#
# Tests that have to do with checking whether characters have (or not have)
# certain Unicode properties; belong (or not belong) to blocks, scripts, etc.
+# including user-defined properties
#
use strict;
use warnings;
use v5.16;
+use utf8;
+
+# To verify that messages containing the expansions work on UTF-8
+my $utf8_comment;
my @warnings;
local $SIG {__WARN__} = sub {push @warnings, "@_"};
my @USER_DEFINED_PROPERTIES;
my @USER_CASELESS_PROPERTIES;
+my @USER_ERROR_PROPERTIES;
my @DEFERRED;
+my $overflow;
BEGIN {
+ $utf8_comment = "#\N{U+30CD}";
+
+ use Config;
+ $overflow = $Config{uvsize} < 8 ? "80000000" : "80000000000000000";
# We defined these at compile time, so that the subroutines that they
# refer to aren't known, so that we can test properties not known until
Dash => ['-'],
ASCII_Hex_Digit => ['!-', 'A'],
IsAsciiHexAndDash => ['-', 'A'],
+ InLatin1 => ['\x{0100}', '!\x{00FF}'],
);
@USER_CASELESS_PROPERTIES = (
'pkg1::pkg2::IsMyLower' => ["a", "!A" ],
);
+ @USER_ERROR_PROPERTIES = (
+ 'IsOverflow' => qr/Code point too large in (?#
+ )"0\t$overflow$utf8_comment" in expansion of (?#
+ )main::IsOverflow/,
+ 'InRecursedA' => qr/Infinite recursion in user-defined property (?#
+ )"main::InRecursedA" in expansion of (?#
+ )main::InRecursedC in expansion of (?#
+ )main::InRecursedB in expansion of (?#
+ )main::InRecursedA/,
+ 'IsRangeReversed' => qr/Illegal range in "200 100$utf8_comment" in (?#
+ )expansion of main::IsRangeReversed/,
+ 'IsNonHex' => qr/Can't find Unicode property definition (?#
+ )"BEEF CAGED" in expansion of main::IsNonHex/,
+
+ # Could have \n, hence /s
+ 'IsDeath' => qr/Died.* in expansion of main::IsDeath/s,
+ );
# Now create a list of properties whose definitions won't be known at
# runtime. The qr// below thus will have forward references to them, and
my @DEFERRABLE_USER_DEFINED_PROPERTIES;
push @DEFERRABLE_USER_DEFINED_PROPERTIES, @USER_DEFINED_PROPERTIES;
push @DEFERRABLE_USER_DEFINED_PROPERTIES, @USER_CASELESS_PROPERTIES;
+ unshift @DEFERRABLE_USER_DEFINED_PROPERTIES, @USER_ERROR_PROPERTIES;
for (my $i = 0; $i < @DEFERRABLE_USER_DEFINED_PROPERTIES; $i+=2) {
my $property = $DEFERRABLE_USER_DEFINED_PROPERTIES[$i];
if ($property =~ / ^ \# /x) {
}
}
-# These override the official ones, so if found before defined, the official
-# ones prevail, so can't test deferred definition
-my @OVERRIDING_USER_DEFINED_PROPERTIES = (
- InLatin1 => ['\x{0100}', '!\x{00FF}'],
-);
-
#
# From the short properties we populate POSIX-like classes.
#
push @CLASSES => "# Short properties" => %SHORT_PROPERTIES,
"# POSIX like properties" => %d,
- "# User defined properties" => @USER_DEFINED_PROPERTIES,
- "# Overriding user defined properties" => @OVERRIDING_USER_DEFINED_PROPERTIES;
+ "# User defined properties" => @USER_DEFINED_PROPERTIES;
#
$count += 4 * @ILLEGAL_PROPERTIES;
$count += 4 * grep {length $_ == 1} @ILLEGAL_PROPERTIES;
$count += 8 * @USER_CASELESS_PROPERTIES;
-$count += 1 * @DEFERRED / 2;
+$count += 1 * (@DEFERRED - @USER_ERROR_PROPERTIES) / 2;
+$count += 1 * @USER_ERROR_PROPERTIES;
+$count += 1; # one bad apple
$count += 1; # No warnings generated
plan(tests => $count);
sub run_tests {
for (my $i = 0; $i < @DEFERRED; $i+=2) {
+ if (ref $DEFERRED[$i+1] eq 'ARRAY') {
my ($str, $name) = get_str_name($DEFERRED[$i+1][0]);
like($str, $DEFERRED[$i],
"$name correctly matched $DEFERRED[$i] (defn. not known until runtime)");
+ }
+ else { # Single entry rhs indicates a property that is an error
+ undef $@;
+
+ # Using block eval causes the pattern to not be recompiled, so it
+ # retains its deferred status until this is executed.
+ eval { 'A' =~ $DEFERRED[$i] };
+ like($@, $DEFERRED[$i+1],
+ "$DEFERRED[$i] gave correct failure message (defn. not known until runtime)");
+ }
}
while (@CLASSES) {
# Verify works as regularly for not /i
match $_, $in_pat, $out_pat for @in;
match $_, $out_pat, $in_pat for @out;
+ }
+ print "# User-defined properties with errors in their definition\n";
+ while (my $error_property = shift @USER_ERROR_PROPERTIES) {
+ my $error_re = shift @USER_ERROR_PROPERTIES;
+ undef $@;
+ eval { 'A' =~ /\p{$error_property}/; };
+ like($@, $error_re, "$error_property gave correct failure message");
}
}
0730 074A
--
+sub InRecursedA {
+ return "+main::InRecursedB\n";
+}
+
+sub InRecursedB {
+ return "+main::InRecursedC\n";
+}
+
+sub InRecursedC {
+ return "+main::InRecursedA\n";
+}
+
sub InGreekSmall {return "03B1\t03C9"}
sub InGreekCapital {return "0391\t03A9\n-03A2"}
}
sub IsMyUpper {
+ use feature 'state';
+
+ state $cased_count = 0;
+ state $caseless_count = 0;
+ my $ret= "+utf8::";
+
my $caseless = shift;
- return "+utf8::"
- . (($caseless)
- ? 'Alphabetic'
- : 'Uppercase')
- . "\n&utf8::ASCII";
+ if($caseless) {
+ die "Called twice" if $caseless_count;
+ $caseless_count++;
+ $ret .= 'Alphabetic'
+ }
+ else {
+ die "Called twice" if $cased_count;
+ $cased_count++;
+ $ret .= 'Uppercase';
+ }
+
+ return $ret . "\n&utf8::ASCII";
}
sub pkg1::pkg2::IsMyLower {
. "\n&utf8::ASCII";
}
+sub IsRangeReversed {
+ return "200 100$utf8_comment";
+}
+
+sub IsNonHex {
+ return "BEEF CAGED$utf8_comment";
+}
+
+sub IsDeath {
+ die;
+}
+
# Verify that can use user-defined properties inside another one
sub IsSyriac1KanaMark {<<'--'}
+main::IsSyriac1
sub Is::foo { die }
sub In::foo { die }
+sub IsOverflow {
+ return "0\t$overflow$utf8_comment";
+}
+
+fresh_perl_like(<<'EOP', qr/Can't find Unicode property definition "F000\\tF010" in expansion of InOneBadApple/, {}, "Just one component bad");
+# Extra backslash converts tab to backslash-t
+sub InOneBadApple { return "0100\t0110\n10000\t10010\nF000\\tF010\n0400\t0410" }
+qr/\p{InOneBadApple}/;
+EOP
+
if (! is(@warnings, 0, "No warnings were generated")) {
diag join "\n", @warnings, "\n";
}