[p5sagit/p5-mst-13.2.git] / t / re / regexp.t

#!./perl

# The tests are in a separate file 't/re/re_tests'.
# Each line in that file is a separate test.
# There are five columns, separated by tabs.
#
# Column 1 contains the pattern, optionally enclosed in C<''>.
# Modifiers can be put after the closing C<'>.
#
# Column 2 contains the string to be matched.
#
# Column 3 contains the expected result:
# 	y	expect a match
# 	n	expect no match
# 	c	expect an error
#	T	the test is a TODO (can be combined with y/n/c)
#	B	test exposes a known bug in Perl, should be skipped
#	b	test exposes a known bug in Perl, should be skipped if noamp
#	t	test exposes a bug with threading, TODO if qr_embed_thr
#
# Columns 4 and 5 are used only if column 3 contains C<y> or C<c>.
#
# Column 4 contains a string, usually C<$&>.
#
# Column 5 contains the expected result of double-quote
# interpolating that string after the match, or start of error message.
#
# Column 6, if present, contains a reason why the test is skipped.
# This is printed with "skipped", for harness to pick up.
#
# \n in the tests are interpolated, as are variables of the form ${\w+}.
#
# Blanks lines are treated as PASSING tests to keep the line numbers
# linked to the test number.
#
# If you want to add a regular expression test that can't be expressed
# in this format, don't add it here: put it in re/pat.t instead.
#
# Note that the inputs get passed on as "m're'", so the re bypasses the lexer.
# This means this file cannot be used for testing anything that the lexer
# handles; in 5.12 this means just \N{NAME} and \N{U+...}.
#
# Note that columns 2,3 and 5 are all enclosed in double quotes and then
# evalled; so something like a\"\x{100}$1 has length 3+length($1).

my $file;
BEGIN {
    $iters = shift || 1;	# Poor man performance suite, 10000 is OK.

    # Do this open before any chdir
    $file = shift;
    if (defined $file) {
	open TESTS, $file or die "Can't open $file";
    }

    chdir 't' if -d 't';
    @INC = '../lib';

    if ($qr_embed_thr) {
	require Config;
	if (!$Config::Config{useithreads}) {
	    print "1..0 # Skip: no ithreads\n";
		exit 0;
	}
	if ($ENV{PERL_CORE_MINITEST}) {
	    print "1..0 # Skip: no dynamic loading on miniperl, no threads\n";
		exit 0;
	}
	require threads;
    }
}

use strict;
use warnings FATAL=>"all";
use vars qw($iters $numtests $bang $ffff $nulnul $OP);
use vars qw($qr $skip_amp $qr_embed $qr_embed_thr); # set by our callers


if (!defined $file) {
    open(TESTS,'re/re_tests') || open(TESTS,'t/re/re_tests')
	|| open(TESTS,':re:re_tests') || die "Can't open re_tests";
}

my @tests = <TESTS>;

close TESTS;

$bang = sprintf "\\%03o", ord "!"; # \41 would not be portable.
$ffff  = chr(0xff) x 2;
$nulnul = "\0" x 2;
$OP = $qr ? 'qr' : 'm';

$| = 1;
printf "1..%d\n# $iters iterations\n", scalar @tests;

my $test;
TEST:
foreach (@tests) {
    $test++;
    if (!/\S/ || /^\s*#/ || /^__END__$/) {
        print "ok $test # (Blank line or comment)\n";
        if (/#/) { print $_ };
        next;
    }
    chomp;
    s/\\n/\n/g;
    my ($pat, $subject, $result, $repl, $expect, $reason) = split(/\t/,$_,6);
    $reason = '' unless defined $reason;
    my $input = join(':',$pat,$subject,$result,$repl,$expect);
    # the double '' below keeps simple syntax highlighters from going crazy
    $pat = "'$pat'" unless $pat =~ /^[:''\/]/; 
    $pat =~ s/(\$\{\w+\})/$1/eeg;
    $pat =~ s/\\n/\n/g;
    $subject = eval qq("$subject"); die $@ if $@;
    $expect  = eval qq("$expect"); die $@ if $@;
    $expect = $repl = '-' if $skip_amp and $input =~ /\$[&\`\']/;
    my $todo_qr = $qr_embed_thr && ($result =~ s/t//);
    my $skip = ($skip_amp ? ($result =~ s/B//i) : ($result =~ s/B//));
    $reason = 'skipping $&' if $reason eq  '' && $skip_amp;
    $result =~ s/B//i unless $skip;
    my $todo= $result =~ s/T// ? " # TODO" : "";
    

    for my $study ('', 'study $subject', 'utf8::upgrade($subject)',
		   'utf8::upgrade($subject); study $subject') {
	# Need to make a copy, else the utf8::upgrade of an alreay studied
	# scalar confuses things.
	my $subject = $subject;
	my $c = $iters;
	my ($code, $match, $got);
        if ($repl eq 'pos') {
            $code= <<EOFCODE;
                $study;
                pos(\$subject)=0;
                \$match = ( \$subject =~ m${pat}g );
                \$got = pos(\$subject);
EOFCODE
        }
        elsif ($qr_embed) {
            $code= <<EOFCODE;
                my \$RE = qr$pat;
                $study;
                \$match = (\$subject =~ /(?:)\$RE(?:)/) while \$c--;
                \$got = "$repl";
EOFCODE
        }
        elsif ($qr_embed_thr) {
            $code= <<EOFCODE;
		# Can't run the match in a subthread, but can do this and
	 	# clone the pattern the other way.
                my \$RE = threads->new(sub {qr$pat})->join();
                $study;
                \$match = (\$subject =~ /(?:)\$RE(?:)/) while \$c--;
                \$got = "$repl";
EOFCODE
        }
        else {
            $code= <<EOFCODE;
                $study;
                \$match = (\$subject =~ $OP$pat) while \$c--;
                \$got = "$repl";
EOFCODE
        }
        #$code.=qq[\n\$expect="$expect";\n];
        #use Devel::Peek;
        #die Dump($code) if $pat=~/\\h/ and $subject=~/\x{A0}/;
	{
	    # Probably we should annotate specific tests with which warnings
	    # categories they're known to trigger, and hence should be
	    # disabled just for that test
	    no warnings qw(uninitialized regexp);
	    eval $code;
	}
	chomp( my $err = $@ );
	if ($result eq 'c') {
	    if ($err !~ m!^\Q$expect!) { print "not ok $test$todo (compile) $input => `$err'\n"; next TEST }
	    last;  # no need to study a syntax error
	}
	elsif ( $skip ) {
	    print "ok $test # skipped", length($reason) ? " $reason" : '', "\n";
	    next TEST;
	}
	elsif ( $todo_qr ) {
	    print "not ok $test # TODO", length($reason) ? " - $reason" : '', "\n";
	    next TEST;
	}
	elsif ($@) {
	    print "not ok $test$todo $input => error `$err'\n$code\n$@\n"; next TEST;
	}
	elsif ($result =~ /^n/) {
	    if ($match) { print "not ok $test$todo ($study) $input => false positive\n"; next TEST }
	}
	else {
	    if (!$match || $got ne $expect) {
	        eval { require Data::Dumper };
		if ($@) {
		    print "not ok $test$todo ($study) $input => `$got', match=$match\n$code\n";
		}
		else { # better diagnostics
		    my $s = Data::Dumper->new([$subject],['subject'])->Useqq(1)->Dump;
		    my $g = Data::Dumper->new([$got],['got'])->Useqq(1)->Dump;
		    print "not ok $test$todo ($study) $input => `$got', match=$match\n$s\n$g\n$code\n";
		}
		next TEST;
	    }
	}
    }
    print "ok $test$todo\n";
}

1;
Commit	Line	Data
378cc40b	1	#!./perl
378cc40b	2
ae34ee58	3	# The tests are in a separate file 't/re/re_tests'.
ad4f75a6	4	# Each line in that file is a separate test.
	5	# There are five columns, separated by tabs.
	6	#
	7	# Column 1 contains the pattern, optionally enclosed in C<''>.
	8	# Modifiers can be put after the closing C<'>.
	9	#
	10	# Column 2 contains the string to be matched.
	11	#
	12	# Column 3 contains the expected result:
	13	# y expect a match
	14	# n expect no match
	15	# c expect an error
24d786f4	16	# T the test is a TODO (can be combined with y/n/c)
cf93c79d	17	# B test exposes a known bug in Perl, should be skipped
cf93c79d	18	# b test exposes a known bug in Perl, should be skipped if noamp
e3faa678	19	# t test exposes a bug with threading, TODO if qr_embed_thr
ad4f75a6	20	#
1b1626e4	21	# Columns 4 and 5 are used only if column 3 contains C<y> or C<c>.
ad4f75a6	22	#
	23	# Column 4 contains a string, usually C<$&>.
	24	#
	25	# Column 5 contains the expected result of double-quote
c277df42	26	# interpolating that string after the match, or start of error message.
c277df42	27	#
ee595aa6	28	# Column 6, if present, contains a reason why the test is skipped.
	29	# This is printed with "skipped", for harness to pick up.
	30	#
9d116dd7	31	# \n in the tests are interpolated, as are variables of the form ${\w+}.
83e898de	32	#
b9b4dddf	33	# Blanks lines are treated as PASSING tests to keep the line numbers
	34	# linked to the test number.
	35	#
8d37f932	36	# If you want to add a regular expression test that can't be expressed
67a2b8c6	37	# in this format, don't add it here: put it in re/pat.t instead.
b2a156bd	38	#
ff3f963a	39	# Note that the inputs get passed on as "m're'", so the re bypasses the lexer.
	40	# This means this file cannot be used for testing anything that the lexer
	41	# handles; in 5.12 this means just \N{NAME} and \N{U+...}.
	42	#
b2a156bd	43	# Note that columns 2,3 and 5 are all enclosed in double quotes and then
b2a156bd	44	# evalled; so something like a\"\x{100}$1 has length 3+length($1).
c277df42	45
1a610890	46	my $file;
e4d48cc9	47	BEGIN {
1a610890	48	$iters = shift \|\| 1; # Poor man performance suite, 10000 is OK.
	49
	50	# Do this open before any chdir
	51	$file = shift;
	52	if (defined $file) {
	53	open TESTS, $file or die "Can't open $file";
	54	}
	55
e4d48cc9	56	chdir 't' if -d 't';
20822f61	57	@INC = '../lib';
e3faa678	58
	59	if ($qr_embed_thr) {
	60	require Config;
	61	if (!$Config::Config{useithreads}) {
	62	print "1..0 # Skip: no ithreads\n";
	63	exit 0;
	64	}
	65	if ($ENV{PERL_CORE_MINITEST}) {
	66	print "1..0 # Skip: no dynamic loading on miniperl, no threads\n";
	67	exit 0;
	68	}
	69	require threads;
	70	}
e4d48cc9	71	}
1a610890	72
1286eaeb	73	use strict;
66fb63c1	74	use warnings FATAL=>"all";
1286eaeb	75	use vars qw($iters $numtests $bang $ffff $nulnul $OP);
e3faa678	76	use vars qw($qr $skip_amp $qr_embed $qr_embed_thr); # set by our callers
e4d48cc9	77
ad4f75a6	78
1a610890	79	if (!defined $file) {
a4499558	80	open(TESTS,'re/re_tests') \|\| open(TESTS,'t/re/re_tests')
a4499558	81	\|\| open(TESTS,':re:re_tests') \|\| die "Can't open re_tests";
1a610890	82	}
	83
	84	my @tests = <TESTS>;
cfa4f241	85
1a610890	86	close TESTS;
378cc40b	87
9d116dd7	88	$bang = sprintf "\\%03o", ord "!"; # \41 would not be portable.
b8c5462f	89	$ffff = chr(0xff) x 2;
b8c5462f	90	$nulnul = "\0" x 2;
7fba1cd6	91	$OP = $qr ? 'qr' : 'm';
9d116dd7	92
1462b684	93	$\| = 1;
1a610890	94	printf "1..%d\n# $iters iterations\n", scalar @tests;
e3faa678	95
1a610890	96	my $test;
cfa4f241	97	TEST:
1a610890	98	foreach (@tests) {
1a610890	99	$test++;
5a51db05	100	if (!/\S/ \|\| /^\s*#/ \|\| /^__END__$/) {
1a610890	101	print "ok $test # (Blank line or comment)\n";
5a51db05	102	if (/#/) { print $_ };
b9b4dddf	103	next;
b9b4dddf	104	}
b85d18e9	105	chomp;
b85d18e9	106	s/\\n/\n/g;
1286eaeb	107	my ($pat, $subject, $result, $repl, $expect, $reason) = split(/\t/,$_,6);
66fb63c1	108	$reason = '' unless defined $reason;
1286eaeb	109	my $input = join(':',$pat,$subject,$result,$repl,$expect);
24d786f4	110	# the double '' below keeps simple syntax highlighters from going crazy
24d786f4	111	$pat = "'$pat'" unless $pat =~ /^[:''\/]/;
9d116dd7	112	$pat =~ s/(\$\{\w+\})/$1/eeg;
b8c5462f	113	$pat =~ s/\\n/\n/g;
1a610890	114	$subject = eval qq("$subject"); die $@ if $@;
1a610890	115	$expect = eval qq("$expect"); die $@ if $@;
c277df42	116	$expect = $repl = '-' if $skip_amp and $input =~ /\$[&\`\']/;
24d786f4	117	my $todo_qr = $qr_embed_thr && ($result =~ s/t//);
1286eaeb	118	my $skip = ($skip_amp ? ($result =~ s/B//i) : ($result =~ s/B//));
906e884f	119	$reason = 'skipping $&' if $reason eq '' && $skip_amp;
cf93c79d	120	$result =~ s/B//i unless $skip;
24d786f4	121	my $todo= $result =~ s/T// ? " # TODO" : "";
24d786f4	122
1de06328	123
52e33015	124	for my $study ('', 'study $subject', 'utf8::upgrade($subject)',
	125	'utf8::upgrade($subject); study $subject') {
	126	# Need to make a copy, else the utf8::upgrade of an alreay studied
	127	# scalar confuses things.
	128	my $subject = $subject;
1286eaeb	129	my $c = $iters;
1286eaeb	130	my ($code, $match, $got);
1de06328	131	if ($repl eq 'pos') {
	132	$code= <<EOFCODE;
	133	$study;
	134	pos(\$subject)=0;
	135	\$match = ( \$subject =~ m${pat}g );
	136	\$got = pos(\$subject);
	137	EOFCODE
	138	}
	139	elsif ($qr_embed) {
	140	$code= <<EOFCODE;
	141	my \$RE = qr$pat;
	142	$study;
	143	\$match = (\$subject =~ /(?:)\$RE(?:)/) while \$c--;
	144	\$got = "$repl";
	145	EOFCODE
	146	}
e3faa678	147	elsif ($qr_embed_thr) {
	148	$code= <<EOFCODE;
	149	# Can't run the match in a subthread, but can do this and
	150	# clone the pattern the other way.
	151	my \$RE = threads->new(sub {qr$pat})->join();
	152	$study;
	153	\$match = (\$subject =~ /(?:)\$RE(?:)/) while \$c--;
	154	\$got = "$repl";
	155	EOFCODE
	156	}
1de06328	157	else {
	158	$code= <<EOFCODE;
	159	$study;
1286eaeb	160	\$match = (\$subject =~ $OP$pat) while \$c--;
1de06328	161	\$got = "$repl";
	162	EOFCODE
	163	}
e1d1eefb	164	#$code.=qq[\n\$expect="$expect";\n];
	165	#use Devel::Peek;
	166	#die Dump($code) if $pat=~/\\h/ and $subject=~/\x{A0}/;
66fb63c1	167	{
	168	# Probably we should annotate specific tests with which warnings
	169	# categories they're known to trigger, and hence should be
	170	# disabled just for that test
	171	no warnings qw(uninitialized regexp);
	172	eval $code;
	173	}
1286eaeb	174	chomp( my $err = $@ );
cfa4f241	175	if ($result eq 'c') {
24d786f4	176	if ($err !~ m!^\Q$expect!) { print "not ok $test$todo (compile) $input => `$err'\n"; next TEST }
cfa4f241	177	last; # no need to study a syntax error
cfa4f241	178	}
cf93c79d	179	elsif ( $skip ) {
1a610890	180	print "ok $test # skipped", length($reason) ? " $reason" : '', "\n";
ee595aa6	181	next TEST;
cf93c79d	182	}
24d786f4	183	elsif ( $todo_qr ) {
e0892690	184	print "not ok $test # TODO", length($reason) ? " - $reason" : '', "\n";
e3faa678	185	next TEST;
e3faa678	186	}
c277df42	187	elsif ($@) {
24d786f4	188	print "not ok $test$todo $input => error `$err'\n$code\n$@\n"; next TEST;
c277df42	189	}
e3faa678	190	elsif ($result =~ /^n/) {
24d786f4	191	if ($match) { print "not ok $test$todo ($study) $input => false positive\n"; next TEST }
378cc40b	192	}
378cc40b	193	else {
cfa4f241	194	if (!$match \|\| $got ne $expect) {
cde0cee5	195	eval { require Data::Dumper };
cde0cee5	196	if ($@) {
24d786f4	197	print "not ok $test$todo ($study) $input => `$got', match=$match\n$code\n";
cde0cee5	198	}
	199	else { # better diagnostics
	200	my $s = Data::Dumper->new([$subject],['subject'])->Useqq(1)->Dump;
	201	my $g = Data::Dumper->new([$got],['got'])->Useqq(1)->Dump;
24d786f4	202	print "not ok $test$todo ($study) $input => `$got', match=$match\n$s\n$g\n$code\n";
cde0cee5	203	}
cfa4f241	204	next TEST;
cfa4f241	205	}
378cc40b	206	}
378cc40b	207	}
24d786f4	208	print "ok $test$todo\n";
378cc40b	209	}
cfa4f241	210
1a610890	211	1;