3 # Checks if the parser behaves correctly in edge cases
4 # (including weird syntax errors)
9 require './charset_tools.pl';
10 skip_all_without_unicode_tables();
16 use open qw( :utf8 :std );
18 is *tèst, "*main::tèst", "sanity check.";
19 ok $::{"tèst"}, "gets the right glob in the stash.";
21 my $glob_by_sub = sub { *main::method }->();
23 is *main::method, "*main::method", "glob stringy works";
24 is "" . *main::method, "*main::method", "glob stringify-through-concat works";
25 is $glob_by_sub, "*main::method", "glob stringy works";
26 is "" . $glob_by_sub, "*main::method", "";
30 is *{$_[0]}, "*main::$_[0]";
34 is "" . gimme_glob("下郎"), "*main::下郎";
36 is "" . $a, "*main::下郎";
38 *{gimme_glob("下郎")} = sub {};
41 ok defined *{"下郎"}{CODE};
42 ok !defined *{"\344\270\213\351\203\216"}{CODE};
46 is ${*Lèon{SCALAR}}, 1, "scalar define in the right glob,";
47 ok !${*{"L\303\250on"}{SCALAR}}, "..and nothing in the wrong one.";
49 my $a = "foo" . chr(190);
50 my $b = $a . chr(256);
51 chop $b; # $b is $a with utf8 on
53 is $a, $b, '$a equals $b';
57 is eval { main->$a }, 5, q!$a can call $b's sub!;
58 ok !$@, "..and there's no error.";
62 ok $b ne $c, '$b unequal $c';
64 ok $@, q!$c can't call $b's sub.!;
66 # Now define another sub under the downgraded name:
69 is eval { main->$a }, 6, "Adding a new sub to *a and calling it works,";
70 ok !$@, "..without errors.";
72 ok $@, "but it's still unreachable through *c";
75 is ${*$a{SCALAR}}, 10;
76 is ${*$b{SCALAR}}, 10;
77 is ${*$c{SCALAR}}, undef;
81 ::ok($::{"FÒÒ"}, "Bareword generates the right glob.");
82 ::ok(!$::{"F\303\222\303\222"});
86 ok $::{"участники"}, "non-const sub declarations generate the right glob";
87 is $::{"участники"}->(), 1;
91 is grep({ $_ eq "\x{539f}" } keys %::), 1, "Constant subs generate the right glob.";
92 is grep({ $_ eq "\345\216\237" } keys %::), 0;
94 #These should probably go elsewhere.
95 eval q{ sub wròng1 (_$); wròng1(1,2) };
96 like( $@, qr/Malformed prototype for main::wròng1/, 'Malformed prototype croak is clean.' );
98 eval q{ sub ча::ики ($__); ча::ики(1,2) };
99 like( $@, qr/Malformed prototype for ча::ики/ );
102 is $問, 10, "our works";
103 is $main::問, 10, "...as does getting the same variable through the fully qualified name";
104 is ${"main::\345\225\217"}, undef, "..and using the encoded form doesn't";
107 use charnames qw( :full );
109 eval qq! my \$\x{30cb} \N{DROMEDARY CAMEL} !;
110 $@ =~ s/eval \d+/eval 11/;
111 is $@, 'Unrecognized character \x{1f42a}; marked by <-- HERE after my $ニ <-- HERE near column 8 at (eval 11) line 1.
112 ', "'Unrecognized character' croak is UTF-8 clean";
114 eval "q\0foobar\0 \x{FFFF}+1";
115 $@ =~ s/eval \d+/eval 11/;
118 "Unrecognized character \\x{ffff}; marked by <-- HERE after q\0foobar\0 <-- HERE near column 11 at (eval 11) line 1.\n",
124 my $f = qq{(?{\$ネ+ 1; \x{1F42A} })};
125 eval { "a" =~ /^a$f/ };
127 $e =~ s/eval \d+/eval 11/;
130 "Unrecognized character \\x{1f42a}; marked by <-- HERE after (?{\$ネ+ 1; <-- HERE near column 13 at (eval 11) line 1.\n",
131 "Messages from a re-eval are UTF-8 clean"
134 $f = qq{(?{q\0foobar\0 \x{FFFF}+1 })};
135 eval { "a" =~ /^a$f/ };
137 $e =~ s/eval \d+/eval 11/;
140 "Unrecognized character \\x{ffff}; marked by <-- HERE after q\x{0}foobar\x{0} <-- HERE near column 16 at (eval 11) line 1.\n",
146 eval qq{\$ネ+ 1; \x{1F42A}};
147 $@ =~ s/eval \d+/eval 11/;
150 "Unrecognized character \\x{1f42a}; marked by <-- HERE after \$ネ+ 1; <-- HERE near column 8 at (eval 11) line 1.\n",
151 "Unrecognized character error doesn't cut off in the middle of characters"
159 for ( qw( my state our ) ) {
161 eval "$_ Foo $x = 1;";
162 like $@, qr/No such class Foo/u, "'No such class' warning for $_ is UTF-8 clean";
168 eval "our \$main::\x{30cb};";
169 like $@, qr!No package name allowed for variable \$main::\x{30cb} in "our"!, "'No such package name allowed for variable' is UTF-8 clean";
175 for ( qw( my state ) ) {
176 eval "$_ \$::\x{30cb};";
177 like $@, qr!"$_" variable \$::\x{30cb} can't be in a package!, qq!'"$_" variable %s can't be in a package' is UTF-8 clean!;
183 eval qq!print \x{30cb}, "comma""!;
184 like $@, qr/No comma allowed after filehandle/, "No comma allowed after filehandle triggers correctly for UTF-8 filehandles.";
187 # tests for "Bad name"
189 like( $@, qr/Bad name after Foo::/, 'Bad name after Foo::' );
191 like( $@, qr/Bad name after Foo'/, 'Bad name after Foo\'' );
195 local $SIG{__WARN__} = sub { }; # The eval will also output a warning,
197 my $malformed_to_be = ($::IS_EBCDIC) # Overlong sequence
200 CORE::evalbytes "use charnames ':full'; use utf8; my \$x = \"\\N{abc$malformed_to_be}\"";
201 like( $@, qr/Malformed UTF-8 character \(fatal\) at /, 'Malformed UTF-8 input to \N{}');
204 # RT# 124216: Perl_sv_clear: Assertion
205 # If a parsing error occurred during a forced token within an interpolated
206 # context, the stack unwinding failed to restore PL_lex_defer and so after
207 # error recovery the state restored after the forced token was processed
208 # was the wrong one, resulting in the lexer thinking we're still inside a
209 # quoted string and things getting freed multiple times.
211 # The \x{3030} char isn't a legal var name, and this triggers the error.
213 # NB: this only failed if the closing quote of the interpolated string is
214 # the last char of the file (i.e. no trailing \n).
217 my $bad = "\x{3030}";
218 # Write out the individual utf8 bytes making up \x{3030}. This
219 # avoids 'Wide char in print' warnings from test.pl. (We may still
220 # get that warning when compiling the prog itself, since the
221 # error it prints to stderr contains a wide char.)
224 fresh_perl_like(qq{use utf8; "\$$bad"},
227 ( \QWide character in print at - line 1.\E\n )?
228 \Qsyntax error at - line 1, near \E"\$.*"\n
229 \QExecution of - aborted due to compilation errors.\E\z
232 {stderr => 1}, "RT# 124216");
238 if ($Config{uvsize} < 8) {
239 skip("test is only valid on 64-bit ints", 4);
245 # This caused a memory fault [perl #128738]
246 $b = byte_utf8a_to_utf8n("\xFE\x82\x80\x80\x80\x80\x80"); # 0x80000000
247 eval "\$a = q ${b}abc${b}";
249 "No errors in eval'ing a string with large code point delimiter";
251 "Got expected result in eval'ing a string with a large code point"
254 $b = byte_utf8a_to_utf8n("\xFE\x83\xBF\xBF\xBF\xBF\xBF"); # 0xFFFFFFFF
255 eval "\$a = q ${b}Hello, \\\\whirled!${b}";
257 "No errors in eval'ing a string with large code point delimiter";
258 is $a, 'Hello, \whirled!',
259 "Got expected result in eval'ing a string with a large code point"
264 fresh_perl_is(<<'EOS', <<'EXPECT', {}, 'no panic in pad_findmy_pvn (#134061)');
266 eval "sort \x{100}%";
269 syntax error at (eval 1) line 1, at EOF
272 # New tests go here ^^^^^
274 # Keep this test last, as it will mess up line number reporting for any
282 is __LINE__, 59, '#line directive and qq with uni delims inside heredoc';
284 # Put new tests above the line number tests.