unicode-password.test 7.52 KB
#!/usr/bin/env perl
require 5.008;
use warnings;
use strict;

unshift(@INC, '.');
require qpdf_test_helpers;

chdir("qpdf") or die "chdir testdir failed: $!\n";

require TestDriver;

cleanup();

my $td = new TestDriver('unicode-password');

my $n_tests = 0;
# $n_tests incremented below

# Files with each of these passwords when properly encoded have been
# tested manually with multiple PDF viewers. Adobe Reader, chrome,
# xpdf, and gv can open all of them except R3 with "single-byte",
# which can be opened by xpdf and gv but not the others. As of
# 2019-01-19, okular and atril (evince) are not able to open R=6 files
# with Unicode passwords as generated by qpdf but can open the R=3
# files.

# [bits, password-or-password-name, write-encoding, actual-encoding, xargs,
#  [[read-encoding, strict?, fail?, tried-others, xargs]]]
my @unicode_pw_cases = (
    [128, 'simple', 'pdf-doc', 'pdf-doc', '',
     [['utf8', 0, 0, 1, ''],
      ['utf8', 1, 1, 0, ''],
      ['pdf-doc', 1, 0, 0, ''],
     ]],
    [128, 'simple', 'utf8', 'utf8', '--password-mode=bytes',
     [['pdf-doc', 0, 0, 1, ''],
      ['pdf-doc', 1, 1, 0, ''],
      ['utf8', 1, 0, 0, ''],
     ]],
    [128, 'simple', 'utf8', 'pdf-doc', '--password-mode=unicode',
     [['pdf-doc', 1, 0, 0, ''],
     ]],
    [128, 'simple', 'utf8', 'pdf-doc', '--password-mode=auto',
     [['pdf-doc', 1, 0, 0, ''],
     ]],
    [128, 'single-byte', 'utf8', 'pdf-doc', '',
     [['pdf-doc', 1, 0, 0, ''],
      ['win-ansi', 0, 0, 1, ''],
     ]],
    [128, 'single-byte', 'utf8', 'pdf-doc', '--password-mode=unicode',
     [['pdf-doc', 1, 0, 0, ''],
      ['win-ansi', 0, 0, 1, ''],
     ]],
    [128, 'single-byte', 'win-ansi', '', '--password-mode=unicode',
     "supplied password is not valid UTF-8\n",
    ],
    [128, 'single-byte', 'win-ansi', 'win-ansi', '',
     [['win-ansi', 1, 0, 0, ''],
     ]],
    [128, 'single-byte', 'pdf-doc', 'pdf-doc', '',
     [['pdf-doc', 1, 0, 0, ''],
      ['win-ansi', 0, 0, 1, ''],
      ['pdf-doc-hex', 1, 0, 0, '--password-mode=hex-bytes'],
     ]],
    [128, 'complex', 'utf8', '', '--password-mode=unicode',
     "supplied password cannot be encoded for 40-bit or" .
     " 128-bit encryption formats\n"
    ],
    [128, 'complex', 'utf8', 'utf8', '--password-mode=bytes',
     [['utf8', 1, 0, 0, ''],
     ]],
    [256, 'single-byte', 'win-ansi', '', '--password-mode=unicode',
     "supplied password is not valid UTF-8\n",
    ],
    [256, 'single-byte', 'win-ansi', '', '--password-mode=auto',
     "supplied password is not a valid Unicode password, which is" .
     " required for 256-bit encryption; to really use this password," .
     " rerun with the --password-mode=bytes option\n",
    ],
    [256, 'single-byte', 'win-ansi', 'win-ansi', '--password-mode=bytes',
     [['utf8', 0, 0, 1, ''],
      ['utf8', 1, 1, 0, ''],
      ['win-ansi', 1, 0, 0, ''],
      ['win-ansi', 0, 0, 0, ''],
      ['pdf-doc', 0, 0, 1, ''],
      ['pdf-doc-hex', 0, 0, 1, '--password-mode=hex-bytes'],
     ]],
    [256, 'complex', 'utf8', 'utf8', '',
     [['utf8', 1, 0, 0, ''],
      ['utf8-hex', 1, 0, 0, '--password-mode=hex-bytes'],
     ]],
    [256, 'complex', 'utf8-hex', 'utf8', '--password-mode=hex-bytes',
     [['utf8', 1, 0, 0, ''],
      ['utf8-hex', 1, 0, 0, '--password-mode=hex-bytes'],
     ]],
    [256, 'complex', 'utf8', 'utf8', '--password-mode=unicode',
     [['utf8', 1, 0, 0, ''],
      ['password-arg-simple-utf8', 0, 1, 1, ''],
     ]],
    );

for my $d (@unicode_pw_cases)
{
    my $decode_cases = $d->[5];
    $n_tests += 1;
    if (ref($decode_cases) eq 'ARRAY')
    {
        $n_tests += scalar(@$decode_cases);
    }
}

foreach my $d (@unicode_pw_cases)
{
    my ($bits, $pw, $w_encoding, $a_encoding, $xargs, $decode_cases) = @$d;
    my $w_pfile = "password-bare-$pw-$w_encoding";
    my $upass;
    if (-f $w_pfile)
    {
        $upass = '@' . $w_pfile;
    }
    else
    {
        $upass = "$pw";
    }
    my $outbase = "unicode-pw-$bits-$pw-$w_encoding-$xargs";
    my $exp = '';
    if (ref($decode_cases) ne 'ARRAY')
    {
        $exp = "qpdf: $decode_cases";
        $decode_cases = [];
    }
    $td->runtest("encode $bits, $pw, $w_encoding",
                 {$td->COMMAND =>
                      "qpdf $xargs --static-id --static-aes-iv" .
                      " --allow-weak-crypto" .
                      " --encrypt $upass o $bits -- minimal.pdf a.pdf"},
                 {$td->STRING => $exp, $td->EXIT_STATUS => ($exp ? 2 : 0)},
                 $td->NORMALIZE_NEWLINES);
    foreach my $d2 (@$decode_cases)
    {
        my ($r_encoding, $strict, $xfail, $tried_others, $r_xargs) = @$d2;
        my $r_pfile = "password-arg-$pw-$r_encoding";
        if (! -f $r_pfile)
        {
            $r_pfile = $r_encoding;
        }
        my $r_output = "";
        $r_output .= "trying other\n" if $tried_others;
        my $arg = "--show-encryption";
        if ($xfail)
        {
            $r_output .= "qpdf: a.pdf: invalid password\n";
            $arg = "--check";
        }
        else
        {
            $r_output .= "R = " . ($bits == 128 ? '3' : '6') . "\n";
            open(F, "<password-bare-$pw-$a_encoding") or die;
            chomp (my $apw = <F>);
            close(F);
            $r_output .= "User password = $apw\n";
        }
        $r_xargs .= $strict ? ' --suppress-password-recovery' : '';
        $td->runtest("decrypt $pw, $r_encoding, strict=$strict",
                     {$td->COMMAND =>
                          "qpdf $arg --verbose" .
                          " $r_xargs a.pdf \@$r_pfile",
                          $td->FILTER => "perl show-unicode-encryption.pl"},
                     {$td->STRING => "$r_output",
                          $td->EXIT_STATUS => ($xfail ? 2 : 0)},
                     $td->NORMALIZE_NEWLINES);
    }
}

$n_tests += 5;

$td->runtest("bytes fallback warning",
             {$td->COMMAND =>
                  "qpdf --allow-weak-crypto" .
                  " --encrypt \@password-bare-complex-utf8 o 128 --" .
                  " minimal.pdf a.pdf"},
             {$td->FILE => "bytes-fallback.out", $td->EXIT_STATUS => 0},
             $td->NORMALIZE_NEWLINES);
{ # local scope
    my $r_output = "";
    $r_output .= "R = 3\n";
    open(F, "<password-bare-complex-utf8") or die;
    chomp (my $apw = <F>);
    close(F);
    $r_output .= "User password = $apw\n";
    $td->runtest("decrypt bytes fallback",
                 {$td->COMMAND =>
                      "qpdf --show-encryption --verbose" .
                      " a.pdf \@password-arg-complex-utf8" .
                      " --password-mode=bytes",
                      $td->FILTER => "perl show-unicode-encryption.pl"},
                 {$td->STRING => "$r_output", $td->EXIT_STATUS => 0},
                 $td->NORMALIZE_NEWLINES);
}

# Exercise passing Unicode passwords via the command line. This tests
# wmain for Windows and assumes a UTF-8 locale for other platforms.
$td->runtest("Unicode at CLI",
             {$td->COMMAND =>
                  "qpdf --encrypt π ʬ 256 --" .
                  " minimal.pdf a.pdf"},
             {$td->STRING => "", $td->EXIT_STATUS => 0},
             $td->NORMALIZE_NEWLINES);
$td->runtest("Decrypt using user password",
             {$td->COMMAND => "qpdf --show-encryption a.pdf --password=π"},
             {$td->FILE => "unicode-up.out", $td->EXIT_STATUS => 0},
             $td->NORMALIZE_NEWLINES);
$td->runtest("Decrypt using owner password",
             {$td->COMMAND => "qpdf --show-encryption a.pdf --password=ʬ"},
             {$td->FILE => "unicode-op.out", $td->EXIT_STATUS => 0},
             $td->NORMALIZE_NEWLINES);

cleanup();
$td->report($n_tests);