package urpm::util; # $Id$ use strict; use Exporter; our @ISA = 'Exporter'; our @EXPORT = qw(min max quotespace unquotespace remove_internal_name reduce_pathname offset_pathname untaint copy_and_own same_size_and_mtime partition put_in_hash uniq uniq_ begins_with difference2 intersection member file_size cat_ cat_utf8 output_safe dirname basename ); (our $VERSION) = q($Revision$) =~ /(\d+)/; sub min { my $n = shift; $_ < $n and $n = $_ foreach @_; $n } sub max { my $n = shift; $_ > $n and $n = $_ foreach @_; $n } #- quoting/unquoting a string that may be containing space chars. sub quotespace { my $x = $_[0] || ''; $x =~ s/(\s)/\\$1/g; $x } sub unquotespace { my $x = $_[0] || ''; $x =~ s/\\(\s)/$1/g; $x } sub remove_internal_name { my $x = $_[0] || ''; $x =~ s/\(\S+\)$/$1/g; $x } sub dirname { local $_ = shift; s|[^/]*/*\s*$||; s|(.)/*$|$1|; $_ || '.' } sub basename { local $_ = shift; s|/*\s*$||; s|.*/||; $_ } #- reduce pathname by removing <something>/.. each time it appears (or . too). sub reduce_pathname { my ($url) = @_; #- clean url to remove any macro (which cannot be solved now). #- take care if this is a true url and not a simple pathname. my ($host, $dir) = $url =~ m|([^:/]*://[^/]*/)?(.*)|; $host = '' if !defined $host; #- remove any multiple /s or trailing /. #- then split all components of pathname. $dir =~ s|/+|/|g; $dir =~ s|/$||; my @paths = split '/', $dir; #- reset $dir, recompose it, and clean trailing / added by algorithm. $dir = ''; foreach (@paths) { if ($_ eq '..') { if ($dir =~ s|([^/]+)/$||) { if ($1 eq '..') { $dir .= "../../"; } } else { $dir .= "../"; } } elsif ($_ ne '.') { $dir .= "$_/"; } } $dir =~ s|/$||; $dir ||= '/'; $host . $dir; } #- offset pathname by returning the right things to add to a relative directory #- to make no change. url is needed to resolve going before to top base. sub offset_pathname { my ($url, $offset) = map { reduce_pathname($_) } @_; #- clean url to remove any macro (which cannot be solved now). #- take care if this is a true url and not a simple pathname. my (undef, $dir) = $url =~ m|([^:/]*://[^/]*/)?(.*)|; my @paths = split '/', $dir; my @offpaths = reverse split '/', $offset; my @corrections; my $result = ''; foreach (@offpaths) { if ($_ eq '..') { push @corrections, pop @paths; } else { $result .= '../'; } } $result . join('/', reverse @corrections); } sub untaint { my @r = map { /(.*)/ } @_; @r == 1 ? $r[0] : @r; } sub copy { my ($file, $dest) = @_; !system("/bin/cp", "-p", "-L", "-R", $file, $dest); } sub copy_and_own { my ($file, $dest_file) = @_; copy($file, $dest_file) && chown(0, 0, $dest_file) == 1; } sub move { my ($file, $dest) = @_; rename($file, $dest) || !system("/bin/mv", "-f", $file, $dest); } #- file_size is useful to write file_size(...) > 32 without having warnings if file doesn't exist sub file_size { my ($file) = @_; -s $file || 0; } sub same_size_and_mtime { my ($f1, $f2) = @_; my @sstat = stat $f1; my @lstat = stat $f2; $sstat[7] == $lstat[7] && $sstat[9] == $lstat[9]; } sub partition(&@) { my $f = shift; my (@a, @b); foreach (@_) { $f->($_) ? push(@a, $_) : push(@b, $_); } \@a, \@b; } sub begins_with { my ($s, $prefix) = @_; index($s, $prefix) == 0; } sub put_in_hash { my ($a, $b) = @_; while (my ($k, $v) = each %{$b || {}}) { $a->{$k} = $v } $a } sub uniq { my %l; $l{$_} = 1 foreach @_; grep { delete $l{$_} } @_ } sub difference2 { my %l; @l{@{$_[1]}} = (); grep { !exists $l{$_} } @{$_[0]} } sub intersection { my (%l, @m); @l{@{shift @_}} = (); foreach (@_) { @m = grep { exists $l{$_} } @$_; %l = (); @l{@m} = () } keys %l } sub member { my $e = shift; foreach (@_) { $e eq $_ and return 1 } 0 } sub cat_ { my @l = map { my $F; open($F, '<', $_) ? <$F> : () } @_; wantarray() ? @l : join '', @l } sub cat_utf8 { my @l = map { my $F; open($F, '<:utf8', $_) ? <$F> : () } @_; wantarray() ? @l : join '', @l } sub uniq_(&@) { my $f = shift; my %l; $l{$f->($_)} = 1 foreach @_; grep { delete $l{$f->($_)} } @_; } sub output_safe { my ($file, $content) = @_; open(my $f, '>', "$file.new") or return; print $f $content or return; close $f or return; rename("$file.new", $file) or return; 1; } 1; __END__ =head1 NAME urpm::util - Misc. utilities subs for urpmi =head1 SYNOPSIS =head1 DESCRIPTION =head1 COPYRIGHT Copyright (C) 2005 MandrakeSoft SA Copyright (C) 2005, 2006 Mandriva SA =cut