#!/usr/bin/perl

=head1 NAME

dh_missing - check for missing files

=cut

use v5.24;
use warnings;
use Debian::Debhelper::Dh_Lib;

our $VERSION = DH_BUILTIN_VERSION;

=head1 SYNOPSIS

B<dh_missing> [B<-X>I<item>] [B<--sourcedir=>I<dir>] [S<I<debhelper options>>]

=head1 DESCRIPTION

B<dh_missing> compares the list of installed files with the files in
the source directory. If any of the files (and symlinks) in the source
directory were not installed to somewhere, it will warn on stderr
about that (B<--list-missing>) or fail (B<--fail-missing>).

Please note that in compat 11 and earlier without either of these
options, B<dh_missing> will silently do nothing.  In compat 12,
B<--list-missing> is the default  In compat 13 and later,
B<--fail-missing> is the default.

This may be useful if you have a large package and want to make sure that
you don't miss installing newly added files in new upstream releases.

Remember to test different kinds of builds (dpkg-buildpackage -A/-B/...) as
you may experience varying results when only a subset of the packages are
built.

=head1 FILES

=over 4

=item debian/not-installed

List the files that are deliberately not installed in I<any> binary
package.  Paths listed in this file are ignored by B<dh_missing>.
However, it is B<not> a method to exclude files from being installed
by any of the debhelper tool.  If you want a tool to not install a
given file, please use its B<--exclude> option (where available).

B<dh_missing> will expand wildcards in this file (since debhelper 11.1).
Wildcards without matches will be ignored.

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

=back

=head1 OPTIONS

=over 4

=item B<--list-missing>

Warn on stderr about source files not installed to somewhere.

Note that many dh-tools acting on a path will mark the path as
installed even if it has been excluded via B<-X> or B<--exclude>.
This is also seen when a dh-tool is acting on a directory and
exclusion is used to ignore some files in the directory.  In either
case, this will make B<dh_missing> silently assume the excluded files
have been handled.

This is the default in compat 12.

=item B<--fail-missing>

This option is like B<--list-missing>, except if a file was missed, it will
not only list the missing files, but also fail with a nonzero exit code.

This is the default in compat 13 and later.

=back

=cut

# INTROSPECTABLE: CONFIG-FILES path(debian/not-installed)

init(options => {
	"list-missing" => \$dh{LIST_MISSING},
	"fail-missing" => \$dh{FAIL_MISSING},
	"sourcedir=s" => \$dh{SOURCEDIR},
});

my (@installed, %helpers);

my $srcdir = '.';
if (defined($dh{SOURCEDIR})) {
	$srcdir = $dh{SOURCEDIR};
	$srcdir =~ s{/+$}{};
	error("Invalid --sourcedir - must not be empty nor /") if not $srcdir;
}

if (!$dh{LIST_MISSING} && !$dh{FAIL_MISSING}) {
	exit 0 if compat(11);
	# --list-missing is the default in compat 12 and --fail-missing in compat 13+
	my $option = compat(12) ? 'LIST_MISSING' : 'FAIL_MISSING';
	$dh{$option} = 1;
}

# . as srcdir makes no sense, so this is a special case.
if ($srcdir eq '.') {
	$srcdir='debian/tmp';
}

if (! -d $srcdir) {
	# If there was no explicit source directory, then we do not care
	# if it is missing.
	exit(0) if not defined $dh{SOURCEDIR};

	if (scalar(getpackages()) == 1 && defined($dh{SOURCEDIR})) {
		warning("$srcdir does not exist and there is only binary package.");
		warning("Assuming everything is installed directly into the package directory.");
		exit(0);
	}
	if (compat(10)) {
		# Prevent "dh $@ --list-missing --destdir=... ..." from failing in compat 10.
		warning("Cannot check if installation is missing files: $srcdir does not exist");
		exit(0);
	} else {
		error("Cannot check if installation is missing files: $srcdir does not exist");
	}
}

for my $file (glob('debian/.debhelper/generated/*/installed-by-*')) {
	my ($target_pkg, $helper) = ('unknown', 'unknown');
	my $had_files = 0;
	my %seen;
	if ($file =~ m@.*/([^/]+)/installed-by-(.*)@) {
		($target_pkg, $helper) = ($1, $2);
	}

	open(my $fh, '<', $file) or error("could not open $file: $!");
	while (my $line = <$fh>) {
		chomp($line);
		next if $line =~ m/^\s*$/;
		next if $seen{$line}++; # Ignore duplicates
		$had_files++;
		push(@installed, $line);
	}
	$helpers{$helper}{$target_pkg} = $had_files;
	close($fh);
}

my @missing;
if ( -f 'debian/not-installed') {
	my @not_installed = filearray('debian/not-installed');
	for my $pattern (@not_installed) {
		my @matches;
		# Add an explicit d/tmp if absent as there is no point in
		# looking outside the debian staging directory
		$pattern =~ s:^\s*:debian/tmp/: unless $pattern =~ m:^\s*debian/tmp/:;
		@matches = glob_expand(['.'], \&glob_expand_error_handler_silently_ignore, $pattern);
		if (@matches) {
			# Assume classify them as installed
			push(@installed, @matches);
		} else {
			# Assume it is not a pattern and classify it as installed
			push(@installed, $pattern);
		}
	}

	push(@installed, @not_installed);
}
my $installed=join("|", map {
	# Kill any extra slashes, for robustness.
	y:/:/:s;
	s:/+$::;
	s:^(\./)*::;
	"\Q$_\E\/.*|\Q$_\E";
} @installed);
$installed=qr{^($installed)$};

# Lazy load File::Find
require File::Find;

File::Find::find(sub {
	# Lazy loading of File::Find makes perl think that File::Find::dir is only used once
	# and we might have typo'ed something
	no warnings qw(once);
	-f || -l || return;
	$_="$File::Find::dir/$_";
	if (! /$installed/ && ! excludefile($_)) {
		my $file=$_;
		$file=~s/^\Q$srcdir\E\///;
		push @missing, $file;
	}
}, $srcdir);
if (@missing) {
	my $had_related_files;
	my %seen_basename = map { basename($_) => $_ } @installed;
	my $multiarch = dpkg_architecture_value("DEB_HOST_MULTIARCH");
	my $seen_ma_value = 0;
	for my $file (sort(@missing)) {
		my $basename = basename($file);
		if (exists($seen_basename{$basename})) {
			my $alt_source = $seen_basename{$basename};
			if ($file eq $alt_source) {
				warning("$file exists in $srcdir but is not installed to anywhere ");
				warning("A file with an equivalent path was found but did not get installed. Note that dh_install prefers paths in ./ over paths in ./debian/tmp.");
			} else {
			$had_related_files //= [$file, $alt_source];
			warning("$file exists in $srcdir but is not installed to anywhere (related file: \"$alt_source\")");
			}
		} else {
			warning("$file exists in $srcdir but is not installed to anywhere ");
		}
		$seen_ma_value = 1 if index($file, $multiarch) > -1;
	}
	if ($had_related_files) {
		my ($missing, $alt_source) = $had_related_files->@*;
		my $error = $dh{FAIL_MISSING} ? 'error' : 'warning';
		nonquiet_print();
		nonquiet_print('While detecting missing files, dh_missing noted some files with a similar name to those');
		nonquiet_print("that were missing.  This ${error} /might/ be resolved by replacing references to the");
		nonquiet_print('missing files with the similarly named ones that dh_missing found - assuming the content');
		nonquiet_print('is identical.');
		nonquiet_print();
		nonquiet_print('As an example, you might want to replace:');
		nonquiet_print(" * ${alt_source}");
		nonquiet_print('with:');
		nonquiet_print(" * ${missing}");
		nonquiet_print('in a file in debian/ or as argument to one of the dh_* tools called from debian/rules.');
		nonquiet_print('(Note it is possible the paths are not used verbatim but instead directories ');
		nonquiet_print('containing or globs matching them are used instead)');
		nonquiet_print();
		nonquiet_print('Alternatively, add the missing file to debian/not-installed if it cannot and should not');
		nonquiet_print('be used.');
		nonquiet_print();
	}
	nonquiet_print("The following debhelper tools have reported what they installed (with files per package)");
	for my $helper (sort(keys(%helpers))) {
		my $pkg_info = $helpers{$helper};
		my @results;
		for my $pkg (sort(keys(%{$pkg_info}))) {
			my $no = $pkg_info->{$pkg};
			push(@results, "${pkg} (${no})")
		}
		nonquiet_print(" * ${helper}: " . join(', ', @results));
	}
	nonquiet_print('If the missing files are installed by another tool, please file a bug against it.');
	nonquiet_print('When filing the report, if the tool is not part of debhelper itself, please reference the');
	nonquiet_print('"Logging helpers and dh_missing" section from the "PROGRAMMING" guide for debhelper (10.6.3+).');
	nonquiet_print('  (in the debhelper package: /usr/share/doc/debhelper/PROGRAMMING.md.gz)');
	nonquiet_print("Be sure to test with dpkg-buildpackage -A/-B as the results may vary when only a subset is built");
	nonquiet_print("If the omission is intentional or no other helper can take care of this consider adding the");
	nonquiet_print("paths to debian/not-installed.");
	if ($seen_ma_value) {
		nonquiet_print();
		nonquiet_print("Remember to be careful with paths containing \"${multiarch}\", where you might need to");
		nonquiet_print("use a wildcard or (assuming compat 13+) e.g. \${DEB_HOST_MULTIARCH} in debian/not-installed");
		nonquiet_print("to ensure it works on all architectures (see #961104).");
	}
	if ($dh{FAIL_MISSING}) {
		error("missing files, aborting");
	}
}

=head1 SEE ALSO

L<debhelper(7)>

This program is a part of debhelper.

=head1 AUTHOR

Michael Stapelberg <stapelberg@debian.org>

=cut