#!/usr/bin/perl

=head1 NAME

dh_installman - install man pages into package build directories

=cut

use strict;
use warnings;
use File::Find;
use Debian::Debhelper::Dh_Lib;

our $VERSION = DH_BUILTIN_VERSION;

=head1 SYNOPSIS

B<dh_installman> [S<I<debhelper options>>] [S<I<manpage> ...>]

=head1 DESCRIPTION

B<dh_installman> is a debhelper program that handles installing man
pages into the correct locations in package build directories.

In compat 10 and earlier, this program was primarily for when
upstream's build system does not properly install them as a part of
its install step (or it does not have an install step).  In compat 11
and later, it also supports the default searchdir plus --sourcedir
like dh_install(1) and has the advantage that it respects the nodoc
build profile (unlike dh_install(1)).

Even if you prefer to use L<dh_install(1)> for installing the manpages,
B<dh_installman> can still be useful for converting the manpage encoding
to UTF-8 and for converting F<.so> links (as described below).  However,
that part happens automatically without any explicit configuration.


You tell B<dh_installman> what man pages go in your packages, and it figures out
where to install them based on the section field in their B<.TH> or
B<.Dt> line. If you have a properly formatted B<.TH> or B<.Dt> line,
your man page will be installed into the right directory, with the
right name (this includes proper handling of pages with a subsection,
like B<3perl>, which are placed in F<man3>, and given an extension of
F<.3perl>). If your B<.TH> or B<.Dt> line is incorrect or missing, the
program may guess wrong based on the file extension.

It also supports translated man pages, by looking for extensions
like F<.ll.8> and F<.ll_LL.8>, or by use of the B<--language> switch.

If B<dh_installman> seems to install a man page into the wrong section or with
the wrong extension, this is because the man page has the wrong section
listed in its B<.TH> or B<.Dt> line. Edit the man page and correct the
section, and B<dh_installman> will follow suit. See L<man(7)> for details
about the B<.TH> section, and L<mdoc(7)> for the B<.Dt> section. If
B<dh_installman> seems to install a man page into a directory
like F</usr/share/man/pl/man1/>, that is because your program has a
name like F<foo.pl>, and B<dh_installman> assumes that means it is translated
into Polish. Use B<--language=C> to avoid this.

After the man page installation step, B<dh_installman> will check to see if
any of the man pages in the temporary directories of any of the packages it
is acting on contain F<.so> links. If so, it changes them to symlinks.

Also, B<dh_installman> will use man to guess the character encoding of each
manual page and convert it to UTF-8. If the guesswork fails for some
reason, you can override it using an encoding declaration. See
L<manconv(1)> for details.

From debhelper compatibility level 11 on, B<dh_install> will fall back to
looking in F<debian/tmp> for files, if it does not find them in the current
directory (or wherever you've told it to look using B<--sourcedir>).

=head1 FILES

=over 4

=item debian/I<package>.manpages

Lists man pages to be installed.

Supports substitution variables in compat 13 and later as
documented in L<debhelper(7)>.

=back

=head1 OPTIONS

=over 4

=item B<-A>, B<--all>

Install all files specified by command line parameters in ALL packages
acted on.

=item B<--language=>I<ll>

Use this to specify that the man pages being acted on are written in the
specified language.

=item B<--sourcedir=>I<dir>

Look in the specified directory for files to be installed.  This option
requires compat 11 or later (it is silently ignored in compat 10 or earlier).

Note that this is not the same as the B<--sourcedirectory> option used
by the B<dh_auto_>I<*> commands. You rarely need to use this option, since
B<dh_installman> automatically looks for files in F<debian/tmp> in debhelper
compatibility level 11 and above.

=item I<manpage> ...

Install these man pages into the first package acted on. (Or in all
packages if B<-A> is specified).

=back

=head1 EXAMPLES

An example F<< debian/I<package>.manpages >> file could look like this:

  doc/man/foo.1
  # Translations
  doc/man/foo.da.1
  doc/man/foo.de.1
  doc/man/foo.fr.1
  # NB: The following line is considered a polish translation
  # of "foo.1" (and not a manpage written in perl called "foo.pl")
  doc/man/foo.pl.1
  # ...

=head1 NOTES

An older version of this program, L<dh_installmanpages(1)>, is still used
by some packages, and so is still included in debhelper.
It is, however, deprecated, due to its counterintuitive and inconsistent
interface. Use this program instead.

=cut

init(options => {
	"language=s" => \$dh{LANGUAGE},
	"sourcedir=s" => \$dh{SOURCEDIR},
});


# PROMISE: DH NOOP WITHOUT pkgfile-logged(manpages) tmp(usr/share/man) cli-options()

my (@sofiles, @sodests);
my @all_packages = getpackages();

my $default_error_handler = compat(10) ? \&glob_expand_error_handler_reject_nomagic_warn_discard : \&glob_expand_error_handler_reject;
my $nodocs = is_build_profile_active('nodoc') || get_buildoption('nodoc') ? 1 : 0;
# We cannot assume documentation is built under nodoc, but if it is we must flag it as handled
# or dh_missing might make noise.
$default_error_handler = \&glob_expand_error_handler_silently_ignore if $nodocs;

on_items_in_parallel(\@all_packages, sub {

	foreach my $package (@_) {
		next if is_udeb($package);

		my $tmp = tmpdir($package);
		my $file = pkgfile($package, "manpages");
		my @manpages;
		my @search_dirs = ('.');
		my $skip_install = process_pkg($package) ? 0 : 1;
		my $error_handler = $skip_install ? \&glob_expand_error_handler_silently_ignore : $default_error_handler;
		@search_dirs = ($dh{SOURCEDIR} // '.', default_sourcedir($package)) if not compat(10);

		@manpages = filearray($file, \@search_dirs, $error_handler) if $file;

		if (($package eq $dh{FIRSTPACKAGE} || $dh{PARAMS_ALL}) && @ARGV) {
			push @manpages, @ARGV;
		}

		log_installed_files($package, @manpages);

		next if $skip_install or $nodocs;

		foreach my $page (@manpages) {
			my $basename = basename($page);

			# Support compressed pages.
			my $gz = '';
			if ($basename =~ m/(.*)(\.gz)/) {
				$basename = $1;
				$gz = $2;
			}

			my ($fd, $section);
			# See if there is a .TH or .Dt entry in the man page. If so,
			# we'll pull the section field from that.
			if ($gz) {
				$fd = open_gz($page) or error("open $page failed: $!");
			}
			else {
				open($fd, '<', $page) or error("open $page failed: $!");
			}
			while (<$fd>) {
				if (/^\.TH\s+\S+\s+"?(\d+[^"\s]*)"?/ ||
					/^\.Dt\s+\S+\s+(\d+[^\s]*)/) {
					$section = $1;
					if ($section =~ m/^\d+[.]\d+/) {
						warning("Ignoring section defined in TH/Dt for ${page} as it looks like a version number: ${section}");
						$section = undef;
					}
					last;
				}
			}
			close($fd);
			# Failing that, we can try to get it from the filename.
			if (!$section) {
				($section) = $basename =~ m/\.([1-9]\w*)$/;
			}

			# Now get the numeric component of the section.
			my ($realsection) = $section =~ m/^(\d+)/ if defined $section;
			if (!$realsection or ($realsection < 0 or $realsection > 9)) {
				warning("Section for ${page} is computed as \"${section}\", which is not a valid section")
					if defined($section);
				error("Could not determine section for $page");
			}

			# Get the man page's name -- everything up to the last dot.
			my ($instname) = $basename =~ m/^(.*)\./;

			my $destdir = "$tmp/usr/share/man/man$realsection/";
			my $langcode;
			if (!defined $dh{LANGUAGE} || !exists $dh{LANGUAGE}) {
				if (not compat(10) and $page =~ m{/man/(?:([a-z][a-z](?:_[A-Z][A-Z])?)(?:\.[^/]+)?)?/man[1-9]/}) {
					# If it looks like it was installed in a proper man dir, assume the language
					# from that is correct.
					$langcode = $1;
				} else {
					# Translated man pages are typically specified by adding the
					# language code to the filename, so detect that and
					# redirect to appropriate directory, stripping the code.
					($langcode) = $basename =~ m/\.([a-z][a-z](?:_[A-Z][A-Z])?)\.(?:[1-9]|man)/;
					# Avoid false positives such as /usr/share/man/man8/libnss_myhostname.so.2.8
					if (defined $langcode && $langcode eq 'so' && $basename =~ /^lib.*\.so(\.[0-9]+)*$/) {
						$langcode = '';
					}
				}
			} elsif ($dh{LANGUAGE} ne 'C') {
				$langcode = $dh{LANGUAGE};
			}

			if (defined $langcode && $langcode ne '') {
				# Strip the language code from the instname.
				$instname =~ s/\.$langcode$//;
			}

			if (defined $langcode && $langcode ne '') {
				$destdir = "$tmp/usr/share/man/$langcode/man$realsection/";
			}
			$destdir =~ tr:/:/:s; # just for looks
			my $instpage = "$destdir$instname.$section";

			next if -l $instpage;

			install_dir($destdir);
			if ($gz) {
				doit({ stdout => $instpage }, 'zcat', $page);
			}
			else {
				install_file($page, $instpage);
			}
		}

		# Now the .so conversion.
		@sofiles = @sodests = ();
		foreach my $dir (qw{usr/share/man}) {
			if (-e "$tmp/$dir") {
				find(\&find_so_man, "$tmp/$dir");
			}
		}
		foreach my $sofile (@sofiles) {
			my $sodest = shift(@sodests);
			rm_files($sofile);
			make_symlink_raw_target($sodest, $sofile);
		}
	}

});

# Now utf-8 conversion.
my $has_man_recode = 0;
$has_man_recode = 1 if has_man_db_tool('man-recode');

if ($has_man_recode || has_man_db_tool('man')) {
	my (@manpages_to_reencode, @issues);
	for my $package (@{$dh{DOPACKAGES}}) {
		next if is_udeb($package);
		my $tmp = tmpdir($package);
		foreach my $dir (qw{usr/share/man}) {
			next unless -e "$tmp/$dir";
			my %seen;
			my $wanted = sub {
				my $path = $File::Find::name;
				return if -l $path || !-f _;
				if ($path =~ m/\.dh-new$/) {
					push(@issues, $path);
					return;
				}
				my $uncompressed_name = $path;
				$uncompressed_name =~ s/\.(?:gz|Z)$//;
				if (exists($seen{$uncompressed_name})) {
					my $msg = "Multiple definitions for manpage ${uncompressed_name} via different compressions.";
					my @values = sort ($path, $seen{$uncompressed_name});
					my $warn_msg = $msg . '  Picking ' . $values[0] . ' as the canonical definition.';
					my $error_msg = $msg . '  Please ensure there is at most one definition.';
					deprecated_functionality($warn_msg, 13, $error_msg);
					$path = $values[0];
					warn("Removing conflicting definition of ${uncompressed_name} (" . $values[1]
						. ') to ensure deterministic behaviour.');
					rm_files($values[1]);
				}
				$seen{$uncompressed_name} = $path;
			};
			find({
					no_chdir => 1,
					wanted => $wanted,
				}, "$tmp/$dir");
			push(@manpages_to_reencode, sort(values(%seen)));
		}

		if (@issues) {
			warning("Removing temporary manpages from another dh_installman instance");
			rm_files(@issues);
			warning("Possibly race-condition detected or left-overs from an interrupted dh_installman (e.g. with ^C)");
			error("Please ensure there are no parallel dh_installman's running (for this pkg) and then re-run dh_installman");
		}
	}
	if (@manpages_to_reencode) {
		on_items_in_parallel(\@manpages_to_reencode, \&reencode_manpages);
	}
} else {
	# Should only occur during debhelper building itself (to avoid a B-D on man-db).
	warning("man is not available.  Skipping re-encode of UTF-8 manpages")
}

# Check if a file is a .so man page, for use by File::Find.
sub find_so_man {
	# The -s test is because a .so file tends to be small. We don't want
	# to open every man page. 1024 is arbitrary.
	if (! -f $_ || -s _ > 1024 || -s _ == 0) {
		return;
	}

	# Test first line of file for the .so thing.
	my $fd;
	if (/\.gz$/) {
		$fd = open_gz($_) or error("open $_ failed: $!");
	}
	else {
		open($fd, '<', $_) || error("open $_ failed: $!");
	}
	my $l = <$fd>;
	close($fd);

	if (! defined $l) {
		error("failed to read $_");
	}

	if ($l=~m/\.so\s+(.*)\s*/) {
		my $solink=$1;
		# This test is here to prevent links like ... man8/../man8/foo.8
		if (basename($File::Find::dir) eq
		    dirname($solink)) {
			$solink=basename($solink);
		}
		# A so link with a path is relative to the base of the man
		# page hierarchy, but without a path, is relative to the
		# current section.
		elsif ($solink =~ m!/!) {
			$solink="../$solink";
		}

		if (-e $solink || -e "$solink.gz") {
			push @sofiles,"$File::Find::dir/$_";
			push @sodests,$solink;
		}
	}
}

sub has_man_db_tool {
	my ($tool) = @_;
	open(my $old_stderr, '>&', *STDERR) or error("dup(STDERR, tmp_fd): $!");
	# Ignore the error; it is intended as noise-reduction.  As long as we can restore
	# the stderr later, the log will just be slightly more noisy than planned.
	open(*STDERR, '>', '/dev/null') or warn("redirect stderr to /dev/null failed: $!");

	my $res = defined(`$tool --version`);
	open(*STDERR, '>&', $old_stderr) or error("dup(tmp_fd, STDERR): $!");
	close($old_stderr);
	return $res;
}

sub reencode_manpages {
	my (@manpages) = @_;
	if ($has_man_recode) {
		xargs(\@manpages, 'man-recode', '--to-code', 'UTF-8', '--suffix', '.dh-new');
	}
	for my $manpage (@manpages) {
		my $manpage_tmp = "${manpage}.dh-new";
		$manpage_tmp =~ s/\.(?:gz|Z)\.dh-new$/.dh-new/;
		if (not $has_man_recode) {
			my $manpage_cmd = ($manpage =~ m{^/}) ? $manpage : "./${manpage}";
			doit({ stdout => $manpage_tmp }, 'man', '-l', '--recode', 'UTF-8', $manpage_cmd);
		}
		# recode uncompresses compressed pages
		my $orig = $manpage;
		rm_files($orig) if $manpage =~ s/\.(gz|Z)$//;
		rename_path($manpage_tmp, $manpage);
	}
	# Bulk reset permissions of all re-encoded files
	xargs(\@manpages, 'chmod', '0644', '--');
}

=head1 SEE ALSO

L<debhelper(7)>

This program is a part of debhelper.

=head1 AUTHOR

Joey Hess <joeyh@debian.org>

=cut