xref: /openbsd/gnu/usr.bin/perl/t/op/pack.t (revision e0680481)
1#!./perl -w
2
3BEGIN {
4    chdir 't' if -d 't';
5    require './test.pl'; require './charset_tools.pl';
6    set_up_inc(qw '../lib ../cpan/Math-BigInt/lib');
7}
8
9plan tests => 14722;
10
11use strict;
12use warnings qw(FATAL all);
13use Config;
14
15my $Perl = which_perl();
16my @valid_errors = (qr/^Invalid type '\w'/);
17
18my $ByteOrder = 'unknown';
19my $maybe_not_avail = '(?:hto[bl]e|[bl]etoh)';
20if ($Config{byteorder} =~ /^1234(?:5678)?$/) {
21  $ByteOrder = 'little';
22  $maybe_not_avail = '(?:htobe|betoh)';
23}
24elsif ($Config{byteorder} =~ /^(?:8765)?4321$/) {
25  $ByteOrder = 'big';
26  $maybe_not_avail = '(?:htole|letoh)';
27}
28else {
29  push @valid_errors, qr/^Can't (?:un)?pack (?:big|little)-endian .*? on this platform/;
30}
31
32for my $size ( 16, 32, 64 ) {
33  if (defined $Config{"u${size}size"} and ($Config{"u${size}size"}||0) != ($size >> 3)) {
34    push @valid_errors, qr/^Perl_my_$maybe_not_avail$size\(\) not available/;
35  }
36}
37
38my $IsTwosComplement = pack('i', -1) eq "\xFF" x $Config{intsize};
39print "# \$IsTwosComplement = $IsTwosComplement\n";
40
41sub is_valid_error
42{
43  my $err = shift;
44
45  for my $e (@valid_errors) {
46    $err =~ $e and return 1;
47  }
48
49  return 0;
50}
51
52sub encode_list {
53  my @result = map {_qq($_)} @_;
54  if (@result == 1) {
55    return @result;
56  }
57  return '(' . join (', ', @result) . ')';
58}
59
60
61sub list_eq ($$) {
62  my ($l, $r) = @_;
63  return 0 unless @$l == @$r;
64  for my $i (0..$#$l) {
65    if (defined $l->[$i]) {
66      return 0 unless defined ($r->[$i]) && $l->[$i] eq $r->[$i];
67    } else {
68      return 0 if defined $r->[$i]
69    }
70  }
71  return 1;
72}
73
74##############################################################################
75#
76# Here starteth the tests
77#
78
79{
80    my $format = "c2 x5 C C x s d i l a6";
81    # Need the expression in here to force ary[5] to be numeric.  This avoids
82    # test2 failing because ary2 goes str->numeric->str and ary doesn't.
83    my @ary = (1,-100,127,128,32767,987.654321098 / 100.0,12345,123456,
84               "abcdef");
85    my $foo = pack($format,@ary);
86    my @ary2 = unpack($format,$foo);
87
88    is($#ary, $#ary2);
89
90    my $out1=join(':',@ary);
91    my $out2=join(':',@ary2);
92    # Using long double NVs may introduce greater accuracy than wanted.
93    $out1 =~ s/:9\.87654321097999\d*:/:9.87654321098:/;
94    $out2 =~ s/:9\.87654321097999\d*:/:9.87654321098:/;
95    is($out1, $out2);
96
97    like($foo, qr/def/);
98}
99# How about counting bits?
100
101{
102    my $x;
103    is( ($x = unpack("%32B*", "\001\002\004\010\020\040\100\200\377")), 16 );
104
105    is( ($x = unpack("%32b69", "\001\002\004\010\020\040\100\200\017")), 12 );
106
107    is( ($x = unpack("%32B69", "\001\002\004\010\020\040\100\200\017")), 9 );
108}
109
110{
111    my $sum = 129; # ASCII
112    $sum = 103 if $::IS_EBCDIC;
113
114    my $x;
115    is( ($x = unpack("%32B*", "Now is the time for all good blurfl")), $sum );
116
117    my $foo;
118    open(BIN, $Perl) || die "Can't open $Perl: $!\n";
119    binmode BIN;
120    sysread BIN, $foo, 8192;
121    close BIN;
122
123    $sum = unpack("%32b*", $foo);
124    my $longway = unpack("b*", $foo);
125    is( $sum, $longway =~ tr/1/1/ );
126}
127
128{
129  my $x;
130  is( ($x = unpack("I",pack("I", 0xFFFFFFFF))), 0xFFFFFFFF );
131}
132
133{
134    note("check 'w'");
135    my @x = (5,130,256,560,32000,3097152,268435455,1073741844, 2**33,
136             '4503599627365785','23728385234614992549757750638446');
137    my $x = pack('w*', @x);
138    my $y = pack 'H*', '0581028200843081fa0081bd8440ffffff7f8480808014A0808'.
139                       '0800087ffffffffffdb19caefe8e1eeeea0c2e1e3e8ede1ee6e';
140
141    is($x, $y);
142
143    my @y = unpack('w*', $y);
144    my $a;
145    while ($a = pop @x) {
146        my $b = pop @y;
147        is($a, $b);
148    }
149
150    @y = unpack('w2', $x);
151
152    is(scalar(@y), 2);
153    is($y[1], 130);
154
155    SKIP: {
156        skip "no Scalar::Util (Math::BigInt prerequisite) on miniperl", 10, if is_miniperl();
157        $x = pack('w*', 5000000000); $y = '';
158        eval q{
159            use Math::BigInt;
160            $y = pack('w*', Math::BigInt::->new(5000000000));
161        };
162        is($x, $y, 'pack');
163
164        $x = pack 'w', ~0;
165        $y = pack 'w', (~0).'';
166        is($x, $y);
167        is(unpack ('w',$x), ~0);
168        is(unpack ('w',$y), ~0);
169
170        $x = pack 'w', ~0 - 1;
171        $y = pack 'w', (~0) - 2;
172
173        if (~0 - 1 == (~0) - 2) {
174            is($x, $y, "NV arithmetic");
175        } else {
176            isnt($x, $y, "IV/NV arithmetic");
177        }
178        cmp_ok(unpack ('w',$x), '==', ~0 - 1);
179        cmp_ok(unpack ('w',$y), '==', ~0 - 2);
180
181        # These should spot that pack 'w' is using NV, not double, on platforms
182        # where IVs are smaller than doubles, and harmlessly pass elsewhere.
183        # (tests for change 16861)
184        my $x0 = 2**54+3;
185        my $y0 = 2**54-2;
186
187        $x = pack 'w', $x0;
188        $y = pack 'w', $y0;
189
190        if ($x0 == $y0) {
191            is($x, $y, "NV arithmetic");
192        } else {
193            isnt($x, $y, "IV/NV arithmetic");
194        }
195        cmp_ok(unpack ('w',$x), '==', $x0);
196        cmp_ok(unpack ('w',$y), '==', $y0);
197    }
198}
199
200
201{
202  print "# test exceptions\n";
203  my $x;
204  eval { $x = unpack 'w', pack 'C*', 0xff, 0xff};
205  like($@, qr/^Unterminated compressed integer/);
206
207  eval { $x = unpack 'w', pack 'C*', 0xff, 0xff, 0xff, 0xff};
208  like($@, qr/^Unterminated compressed integer/);
209
210  eval { $x = unpack 'w', pack 'C*', 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff};
211  like($@, qr/^Unterminated compressed integer/);
212
213  eval { $x = pack 'w', -1 };
214  like ($@, qr/^Cannot compress negative numbers/);
215
216  eval { $x = pack 'w', '1'x(1 + length ~0) . 'e0' };
217  like ($@, qr/^Can only compress unsigned integers/);
218
219  # Check that the warning behaviour on the modifiers !, < and > is as we
220  # expect it for this perl.
221  my $can_endian = 'sSiIlLqQjJfFdDpP';
222  my $can_shriek = 'sSiIlLnNvV';
223  # h and H can't do either, so act as sanity checks in blead
224  foreach my $base (split '', 'hHsSiIlLqQjJfFdDpPnNvV') {
225    foreach my $mod ('', '<', '>', '!', '<!', '>!', '!<', '!>') {
226    SKIP: {
227	# Avoid void context warnings.
228	my $a = eval {pack "$base$mod"};
229	skip "pack can't $base", 1 if $@ =~ /^Invalid type '\w'/;
230	# Which error you get when 2 would be possible seems to be emergent
231	# behaviour of pack's format parser.
232
233	my $fails_shriek = $mod =~ /!/ && index ($can_shriek, $base) == -1;
234	my $fails_endian = $mod =~ /[<>]/ && index ($can_endian, $base) == -1;
235	my $shriek_first = $mod =~ /^!/;
236
237	if ($fails_shriek and $fails_endian) {
238	  if ($shriek_first) {
239	    undef $fails_endian;
240	  }
241	}
242
243	if ($fails_endian) {
244          like ($@, qr/^'[<>]' allowed only after types/,
245                "pack can't $base$mod");
246	} elsif ($fails_shriek) {
247	  like ($@, qr/^'!' allowed only after types/,
248		"pack can't $base$mod");
249	} else {
250	  is ($@, '', "pack can $base$mod");
251	}
252      }
253    }
254  }
255
256  for my $mod (qw( ! < > )) {
257    eval { $x = pack "a$mod", 42 };
258    like ($@, qr/^'$mod' allowed only after types \S+ in pack/);
259
260    eval { $x = unpack "a$mod", 'x'x8 };
261    like ($@, qr/^'$mod' allowed only after types \S+ in unpack/);
262  }
263
264  for my $mod (qw( <> >< !<> !>< <!> >!< <>! ><! )) {
265    eval { $x = pack "sI${mod}s", 42, 47, 11 };
266    like ($@, qr/^Can't use both '<' and '>' after type 'I' in pack/);
267
268    eval { $x = unpack "sI${mod}s", 'x'x16 };
269    like ($@, qr/^Can't use both '<' and '>' after type 'I' in unpack/);
270  }
271
272 SKIP: {
273    # Is this a stupid thing to do on VMS, VOS and other unusual platforms?
274
275    skip("-- the IEEE infinity model is unavailable in this configuration.", 1)
276       if (($^O eq 'VMS') && !defined($Config{useieee}) || !$Config{d_double_has_inf});
277
278    skip("-- $^O has serious fp indigestion on w-packed infinities", 1)
279       if $^O =~ /^svr4/ && -f "/etc/issue" && -f "/etc/.relid";  # NCR MP-RAS
280
281    my $inf = eval '2**1000000';
282
283    skip("Couldn't generate infinity - got error '$@'", 1)
284      unless defined $inf and $inf == $inf / 2 and $inf + 1 == $inf;
285
286    local our $TODO;
287    $TODO = "VOS needs a fix for posix-1022 to pass this test."
288      if ($^O eq 'vos');
289
290    eval { $x = pack 'w', $inf };
291    like ($@, qr/^Cannot compress Inf/, "Cannot compress infinity");
292  }
293
294 SKIP: {
295
296    skip("-- the full range of an IEEE double may not be available in this configuration.", 3)
297       if (($^O eq 'VMS') && !defined($Config{useieee}) || !$Config{d_double_style_ieee});
298
299    # This should be about the biggest thing possible on an IEEE double
300    my $big = eval '2**1023';
301
302    skip("Couldn't generate 2**1023 - got error '$@'", 3)
303      unless defined $big and $big != $big / 2;
304
305    eval { $x = pack 'w', $big };
306    is ($@, '', "Should be able to pack 'w', $big # 2**1023");
307
308    my $y = eval {unpack 'w', $x};
309    is ($@, '',
310	"Should be able to unpack 'w' the result of pack 'w', $big # 2**1023");
311
312    # I'm getting about 1e-16 on FreeBSD
313    my $quotient = int (100 * ($y - $big) / $big);
314    ok($quotient < 2 && $quotient > -2,
315       "Round trip pack, unpack 'w' of $big is within 1% ($quotient%)");
316  }
317
318}
319
320print "# test the 'p' template\n";
321
322# literals
323is(unpack("p",pack("p","foo")), "foo");
324SKIP: {
325  is(unpack("p<",pack("p<","foo")), "foo");
326  is(unpack("p>",pack("p>","foo")), "foo");
327}
328# scalars
329is(unpack("p",pack("p",239)), 239);
330is(unpack("p<",pack("p<",239)), 239);
331is(unpack("p>",pack("p>",239)), 239);
332
333# temps
334sub foo { my $a = "a"; return $a . $a++ . $a++ }
335{
336  use warnings qw(NONFATAL all);;
337  my $warning;
338  local $SIG{__WARN__} = sub {
339      $warning = $_[0];
340  };
341  my $junk = pack("p", &foo);
342
343  like($warning, qr/temporary val/);
344}
345
346# undef should give null pointer
347like(pack("p", undef), qr/^\0+$/);
348like(pack("p<", undef), qr/^\0+$/);
349like(pack("p>", undef), qr/^\0+$/);
350
351# Check for optimizer bug (e.g.  Digital Unix GEM cc with -O4 on DU V4.0B gives
352#                                4294967295 instead of -1)
353#				 see #ifdef __osf__ in pp.c pp_unpack
354is((unpack("i",pack("i",-1))), -1);
355
356print "# test the pack lengths of s S i I l L n N v V + modifiers\n";
357
358my @lengths = (
359  qw(s 2 S 2 i -4 I -4 l 4 L 4 n 2 N 4 v 2 V 4 n! 2 N! 4 v! 2 V! 4),
360  's!'  => $Config{shortsize}, 'S!'  => $Config{shortsize},
361  'i!'  => $Config{intsize},   'I!'  => $Config{intsize},
362  'l!'  => $Config{longsize},  'L!'  => $Config{longsize},
363);
364
365while (my ($base, $expect) = splice @lengths, 0, 2) {
366  my @formats = ($base);
367  $base =~ /^[nv]/i or push @formats, "$base>", "$base<";
368  for my $format (@formats) {
369    my $len = length(pack($format, 0));
370    if ($expect > 0) {
371      is($expect, $len, "format '$format'");
372    } else {
373      $expect = -$expect;
374      ok ($len >= $expect, "format '$format'") ||
375	  print "# format '$format' has length $len, expected >= $expect\n";
376    }
377  }
378}
379
380
381print "# test unpack-pack lengths\n";
382
383my @templates = qw(c C W i I s S l L n N v V f d q Q);
384
385foreach my $base (@templates) {
386    my @tmpl = ($base);
387    $base =~ /^[cwnv]/i or push @tmpl, "$base>", "$base<";
388    foreach my $t (@tmpl) {
389        SKIP: {
390            my @t = eval { unpack("$t*", pack("$t*", 12, 34)) };
391
392            skip "cannot pack '$t' on this perl", 4
393              if is_valid_error($@);
394
395            is( $@, '', "Template $t works");
396            is(scalar @t, 2);
397
398            is($t[0], 12);
399            is($t[1], 34);
400        }
401    }
402}
403
404{
405    # uuencode/decode
406
407    # Note that first uuencoding known 'text' data and then checking the
408    # binary values of the uuencoded version would not be portable between
409    # character sets.  Uuencoding is meant for encoding binary data, not
410    # text data.
411
412    my $in = pack 'C*', 0 .. 255;
413
414    # just to be anal, we do some random tr/`/ /
415    my $uu = <<'EOUU';
416M` $"`P0%!@<("0H+# T.#Q`1$A,4%187&!D:&QP='A\@(2(C)"4F)R@I*BLL
417M+2XO,#$R,S0U-C<X.3H[/#T^/T!!0D-$149'2$E*2TQ-3D]045)35%565UA9
418M6EM<75Y?8&%B8V1E9F=H:6IK;&UN;W!Q<G-T=79W>'EZ>WQ]?G^`@8*#A(6&
419MAXB)BHN,C8Z/D)&2DY25EI>8F9J;G)V>GZ"AHJ.DI::GJ*FJJZRMKJ^PL;*S
420MM+6VM[BYNKN\O;Z_P,'"P\3%QL?(R<K+S,W.S]#1TM/4U=;7V-G:V]S=WM_@
421?X>+CY.7FY^CIZNOL[>[O\/'R\_3U]O?X^?K[_/W^_P `
422EOUU
423
424    $_ = $uu;
425    tr/ /`/;
426
427    is(pack('u', $in), $_);
428
429    is(unpack('u', $uu), $in);
430
431    $in = "\x1f\x8b\x08\x08\x58\xdc\xc4\x35\x02\x03\x4a\x41\x50\x55\x00\xf3\x2a\x2d\x2e\x51\x48\xcc\xcb\x2f\xc9\x48\x2d\x52\x08\x48\x2d\xca\x51\x28\x2d\x4d\xce\x4f\x49\x2d\xe2\x02\x00\x64\x66\x60\x5c\x1a\x00\x00\x00";
432    $uu = <<'EOUU';
433M'XL("%C<Q#4"`TI!4%4`\RHM+E%(S,LOR4@M4@A(+<I1*"U-SD])+>("`&1F
434&8%P:````
435EOUU
436
437    is(unpack('u', $uu), $in);
438
439# This is identical to the above except that backquotes have been
440# changed to spaces
441
442    $uu = <<'EOUU';
443M'XL("%C<Q#4" TI!4%4 \RHM+E%(S,LOR4@M4@A(+<I1*"U-SD])+>(" &1F
444&8%P:
445EOUU
446
447    # ' # Grr
448    is(unpack('u', $uu), $in);
449
450}
451
452# test the ascii template types (A, a, Z)
453
454foreach (
455['p', 'A*',  "foo\0bar\0 ", "foo\0bar\0 "],
456['p', 'A11', "foo\0bar\0 ", "foo\0bar\0   "],
457['u', 'A*',  "foo\0bar \0", "foo\0bar"],
458['u', 'A8',  "foo\0bar \0", "foo\0bar"],
459['p', 'a*',  "foo\0bar\0 ", "foo\0bar\0 "],
460['p', 'a11', "foo\0bar\0 ", "foo\0bar\0 \0\0"],
461['u', 'a*',  "foo\0bar \0", "foo\0bar \0"],
462['u', 'a8',  "foo\0bar \0", "foo\0bar "],
463['p', 'Z*',  "foo\0bar\0 ", "foo\0bar\0 \0"],
464['p', 'Z11', "foo\0bar\0 ", "foo\0bar\0 \0\0"],
465['p', 'Z3',  "foo",         "fo\0"],
466['u', 'Z*',  "foo\0bar \0", "foo"],
467['u', 'Z8',  "foo\0bar \0", "foo"],
468)
469{
470    my ($what, $template, $in, $out) = @$_;
471    my $got = $what eq 'u' ? (unpack $template, $in) : (pack $template, $in);
472    unless (is($got, $out)) {
473        my $un = $what eq 'u' ? 'un' : '';
474        print "# ${un}pack ('$template', "._qq($in).') gave '._qq($out).
475            ' not '._qq($got)."\n";
476    }
477}
478
479print "# packing native shorts/ints/longs\n";
480
481is(length(pack("s!", 0)), $Config{shortsize});
482is(length(pack("i!", 0)), $Config{intsize});
483is(length(pack("l!", 0)), $Config{longsize});
484ok(length(pack("s!", 0)) <= length(pack("i!", 0)));
485ok(length(pack("i!", 0)) <= length(pack("l!", 0)));
486is(length(pack("i!", 0)), length(pack("i", 0)));
487
488sub numbers {
489  my $base = shift;
490  my @formats = ($base);
491  $base =~ /^[silqjfdp]/i and push @formats, "$base>", "$base<";
492  for my $format (@formats) {
493    numbers_with_total ($format, undef, @_);
494  }
495}
496
497sub numbers_with_total {
498  my $format = shift;
499  my $total = shift;
500  if (!defined $total) {
501    foreach (@_) {
502      $total += $_;
503    }
504  }
505  print "# numbers test for $format\n";
506  foreach (@_) {
507    SKIP: {
508        my $out = eval {unpack($format, pack($format, $_))};
509        skip "cannot pack '$format' on this perl", 2
510          if is_valid_error($@);
511
512        is($@, '', "no error");
513        is($out, $_, "unpack pack $format $_");
514    }
515  }
516
517  my $skip_if_longer_than = ~0; # "Infinity"
518  if (~0 - 1 == ~0) {
519    # If we're running with -DNO_PERLPRESERVE_IVUV and NVs don't preserve all
520    # UVs (in which case ~0 is NV, ~0-1 will be the same NV) then we can't
521    # correctly in perl calculate UV totals for long checksums, as pp_unpack
522    # is using UV maths, and we've only got NVs.
523    $skip_if_longer_than = $Config{nv_preserves_uv_bits};
524  }
525
526  foreach ('', 1, 2, 3, 15, 16, 17, 31, 32, 33, 53, 54, 63, 64, 65) {
527    SKIP: {
528      my $sum = eval {unpack "%$_$format*", pack "$format*", @_};
529      skip "cannot pack '$format' on this perl", 3
530        if is_valid_error($@);
531
532      is($@, '', "no error");
533      ok(defined $sum, "sum bits $_, format $format defined");
534
535      my $len = $_; # Copy, so that we can reassign ''
536      $len = 16 unless length $len;
537
538      SKIP: {
539        skip "cannot test checksums over $skip_if_longer_than bits", 1
540          if $len > $skip_if_longer_than;
541
542        # Our problem with testing this portably is that the checksum code in
543        # pp_unpack is able to cast signed to unsigned, and do modulo 2**n
544        # arithmetic in unsigned ints, which perl has no operators to do.
545        # (use integer; does signed ints, which won't wrap on UTS, which is just
546        # fine with ANSI, but not with most people's assumptions.
547        # This is why we need to supply the totals for 'Q' as there's no way in
548        # perl to calculate them, short of unpack '%0Q' (is that documented?)
549        # ** returns NVs; make sure it's IV.
550        my $max = 1 + 2 * (int (2 ** ($len-1))-1); # The max possible checksum
551        my $max_p1 = $max + 1;
552        my ($max_is_integer, $max_p1_is_integer);
553        $max_p1_is_integer = 1 unless $max_p1 + 1 == $max_p1;
554        $max_is_integer = 1 if $max - 1 < ~0;
555
556        my $calc_sum;
557        if (ref $total) {
558            $calc_sum = &$total($len);
559        } else {
560            $calc_sum = $total;
561            # Shift into range by some multiple of the total
562            my $mult = $max_p1 ? int ($total / $max_p1) : undef;
563            # Need this to make sure that -1 + (~0+1) is ~0 (ie still integer)
564            $calc_sum = $total - $mult;
565            $calc_sum -= $mult * $max;
566            if ($calc_sum < 0) {
567                $calc_sum += 1;
568                $calc_sum += $max;
569            }
570        }
571        if ($calc_sum == $calc_sum - 1 && $calc_sum == $max_p1) {
572            # we're into floating point (either by getting out of the range of
573            # UV arithmetic, or because we're doing a floating point checksum)
574            # and our calculation of the checksum has become rounded up to
575            # max_checksum + 1
576            $calc_sum = 0;
577        }
578
579        if ($calc_sum == $sum) { # HAS to be ==, not eq (so no is()).
580            pass ("unpack '%$_$format' gave $sum");
581        } else {
582            my $delta = 1.000001;
583            if ($format =~ tr /dDfF//
584                && ($calc_sum <= $sum * $delta && $calc_sum >= $sum / $delta)) {
585                pass ("unpack '%$_$format' gave $sum, expected $calc_sum");
586            } else {
587                my $text = ref $total ? &$total($len) : $total;
588                fail;
589                print "# For list (" . join (", ", @_) . ") (total $text)"
590                    . " packed with $format unpack '%$_$format' gave $sum,"
591                    . " expected $calc_sum\n";
592            }
593        }
594      }
595    }
596  }
597}
598
599numbers ('c', -128, -1, 0, 1, 127);
600numbers ('C', 0, 1, 127, 128, 255);
601numbers ('W', 0, 1, 127, 128, 255, 256, 0x7ff, 0x800, 0xfffd);
602numbers ('s', -32768, -1, 0, 1, 32767);
603numbers ('S', 0, 1, 32767, 32768, 65535);
604numbers ('i', -2147483648, -1, 0, 1, 2147483647);
605numbers ('I', 0, 1, 2147483647, 2147483648, 4294967295);
606numbers ('l', -2147483648, -1, 0, 1, 2147483647);
607numbers ('L', 0, 1, 2147483647, 2147483648, 4294967295);
608numbers ('s!', -32768, -1, 0, 1, 32767);
609numbers ('S!', 0, 1, 32767, 32768, 65535);
610numbers ('i!', -2147483648, -1, 0, 1, 2147483647);
611numbers ('I!', 0, 1, 2147483647, 2147483648, 4294967295);
612numbers ('l!', -2147483648, -1, 0, 1, 2147483647);
613numbers ('L!', 0, 1, 2147483647, 2147483648, 4294967295);
614numbers ('n', 0, 1, 32767, 32768, 65535);
615numbers ('v', 0, 1, 32767, 32768, 65535);
616numbers ('N', 0, 1, 2147483647, 2147483648, 4294967295);
617numbers ('V', 0, 1, 2147483647, 2147483648, 4294967295);
618numbers ('n!', -32768, -1, 0, 1, 32767);
619numbers ('v!', -32768, -1, 0, 1, 32767);
620numbers ('N!', -2147483648, -1, 0, 1, 2147483647);
621numbers ('V!', -2147483648, -1, 0, 1, 2147483647);
622# All these should have exact binary representations:
623numbers ('f', -1, 0, 0.5, 42, 2**34);
624numbers ('d', -(2**34), -1, 0, 1, 2**34);
625## These don't, but 'd' is NV.  XXX wrong, it's double
626#numbers ('d', -1, 0, 1, 1-exp(-1), -exp(1));
627
628numbers_with_total ('q', -1,
629                    -9223372036854775808, -1, 0, 1,9223372036854775807);
630# This total is icky, but the true total is 2**65-1, and need a way to generate
631# the expected checksum on any system including those where NVs can preserve
632# 65 bits. (long double is 128 bits on sparc, so they certainly can)
633# or where rounding is down not up on binary conversion (crays)
634numbers_with_total ('Q', sub {
635                      my $len = shift;
636                      $len = 65 if $len > 65; # unmasked total is 2**65-1 here
637                      my $total = 1 + 2 * (int (2**($len - 1)) - 1);
638                      return 0 if $total == $total - 1; # Overflowed integers
639                      return $total; # NVs still accurate to nearest integer
640                    },
641                    0, 1,9223372036854775807, 9223372036854775808,
642                    18446744073709551615);
643
644print "# pack nvNV byteorders\n";
645
646is(pack("n", 0xdead), "\xde\xad");
647is(pack("v", 0xdead), "\xad\xde");
648is(pack("N", 0xdeadbeef), "\xde\xad\xbe\xef");
649is(pack("V", 0xdeadbeef), "\xef\xbe\xad\xde");
650
651is(pack("n!", 0xdead), "\xde\xad");
652is(pack("v!", 0xdead), "\xad\xde");
653is(pack("N!", 0xdeadbeef), "\xde\xad\xbe\xef");
654is(pack("V!", 0xdeadbeef), "\xef\xbe\xad\xde");
655
656print "# test big-/little-endian conversion\n";
657
658sub byteorder
659{
660  my $format = shift;
661  print "# byteorder test for $format\n";
662  for my $value (@_) {
663    SKIP: {
664      my($nat,$be,$le) = eval { map { pack $format.$_, $value } '', '>', '<' };
665      skip "cannot pack '$format' on this perl", 5
666        if is_valid_error($@);
667
668      {
669        use warnings qw(NONFATAL utf8);
670        print "# [$value][$nat][$be][$le][$@]\n";
671      }
672
673      SKIP: {
674        skip "cannot compare native byteorder with big-/little-endian", 1
675            if $ByteOrder eq 'unknown';
676
677        is($nat, $ByteOrder eq 'big' ? $be : $le);
678      }
679      is($be, reverse($le));
680      my @x = eval { unpack "$format$format>$format<", $nat.$be.$le };
681
682      print "# [$value][", join('][', @x), "][$@]\n";
683
684      is($@, '');
685      is($x[0], $x[1]);
686      is($x[0], $x[2]);
687    }
688  }
689}
690
691byteorder('s', -32768, -1, 0, 1, 32767);
692byteorder('S', 0, 1, 32767, 32768, 65535);
693byteorder('i', -2147483648, -1, 0, 1, 2147483647);
694byteorder('I', 0, 1, 2147483647, 2147483648, 4294967295);
695byteorder('l', -2147483648, -1, 0, 1, 2147483647);
696byteorder('L', 0, 1, 2147483647, 2147483648, 4294967295);
697byteorder('j', -2147483648, -1, 0, 1, 2147483647);
698byteorder('J', 0, 1, 2147483647, 2147483648, 4294967295);
699byteorder('s!', -32768, -1, 0, 1, 32767);
700byteorder('S!', 0, 1, 32767, 32768, 65535);
701byteorder('i!', -2147483648, -1, 0, 1, 2147483647);
702byteorder('I!', 0, 1, 2147483647, 2147483648, 4294967295);
703byteorder('l!', -2147483648, -1, 0, 1, 2147483647);
704byteorder('L!', 0, 1, 2147483647, 2147483648, 4294967295);
705byteorder('q', -9223372036854775808, -1, 0, 1, 9223372036854775807);
706byteorder('Q', 0, 1, 9223372036854775807, 9223372036854775808, 18446744073709551615);
707byteorder('f', -1, 0, 0.5, 42, 2**34);
708byteorder('F', -1, 0, 0.5, 42, 2**34);
709byteorder('d', -(2**34), -1, 0, 1, 2**34);
710byteorder('D', -(2**34), -1, 0, 1, 2**34);
711
712print "# test negative numbers\n";
713
714SKIP: {
715  skip "platform is not using two's complement for negative integers", 120
716    unless $IsTwosComplement;
717
718  for my $format (qw(s i l j s! i! l! q)) {
719    SKIP: {
720      my($nat,$be,$le) = eval { map { pack $format.$_, -1 } '', '>', '<' };
721      skip "cannot pack '$format' on this perl", 15
722        if is_valid_error($@);
723
724      my $len = length $nat;
725      is($_, "\xFF"x$len) for $nat, $be, $le;
726
727      my(@val,@ref);
728      if ($len >= 8) {
729        @val = (-2, -81985529216486896, -9223372036854775808);
730        @ref = ("\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFE",
731                "\xFE\xDC\xBA\x98\x76\x54\x32\x10",
732                "\x80\x00\x00\x00\x00\x00\x00\x00");
733      }
734      elsif ($len >= 4) {
735        @val = (-2, -19088744, -2147483648);
736        @ref = ("\xFF\xFF\xFF\xFE",
737                "\xFE\xDC\xBA\x98",
738                "\x80\x00\x00\x00");
739      }
740      else {
741        @val = (-2, -292, -32768);
742        @ref = ("\xFF\xFE",
743                "\xFE\xDC",
744                "\x80\x00");
745      }
746      for my $x (@ref) {
747        if ($len > length $x) {
748          $x = $x . "\xFF" x ($len - length $x);
749        }
750      }
751
752      for my $i (0 .. $#val) {
753        my($nat,$be,$le) = eval { map { pack $format.$_, $val[$i] } '', '>', '<' };
754        is($@, '');
755
756        SKIP: {
757          skip "cannot compare native byteorder with big-/little-endian", 1
758              if $ByteOrder eq 'unknown';
759
760          is($nat, $ByteOrder eq 'big' ? $be : $le);
761        }
762
763        is($be, $ref[$i]);
764        is($be, reverse($le));
765      }
766    }
767  }
768}
769
770{
771  # /
772
773  my ($x, $y, $z, @a);
774  eval { ($x) = unpack '/a*','hello' };
775  like($@, qr!'/' must follow a numeric type!);
776  undef $x;
777  eval { $x = unpack '/a*','hello' };
778  like($@, qr!'/' must follow a numeric type!);
779
780  # [perl #60204] Unhelpful error message from unpack
781  eval { @a = unpack 'v/a*','h' };
782  is($@, '');
783  is(scalar @a, 0);
784  eval { $x = unpack 'v/a*','h' };
785  is($@, '');
786  is($x, undef);
787
788  undef $x;
789  eval { ($z,$x,$y) = unpack 'a3/A C/a* C/Z', "003ok \003yes\004z\000abc" };
790  is($@, '');
791  is($z, 'ok');
792  is($x, 'yes');
793  is($y, 'z');
794  undef $z;
795  eval { $z = unpack 'a3/A C/a* C/Z', "003ok \003yes\004z\000abc" };
796  is($@, '');
797  is($z, 'ok');
798
799
800  undef $x;
801  eval { ($x) = pack '/a*','hello' };
802  like($@,  qr!Invalid type '/'!);
803  undef $x;
804  eval { $x = pack '/a*','hello' };
805  like($@,  qr!Invalid type '/'!);
806
807  $z = pack 'n/a* N/Z* w/A*','string','hi there ','etc';
808  my $expect = "\000\006string\0\0\0\012hi there \000\003etc";
809  is($z, $expect);
810
811  undef $x;
812  $expect = 'hello world';
813  eval { ($x) = unpack ("w/a", chr (11) . "hello world!")};
814  is($x, $expect);
815  is($@, '');
816
817  undef $x;
818  # Doing this in scalar context used to fail.
819  eval { $x = unpack ("w/a", chr (11) . "hello world!")};
820  is($@, '');
821  is($x, $expect);
822
823  foreach (
824           ['a/a*/a*', '212ab345678901234567','ab3456789012'],
825           ['a/a*/a*', '3012ab345678901234567', 'ab3456789012'],
826           ['a/a*/b*', '212ab', $::IS_EBCDIC ? '100000010100' : '100001100100'],
827  )
828  {
829    my ($pat, $in, $expect) = @$_;
830    undef $x;
831    eval { ($x) = unpack $pat, $in };
832    is($@, '');
833    is($x, $expect) ||
834      printf "# list unpack ('$pat', '$in') gave %s, expected '$expect'\n",
835             encode_list ($x);
836
837    undef $x;
838    eval { $x = unpack $pat, $in };
839    is($@, '');
840    is($x, $expect) ||
841      printf "# scalar unpack ('$pat', '$in') gave %s, expected '$expect'\n",
842             encode_list ($x);
843  }
844
845  # / with #
846
847  my $pattern = <<'EOU';
848 a3/A			# Count in ASCII
849 C/a*			# Count in a C char
850 C/Z			# Count in a C char but skip after \0
851EOU
852
853  $x = $y = $z =undef;
854  eval { ($z,$x,$y) = unpack $pattern, "003ok \003yes\004z\000abc" };
855  is($@, '');
856  is($z, 'ok');
857  is($x, 'yes');
858  is($y, 'z');
859  undef $x;
860  eval { $z = unpack $pattern, "003ok \003yes\004z\000abc" };
861  is($@, '');
862  is($z, 'ok');
863
864  $pattern = <<'EOP';
865  n/a*			# Count as network short
866  w/A*			# Count a  BER integer
867EOP
868  $expect = "\000\006string\003etc";
869  $z = pack $pattern,'string','etc';
870  is($z, $expect);
871}
872
873
874{
875    is("1.20.300.4000", sprintf "%vd", pack("U*",1,20,300,4000));
876    is("1.20.300.4000", sprintf "%vd", pack("  U*",1,20,300,4000));
877}
878isnt(v1.20.300.4000, sprintf "%vd", pack("C0U*",1,20,300,4000));
879
880my $rslt = join " ", map { ord } split "", byte_utf8a_to_utf8n("\xc7\xa2");
881# The ASCII UTF-8 of U+1E2 is "\xc7\xa2"
882is(join(" ", unpack("U0 C*", chr(0x1e2))), $rslt);
883
884# does pack U create Unicode?
885is(ord(pack('U', 300)), 300);
886
887# does unpack U deref Unicode?
888is((unpack('U', chr(300)))[0], 300);
889
890# is unpack U the reverse of pack U for Unicode string?
891is("@{[unpack('U*', pack('U*', 100, 200, 300))]}", "100 200 300");
892
893# is unpack U the reverse of pack U for byte string?
894is("@{[unpack('U*', pack('U*', 100, 200))]}", "100 200");
895
896{
897    # does pack U0C create Unicode?
898    my $cp202 = chr(202);
899    utf8::upgrade $cp202;
900    my @bytes202;
901    {   # This is portable across character sets
902        use bytes;
903        @bytes202 = map { ord } split "", $cp202;
904    }
905
906    # This test requires the first number to be invariant; 64 is invariant on
907    # ASCII and EBCDIC.
908    is("@{[pack('U0C*', 64, @bytes202)]}", v64.v202);
909
910    # does pack C0U create characters?
911    # The U* is expecting Unicode, so convert to that.
912    is("@{[pack('C0U*', map { $_ } 64, 202)]}",
913       pack("C*", 64, @bytes202));
914
915    # does unpack U0U on byte data fail?
916    fresh_perl_like('my $bad = pack("U0C", 202); my @null = unpack("U0U", $bad);',
917                    qr/^Malformed UTF-8 character: /,
918                    {},
919                    "pack doesn't return malformed UTF-8");
920}
921
922{
923  my $p = pack 'i*', -2147483648, ~0, 0, 1, 2147483647;
924  my (@a);
925  # bug - % had to be at the start of the pattern, no leading whitespace or
926  # comments. %i! didn't work at all.
927  foreach my $pat ('%32i*', ' %32i*', "# Muhahahaha\n%32i*", '%32i*  ',
928                   '%32i!*', ' %32i!*', "\n#\n#\n\r \t\f%32i!*", '%32i!*#') {
929    @a = unpack $pat, $p;
930    is($a[0], 0xFFFFFFFF) || print "# $pat\n";
931    @a = scalar unpack $pat, $p;
932    is($a[0], 0xFFFFFFFF) || print "# $pat\n";
933  }
934
935
936  $p = pack 'I*', 42, 12;
937  # Multiline patterns in scalar context failed.
938  foreach my $pat ('I', <<EOPOEMSNIPPET, 'I#I', 'I # I', 'I # !!!') {
939# On the Ning Nang Nong
940# Where the Cows go Bong!
941# And the Monkeys all say Boo!
942I
943EOPOEMSNIPPET
944    @a = unpack $pat, $p;
945    is(scalar @a, 1);
946    is($a[0], 42);
947    @a = scalar unpack $pat, $p;
948    is(scalar @a, 1);
949    is($a[0], 42);
950  }
951
952  # shorts (of all flavours) didn't calculate checksums > 32 bits with floating
953  # point, so a pathologically long pattern would wrap at 32 bits.
954  my $pat = "\xff\xff"x65538; # Start with it long, to save any copying.
955  foreach (4,3,2,1,0) {
956    my $len = 65534 + $_;
957    is(unpack ("%33n$len", $pat), 65535 * $len);
958  }
959}
960
961
962# pack x X @
963foreach (
964         ['x', "N", "\0"],
965         ['x4', "N", "\0"x4],
966         ['xX', "N", ""],
967         ['xXa*', "Nick", "Nick"],
968         ['a5Xa5', "cameL", "llama", "camellama"],
969         ['@4', 'N', "\0"x4],
970         ['a*@8a*', 'Camel', 'Dromedary', "Camel\0\0\0Dromedary"],
971         ['a*@4a', 'Perl rules', '!', 'Perl!'],
972)
973{
974  my ($template, @in) = @$_;
975  my $out = pop @in;
976  my $got = eval {pack $template, @in};
977  is($@, '');
978  is($out, $got) ||
979    printf "# pack ('$template', %s) gave %s expected %s\n",
980           encode_list (@in), encode_list ($got), encode_list ($out);
981}
982
983# unpack x X @
984foreach (
985         ['x', "N"],
986         ['xX', "N"],
987         ['xXa*', "Nick", "Nick"],
988         ['a5Xa5', "camellama", "camel", "llama"],
989         ['@3', "ice"],
990         ['@2a2', "water", "te"],
991         ['a*@1a3', "steam", "steam", "tea"],
992)
993{
994  my ($template, $in, @out) = @$_;
995  my @got = eval {unpack $template, $in};
996  is($@, '');
997  ok (list_eq (\@got, \@out)) ||
998    printf "# list unpack ('$template', %s) gave %s expected %s\n",
999           _qq($in), encode_list (@got), encode_list (@out);
1000
1001  my $got = eval {unpack $template, $in};
1002  is($@, '');
1003  @out ? is( $got, $out[0] ) # 1 or more items; should get first
1004       : ok( !defined $got ) # 0 items; should get undef
1005    or printf "# scalar unpack ('$template', %s) gave %s expected %s\n",
1006              _qq($in), encode_list ($got), encode_list ($out[0]);
1007}
1008
1009{
1010    my $t = 'Z*Z*';
1011    my ($u, $v) = qw(foo xyzzy);
1012    my $p = pack($t, $u, $v);
1013    my @u = unpack($t, $p);
1014    is(scalar @u, 2);
1015    is($u[0], $u);
1016    is($u[1], $v);
1017}
1018
1019{
1020    is((unpack("w/a*", "\x02abc"))[0], "ab");
1021
1022    # "w/a*" should be seen as one unit
1023
1024    is(scalar unpack("w/a*", "\x02abc"), "ab");
1025}
1026
1027print "# group modifiers\n";
1028
1029for my $t (qw{ (s<)< (sl>s)> (s(l(sl)<l)s)< }) {
1030  print "# testing pattern '$t'\n";
1031  eval { ($_) = unpack($t, 'x'x18); };
1032  is($@, '');
1033  eval { $_ = pack($t, (0)x6); };
1034  is($@, '');
1035}
1036
1037for my $t (qw{ (s<)> (sl>s)< (s(l(sl)<l)s)> }) {
1038  print "# testing pattern '$t'\n";
1039  eval { ($_) = unpack($t, 'x'x18); };
1040  like($@, qr/Can't use '[<>]' in a group with different byte-order in unpack/);
1041  eval { $_ = pack($t, (0)x6); };
1042  like($@, qr/Can't use '[<>]' in a group with different byte-order in pack/);
1043}
1044
1045is(pack('L<L>', (0x12345678)x2),
1046   pack('(((L1)1)<)(((L)1)1)>1', (0x12345678)x2));
1047
1048{
1049  sub compress_template {
1050    my $t = shift;
1051    for my $mod (qw( < > )) {
1052      $t =~ s/((?:(?:[SILQJFDP]!?$mod|[^SILQJFDP\W]!?)(?:\d+|\*|\[(?:[^]]+)\])?\/?){2,})/
1053              my $x = $1; $x =~ s!$mod!!g ? "($x)$mod" : $x /ieg;
1054    }
1055    return $t;
1056  }
1057
1058  my %templates = (
1059    's<'                  => [-42],
1060    's<c2x![S]S<'         => [-42, -11, 12, 4711],
1061    '(i<j<[s]l<)3'        => [-11, -22, -33, 1000000, 1100, 2201, 3302,
1062                              -1000000, 32767, -32768, 1, -123456789 ],
1063    '(I!<4(J<2L<)3)5'     => [1 .. 65],
1064    'q<Q<'                => [-50000000005, 60000000006],
1065    'f<F<d<'              => [3.14159, 111.11, 2222.22],
1066    'D<cCD<'              => [1e42, -128, 255, 1e-42],
1067    'n/a*'                => ['/usr/bin/perl'],
1068    'C/a*S</A*L</Z*I</a*' => [qw(Just another Perl hacker)],
1069  );
1070
1071  for my $tle (sort keys %templates) {
1072    my @d = @{$templates{$tle}};
1073    my $tbe = $tle;
1074    $tbe =~ y/</>/;
1075    for my $t ($tbe, $tle) {
1076      my $c = compress_template($t);
1077      print "# '$t' -> '$c'\n";
1078      SKIP: {
1079        my $p1 = eval { pack $t, @d };
1080        skip "cannot pack '$t' on this perl", 5 if is_valid_error($@);
1081        my $p2 = eval { pack $c, @d };
1082        is($@, '');
1083        is($p1, $p2);
1084        s!(/[aAZ])\*!$1!g for $t, $c;
1085        my @u1 = eval { unpack $t, $p1 };
1086        is($@, '');
1087        my @u2 = eval { unpack $c, $p2 };
1088        is($@, '');
1089        is(join('!', @u1), join('!', @u2));
1090      }
1091    }
1092  }
1093}
1094
1095{
1096    # from Wolfgang Laun: fix in change #13163
1097
1098    my $s = 'ABC' x 10;
1099    my $t = '*';
1100    my $x = ord($t);
1101    my $buf = pack( 'Z*/A* C',  $s, $x );
1102    my $y;
1103
1104    my $h = $buf;
1105    $h =~ s/[^[:print:]]/./g;
1106    ( $s, $y ) = unpack( "Z*/A* C", $buf );
1107    is($h, "30.ABCABCABCABCABCABCABCABCABCABC$t");
1108    is(length $buf, 34);
1109    is($s, "ABCABCABCABCABCABCABCABCABCABC");
1110    is($y, $x);
1111}
1112
1113{
1114    # from Wolfgang Laun: fix in change #13288
1115
1116    eval { my $t=unpack("P*", "abc") };
1117    like($@, qr/'P' must have an explicit size/);
1118}
1119
1120{   # Grouping constructs
1121    my (@a, @b);
1122    @a = unpack '(SL)',   pack 'SLSLSL', 67..90;
1123    is("@a", "67 68");
1124    @a = unpack '(SL)3',   pack 'SLSLSL', 67..90;
1125    @b = (67..72);
1126    is("@a", "@b");
1127    @a = unpack '(SL)3',   pack 'SLSLSLSL', 67..90;
1128    is("@a", "@b");
1129    @a = unpack '(SL)[3]', pack 'SLSLSLSL', 67..90;
1130    is("@a", "@b");
1131    @a = unpack '(SL)[2] SL', pack 'SLSLSLSL', 67..90;
1132    is("@a", "@b");
1133    @a = unpack 'A/(SL)',  pack 'ASLSLSLSL', 3, 67..90;
1134    is("@a", "@b");
1135    @a = unpack 'A/(SL)SL',  pack 'ASLSLSLSL', 2, 67..90;
1136    is("@a", "@b");
1137    @a = unpack '(SL)*',   pack 'SLSLSLSL', 67..90;
1138    @b = (67..74);
1139    is("@a", "@b");
1140    @a = unpack '(SL)*SL',   pack 'SLSLSLSL', 67..90;
1141    is("@a", "@b");
1142    eval { @a = unpack '(*SL)',   '' };
1143    like($@, qr/\(\)-group starts with a count/);
1144    eval { @a = unpack '(3SL)',   '' };
1145    like($@, qr/\(\)-group starts with a count/);
1146    eval { @a = unpack '([3]SL)',   '' };
1147    like($@, qr/\(\)-group starts with a count/);
1148    eval { @a = pack '(*SL)' };
1149    like($@, qr/\(\)-group starts with a count/);
1150    @a = unpack '(SL)3 SL',   pack '(SL)4', 67..74;
1151    is("@a", "@b");
1152    @a = unpack '(SL)3 SL',   pack '(SL)[4]', 67..74;
1153    is("@a", "@b");
1154    @a = unpack '(SL)3 SL',   pack '(SL)*', 67..74;
1155    is("@a", "@b");
1156}
1157
1158{  # more on grouping (W.Laun)
1159  # @ absolute within ()-group
1160  my $badc = pack( '(a)*', unpack( '(@1a @0a @2)*', 'abcd' ) );
1161  is( $badc, 'badc' );
1162  my @b = ( 1, 2, 3 );
1163  my $buf = pack( '(@1c)((@2C)@3c)', @b );
1164  is( $buf, "\0\1\0\0\2\3" );
1165  my @a = unpack( '(@1c)((@2c)@3c)', $buf );
1166  is( "@a", "@b" );
1167
1168  # various unpack count/code scenarios
1169  my @Env = ( a => 'AAA', b => 'BBB' );
1170  my $env = pack( 'S(S/A*S/A*)*', @Env/2, @Env );
1171
1172  # unpack full length - ok
1173  my @pup = unpack( 'S/(S/A* S/A*)', $env );
1174  is( "@pup", "@Env" );
1175
1176  # warn when count/code goes beyond end of string
1177  # \0002 \0001 a \0003 AAA \0001 b \0003 BBB
1178  #     2     4 5     7  10    1213
1179  eval { @pup = unpack( 'S/(S/A* S/A*)', substr( $env, 0, 13 ) ) };
1180  like( $@, qr{length/code after end of string} );
1181
1182  # postfix repeat count
1183  $env = pack( '(S/A* S/A*)' . @Env/2, @Env );
1184
1185  # warn when count/code goes beyond end of string
1186  # \0001 a \0003 AAA \0001  b \0003 BBB
1187  #     2 3c    5   8    10 11    13  16
1188  eval { @pup = unpack( '(S/A* S/A*)' . @Env/2, substr( $env, 0, 11 ) ) };
1189  like( $@, qr{length/code after end of string} );
1190
1191  # catch stack overflow/segfault
1192  eval { $_ = pack( ('(' x 105) . 'A' . (')' x 105) ); };
1193  like( $@, qr{Too deeply nested \(\)-groups} );
1194}
1195
1196{ # syntax checks (W.Laun)
1197  use warnings qw(NONFATAL all);;
1198  my @warning;
1199  local $SIG{__WARN__} = sub {
1200      push( @warning, $_[0] );
1201  };
1202  eval { my $s = pack( 'Ax![4c]A', 1..5 ); };
1203  like( $@, qr{Malformed integer in \[\]} );
1204
1205  eval { my $buf = pack( '(c/*a*)', 'AAA', 'BB' ); };
1206  like( $@, qr{'/' does not take a repeat count} );
1207
1208  eval { my @inf = unpack( 'c/1a', "\x03AAA\x02BB" ); };
1209  like( $@, qr{'/' does not take a repeat count} );
1210
1211  eval { my @inf = unpack( 'c/*a', "\x03AAA\x02BB" ); };
1212  like( $@, qr{'/' does not take a repeat count} );
1213
1214  # white space where possible
1215  my @Env = ( a => 'AAA', b => 'BBB' );
1216  my $env = pack( ' S ( S / A*   S / A* )* ', @Env/2, @Env );
1217  my @pup = unpack( ' S / ( S / A*   S / A* ) ', $env );
1218  is( "@pup", "@Env" );
1219
1220  # white space in 4 wrong places
1221  for my $temp (  'A ![4]', 'A [4]', 'A *', 'A 4' ){
1222      eval { my $s = pack( $temp, 'B' ); };
1223      like( $@, qr{Invalid type } );
1224  }
1225
1226  # warning for commas
1227  @warning = ();
1228  my $x = pack( 'I,A', 4, 'X' );
1229  like( $warning[0], qr{Invalid type ','} );
1230  is($x, pack( 'IA', 4, 'X' ), "Comma was ignored in pack string");
1231
1232  # comma warning only once
1233  @warning = ();
1234  $x = pack( 'C(C,C)C,C', 65..71  );
1235  cmp_ok( scalar(@warning), '==', 1 );
1236  is(join(",", unpack 'C(C,,,C),C,,C', $x), join(",", 65..69),
1237     "Comma was ignored in unpack string");
1238
1239  # forbidden code in []
1240  eval { my $x = pack( 'A[@4]', 'XXXX' ); };
1241  like( $@, qr{Within \[\]-length '\@' not allowed} );
1242
1243  # @ repeat default 1
1244  my $s = pack( 'AA@A', 'A', 'B', 'C' );
1245  my @c = unpack( 'AA@A', $s );
1246  is( $s, 'AC' );
1247  is( "@c", "A C C" );
1248
1249  # no unpack code after /
1250  eval { my @a = unpack( "C/", "\3" ); };
1251  like( $@, qr{Code missing after '/'} );
1252
1253  # modifier warnings
1254  @warning = ();
1255  $x = pack "I>>s!!", 47, 11;
1256  ($x) = unpack "I<<l!>!>", 'x'x20;
1257  is(scalar @warning, 5);
1258  like($warning[0], qr/Duplicate modifier '>' after 'I' in pack/);
1259  like($warning[1], qr/Duplicate modifier '!' after 's' in pack/);
1260  like($warning[2], qr/Duplicate modifier '<' after 'I' in unpack/);
1261  like($warning[3], qr/Duplicate modifier '!' after 'l' in unpack/);
1262  like($warning[4], qr/Duplicate modifier '>' after 'l' in unpack/);
1263}
1264
1265{  # Repeat count [SUBEXPR]
1266   my @codes = qw( x A Z a c C W B b H h s v n S i I l V N L p P f F d
1267		   s! S! i! I! l! L! j J);
1268   my $G;
1269   if (eval { pack 'q', 1 } ) {
1270     push @codes, qw(q Q);
1271   } else {
1272     push @codes, qw(s S);	# Keep the count the same
1273   }
1274   if (eval { pack 'D', 1 } ) {
1275     push @codes, 'D';
1276   } else {
1277     push @codes, 'd';	# Keep the count the same
1278   }
1279
1280   push @codes, map { /^[silqjfdp]/i ? ("$_<", "$_>") : () } @codes;
1281
1282   my %val;
1283   @val{@codes} = map { / [Xx]  (?{ undef })
1284			| [AZa] (?{ 'something' })
1285			| C     (?{ 214 })
1286			| W     (?{ 8188 })
1287			| c     (?{ 114 })
1288			| [Bb]  (?{ '101' })
1289			| [Hh]  (?{ 'b8' })
1290			| [svnSiIlVNLqQjJ]  (?{ 10111 })
1291			| [FfDd]  (?{ 1.36514538e37 })
1292			| [pP]  (?{ "try this buffer" })
1293			/x; $^R } @codes;
1294   my @end = (0x12345678, 0x23456781, 0x35465768, 0x15263748);
1295   my $end = "N4";
1296
1297   for my $type (@codes) {
1298     my @list = $val{$type};
1299     @list = () unless defined $list[0];
1300     for my $count ('', '3', '[11]') {
1301       my $c = 1;
1302       $c = $1 if $count =~ /(\d+)/;
1303       my @list1 = @list;
1304       @list1 = (@list1) x $c unless $type =~ /[XxAaZBbHhP]/;
1305       for my $groupend ('', ')2', ')[8]') {
1306	   my $groupbegin = ($groupend ? '(' : '');
1307	   $c = 1;
1308	   $c = $1 if $groupend =~ /(\d+)/;
1309	   my @list2 = (@list1) x $c;
1310
1311           SKIP: {
1312	     my $junk1 = "$groupbegin $type$count $groupend";
1313	     # print "# junk1=$junk1\n";
1314	     my $p = eval { pack $junk1, @list2 };
1315             skip "cannot pack '$type' on this perl", 12
1316               if is_valid_error($@);
1317	     die "pack $junk1 failed: $@" if $@;
1318
1319	     my $half = int( (length $p)/2 );
1320	     for my $move ('', "X$half", "X!$half", 'x1', 'x!8', "x$half") {
1321	       my $junk = "$junk1 $move";
1322	       # print "# junk='$junk', list=(@list2)\n";
1323	       $p = pack "$junk $end", @list2, @end;
1324	       my @l = unpack "x[$junk] $end", $p;
1325	       is(scalar @l, scalar @end);
1326	       is("@l", "@end", "skipping x[$junk]");
1327	     }
1328           }
1329       }
1330     }
1331   }
1332}
1333
1334# / is recognized after spaces in scalar context
1335# XXXX no spaces are allowed in pack...  In pack only before the slash...
1336is(scalar unpack('A /A Z20', pack 'A/A* Z20', 'bcde', 'xxxxx'), 'bcde');
1337is(scalar unpack('A /A /A Z20', '3004bcde'), 'bcde');
1338
1339{ # X! and x!
1340  my $t = 'C[3]  x!8 C[2]';
1341  my @a = (0x73..0x77);
1342  my $p = pack($t, @a);
1343  is($p, "\x73\x74\x75\0\0\0\0\0\x76\x77");
1344  my @b = unpack $t, $p;
1345  is(scalar @b, scalar @a);
1346  is("@b", "@a", 'x!8');
1347  $t = 'x[5] C[6] X!8 C[2]';
1348  @a = (0x73..0x7a);
1349  $p = pack($t, @a);
1350  is($p, "\0\0\0\0\0\x73\x74\x75\x79\x7a");
1351  @b = unpack $t, $p;
1352  @a = (0x73..0x75, 0x79, 0x7a, 0x79, 0x7a);
1353  is(scalar @b, scalar @a);
1354  is("@b", "@a");
1355}
1356
1357{ # struct {char c1; double d; char cc[2];}
1358  my $t = 'C x![d] d C[2]';
1359  my @a = (173, 1.283476517e-45, 42, 215);
1360  my $p = pack $t, @a;
1361  ok( length $p);
1362  my @b = unpack "$t X[$t] $t", $p;	# Extract, step back, extract again
1363  is(scalar @b, 2 * scalar @a);
1364  $b = "@b";
1365  $b =~ s/(?:17000+|16999+)\d+(e-0?45) /17$1 /gi; # stringification is gamble
1366  is($b, "@a @a");
1367
1368  use warnings qw(NONFATAL all);;
1369  my $warning;
1370  local $SIG{__WARN__} = sub {
1371      $warning = $_[0];
1372  };
1373  @b = unpack "x[C] x[$t] X[$t] X[C] $t", "$p\0";
1374
1375  is($warning, undef);
1376  is(scalar @b, scalar @a);
1377  $b = "@b";
1378  $b =~ s/(?:17000+|16999+)\d+(e-0?45) /17$1 /gi; # stringification is gamble
1379  is($b, "@a");
1380}
1381
1382is(length(pack("j", 0)), $Config{ivsize});
1383is(length(pack("J", 0)), $Config{uvsize});
1384is(length(pack("F", 0)), $Config{nvsize});
1385
1386numbers ('j', -2147483648, -1, 0, 1, 2147483647);
1387numbers ('J', 0, 1, 2147483647, 2147483648, 4294967295);
1388numbers ('F', -(2**34), -1, 0, 1, 2**34);
1389SKIP: {
1390    my $t = eval { unpack("D*", pack("D", 12.34)) };
1391
1392    skip "Long doubles not in use", 166 if $@ =~ /Invalid type/;
1393
1394    is(length(pack("D", 0)), $Config{longdblsize});
1395    numbers ('D', -(2**34), -1, 0, 1, 2**34);
1396}
1397
1398# Maybe this knowledge needs to be "global" for all of pack.t
1399# Or a "can checksum" which would effectively be all the number types"
1400my %cant_checksum = map {$_=> 1} qw(A Z u w);
1401# not a b B h H
1402foreach my $template (qw(A Z c C s S i I l L n N v V q Q j J f d F D u U w)) {
1403  SKIP: {
1404    my $packed = eval {pack "${template}4", 1, 4, 9, 16};
1405    if ($@) {
1406      die unless $@ =~ /Invalid type '$template'/;
1407      skip ("$template not supported on this perl",
1408            $cant_checksum{$template} ? 4 : 8);
1409    }
1410    my @unpack4 = unpack "${template}4", $packed;
1411    my @unpack = unpack "${template}*", $packed;
1412    my @unpack1 = unpack "${template}", $packed;
1413    my @unpack1s = scalar unpack "${template}", $packed;
1414    my @unpack4s = scalar unpack "${template}4", $packed;
1415    my @unpacks = scalar unpack "${template}*", $packed;
1416
1417    my @tests = ( ["${template}4 vs ${template}*", \@unpack4, \@unpack],
1418                  ["scalar ${template} ${template}", \@unpack1s, \@unpack1],
1419                  ["scalar ${template}4 vs ${template}", \@unpack4s, \@unpack1],
1420                  ["scalar ${template}* vs ${template}", \@unpacks, \@unpack1],
1421                );
1422
1423    unless ($cant_checksum{$template}) {
1424      my @unpack4_c = unpack "\%${template}4", $packed;
1425      my @unpack_c = unpack "\%${template}*", $packed;
1426      my @unpack1_c = unpack "\%${template}", $packed;
1427      my @unpack1s_c = scalar unpack "\%${template}", $packed;
1428      my @unpack4s_c = scalar unpack "\%${template}4", $packed;
1429      my @unpacks_c = scalar unpack "\%${template}*", $packed;
1430
1431      push @tests,
1432        ( ["% ${template}4 vs ${template}*", \@unpack4_c, \@unpack_c],
1433          ["% scalar ${template} ${template}", \@unpack1s_c, \@unpack1_c],
1434          ["% scalar ${template}4 vs ${template}*", \@unpack4s_c, \@unpack_c],
1435          ["% scalar ${template}* vs ${template}*", \@unpacks_c, \@unpack_c],
1436        );
1437    }
1438    foreach my $test (@tests) {
1439      ok (list_eq ($test->[1], $test->[2]), $test->[0]) ||
1440        printf "# unpack gave %s expected %s\n",
1441          encode_list (@{$test->[1]}), encode_list (@{$test->[2]});
1442    }
1443  }
1444}
1445
1446ok(pack('u2', 'AA'), "[perl #8026]"); # used to hang and eat RAM in perl 5.7.2
1447
1448$_ = pack('c', 65); # 'A' would not be EBCDIC-friendly
1449is(unpack('c'), 65, "one-arg unpack (change #18751)"); # defaulting to $_
1450
1451{
1452    my $a = "X\x0901234567\n" x 100; # \t would not be EBCDIC TAB
1453    my @a = unpack("(a1 c/a)*", $a);
1454    is(scalar @a, 200,       "[perl #15288]");
1455    is($a[-1], "01234567\n", "[perl #15288]");
1456    is($a[-2], "X",          "[perl #15288]");
1457}
1458
1459{
1460    use warnings qw(NONFATAL all);;
1461    my $warning;
1462    local $SIG{__WARN__} = sub {
1463        $warning = $_[0];
1464    };
1465
1466    # This test is looking for the encoding of the bit pattern "\x66\x6f\x6f",
1467    # which is ASCII "foo"
1468    my $out = pack("u99", native_to_uni("foo") x 99);
1469    like($warning, qr/Field too wide in 'u' format in pack at /,
1470         "Warn about too wide uuencode");
1471    is($out, ("_" . "9F]O" x 21 . "\n") x 4 . "M" . "9F]O" x 15 . "\n",
1472       "Use max width in case of too wide uuencode");
1473}
1474
1475# checksums
1476{
1477    # verify that unpack advances correctly wrt a checksum
1478    my (@x) = unpack("b10a", "abcd");
1479    my (@y) = unpack("%b10a", "abcd");
1480    is($x[1], $y[1], "checksum advance ok");
1481
1482    SKIP: {
1483        skip("-- non-IEEE float", 1) if !$Config{d_double_style_ieee};
1484        # verify that the checksum is not overflowed with C0
1485        is(unpack("C0%128U", "abcd"), unpack("U0%128U", "abcd"), "checksum not overflowed");
1486    }
1487}
1488
1489my $U_1FFC_bytes = byte_utf8a_to_utf8n("\341\277\274");
1490{
1491    # U0 and C0 must be scoped
1492    my (@x) = unpack("a(U0)U", "b$U_1FFC_bytes");
1493    is($x[0], 'b', 'before scope');
1494    is($x[1], 8188, 'after scope');
1495
1496    is(pack("a(U0)U", "b", 8188), "b$U_1FFC_bytes");
1497}
1498
1499{
1500    # counted length prefixes shouldn't change C0/U0 mode
1501    # (note the length is actually 0 in this test, as the C/ is replaced by C0
1502    # due to the \0 in the string)
1503    is(join(',', unpack("aC/UU",   "b\0$U_1FFC_bytes")), 'b,8188');
1504    is(join(',', unpack("aC/CU",   "b\0$U_1FFC_bytes")), 'b,8188');
1505
1506    # The U expects Unicode, so convert from native
1507    my $first_byte = ord substr($U_1FFC_bytes, 0, 1);
1508
1509    is(join(',', unpack("aU0C/UU", "b\0$U_1FFC_bytes")), "b,$first_byte");
1510    is(join(',', unpack("aU0C/CU", "b\0$U_1FFC_bytes")), "b,$first_byte");
1511}
1512
1513{
1514    # "Z0" (bug #34062)
1515    my (@x) = unpack("C*", pack("CZ0", 1, "b"));
1516    is(join(',', @x), '1', 'pack Z0 doesn\'t destroy the character before');
1517}
1518
1519{
1520    # Encoding neutrality
1521    # String we will pull apart and rebuild in several ways:
1522    my $down = "\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff\x05\x06";
1523    my $up   = $down;
1524    utf8::upgrade($up);
1525
1526    my %expect =
1527        # [expected result,
1528        #  how many chars it should progress,
1529        #  (optional) expected result of pack]
1530        (a5 => ["\xf8\xf9\xfa\xfb\xfc", 5],
1531         A5 => ["\xf8\xf9\xfa\xfb\xfc", 5],
1532         Z5 => ["\xf8\xf9\xfa\xfb\xfc", 5, "\xf8\xf9\xfa\xfb\x00\xfd"],
1533         b21 => ["000111111001111101011", 3, "\xf8\xf9\x1a\xfb"],
1534         B21 => ["111110001111100111111", 3, "\xf8\xf9\xf8\xfb"],
1535         H5 => ["f8f9f", 3, "\xf8\xf9\xf0\xfb"],
1536         h5 => ["8f9fa", 3, "\xf8\xf9\x0a\xfb"],
1537         "s<"  => [-1544, 2],
1538         "s>"  => [-1799, 2],
1539         "S<"  => [0xf9f8, 2],
1540         "S>"  => [0xf8f9, 2],
1541         "l<"  => [-67438088, 4],
1542         "l>"  => [-117835013, 4],
1543         "L>"  => [0xf8f9fafb, 4],
1544         "L<"  => [0xfbfaf9f8, 4],
1545         n     => [0xf8f9, 2],
1546         N     => [0xf8f9fafb, 4],
1547         v     => [63992, 2],
1548         V     => [0xfbfaf9f8, 4],
1549         c     => [-8, 1],
1550         U0U   => [0xf8, 1],
1551         w     => ["8715569050387726213", 9],
1552         q     => ["-283686952306184", 8],
1553         Q     => ["18446460386757245432", 8],
1554         );
1555
1556    for my $string ($down, $up) {
1557        for my $format (sort {lc($a) cmp lc($b) || $a cmp $b } keys %expect) {
1558          SKIP: {
1559              my $expect = $expect{$format};
1560              # unpack upgraded and downgraded string
1561              my @result = eval { unpack("$format C0 W", $string) };
1562              skip "cannot pack/unpack '$format C0 W' on this perl", 5 if
1563                  $@ && is_valid_error($@);
1564              is(@result, 2, "Two results from unpack $format C0 W");
1565
1566              # pack to downgraded
1567              my $new = pack("$format C0 W", @result);
1568              is(length($new), $expect->[1]+1,
1569                 "pack $format C0 W should give $expect->[1]+1 chars");
1570              is($new, $expect->[2] || substr($string, 0, length $new),
1571                 "pack $format C0 W returns expected value");
1572
1573              # pack to upgraded
1574              $new = pack("a0 $format C0 W", chr(256), @result);
1575              is(length($new), $expect->[1]+1,
1576                 "pack a0 $format C0 W should give $expect->[1]+1 chars");
1577              is($new, $expect->[2] || substr($string, 0, length $new),
1578                 "pack a0 $format C0 W returns expected value");
1579          }
1580        }
1581    }
1582}
1583
1584{
1585    # Encoding neutrality, numbers
1586    my $val = -2.68;
1587    for my $format (qw(s S i I l L j J f d F D q Q
1588                       s! S! i! I! l! L! n! N! v! V!)) {
1589      SKIP: {
1590          my $down = eval { pack($format, $val) };
1591          skip "cannot pack/unpack $format on this perl", 9 if
1592              $@ && is_valid_error($@);
1593          ok(!utf8::is_utf8($down), "Simple $format pack doesn't get upgraded");
1594          my $up = pack("a0 $format", chr(256), $val);
1595          ok(utf8::is_utf8($up), "a0 $format with high char leads to upgrade");
1596          is($down, $up, "$format generated strings are equal though");
1597          my @down_expanded = unpack("$format W", $down . chr(0xce));
1598          is(@down_expanded, 2, "Expand to two values");
1599          is($down_expanded[1], 0xce,
1600             "unpack $format left us at the expected position");
1601          my @up_expanded   = unpack("$format W", $up   . chr(0xce));
1602          is(@up_expanded, 2, "Expand to two values");
1603          is($up_expanded[1], 0xce,
1604             "unpack $format left us at the expected position");
1605          is($down_expanded[0], $up_expanded[0], "$format unpack was neutral");
1606          is(pack($format, $down_expanded[0]), $down, "Pack $format undoes unpack $format");
1607      }
1608    }
1609}
1610
1611{
1612    # C *is* neutral
1613    my $down = "\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff\x05\x06";
1614    my $up   = $down;
1615    utf8::upgrade($up);
1616    my @down = unpack("C*", $down);
1617    my @expect_down = (0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xff, 0x05, 0x06);
1618    is("@down", "@expect_down", "byte expand");
1619    is(pack("C*", @down), $down, "byte join");
1620
1621    my @up   = unpack("C*", $up);
1622    my @expect_up = (0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xff, 0x05, 0x06);
1623    is("@up", "@expect_up", "UTF-8 expand");
1624    is(pack("U0C0C*", @up), $up, "UTF-8 join");
1625}
1626
1627{
1628    # Harder cases for the neutrality test
1629
1630    # u format
1631    my $down = "\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff\x05\x06";
1632    my $up   = $down;
1633    utf8::upgrade($up);
1634    is(pack("u", $down), pack("u", $up), "u pack is neutral");
1635    is(unpack("u", pack("u", $down)), $down, "u unpack to downgraded works");
1636    is(unpack("U0C0u", pack("u", $down)), $up, "u unpack to upgraded works");
1637
1638    # p/P format
1639    # This actually only tests something if the address contains a byte >= 0x80
1640    my $str = "abc\xa5\x00\xfede";
1641    $down = pack("p", $str);
1642    is(pack("P", $str), $down);
1643    is(pack("U0C0p", $str), $down);
1644    is(pack("U0C0P", $str), $down);
1645    is(unpack("p", $down), "abc\xa5", "unpack p downgraded");
1646    $up   = $down;
1647    utf8::upgrade($up);
1648    is(unpack("p", $up), "abc\xa5", "unpack p upgraded");
1649
1650    is(unpack("P7", $down), "abc\xa5\x00\xfed", "unpack P downgraded");
1651    is(unpack("P7", $up),   "abc\xa5\x00\xfed", "unpack P upgraded");
1652
1653    # x, X and @
1654    $down = "\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff\x05\x06";
1655    $up   = $down;
1656    utf8::upgrade($up);
1657
1658    is(unpack('@4W', $down), 0xfc, "\@positioning on downgraded string");
1659    is(unpack('@4W', $up),   0xfc, "\@positioning on upgraded string");
1660
1661    is(unpack('@4x2W', $down), 0xfe, "x moving on downgraded string");
1662    is(unpack('@4x2W', $up),   0xfe, "x moving on upgraded string");
1663    is(unpack('@4x!4W', $down), 0xfc, "x! moving on downgraded string");
1664    is(unpack('@4x!4W', $up),   0xfc, "x! moving on upgraded string");
1665    is(unpack('@5x!4W', $down), 0x05, "x! moving on downgraded string");
1666    is(unpack('@5x!4W', $up),   0x05, "x! moving on upgraded string");
1667
1668    is(unpack('@4X2W', $down), 0xfa, "X moving on downgraded string");
1669    is(unpack('@4X2W', $up),   0xfa, "X moving on upgraded string");
1670    is(unpack('@4X!4W', $down), 0xfc, "X! moving on downgraded string");
1671    is(unpack('@4X!4W', $up),   0xfc, "X! moving on upgraded string");
1672    is(unpack('@5X!4W', $down), 0xfc, "X! moving on downgraded string");
1673    is(unpack('@5X!4W', $up),   0xfc, "X! moving on upgraded string");
1674    is(unpack('@5X!8W', $down), 0xf8, "X! moving on downgraded string");
1675    is(unpack('@5X!8W', $up),   0xf8, "X! moving on upgraded string");
1676
1677    is(pack("W2x", 0xfa, 0xe3), "\xfa\xe3\x00", "x on downgraded string");
1678    is(pack("W2x!4", 0xfa, 0xe3), "\xfa\xe3\x00\x00",
1679       "x! on downgraded string");
1680    is(pack("W2x!2", 0xfa, 0xe3), "\xfa\xe3", "x! on downgraded string");
1681    is(pack("U0C0W2x", 0xfa, 0xe3), "\xfa\xe3\x00", "x on upgraded string");
1682    is(pack("U0C0W2x!4", 0xfa, 0xe3), "\xfa\xe3\x00\x00",
1683       "x! on upgraded string");
1684    is(pack("U0C0W2x!2", 0xfa, 0xe3), "\xfa\xe3", "x! on upgraded string");
1685    is(pack("W2X", 0xfa, 0xe3), "\xfa", "X on downgraded string");
1686    is(pack("U0C0W2X", 0xfa, 0xe3), "\xfa", "X on upgraded string");
1687    is(pack("W2X!2", 0xfa, 0xe3), "\xfa\xe3", "X! on downgraded string");
1688    is(pack("U0C0W2X!2", 0xfa, 0xe3), "\xfa\xe3", "X! on upgraded string");
1689    is(pack("W3X!2", 0xfa, 0xe3, 0xa6), "\xfa\xe3", "X! on downgraded string");
1690    is(pack("U0C0W3X!2", 0xfa, 0xe3, 0xa6), "\xfa\xe3",
1691       "X! on upgraded string");
1692
1693    # backward eating through a ( moves the group starting point backwards
1694    is(pack("a*(Xa)", "abc", "q"), "abq",
1695       "eating before strbeg moves it back");
1696    is(pack("a*(Xa)", "ab" . chr(512), "q"), "abq",
1697       "eating before strbeg moves it back");
1698
1699    # Check marked_upgrade
1700    is(pack('W(W(Wa@3W)@6W)@9W', 0xa1, 0xa2, 0xa3, "a", 0xa4, 0xa5, 0xa6),
1701       "\xa1\xa2\xa3a\x00\xa4\x00\xa5\x00\xa6");
1702    $up = "a";
1703    utf8::upgrade($up);
1704    is(pack('W(W(Wa@3W)@6W)@9W', 0xa1, 0xa2, 0xa3, $up, 0xa4, 0xa5, 0xa6),
1705       "\xa1\xa2\xa3a\x00\xa4\x00\xa5\x00\xa6", "marked upgrade caused by a");
1706    is(pack('W(W(WW@3W)@6W)@9W', 0xa1, 0xa2, 0xa3, 256, 0xa4, 0xa5, 0xa6),
1707       "\xa1\xa2\xa3\x{100}\x00\xa4\x00\xa5\x00\xa6",
1708       "marked upgrade caused by W");
1709    is(pack('W(W(WU0aC0@3W)@6W)@9W', 0xa1, 0xa2, 0xa3, "a", 0xa4, 0xa5, 0xa6),
1710       "\xa1\xa2\xa3a\x00\xa4\x00\xa5\x00\xa6", "marked upgrade caused by U0");
1711
1712    # a, A and Z
1713    $down = "\xa4\xa6\xa7";
1714    $up   = $down;
1715    utf8::upgrade($up);
1716    utf8::upgrade(my $high = "\xfeb");
1717
1718    for my $format ("a0", "A0", "Z0", "U0a0C0", "U0A0C0", "U0Z0C0") {
1719        is(pack("a* $format a*", "ab", $down, "cd"), "abcd",
1720           "$format format on plain string");
1721        is(pack("a* $format a*", "ab", $up,   "cd"), "abcd",
1722           "$format format on upgraded string");
1723        is(pack("a* $format a*", $high, $down, "cd"), "\xfebcd",
1724           "$format format on plain string");
1725        is(pack("a* $format a*", $high, $up,   "cd"), "\xfebcd",
1726           "$format format on upgraded string");
1727        my @down = unpack("a1 $format a*", "\xfeb");
1728        is("@down", "\xfe  b", "unpack $format");
1729        my @up = unpack("a1 $format a*", $high);
1730        is("@up", "\xfe  b", "unpack $format");
1731    }
1732    is(pack("a1", $high), "\xfe");
1733    is(pack("A1", $high), "\xfe");
1734    is(pack("Z1", $high), "\x00");
1735    is(pack("a2", $high), "\xfeb");
1736    is(pack("A2", $high), "\xfeb");
1737    is(pack("Z2", $high), "\xfe\x00");
1738    is(pack("a5", $high), "\xfeb\x00\x00\x00");
1739    is(pack("A5", $high), "\xfeb   ");
1740    is(pack("Z5", $high), "\xfeb\x00\x00\x00");
1741    is(pack("a*", $high), "\xfeb");
1742    is(pack("A*", $high), "\xfeb");
1743    is(pack("Z*", $high), "\xfeb\x00");
1744
1745    utf8::upgrade($high = byte_utf8a_to_utf8n("\xc3\xbe") . "b");
1746    is(pack("U0a2", $high), uni_to_native("\xfe"));
1747    is(pack("U0A2", $high), uni_to_native("\xfe"));
1748    is(pack("U0Z1", $high), uni_to_native("\x00"));
1749    is(pack("U0a3", $high), uni_to_native("\xfe") . "b");
1750    is(pack("U0A3", $high), uni_to_native("\xfe") . "b");
1751    is(pack("U0Z3", $high), uni_to_native("\xfe\x00"));
1752    is(pack("U0a6", $high), uni_to_native("\xfe") . "b" . uni_to_native("\x00\x00\x00"));
1753    is(pack("U0A6", $high), uni_to_native("\xfe") . "b   ");
1754    is(pack("U0Z6", $high), uni_to_native("\xfe") . "b" . uni_to_native("\x00\x00\x00"));
1755    is(pack("U0a*", $high), uni_to_native("\xfe") . "b");
1756    is(pack("U0A*", $high), uni_to_native("\xfe") . "b");
1757    is(pack("U0Z*", $high), uni_to_native("\xfe") . "b" . uni_to_native("\x00"));
1758}
1759{
1760    # pack /
1761    my @array = 1..14;
1762    my @out = unpack("N/S", pack("N/S", @array) . "abcd");
1763    is("@out", "@array", "pack N/S works");
1764    @out = unpack("N/S*", pack("N/S*", @array) . "abcd");
1765    is("@out", "@array", "pack N/S* works");
1766    @out = unpack("N/S*", pack("N/S14", @array) . "abcd");
1767    is("@out", "@array", "pack N/S14 works");
1768    @out = unpack("N/S*", pack("N/S15", @array) . "abcd");
1769    is("@out", "@array", "pack N/S15 works");
1770    @out = unpack("N/S*", pack("N/S13", @array) . "abcd");
1771    is("@out", "@array[0..12]", "pack N/S13 works");
1772    @out = unpack("N/S*", pack("N/S0", @array) . "abcd");
1773    is("@out", "", "pack N/S0 works");
1774    is(pack("Z*/a0", "abc"), "0\0", "pack Z*/a0 makes a short string");
1775    is(pack("Z*/Z0", "abc"), "0\0", "pack Z*/Z0 makes a short string");
1776    is(pack("Z*/a3", "abc"), "3\0abc", "pack Z*/a3 makes a full string");
1777    is(pack("Z*/Z3", "abc"), "3\0ab\0", "pack Z*/Z3 makes a short string");
1778    is(pack("Z*/a5", "abc"), "5\0abc\0\0", "pack Z*/a5 makes a long string");
1779    is(pack("Z*/Z5", "abc"), "5\0abc\0\0", "pack Z*/Z5 makes a long string");
1780    is(pack("Z*/Z"), "1\0\0", "pack Z*/Z makes an extended string");
1781    is(pack("Z*/Z", ""), "1\0\0", "pack Z*/Z makes an extended string");
1782    is(pack("Z*/a", ""), "0\0", "pack Z*/a makes an extended string");
1783}
1784{
1785    # unpack("A*", $unicode) strips general unicode spaces
1786    is(unpack("A*", "ab \n" . uni_to_native("\xa0") . " \0"), "ab \n" . uni_to_native("\xa0"),
1787       'normal A* strip leaves \xa0');
1788    is(unpack("U0C0A*", "ab \n" . uni_to_native("\xa0") . " \0"), "ab \n" . uni_to_native("\xa0"),
1789       'normal A* strip leaves \xa0 even if it got upgraded for technical reasons');
1790    is(unpack("A*", pack("a*(U0U)a*", "ab \n", utf8::unicode_to_native(0xa0), " \0")), "ab",
1791       'upgraded strings A* removes \xa0');
1792    is(unpack("A*", pack("a*(U0UU)a*", "ab \n", utf8::unicode_to_native(0xa0), 0x1680, " \0")), "ab",
1793       'upgraded strings A* removes all unicode whitespace');
1794    is(unpack("A5", pack("a*(U0U)a*", "ab \n", 0x1680, "def", "ab")), "ab",
1795       'upgraded strings A5 removes all unicode whitespace');
1796    is(unpack("A*", pack("U", 0x1680)), "",
1797       'upgraded strings A* with nothing left');
1798}
1799{
1800    # Testing unpack . and .!
1801    is(unpack(".", "ABCD"), 0, "offset at start of string is 0");
1802    is(unpack(".", ""), 0, "offset at start of empty string is 0");
1803    is(unpack("x3.", "ABCDEF"), 3, "simple offset works");
1804    is(unpack("x3.", "ABC"), 3, "simple offset at end of string works");
1805    is(unpack("x3.0", "ABC"), 0, "self offset is 0");
1806    is(unpack("x3(x2.)", "ABCDEF"), 2, "offset is relative to inner group");
1807    is(unpack("x3(X2.)", "ABCDEF"), -2,
1808       "negative offset relative to inner group");
1809    is(unpack("x3(X2.2)", "ABCDEF"), 1, "offset is relative to inner group");
1810    is(unpack("x3(x2.0)", "ABCDEF"), 0, "self offset in group is still 0");
1811    is(unpack("x3(x2.2)", "ABCDEF"), 5, "offset counts groups");
1812    is(unpack("x3(x2.*)", "ABCDEF"), 5, "star offset is relative to start");
1813
1814    my $high = chr(8188) x 6;
1815    is(unpack("x3(x2.)", $high), 2, "utf8 offset is relative to inner group");
1816    is(unpack("x3(X2.)", $high), -2,
1817       "utf8 negative offset relative to inner group");
1818    is(unpack("x3(X2.2)", $high), 1, "utf8 offset counts groups");
1819    is(unpack("x3(x2.0)", $high), 0, "utf8 self offset in group is still 0");
1820    is(unpack("x3(x2.2)", $high), 5, "utf8 offset counts groups");
1821    is(unpack("x3(x2.*)", $high), 5, "utf8 star offset is relative to start");
1822
1823    is(unpack("U0x3(x2.)", $high), 2,
1824       "U0 mode utf8 offset is relative to inner group");
1825    is(unpack("U0x3(X2.)", $high), -2,
1826       "U0 mode utf8 negative offset relative to inner group");
1827    is(unpack("U0x3(X2.2)", $high), 1,
1828       "U0 mode utf8 offset counts groups");
1829    is(unpack("U0x3(x2.0)", $high), 0,
1830       "U0 mode utf8 self offset in group is still 0");
1831    is(unpack("U0x3(x2.2)", $high), 5,
1832       "U0 mode utf8 offset counts groups");
1833    is(unpack("U0x3(x2.*)", $high), 5,
1834       "U0 mode utf8 star offset is relative to start");
1835
1836    is(unpack("x3(x2.!)", $high), 2*3,
1837       "utf8 offset is relative to inner group");
1838    is(unpack("x3(X2.!)", $high), -2*3,
1839       "utf8 negative offset relative to inner group");
1840    is(unpack("x3(X2.!2)", $high), 1*3,
1841       "utf8 offset counts groups");
1842    is(unpack("x3(x2.!0)", $high), 0,
1843       "utf8 self offset in group is still 0");
1844    is(unpack("x3(x2.!2)", $high), 5*3,
1845       "utf8 offset counts groups");
1846    is(unpack("x3(x2.!*)", $high), 5*3,
1847       "utf8 star offset is relative to start");
1848
1849    is(unpack("U0x3(x2.!)", $high), 2,
1850       "U0 mode utf8 offset is relative to inner group");
1851    is(unpack("U0x3(X2.!)", $high), -2,
1852       "U0 mode utf8 negative offset relative to inner group");
1853    is(unpack("U0x3(X2.!2)", $high), 1,
1854       "U0 mode utf8 offset counts groups");
1855    is(unpack("U0x3(x2.!0)", $high), 0,
1856       "U0 mode utf8 self offset in group is still 0");
1857    is(unpack("U0x3(x2.!2)", $high), 5,
1858       "U0 mode utf8 offset counts groups");
1859    is(unpack("U0x3(x2.!*)", $high), 5,
1860       "U0 mode utf8 star offset is relative to start");
1861}
1862{
1863    # Testing pack . and .!
1864    is(pack("(a)5 .", 1..5, 3), "123", ". relative to string start, shorten");
1865    eval { () = pack("(a)5 .", 1..5, -3) };
1866    like($@, qr{'\.' outside of string in pack}, "Proper error message");
1867    is(pack("(a)5 .", 1..5, 8), "12345\x00\x00\x00",
1868       ". relative to string start, extend");
1869    is(pack("(a)5 .", 1..5, 5), "12345", ". relative to string start, keep");
1870
1871    is(pack("(a)5 .0", 1..5, -3), "12",
1872       ". relative to string current, shorten");
1873    is(pack("(a)5 .0", 1..5, 2), "12345\x00\x00",
1874       ". relative to string current, extend");
1875    is(pack("(a)5 .0", 1..5, 0), "12345",
1876       ". relative to string current, keep");
1877
1878    is(pack("(a)5 (.)", 1..5, -3), "12",
1879       ". relative to group, shorten");
1880    is(pack("(a)5 (.)", 1..5, 2), "12345\x00\x00",
1881       ". relative to group, extend");
1882    is(pack("(a)5 (.)", 1..5, 0), "12345",
1883       ". relative to group, keep");
1884
1885    is(pack("(a)3 ((a)2 .)", 1..5, -2), "1",
1886       ". relative to group, shorten");
1887    is(pack("(a)3 ((a)2 .)", 1..5, 2), "12345",
1888       ". relative to group, keep");
1889    is(pack("(a)3 ((a)2 .)", 1..5, 4), "12345\x00\x00",
1890       ". relative to group, extend");
1891
1892    is(pack("(a)3 ((a)2 .2)", 1..5, 2), "12",
1893       ". relative to counted group, shorten");
1894    is(pack("(a)3 ((a)2 .2)", 1..5, 7), "12345\x00\x00",
1895       ". relative to counted group, extend");
1896    is(pack("(a)3 ((a)2 .2)", 1..5, 5), "12345",
1897       ". relative to counted group, keep");
1898
1899    is(pack("(a)3 ((a)2 .*)", 1..5, 2), "12",
1900       ". relative to start, shorten");
1901    is(pack("(a)3 ((a)2 .*)", 1..5, 7), "12345\x00\x00",
1902       ". relative to start, extend");
1903    is(pack("(a)3 ((a)2 .*)", 1..5, 5), "12345",
1904       ". relative to start, keep");
1905
1906    is(pack('(a)5 (. @2 a)', 1..5, -3, "a"), "12\x00\x00a",
1907       ". based shrink properly updates group starts");
1908
1909    is(pack("(W)3 ((W)2 .)", 0x301..0x305, -2), "\x{301}",
1910       "utf8 . relative to group, shorten");
1911    is(pack("(W)3 ((W)2 .)", 0x301..0x305, 2),
1912       "\x{301}\x{302}\x{303}\x{304}\x{305}",
1913       "utf8 . relative to group, keep");
1914    is(pack("(W)3 ((W)2 .)", 0x301..0x305, 4),
1915       "\x{301}\x{302}\x{303}\x{304}\x{305}\x00\x00",
1916       "utf8 . relative to group, extend");
1917
1918    is(pack("(W)3 ((W)2 .!)", 0x301..0x305, -2), "\x{301}\x{302}",
1919       "utf8 . relative to group, shorten");
1920    is(pack("(W)3 ((W)2 .!)", 0x301..0x305, 4),
1921       "\x{301}\x{302}\x{303}\x{304}\x{305}",
1922       "utf8 . relative to group, keep");
1923    is(pack("(W)3 ((W)2 .!)", 0x301..0x305, 6),
1924       "\x{301}\x{302}\x{303}\x{304}\x{305}\x00\x00",
1925       "utf8 . relative to group, extend");
1926
1927    is(pack('(W)5 (. @2 a)', 0x301..0x305, -3, "a"),
1928       "\x{301}\x{302}\x00\x00a",
1929       "utf8 . based shrink properly updates group starts");
1930}
1931{
1932    # Testing @!
1933    is(pack('a* @3',  "abcde"), "abc", 'Test basic @');
1934    is(pack('a* @!3', "abcde"), "abc", 'Test basic @!');
1935    is(pack('a* @2', "\x{301}\x{302}\x{303}\x{304}\x{305}"), "\x{301}\x{302}",
1936       'Test basic utf8 @');
1937    is(pack('a* @!2', "\x{301}\x{302}\x{303}\x{304}\x{305}"), "\x{301}",
1938       'Test basic utf8 @!');
1939
1940    is(unpack('@4 a*',  "abcde"), "e", 'Test basic @');
1941    is(unpack('@!4 a*', "abcde"), "e", 'Test basic @!');
1942    is(unpack('@4 a*',  "\x{301}\x{302}\x{303}\x{304}\x{305}"), "\x{305}",
1943       'Test basic utf8 @');
1944    is(unpack('@!4 a*', "\x{301}\x{302}\x{303}\x{304}\x{305}"),
1945       "\x{303}\x{304}\x{305}", 'Test basic utf8 @!');
1946}
1947{
1948    #50256
1949    # This test is for the bit pattern "\x61\x62", which is ASCII "ab"
1950    my ($v) = split //, unpack ('(B)*', native_to_uni('ab'));
1951    is($v, 0); # Doesn't SEGV :-)
1952}
1953{
1954    #73814
1955    my $x = runperl( prog => 'print split( /,/, unpack(q(%2H*), q(hello world))), qq(\n)' );
1956    is($x, "0\n", "split /a/, unpack('%2H*'...) didn't crash");
1957
1958    my $y = runperl( prog => 'print split( /,/, unpack(q(%32u*), q(#,3,Q)), qq(\n)), qq(\n)' );
1959    is($y, "0\n", "split /a/, unpack('%32u*'...) didn't crash");
1960}
1961
1962#90160
1963is(eval { () = unpack "C0 U*", ""; "ok" }, "ok",
1964  'medial U* on empty string');
1965
1966package o {
1967    use overload
1968        '""' => sub { ++$o::str; "42" },
1969        '0+' => sub { ++$o::num; 42 };
1970}
1971is pack("c", bless [], "o"), chr(42), 'overloading called';
1972is $o::str, undef, 'pack "c" does not call string overloading';
1973is $o::num, 1,     'pack "c" does call num overloading';
1974
1975#[perl #123874]: argument underflow leads to corrupt length
1976eval q{ pack "pi/x" };
1977ok(1, "argument underflow did not crash");
1978
1979{
1980    # [perl #126325] pack [hH] with a unicode string
1981    # the hex encoders would read past the end of the string, using
1982    # invalid source bytes
1983    my $twenty_nuls = "\0" x 20;
1984    # This is the case that failed
1985    is(pack("WH40", 0x100, ""), "\x{100}$twenty_nuls",
1986       "check pack H zero fills (utf8 target)");
1987    my $up_nul = "\0";
1988
1989    utf8::upgrade($up_nul);
1990    # check the other combinations too
1991    is(pack("WH40", 0x100, $up_nul), "\x{100}$twenty_nuls",
1992       "check pack H zero fills (utf8 target/source)");
1993    is(pack("H40", ""), $twenty_nuls,
1994       "check pack H zero fills (utf8 none)");
1995    is(pack("H40", $up_nul), $twenty_nuls,
1996       "check pack H zero fills (utf8 source)");
1997}
1998
1999SKIP:
2000{
2001    # [perl #129149] the code below would write one past the end of the output
2002    # buffer, only detected by ASAN, not by valgrind
2003    skip "ASCII-centric test",1 if $::IS_EBCDIC;
2004    $Config{ivsize} >= 8
2005      or skip "[perl #129149] need 64-bit for this test", 1;
2006    fresh_perl_is(<<'EOS', "ok\n", { stderr => 1 }, "pack W overflow");
2007print pack("ucW", "0000", 0, 140737488355327) eq "\$,#`P,```\n\0\x{7fffffffffff}"
2008 ? "ok\n" : "not ok\n";
2009EOS
2010}
2011
2012SKIP:
2013{
2014  # [perl #131844] pointer addition overflow
2015    $Config{ptrsize} == 4
2016      or skip "[perl #131844] need 32-bit build for this test", 4;
2017    # prevent ASAN just crashing on the allocation failure
2018    local $ENV{ASAN_OPTIONS} = $ENV{ASAN_OPTIONS};
2019    $ENV{ASAN_OPTIONS} .= ",allocator_may_return_null=1";
2020    fresh_perl_like('pack "f999999999"', qr/Out of memory during pack/, { stderr => 1 },
2021		    "pointer addition overflow");
2022
2023    # integer (STRLEN) overflow from addition of glen to current length
2024    fresh_perl_like('pack "c10f1073741823"', qr/Out of memory during pack/, { stderr => 1 },
2025		    "integer overflow calculating allocation (addition)");
2026
2027    fresh_perl_like('pack "W10f536870913", 256', qr/Out of memory during pack/, { stderr => 1 },
2028		    "integer overflow calculating allocation (utf8)");
2029
2030    fresh_perl_like('pack "c10f1073741824"', qr/Out of memory during pack/, { stderr => 1 },
2031		    "integer overflow calculating allocation (multiply)");
2032}
2033
2034{
2035    # [perl #132655] heap-buffer-overflow READ of size 11
2036    # only expect failure under ASAN (and maybe valgrind)
2037    fresh_perl_is('0.0 + unpack("u", "ab")', "", { stderr => 1 },
2038                  "ensure unpack u of invalid data nul terminates result");
2039}
2040
2041{
2042	# [GH #16319] SEGV caused by recursion
2043	my $x = eval { pack "[" x 1_000_000 };
2044	like("$@", qr{No group ending character \Q']'\E found in template},
2045			"many opening brackets should not smash the stack");
2046
2047	$x = eval { pack "[(][)]" };
2048	like("$@", qr{Mismatched brackets in template},
2049			"should match brackets correctly even without recursion");
2050}
2051