.gz of the sources.)

As implied above, by default monolithic Info files are generated.
If you want split Info, or other Info options, use --info to override.

You can set the environment variables MAKEINFO, TEXI2DVI, TEXI2HTML,
and PERL to control the programs that get executed, and
GENDOCS_TEMPLATE_DIR to control where the gendocs_template file is
looked for.  With --docbook, the environment variables DOCBOOK2HTML,
DOCBOOK2PDF, and DOCBOOK2TXT are also consulted.

By default, makeinfo and texi2dvi are run in the default (English)
locale, since that's the language of most Texinfo manuals.  If you
happen to have a non-English manual and non-English web site, see the
SETLANG setting in the source.

Email bug reports or enhancement requests to bug-gnulib@gnu.org.
""

setvar MANUAL_TITLE = ''
setvar PACKAGE = ''
setvar EMAIL = "webmasters@gnu.org"  # please override with --email
setvar commonarg = '' # passed to all makeinfo/texi2html invcations.
setvar dirargs = ''   # passed to all tools (-I dir).
setvar dirs = ''      # -I directories.
setvar htmlarg = "--css-ref=/software/gnulib/manual.css"
setvar infoarg = '--no-split'
setvar generate_ascii = 'true'
setvar generate_html = 'true'
setvar generate_info = 'true'
setvar generate_tex = 'true'
setvar outdir = 'manual'
setvar source_extra = ''
setvar split = 'node'
setvar srcfile = ''
setvar texarg = ""-t @finalout""

while test $Argc -gt 0 {
  match $1 {
    with -s          shift; setvar srcfile = "$1"
    with -o          shift; setvar outdir = "$1"
    with -I          shift; setvar dirargs = ""$dirargs -I '$1'""; setvar dirs = ""$dirs $1""
    with --common    shift; setvar commonarg = "$1"
    with --docbook   setvar docbook = 'yes'
    with --email     shift; setvar EMAIL = "$1"
    with --html      shift; setvar htmlarg = "$1"
    with --info      shift; setvar infoarg = "$1"
    with --no-ascii  setvar generate_ascii = 'false'
    with --no-html   setvar generate_ascii = 'false'
    with --no-info   setvar generate_info = 'false'
    with --no-tex    setvar generate_tex = 'false'
    with --source    shift; setvar source_extra = "$1"
    with --split     shift; setvar split = "$1"
    with --tex       shift; setvar texarg = "$1"
    with --texi2html setvar use_texi2html = '1'

    with --help      echo $usage; exit 0
    with --version   echo $version; exit 0
    with -*
      echo "$0: Unknown option \`$1'." >&2
      echo "$0: Try \`--help' for more information." >&2
      exit 1
    with *
      if test -z $PACKAGE {
        setvar PACKAGE = "$1"
      } elif test -z $MANUAL_TITLE {
        setvar MANUAL_TITLE = "$1"
      } else {
        echo "$0: extra non-option argument \`$1'." >&2
        exit 1
      }
  }
  shift
}

# makeinfo uses the dirargs, but texi2dvi doesn't.
setvar commonarg = "" $dirargs $commonarg""

# For most of the following, the base name is just $PACKAGE
setvar base = "$PACKAGE"

if test -n $srcfile {
  # but here, we use the basename of $srcfile
  setvar base = $(basename $srcfile)
  match $base {
    with *.txi|*.texi|*.texinfo setvar base = $(echo $base|sed 's/\.[texinfo]*$//)
  }
  setvar PACKAGE = "$base"
} elif test -s "$srcdir/$PACKAGE.texinfo" {
  setvar srcfile = "$srcdir/$PACKAGE.texinfo"
} elif test -s "$srcdir/$PACKAGE.texi" {
  setvar srcfile = "$srcdir/$PACKAGE.texi"
} elif test -s "$srcdir/$PACKAGE.txi" {
  setvar srcfile = "$srcdir/$PACKAGE.txi"
} else {
  echo "$0: cannot find .texinfo or .texi or .txi for $PACKAGE in $srcdir." >&2
  exit 1
}

if test ! -r $GENDOCS_TEMPLATE_DIR/gendocs_template {
  echo "$0: cannot read $GENDOCS_TEMPLATE_DIR/gendocs_template." >&2
  echo "$0: it is available from $templateurl." >&2
  exit 1
}

# Function to return size of $1 in something resembling kilobytes.
proc calcsize
{
  setvar size = $(ls -ksl $1 | awk '{print $1})
  echo $size
}

# copy_images OUTDIR HTML-FILE...
# -------------------------------
# Copy all the images needed by the HTML-FILEs into OUTDIR.
# Look for them in . and the -I directories; this is simpler than what
# makeinfo supports with -I, but hopefully it will suffice.
proc copy_images
{
  local odir
  setvar odir = "$1"
  shift
  $PERL -n -e "
BEGIN {
  \$me = '$prog';
  \$odir = '$odir';
  @dirs = qw(. $dirs);
}
" -e '
/<img src="(.*?)"/g && ++$need{$1};

END {
  #print "$me: @{[keys %need]}\n";  # for debugging, show images found.
  FILE: for my $f (keys %need) {
    for my $d (@dirs) {
      if (-f "$d/$f") {
        use File::Basename;
        my $dest = dirname ("$odir/$f");
        #
        use File::Path;
        -d $dest || mkpath ($dest)
          || die "$me: cannot mkdir $dest: $!\n";
        #
        use File::Copy;
        copy ("$d/$f", $dest)
          || die "$me: cannot copy $d/$f to $dest: $!\n";
        next FILE;
      }
    }
    die "$me: $ARGV: cannot find image $f\n";
  }
}
' -- @ARGV || exit 1
}

match $outdir {
  with /* setvar abs_outdir = "$outdir"
  with *  setvar abs_outdir = "$srcdir/$outdir"
}

echo "Making output for $srcfile"
echo " in $(pwd)"
mkdir -p "$outdir/"

# 
if $generate_info {
  setvar cmd = ""$SETLANG $MAKEINFO -o $PACKAGE.info $commonarg $infoarg \"$srcfile"\""
  echo "Generating info... ($cmd)"
  rm -f $PACKAGE.info* # get rid of any strays
  eval $cmd
  tar czf "$outdir/$PACKAGE.info.tar.gz" $PACKAGE.info*
  ls -l "$outdir/$PACKAGE.info.tar.gz"
  setvar info_tgz_size = $(calcsize "$outdir/$PACKAGE.info.tar.gz)
  # do not mv the info files, there's no point in having them available
  # separately on the web.
}  # end info

# 
if $generate_tex {
  setvar cmd = ""$SETLANG $TEXI2DVI $dirargs $texarg \"$srcfile"\""
  printf "\nGenerating dvi... ($cmd)\n"
  eval $cmd
  # compress/finish dvi:
  gzip -f -9 $PACKAGE.dvi
  setvar dvi_gz_size = $(calcsize $PACKAGE.dvi.gz)
  mv $PACKAGE.dvi.gz "$outdir/"
  ls -l "$outdir/$PACKAGE.dvi.gz"

  setvar cmd = ""$SETLANG $TEXI2DVI --pdf $dirargs $texarg \"$srcfile"\""
  printf "\nGenerating pdf... ($cmd)\n"
  eval $cmd
  setvar pdf_size = $(calcsize $PACKAGE.pdf)
  mv $PACKAGE.pdf "$outdir/"
  ls -l "$outdir/$PACKAGE.pdf"
} # end tex (dvi + pdf)

# 
if $generate_ascii {
  setvar opt = ""-o $PACKAGE.txt --no-split --no-headers $commonarg""
  setvar cmd = ""$SETLANG $MAKEINFO $opt \"$srcfile"\""
  printf "\nGenerating ascii... ($cmd)\n"
  eval $cmd
  setvar ascii_size = $(calcsize $PACKAGE.txt)
  gzip -f -9 -c $PACKAGE.txt >"$outdir/$PACKAGE.txt.gz"
  setvar ascii_gz_size = $(calcsize "$outdir/$PACKAGE.txt.gz)
  mv $PACKAGE.txt "$outdir/"
  ls -l "$outdir/$PACKAGE.txt" "$outdir/$PACKAGE.txt.gz"
}

# 

if $generate_html {
# Split HTML at level $1.  Used for texi2html.
proc html_split
{
  setvar opt = ""--split=$1 --node-files $commonarg $htmlarg""
  setvar cmd = ""$SETLANG $TEXI2HTML --output $PACKAGE.html $opt \"$srcfile"\""
  printf "\nGenerating html by $1... ($cmd)\n"
  eval $cmd
  setvar split_html_dir = "$PACKAGE.html"
  shell {
    cd ${split_html_dir} || exit 1
    ln -sf ${PACKAGE}.html index.html
    tar -czf "$abs_outdir/${PACKAGE}.html_$1.tar.gz" -- *.html
  }
  eval html_$1_tgz_size=$(calcsize "$outdir/${PACKAGE}.html_$1.tar.gz)
  rm -f "$outdir"/html_$1/*.html
  mkdir -p "$outdir/html_$1/"
  mv ${split_html_dir}/*.html "$outdir/html_$1/"
  rmdir ${split_html_dir}
}

if test -z $use_texi2html {
  setvar opt = ""--no-split --html -o $PACKAGE.html $commonarg $htmlarg""
  setvar cmd = ""$SETLANG $MAKEINFO $opt \"$srcfile"\""
  printf "\nGenerating monolithic html... ($cmd)\n"
  rm -rf $PACKAGE.html  # in case a directory is left over
  eval $cmd
  setvar html_mono_size = $(calcsize $PACKAGE.html)
  gzip -f -9 -c $PACKAGE.html >"$outdir/$PACKAGE.html.gz"
  setvar html_mono_gz_size = $(calcsize "$outdir/$PACKAGE.html.gz)
  copy_images "$outdir/" $PACKAGE.html
  mv $PACKAGE.html "$outdir/"
  ls -l "$outdir/$PACKAGE.html" "$outdir/$PACKAGE.html.gz"

  # Before Texinfo 5.0, makeinfo did not accept a --split=HOW option,
  # it just always split by node.  So if we're splitting by node anyway,
  # leave it out.
  if test "x$split" = xnode {
    setvar split_arg = ''
  } else {
    setvar split_arg = "--split=$split"
  }
  #
  setvar opt = ""--html -o $PACKAGE.html $split_arg $commonarg $htmlarg""
  setvar cmd = ""$SETLANG $MAKEINFO $opt \"$srcfile"\""
  printf "\nGenerating html by $split... ($cmd)\n"
  eval $cmd
  setvar split_html_dir = "$PACKAGE.html"
  copy_images $split_html_dir/ $split_html_dir/*.html
  shell {
    cd $split_html_dir || exit 1
    tar -czf "$abs_outdir/$PACKAGE.html_$split.tar.gz" -- *
  }
  eval \
    html_${split}_tgz_size=$(calcsize "$outdir/$PACKAGE.html_$split.tar.gz)
  rm -rf "$outdir/html_$split/"
  mv $split_html_dir "$outdir/html_$split/"
  du -s "$outdir/html_$split/"
  ls -l "$outdir/$PACKAGE.html_$split.tar.gz"

} else { # use texi2html:
  setvar opt = ""--output $PACKAGE.html $commonarg $htmlarg""
  setvar cmd = ""$SETLANG $TEXI2HTML $opt \"$srcfile"\""
  printf "\nGenerating monolithic html with texi2html... ($cmd)\n"
  rm -rf $PACKAGE.html  # in case a directory is left over
  eval $cmd
  setvar html_mono_size = $(calcsize $PACKAGE.html)
  gzip -f -9 -c $PACKAGE.html >"$outdir/$PACKAGE.html.gz"
  setvar html_mono_gz_size = $(calcsize "$outdir/$PACKAGE.html.gz)
  mv $PACKAGE.html "$outdir/"

  html_split node
  html_split chapter
  html_split section
}
} # end html

# 
printf "\nMaking .tar.gz for sources...\n"
setvar d = $(dirname $srcfile)
shell {
  cd $d
  setvar srcfiles = $(ls -d *.texinfo *.texi *.txi *.eps $source_extra ) || true
  tar czfh "$abs_outdir/$PACKAGE.texi.tar.gz" $srcfiles
  ls -l "$abs_outdir/$PACKAGE.texi.tar.gz"
}
setvar texi_tgz_size = $(calcsize "$outdir/$PACKAGE.texi.tar.gz)

# 
# Do everything again through docbook.
if test -n $docbook {
  setvar opt = ""-o - --docbook $commonarg""
  setvar cmd = ""$SETLANG $MAKEINFO $opt \"$srcfile\" >${srcdir}/$PACKAGE-db.xml""
  printf "\nGenerating docbook XML... ($cmd)\n"
  eval $cmd
  setvar docbook_xml_size = $(calcsize $PACKAGE-db.xml)
  gzip -f -9 -c $PACKAGE-db.xml >"$outdir/$PACKAGE-db.xml.gz"
  setvar docbook_xml_gz_size = $(calcsize "$outdir/$PACKAGE-db.xml.gz)
  mv $PACKAGE-db.xml "$outdir/"

  setvar split_html_db_dir = 'html_node_db'
  setvar opt = ""$commonarg -o $split_html_db_dir""
  setvar cmd = ""$DOCBOOK2HTML $opt \"${outdir}/$PACKAGE-db.xml"\""
  printf "\nGenerating docbook HTML... ($cmd)\n"
  eval $cmd
  shell {
    cd ${split_html_db_dir} || exit 1
    tar -czf "$abs_outdir/${PACKAGE}.html_node_db.tar.gz" -- *.html
  }
  setvar html_node_db_tgz_size = $(calcsize "$outdir/${PACKAGE}.html_node_db.tar.gz)
  rm -f "$outdir"/html_node_db/*.html
  mkdir -p "$outdir/html_node_db"
  mv ${split_html_db_dir}/*.html "$outdir/html_node_db/"
  rmdir ${split_html_db_dir}

  setvar cmd = ""$DOCBOOK2TXT \"${outdir}/$PACKAGE-db.xml"\""
  printf "\nGenerating docbook ASCII... ($cmd)\n"
  eval $cmd
  setvar docbook_ascii_size = $(calcsize $PACKAGE-db.txt)
  mv $PACKAGE-db.txt "$outdir/"

  setvar cmd = ""$DOCBOOK2PDF \"${outdir}/$PACKAGE-db.xml"\""
  printf "\nGenerating docbook PDF... ($cmd)\n"
  eval $cmd
  setvar docbook_pdf_size = $(calcsize $PACKAGE-db.pdf)
  mv $PACKAGE-db.pdf "$outdir/"
}

# 
printf "\nMaking index.html for $PACKAGE...\n"
if test -z $use_texi2html {
  setvar CONDS = ""/%%IF  *HTML_SECTION%%/,/%%ENDIF  *HTML_SECTION%%/d;\
         /%%IF  *HTML_CHAPTER%%/,/%%ENDIF  *HTML_CHAPTER%%/d""
} else {
  # should take account of --split here.
  setvar CONDS = ""/%%ENDIF.*%%/d;/%%IF  *HTML_SECTION%%/d;/%%IF  *HTML_CHAPTER%%/d""
}

setvar curdate = $($SETLANG date '+%B %d, %Y)
sed \
   -e "s!%%TITLE%%!$MANUAL_TITLE!g" \
   -e "s!%%EMAIL%%!$EMAIL!g" \
   -e "s!%%PACKAGE%%!$PACKAGE!g" \
   -e "s!%%DATE%%!$curdate!g" \
   -e "s!%%HTML_MONO_SIZE%%!$html_mono_size!g" \
   -e "s!%%HTML_MONO_GZ_SIZE%%!$html_mono_gz_size!g" \
   -e "s!%%HTML_NODE_TGZ_SIZE%%!$html_node_tgz_size!g" \
   -e "s!%%HTML_SECTION_TGZ_SIZE%%!$html_section_tgz_size!g" \
   -e "s!%%HTML_CHAPTER_TGZ_SIZE%%!$html_chapter_tgz_size!g" \
   -e "s!%%INFO_TGZ_SIZE%%!$info_tgz_size!g" \
   -e "s!%%DVI_GZ_SIZE%%!$dvi_gz_size!g" \
   -e "s!%%PDF_SIZE%%!$pdf_size!g" \
   -e "s!%%ASCII_SIZE%%!$ascii_size!g" \
   -e "s!%%ASCII_GZ_SIZE%%!$ascii_gz_size!g" \
   -e "s!%%TEXI_TGZ_SIZE%%!$texi_tgz_size!g" \
   -e "s!%%DOCBOOK_HTML_NODE_TGZ_SIZE%%!$html_node_db_tgz_size!g" \
   -e "s!%%DOCBOOK_ASCII_SIZE%%!$docbook_ascii_size!g" \
   -e "s!%%DOCBOOK_PDF_SIZE%%!$docbook_pdf_size!g" \
   -e "s!%%DOCBOOK_XML_SIZE%%!$docbook_xml_size!g" \
   -e "s!%%DOCBOOK_XML_GZ_SIZE%%!$docbook_xml_gz_size!g" \
   -e "s,%%SCRIPTURL%%,$scripturl,g" \
   -e "s!%%SCRIPTNAME%%!$prog!g" \
   -e $CONDS \
$GENDOCS_TEMPLATE_DIR/gendocs_template >"$outdir/index.html"

echo "Done, see $outdir/ subdirectory for new files."

# Local variables:
# eval: (add-hook 'write-file-hooks 'time-stamp)
# time-stamp-start: "scriptversion="
# time-stamp-format: "%:y-%02m-%02d.%02H"
# time-stamp-end: "$"
# End: