#!/usr/bin/perl

=head1 NAME

dh_strip - strip executables, shared libraries, and some static libraries

=cut

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

our $VERSION = DH_BUILTIN_VERSION;

=head1 SYNOPSIS

B<dh_strip> [S<I<debhelper options>>] [B<-X>I<item>] [B<--dbg-package=>I<package>] [B<--keep-debug>]

=head1 DESCRIPTION

B<dh_strip> is a debhelper program that is responsible for stripping
executables, shared libraries, and static libraries that are not used for
debugging.

This program examines your package build directories and works out what
to strip on its own. It uses L<file(1)> and file permissions and filenames
to figure out what files are shared libraries (F<*.so>), executable binaries,
and static (F<lib*.a>) and debugging libraries (F<lib*_g.a>, F<debug/*.so>), and
strips each as much as is possible. (Which is not at all for debugging
libraries.) In general it seems to make very good guesses, and will do the
right thing in almost all cases.

Since it is very hard to automatically guess if a file is a
module, and hard to determine how to strip a module, B<dh_strip> does not
currently deal with stripping binary modules such as F<.o> files.

=head1 OPTIONS

=over 4

=item B<-X>I<item>, B<--exclude=>I<item>

Exclude files that contain I<item> anywhere in their filename from being
stripped. You may use this option multiple times to build up a list of
things to exclude.

=item B<--dbg-package=>I<package>

B<This option is a now special purpose option that you normally do not
need>.  In most cases, there should be little reason to use this
option for new source packages as debhelper automatically generates
debug packages ("dbgsym packages").  B<If you have a manual
--dbg-package> that you want to replace with an automatically
generated debug symbol package, please see the B<--dbgsym-migration>
option.

Causes B<dh_strip> to save debug symbols stripped from the packages it acts on
as independent files in the package build directory of the specified debugging
package.

For example, if your packages are libfoo and foo and you want to include a
I<foo-dbg> package with debugging symbols, use B<dh_strip --dbg-package=>I<foo-dbg>.

This option implies B<--no-automatic-dbgsym> and I<cannot> be used
with B<--automatic-dbgsym> or B<--dbgsym-migration>.

=item B<-k>, B<--keep-debug>

B<This option is a now special purpose option that you normally do not
need>.  In most cases, there should be little reason to use this
option for new source packages as debhelper automatically generates
debug packages ("dbgsym packages").  B<If you have a manual
--dbg-package> that you want to replace with an automatically
generated debug symbol package, please see the B<--dbgsym-migration>
option.

Debug symbols will be retained, but split into an independent
file in F<usr/lib/debug/> in the package build directory. B<--dbg-package>
is easier to use than this option, but this option is more flexible.

This option implies B<--no-automatic-dbgsym> and I<cannot> be used
with B<--automatic-dbgsym>.

=item B<--dbgsym-migration=>I<package-relation>

This option is used to migrate from a manual "-dbg" package (created
with B<--dbg-package>) to an automatic generated debug symbol
package. This option should describe a valid B<Replaces>- and
B<Breaks>-relation, which will be added to the debug symbol package to
avoid file conflicts with the (now obsolete) -dbg package.

This option implies B<--automatic-dbgsym> and I<cannot> be used with
B<--keep-debug>, B<--dbg-package> or B<--no-automatic-dbgsym>.

Examples:

  dh_strip --dbgsym-migration='libfoo-dbg (<< 2.1-3~)'

  dh_strip --dbgsym-migration='libfoo-tools-dbg (<< 2.1-3~), libfoo2-dbg (<< 2.1-3~)'

=item B<--automatic-dbgsym>, B<--no-automatic-dbgsym>

Control whether B<dh_strip> should be creating debug symbol packages
when possible.

The default is to create debug symbol packages.

=item B<--ddebs>, B<--no-ddebs>

Historical name for B<--automatic-dbgsym> and B<--no-automatic-dbgsym>.

=item B<--ddeb-migration=>I<package-relation>

Historical name for B<--dbgsym-migration>.

=item B<--keep-lto>

Sections containing information generated by B<-flto -ffat-lto-objects>
are normally removed from static archives, because the streaming
format changes even in minor GCC versions, and because cross package
link time optimizations are normally not desired.  This option allows
to keep this information in the static libraries.

=back

=head1 NOTES

If the B<DEB_BUILD_OPTIONS> environment variable contains B<nostrip>,
nothing will be stripped, in accordance with Debian policy (section
10.1 "Binaries").  This will also inhibit the automatic creation of
debug symbol packages.

The automatic creation of debug symbol packages can also be prevented
by adding B<noautodbgsym> to the B<DEB_BUILD_OPTIONS> environment
variable.  However, B<dh_strip> will still add debuglinks to ELF
binaries when this flag is set.  This is to ensure that the regular
deb package will be identical with and without this flag (assuming it
is otherwise "bit-for-bit" reproducible).

=head1 CONFORMS TO

Debian policy, version 3.0.1

=cut

init(options => {
	'keep-debug|keep|k'  => \$dh{K_FLAG},
	'keep-lto'           => \$dh{KEEP_LTO},
	'dbgsym-migration=s' => \$dh{MIGRATE_DBGSYM},
	'automatic-dbgsym!'  => \$dh{ENABLE_DBGSYM},
    # Deprecated variants
	'ddeb-migration=s'   => \$dh{MIGRATE_DBGSYM},
	'ddebs!'             => \$dh{ENABLE_DBGSYM},

});

if ($dh{MIGRATE_DBGSYM}) {
	error("--keep-debug and --dbgsym-migration are mutually exclusive") if ($dh{K_FLAG});
	error("--dbg-package and --dbgsym-migration are mutually exclusive") if ($dh{DEBUGPACKAGE});
}

if ($dh{ENABLE_DBGSYM}) {
	error("--keep-debug and explicit --automatic-dbgsym are mutually exclusive") if ($dh{K_FLAG});
	error("--dbg-package and explicit --automatic-dbgsym are mutually exclusive") if ($dh{DEBUGPACKAGE});
}

$dh{ENABLE_DBGSYM} = 1 if not defined($dh{ENABLE_DBGSYM});

if ($dh{MIGRATE_DBGSYM} and not $dh{ENABLE_DBGSYM}) {
	error("--dbgsym-migration and --no-automatic-dbgsym are mutually exclusive");
}

# This variable can be used to turn off stripping (see Policy).
exit 0 if (get_buildoption('nostrip'));

my $no_auto_dbgsym = 0;
$no_auto_dbgsym = 1 if get_buildoption('noautodbgsym') or get_buildoption('noddebs');

# Check if a file is an elf binary, shared library, or static library,
# for use by File::Find. It'll fill the 3 first arrays with anything
# it finds.  The @build_ids will be the collected build-ids (if any)
my (@shared_libs, @executables, @static_libs, @build_ids, %file_output);
sub testfile {
	my $fn = $_;
	return if -l $fn; # Always skip symlinks.
	
	# See if we were asked to exclude this file.
	# Note that we have to test on the full filename, including directory.
	if (excludefile($fn)) {
		$File::Find::prune = 1 if -d _;
		return;
	}
	# Ignore the .../debug/.build-id/ directory.  It is not really helpful
	# to strip debug symbols.
	$File::Find::prune = 1 if -d _ && index($fn, '/debug/.build-id/') > -1;
	return if -d _;

	# Is it a debug library in a debug subdir?
	return if $fn=~m{debug/.*\.so};
	return if $fn=~m{/guile/.*\.go$};

	# Exploit the previous stat call to get the $mode, so we can check
	# later if it is executable or not.
	#
	# NB: compat() can issue a stat, so we /should/ do this now
	my (undef, undef, $mode, undef) = stat(_);

	if (compat(10)) {
		# In compat 10 and earlier, we used filenames and file(1)

		# Does its filename look like a shared library?
		#  - *.cmxs are OCaml native code shared libraries
		#  - *.node are also native ELF binaries (for node-js)
		if ($fn =~ m/\.(?:so.*?|cmxs|node)$/) {
			# Ok, do the expensive test.
			my $type = get_file_type($fn, 1);
			if ($type =~ m/ELF.*shared/) {
				push @shared_libs, $fn;
				return;
			}
		}

		# -x is not good enough for this test
		if ($mode & 0111) {
			# Ok, expensive test.
			my $type = get_file_type($fn, 1);
			if ($type =~ m/ELF.*(executable|shared)/) {
				push(@executables, $fn);
				return;
			}
		}
	} else {
		# In compat 11, we check the ELF header manually (because bulking file(1) is a pain and
		# it is too slow otherwise)

		if (is_so_or_exec_elf_file($fn)) {
			# -x is not good enough for this test
			if ($mode & 0111) {
				push(@executables, $fn);
			} else {
				push(@shared_libs, $fn);
			}
			return;
		}
	}
	# Is it a static library, and not a debug library?
	if ($fn =~ m/\/lib[^\/]*\.a$/ && $fn !~ m/.*_g\.a$/) {
		# Is it a binary file, or something else (maybe a linker
		# script on Hurd, for example? I don't use file, because
		# file returns a variety of things on static libraries.
		if (-B $fn) {
			push @static_libs, $fn;
			return;
		}
	}
}

sub write_buildid_file {
	my ($package, $build_ids) = @_;
	my $dir = "debian/.debhelper/${package}";
	my $path = "${dir}/dbgsym-build-ids";
	install_dir($dir);
	open(my $fd, '>>', $path) or error("open $path failed: $!");
	print {$fd} join(q{ }, sort(@{$build_ids})) . ' ';
	close($fd) or error("close $path failed: $!");
}

# I could just use `file $_[0]`, but this is safer
sub get_file_type {
	my ($file, $cache_ok) = @_;
	return $file_output{$file} if $cache_ok && $file_output{$file};
	my @file_args = Debian::Debhelper::Dh_Lib::_internal_optional_file_args();
	my @cmdline = ('file', @file_args, '--brief', '-e', 'apptype', '-e', 'ascii', '-e', 'encoding', '-e', 'cdf',
				   '-e', 'compress', '-e', 'tar', '--', $file);

	open(my $fd, '-|', @cmdline) // error("cannot fork+exec file: $!");
	my $type = <$fd>;
	close($fd) || error_exitcode(escape_shell(@cmdline));

	error("file(1) gave no result for $file!?") if (not $type) ;
	return $file_output{$file} = $type;
}

sub make_debug {
	my ($objcopy, $file, $tmp, $desttmp, $use_build_id) = @_;
	my ($debug_path, $debug_build_id);

	# Don't try to copy debug symbols out if the file is already
	# stripped.
	#
	# Disable caching for non-build-id based extractions.
	# Unfortunately, it breaks when there are hardlinks to the same
	# ELF files.
	my $file_info = get_file_type($file, $use_build_id ? 1 : 0);
	return unless $file_info =~ /not stripped/;

	if ($use_build_id) {
		if ($file_info =~ m/BuildID\[sha1]\s*=\s*([0-9a-f]{2})([0-9a-f]+)/ or
			  `LC_ALL=C readelf -n $file`=~ /^\s+Build ID: ([0-9a-f]{2})([0-9a-f]+)$/m) {
			$debug_path=$desttmp."/usr/lib/debug/.build-id/$1/$2.debug";
			$debug_build_id="${1}${2}";
			push(@build_ids, $debug_build_id);
		} else {
			# For dbgsyms, we need build-id (else it will not be
			# co-installable).
			warning("Could not find the BuildID in $file");
			return if $use_build_id > 1;
		}
	}
	if (not $debug_path) {
		# Either not using build_id OR no build-id available
		my ($base_file)=$file=~/^\Q$tmp\E(.*)/;
		$debug_path=$desttmp."/usr/lib/debug/".$base_file;
	}
	install_dir(dirname($debug_path));
	if (compat(8) && $use_build_id < 2) {
		doit($objcopy, "--only-keep-debug", $file, $debug_path);
	}
	else {
		# Compat 9 OR a dbgsym package.
		doit($objcopy, "--only-keep-debug", "--compress-debug-sections", $file, $debug_path) unless -e $debug_path;
	}

	# No reason for this to be executable.
	reset_perm_and_owner(0644, $debug_path);
	return $debug_path;
}

sub attach_debug {
	my ($objcopy, $file, $debug_path) = @_;
	doit($objcopy, "--add-gnu-debuglink", $debug_path, $file);
}

my %all_packages = map { $_ => 1 } getpackages();

sub process_packages {
	foreach my $package (@_) {
		my $tmp=tmpdir($package);
		my $objcopy = cross_command($package, "objcopy");
		my $strip = cross_command($package, "strip");

		# Support for keeping the debugging symbols in a detached file.
		my $keep_debug=$dh{K_FLAG};
		my $debugtmp=$tmp;
		my $use_build_id = compat(8) ? 0 : 1;
		if ($dh{DEBUGPACKAGE}) {
			$keep_debug=1;
			my $debugpackage=$dh{DEBUGPACKAGE};
			error("debug package $debugpackage is not listed in the control file") if (!$all_packages{$debugpackage});
			$debugtmp=tmpdir($debugpackage);
		}
		# Temporary workaround: Do not build dbgsym packages for udebs as
		# dpkg-gencontrol and dpkg-deb does not agree on the file
		# extension.
		if ($dh{ENABLE_DBGSYM} and not $keep_debug and not package_is_arch_all($package) and not is_udeb($package)) {
			# Avoid creating a dbgsym that would clash with a registered
			# package or looks like a manual -dbg package.
			if (not $all_packages{"${package}-dbgsym"} and $package !~ m/-dbg(?:sym)?$/) {
				$debugtmp = dbgsym_tmpdir($package);
				$keep_debug = 1;
				$use_build_id = 2;
			}
		}
		%file_output=@shared_libs=@executables=@static_libs=();
		find({
			wanted => \&testfile,
			no_chdir => 1,
		}, $tmp);

		foreach (@shared_libs) {
			my $debug_path = make_debug($objcopy, $_, $tmp, $debugtmp, $use_build_id) if $keep_debug;
			# Note that all calls to strip on shared libs
			# *must* include the --strip-unneeded.
			doit($strip, "--remove-section=.comment", "--remove-section=.note", "--strip-unneeded", $_);
			attach_debug($objcopy, $_, $debug_path) if defined $debug_path;
		}

		foreach (@executables) {
			my $debug_path = make_debug($objcopy, $_, $tmp, $debugtmp, $use_build_id) if $keep_debug;
			doit($strip, "--remove-section=.comment", "--remove-section=.note", $_);
			attach_debug($objcopy, $_, $debug_path) if defined $debug_path;
		}

		foreach (@static_libs) {
			# NB: The short variant (-D) is broken in Jessie
			# (binutils/2.25-3)
			if ($dh{KEEP_LTO}) {
				doit($strip, '--strip-debug', '--remove-section=.comment',
					 '--remove-section=.note', '--enable-deterministic-archives', $_);
			} else {
				doit($strip, '--strip-debug', '--remove-section=.comment',
					 '--remove-section=.note', '--enable-deterministic-archives',
					 '-R', '.gnu.lto_*', '-R', '.gnu.debuglto_*',
					 '-N', '__gnu_lto_slim', '-N', '__gnu_lto_v1', $_);
			}
		}
		if (-d "$tmp/usr/lib/debug/.dwz" and ($use_build_id > 1 or ($dh{DEBUGPACKAGE} and $dh{DEBUGPACKAGE} ne $package))) {
			my @files = glob_expand(["$tmp/usr/lib/debug/.dwz"], \&glob_expand_error_handler_reject, '*');
			install_dir("$debugtmp/usr/lib/debug/.dwz");
			xargs(\@files, 'cp', '--reflink=auto', "-a", XARGS_INSERT_PARAMS_HERE, "$debugtmp/usr/lib/debug/.dwz");
			doit('rm', '-fr', "$tmp/usr/lib/debug/.dwz");
			doit('rmdir', '-p', '--ignore-fail-on-non-empty', "$tmp/usr/lib/debug");
		}

		if ($no_auto_dbgsym and $use_build_id > 1) {
			# When DEB_BUILD_OPTIONS contains noautodbgsym, remove the
			# dbgsym dir and clear the build-ids.
			#
			# Note we have to extract the dbg symbols as usual, since
			# attach_debug (objcopy --add-gnu-debuglink) requires the dbg
			# file to exist.
			doit('rm', '-fr', $debugtmp);
			@build_ids = ();
		}
		if ($use_build_id > 1 and -d $debugtmp) {
			my $dbgsym_docdir = "${debugtmp}/usr/share/doc";
			my $doc_symlink = "${dbgsym_docdir}/${package}-dbgsym";
			if ( not -l $doc_symlink and not -e _ ) {
				install_dir($dbgsym_docdir);
				make_symlink_raw_target($package, $doc_symlink);
			}
			if ($dh{MIGRATE_DBGSYM}) {
				my $path = "debian/.debhelper/${package}/dbgsym-migration";
				open(my $fd, '>', $path) or error("open $path failed: $!");
				print {$fd} "$dh{MIGRATE_DBGSYM}\n";
				close($fd) or error("close $path failed: $!");
			}
		}
		if ($use_build_id > 1 and @build_ids) {
			write_buildid_file($package, \@build_ids);
			@build_ids = ();
		}
	}
	if (@build_ids and $dh{DEBUGPACKAGE}) {
		write_buildid_file($dh{DEBUGPACKAGE}, \@build_ids);
	}
}

if ($dh{DEBUGPACKAGE}) {
	# Non-deterministic issues with --dbg-package and parallelism (see
	# #872007).  Analysis and patches welcome for this case.
	process_packages(@{$dh{DOPACKAGES}});
} else {
	on_pkgs_in_parallel(\&process_packages);
}

=head1 SEE ALSO

L<debhelper(7)>

This program is a part of debhelper.

=head1 AUTHOR

Joey Hess <joeyh@debian.org>

=cut