Not having such space has been deprecated since v5.14.0.
refers to U+1F514, and the name for U+0007 is "ALERT". All the
functions in L<charnames> have been correspondingly updated.
+=head2 Alphanumeric operators must now be separated from the closing
+delimter of regular expressions
+
+You may no longer write something like:
+
+ m/a/and 1
+
+Instead you must write
+
+ m/a/ and 1
+
+with whitespace separating the operator from the closing delimiter of
+the regular expression. Not having whitespace has resulted in a
+deprecated warning since Perl v5.14.0.
+
=head1 Deprecations
XXX Any deprecated features, syntax, modules etc. should be listed here.
the string, just write C<"-foo">. If you meant the function call,
write C<-foo()>.
-=item Ambiguous use of 's//le...' resolved as 's// le...'; Rewrite as 's//el' if you meant 'use locale rules and evaluate rhs as an expression'. In Perl 5.18, it will be resolved the other way
-
-(W deprecated, ambiguous) You wrote a pattern match with substitution
-immediately followed by "le". In Perl 5.16 and earlier, this is
-resolved as meaning to take the result of the substitution, and see if
-it is stringwise less-than-or-equal-to what follows in the expression.
-Having the "le" immediately following a pattern is deprecated behavior,
-so in Perl 5.18, this expression will be resolved as meaning to do the
-pattern match using the rules of the current locale, and evaluate the
-rhs as an expression when doing the substitution. In 5.14, and 5.16 if
-you want the latter interpretation, you can simply write "el" instead.
-But note that the C</l> modifier should not be used explicitly anyway;
-you should use C<use locale> instead. See L<perllocale>.
-
=item '|' and '<' may not both be specified on command line
(F) An error peculiar to VMS. Perl does its own command line
(F) The parser has given up trying to parse the program after 10 errors.
Further error messages would likely be uninformative.
-=item Having no space between pattern and following word is deprecated
-
-(D syntax)
-
-You had a word that isn't a regex modifier immediately following
-a pattern without an intervening space. If you are trying to use
-the C</le> flags on a substitution, use C</el> instead. Otherwise, add
-white space between the pattern and following word to eliminate
-the warning. As an example of the latter, the two constructs:
-
-
- $a =~ m/$foo/sand $bar
- $a =~ m/$foo/s and $bar
-
-both currently mean the same thing, but it is planned to disallow
-the first form in Perl 5.18. And,
-
- $a =~ m/$foo/and $bar
-
-will be disallowed too.
-
=item Hexadecimal number > 0xffffffff non-portable
(W portable) The hexadecimal number you specified is larger than 2**32-1
(W) You tried to use an unknown subpragma of the "re" pragma.
+=item Unknown regex modifier "%s"
+
+(F) Alphanumerics immediately following the closing delimiter
+of a regular expression pattern are interpreted by Perl as modifier
+flags for the regex. One of the ones you specified is invalid. One way
+this can happen is if you didn't put in white space between the end of
+the regex and a following alphanumeric operator:
+
+ if ($a =~ /foo/and $bar == 3) { ... }
+
+The C<"a"> is a valid modifier flag, but the C<"n"> is not, and raises
+this error. Likely what was meant instead was:
+
+ if ($a =~ /foo/ and $bar == 3) { ... }
+
=item Unknown switch condition (?(%s in regex; marked by <-- HERE in m/%s/
(F) The condition part of a (?(condition)if-clause|else-clause) construct
# expand tabs to 8-column spacing
1 while s/\t+/' ' x (length($&)*8 - length($`)%8)/e;
-C<s///le> is treated as a substitution followed by the C<le> operator, not
-the C</le> flags. This may change in a future version of Perl. It
-produces a warning if warnings are enabled. To disambiguate, use a space
-or change the order of the flags:
-
- s/foo/bar/ le 5; # "le" infix operator
- s/foo/bar/el; # "e" and "l" flags
-
=back
=head2 Quote-Like Operators
Use of comma-less variable list is deprecated at - line 4.
########
# toke.c
-$a =~ m/$foo/sand $bar;
-$a =~ s/$foo/fool/sand $bar;
$a = <<;
no warnings 'deprecated' ;
-$a =~ m/$foo/sand $bar;
-$a =~ s/$foo/fool/sand $bar;
$a = <<;
EXPECT
-Having no space between pattern and following word is deprecated at - line 2.
-Having no space between pattern and following word is deprecated at - line 3.
-Use of bare << to mean <<"" is deprecated at - line 4.
+Use of bare << to mean <<"" is deprecated at - line 2.
+########
+# toke.c
+$a =~ m/$foo/eq;
+$a =~ s/$foo/fool/seq;
+
+EXPECT
+OPTION fatal
+Unknown regexp modifier "/e" at - line 2, near "=~ "
+Unknown regexp modifier "/q" at - line 2, near "=~ "
+Unknown regexp modifier "/q" at - line 3, near "=~ "
+Execution of - aborted due to compilation errors.
########
# toke.c
use warnings 'syntax' ;
require './test.pl';
}
-plan tests => 474; # Update this when adding/deleting tests.
+plan tests => 436; # Update this when adding/deleting tests.
run_tests() unless caller;
}
{
- # Test that a regex followed by an operator and/or a statement modifier work
- # These tests use string-eval so that it reports a clean error when it fails
- # (without the string eval the test script might be unparseable)
-
- # Note: these test check the behaviour that currently is valid syntax
- # If a new regex modifier is added and a test fails then there is a backwards-compatibility issue
- # Note-2: a new deprecate warning was added for this with commit e6897b1a5db0410e387ccbf677e89fc4a1d8c97a
- # which indicate that this syntax will be removed in 5.16.
- # When this happens the tests can be removed
-
- foreach (['my $r = "a" =~ m/a/lt 2', 'm', 'lt'],
- ['my $r = "a" =~ m/a/le 1', 'm', 'le'],
- ['my $r = "a" =~ m/a/eq 1', 'm', 'eq'],
- ['my $r = "a" =~ m/a/ne 0', 'm', 'ne'],
- ['my $r = "a" =~ m/a/and 1', 'm', 'and'],
- ['my $r = "a" =~ m/a/unless 0', 'm', 'unless'],
- ['my $c = 1; my $r; $r = "a" =~ m/a/while $c--', 'm', 'while'],
- ['my $c = 0; my $r; $r = "a" =~ m/a/until $c++', 'm', 'until'],
- ['my $r; $r = "a" =~ m/a/for 1', 'm', 'for'],
- ['my $r; $r = "a" =~ m/a/foreach 1', 'm', 'foreach'],
-
- ['my $t = "a"; my $r = $t =~ s/a//lt 2', 's', 'lt'],
- ['my $t = "a"; my $r = $t =~ s/a//le 1', 's', 'le'],
- ['my $t = "a"; my $r = $t =~ s/a//ne 0', 's', 'ne'],
- ['my $t = "a"; my $r = $t =~ s/a//and 1', 's', 'and'],
- ['my $t = "a"; my $r = $t =~ s/a//unless 0', 's', 'unless'],
-
- ['my $c = 1; my $r; my $t = "a"; $r = $t =~ s/a//while $c--', 's', 'while'],
- ['my $c = 0; my $r; my $t = "a"; $r = $t =~ s/a//until $c++', 's', 'until'],
- ['my $r; my $t = "a"; $r = $t =~ s/a//for 1', 's', 'for'],
- ['my $r; my $t = "a"; $r = $t =~ s/a//for 1', 's', 'foreach'],
- ) {
- my $message = sprintf 'regex (%s) followed by $_->[2]',
- $_->[1] eq 'm' ? 'm//' : 's///';
- my $code = "$_->[0]; 'eval_ok ' . \$r";
- my $result = do {
- no warnings 'syntax';
- eval $code;
- };
- is($@, '', $message);
- is($result, 'eval_ok 1', $message);
- }
- }
-
- {
my $str= "\x{100}";
chop $str;
my $qr= qr/$str/;
* current regex. This routine will set it to any charset modifier found.
* The caller shouldn't change it. This way, another charset modifier
* encountered in the parse can be detected as an error, as we have decided
- * allow only one */
+ * to allow only one */
const char c = **s;
if (! strchr(valid_flags, c)) {
if (isALNUM(c)) {
- goto deprecate;
+ yyerror(Perl_form(aTHX_ "Unknown regexp modifier \"/%c\"", c));
+ (*s)++;
+ return TRUE;
}
return FALSE;
}
case KEEPCOPY_PAT_MOD: *pmfl |= RXf_PMf_KEEPCOPY; break;
case NONDESTRUCT_PAT_MOD: *pmfl |= PMf_NONDESTRUCT; break;
case LOCALE_PAT_MOD:
-
- /* In 5.14, qr//lt is legal but deprecated; the 't' means they
- * can't be regex modifiers.
- * In 5.14, s///le is legal and ambiguous. Try to disambiguate as
- * much as easily done. s///lei, for example, has to mean regex
- * modifiers if it's not an error (as does any word character
- * following the 'e'). Otherwise, we resolve to the backwards-
- * compatible, but less likely 's/// le ...', i.e. as meaning
- * less-than-or-equal. The reason it's not likely is that s//
- * returns a number for code in the field (/r returns a string, but
- * that wasn't added until the 5.13 series), and so '<=' should be
- * used for comparing, not 'le'. */
- if (*((*s) + 1) == 't') {
- goto deprecate;
- }
- else if (*((*s) + 1) == 'e' && ! isALNUM(*((*s) + 2))) {
-
- /* 'e' is valid only for substitutes, s///e. If it is not
- * valid in the current context, then 'm//le' must mean the
- * comparison operator, so use the regular deprecation message.
- */
- if (! strchr(valid_flags, 'e')) {
- goto deprecate;
- }
- Perl_ck_warner_d(aTHX_ packWARN(WARN_AMBIGUOUS),
- "Ambiguous use of 's//le...' resolved as 's// le...'; Rewrite as 's//el' if you meant 'use locale rules and evaluate rhs as an expression'. In Perl 5.18, it will be resolved the other way");
- return FALSE;
- }
if (*charset) {
goto multiple_charsets;
}
*charset = c;
break;
case UNICODE_PAT_MOD:
- /* In 5.14, qr//unless and qr//until are legal but deprecated; the
- * 'n' means they can't be regex modifiers */
- if (*((*s) + 1) == 'n') {
- goto deprecate;
- }
if (*charset) {
goto multiple_charsets;
}
*charset = c;
break;
case ASCII_RESTRICT_PAT_MOD:
- /* In 5.14, qr//and is legal but deprecated; the 'n' means they
- * can't be regex modifiers */
- if (*((*s) + 1) == 'n') {
- goto deprecate;
- }
-
if (! *charset) {
set_regex_charset(pmfl, REGEX_ASCII_RESTRICTED_CHARSET);
}
(*s)++;
return TRUE;
- deprecate:
- Perl_ck_warner_d(aTHX_ packWARN(WARN_SYNTAX),
- "Having no space between pattern and following word is deprecated");
- return FALSE;
-
multiple_charsets:
if (*charset != c) {
yyerror(Perl_form(aTHX_ "Regexp modifiers \"/%c\" and \"/%c\" are mutually exclusive", *charset, c));
https://perl5.git.perl.org / perl5.git / commitdiff