PATCH: [perl #122671] Many warnings in regcomp.c can occur twice

This solves the problem by moving the warnings to be output only in
pass2 of compilation.  The problem arises because almost all of pass1
can be repeated under certain circumstances described in the ticket and
the added comments of this patch.

4 files changed, 126 insertions, 80 deletions

diff --git a/pod/perldelta.pod b/pod/perldelta.pod
index 48e9c58002..016c1bfd48 100644
--- a/pod/perldelta.pod
+++ b/pod/perldelta.pod
@@ -437,6 +437,12 @@ Stub declarations like C<sub f;> and C<sub f ();> no longer wipe out
 constants of the same name declared by C<use constant>.  This bug was
 introduced in perl 5.10.0.
 
+=item *
+
+Under some conditions a warning raised in compilation of regular
+expression patterns could be displayed multiple times.  This is now
+fixed.
+
 =back
 
 =head1 Known Problems
diff --git a/regcomp.c b/regcomp.c
index 1730724dc1..92dc395b50 100644
--- a/regcomp.c
+++ b/regcomp.c
@@ -570,80 +570,85 @@ static const scan_data_t zero_scan_data =
             REPORT_LOCATION_ARGS(offset));         \
 } STMT_END
 
+/* These have asserts in them because of [perl #122671] Many warnings in
+ * regcomp.c can occur twice.  If they get output in pass1 and later in that
+ * pass, the pattern has to be converted to UTF-8 and the pass restarted, they
+ * would get output again.  So they should be output in pass2, and these
+ * asserts make sure new warnings follow that paradigm. */
 
 /* m is not necessarily a "literal string", in this macro */
 #define reg_warn_non_literal_string(loc, m) STMT_START {                \
     const IV offset = loc - RExC_precomp;                               \
-    Perl_warner(aTHX_ packWARN(WARN_REGEXP), "%s" REPORT_LOCATION,      \
+    __ASSERT_(PASS2) Perl_warner(aTHX_ packWARN(WARN_REGEXP), "%s" REPORT_LOCATION,      \
             m, REPORT_LOCATION_ARGS(offset));       \
 } STMT_END
 
 #define	ckWARNreg(loc,m) STMT_START {					\
     const IV offset = loc - RExC_precomp;				\
-    Perl_ck_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION,	\
+    __ASSERT_(PASS2) Perl_ck_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION,	\
 	    REPORT_LOCATION_ARGS(offset));		\
 } STMT_END
 
 #define	vWARN_dep(loc, m) STMT_START {				        \
     const IV offset = loc - RExC_precomp;				\
-    Perl_warner(aTHX_ packWARN(WARN_DEPRECATED), m REPORT_LOCATION,	\
+    __ASSERT_(PASS2) Perl_warner(aTHX_ packWARN(WARN_DEPRECATED), m REPORT_LOCATION,	\
 	    REPORT_LOCATION_ARGS(offset));	        \
 } STMT_END
 
 #define	ckWARNdep(loc,m) STMT_START {				        \
     const IV offset = loc - RExC_precomp;				\
-    Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED),	                \
+    __ASSERT_(PASS2) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED),	                \
 	    m REPORT_LOCATION,						\
 	    REPORT_LOCATION_ARGS(offset));		\
 } STMT_END
 
 #define	ckWARNregdep(loc,m) STMT_START {				\
     const IV offset = loc - RExC_precomp;				\
-    Perl_ck_warner_d(aTHX_ packWARN2(WARN_DEPRECATED, WARN_REGEXP),	\
+    __ASSERT_(PASS2) Perl_ck_warner_d(aTHX_ packWARN2(WARN_DEPRECATED, WARN_REGEXP),	\
 	    m REPORT_LOCATION,						\
 	    REPORT_LOCATION_ARGS(offset));		\
 } STMT_END
 
 #define	ckWARN2reg_d(loc,m, a1) STMT_START {				\
     const IV offset = loc - RExC_precomp;				\
-    Perl_ck_warner_d(aTHX_ packWARN(WARN_REGEXP),			\
+    __ASSERT_(PASS2) Perl_ck_warner_d(aTHX_ packWARN(WARN_REGEXP),			\
 	    m REPORT_LOCATION,						\
 	    a1, REPORT_LOCATION_ARGS(offset));	\
 } STMT_END
 
 #define	ckWARN2reg(loc, m, a1) STMT_START {				\
     const IV offset = loc - RExC_precomp;				\
-    Perl_ck_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION,	\
+    __ASSERT_(PASS2) Perl_ck_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION,	\
 	    a1, REPORT_LOCATION_ARGS(offset));	\
 } STMT_END
 
 #define	vWARN3(loc, m, a1, a2) STMT_START {				\
     const IV offset = loc - RExC_precomp;				\
-    Perl_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION,		\
+    __ASSERT_(PASS2) Perl_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION,		\
 	    a1, a2, REPORT_LOCATION_ARGS(offset));	\
 } STMT_END
 
 #define	ckWARN3reg(loc, m, a1, a2) STMT_START {				\
     const IV offset = loc - RExC_precomp;				\
-    Perl_ck_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION,	\
+    __ASSERT_(PASS2) Perl_ck_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION,	\
 	    a1, a2, REPORT_LOCATION_ARGS(offset));	\
 } STMT_END
 
 #define	vWARN4(loc, m, a1, a2, a3) STMT_START {				\
     const IV offset = loc - RExC_precomp;				\
-    Perl_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION,		\
+    __ASSERT_(PASS2) Perl_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION,		\
 	    a1, a2, a3, REPORT_LOCATION_ARGS(offset)); \
 } STMT_END
 
 #define	ckWARN4reg(loc, m, a1, a2, a3) STMT_START {			\
     const IV offset = loc - RExC_precomp;				\
-    Perl_ck_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION,	\
+    __ASSERT_(PASS2) Perl_ck_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION,	\
 	    a1, a2, a3, REPORT_LOCATION_ARGS(offset)); \
 } STMT_END
 
 #define	vWARN5(loc, m, a1, a2, a3, a4) STMT_START {			\
     const IV offset = loc - RExC_precomp;				\
-    Perl_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION,		\
+    __ASSERT_(PASS2) Perl_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION,		\
 	    a1, a2, a3, a4, REPORT_LOCATION_ARGS(offset)); \
 } STMT_END
 
@@ -9380,7 +9385,7 @@ S_parse_lparen_question_flags(pTHX_ RExC_state_t *pRExC_state)
                 /*NOTREACHED*/
             case ONCE_PAT_MOD: /* 'o' */
             case GLOBAL_PAT_MOD: /* 'g' */
-                if (SIZE_ONLY && ckWARN(WARN_REGEXP)) {
+                if (PASS2 && ckWARN(WARN_REGEXP)) {
                     const I32 wflagbit = *RExC_parse == 'o'
                                          ? WASTED_O
                                          : WASTED_G;
@@ -9400,7 +9405,7 @@ S_parse_lparen_question_flags(pTHX_ RExC_state_t *pRExC_state)
                 break;
 
             case CONTINUE_PAT_MOD: /* 'c' */
-                if (SIZE_ONLY && ckWARN(WARN_REGEXP)) {
+                if (PASS2 && ckWARN(WARN_REGEXP)) {
                     if (! (wastedflags & WASTED_C) ) {
                         wastedflags |= WASTED_GC;
 			/* diag_listed_as: Useless (?-%s) - don't use /%s modifier in regex; marked by <-- HERE in m/%s/ */
@@ -9415,7 +9420,7 @@ S_parse_lparen_question_flags(pTHX_ RExC_state_t *pRExC_state)
                 break;
             case KEEPCOPY_PAT_MOD: /* 'p' */
                 if (flagsp == &negflags) {
-                    if (SIZE_ONLY)
+                    if (PASS2)
                         ckWARNreg(RExC_parse + 1,"Useless use of (?-p)");
                 } else {
                     *flagsp |= RXf_PMf_KEEPCOPY;
@@ -10547,7 +10552,6 @@ S_regpiece(pTHX_ RExC_state_t *pRExC_state, I32 *flagp, U32 depth)
             if (max < min) {    /* If can't match, warn and optimize to fail
                                    unconditionally */
                 if (SIZE_ONLY) {
-                    ckWARNreg(RExC_parse, "Quantifier {n,m} with n > m can't match");
 
                     /* We can't back off the size because we have to reserve
                      * enough space for all the things we are about to throw
@@ -10556,6 +10560,7 @@ S_regpiece(pTHX_ RExC_state_t *pRExC_state, I32 *flagp, U32 depth)
                     RExC_size = PREVOPER(RExC_size) - regarglen[(U8)OPFAIL];
                 }
                 else {
+                    ckWARNreg(RExC_parse, "Quantifier {n,m} with n > m can't match");
                     RExC_emit = orig_emit;
                 }
                 ret = reg_node(pRExC_state, OPFAIL);
@@ -10565,7 +10570,7 @@ S_regpiece(pTHX_ RExC_state_t *pRExC_state, I32 *flagp, U32 depth)
                      && RExC_parse < RExC_end
                      && (*RExC_parse == '?' || *RExC_parse == '+'))
             {
-                if (SIZE_ONLY) {
+                if (PASS2) {
                     ckWARN2reg(RExC_parse + 1,
                                "Useless use of greediness modifier '%c'",
                                *RExC_parse);
@@ -10837,7 +10842,7 @@ S_grok_bslash_N(pTHX_ RExC_state_t *pRExC_state, regnode** node_p,
 	    *node_p = reg_node(pRExC_state,NOTHING);
 	}
         else if (in_char_class) {
-            if (SIZE_ONLY && in_char_class) {
+            if (PASS2 && in_char_class) {
                 if (strict) {
                     RExC_parse++;   /* Position after the "}" */
                     vFAIL("Zero length \\N{}");
@@ -11430,7 +11435,7 @@ tryagain:
 	    ret = reg_node(pRExC_state, CANY);
             RExC_seen |= REG_CANY_SEEN;
 	    *flagp |= HASWIDTH|SIMPLE;
-            if (SIZE_ONLY) {
+            if (PASS2) {
                 ckWARNdep(RExC_parse+1, "\\C is deprecated");
             }
 	    goto finish_meta_pat;
@@ -11930,7 +11935,7 @@ tryagain:
 			    bool valid = grok_bslash_o(&p,
 						       &result,
 						       &error_msg,
-						       TRUE, /* out warnings */
+						       PASS2, /* out warnings */
                                                        FALSE, /* not strict */
                                                        TRUE, /* Output warnings
                                                                 for non-
@@ -11959,7 +11964,7 @@ tryagain:
 			    bool valid = grok_bslash_x(&p,
 						       &result,
 						       &error_msg,
-						       TRUE, /* out warnings */
+						       PASS2, /* out warnings */
                                                        FALSE, /* not strict */
                                                        TRUE, /* Output warnings
                                                                 for non-
@@ -11982,7 +11987,7 @@ tryagain:
 			}
 		    case 'c':
 			p++;
-			ender = grok_bslash_c(*p++, SIZE_ONLY);
+			ender = grok_bslash_c(*p++, PASS2);
 			break;
                     case '8': case '9': /* must be a backreference */
                         --p;
@@ -12021,7 +12026,7 @@ tryagain:
 				REQUIRE_UTF8;
 			    }
 			    p += numlen;
-                            if (SIZE_ONLY   /* like \08, \178 */
+                            if (PASS2   /* like \08, \178 */
                                 && numlen < 3
                                 && p < RExC_end
                                 && isDIGIT(*p) && ckWARN(WARN_REGEXP))
@@ -12038,7 +12043,7 @@ tryagain:
 			if (! RExC_override_recoding) {
 			    SV* enc = PL_encoding;
 			    ender = reg_recode((const char)(U8)ender, &enc);
-			    if (!enc && SIZE_ONLY)
+			    if (!enc && PASS2)
 				ckWARNreg(p, "Invalid escape in the specified encoding");
 			    REQUIRE_UTF8;
 			}
@@ -12772,9 +12777,7 @@ S_handle_regex_sets(pTHX_ RExC_state_t *pRExC_state, SV** return_invlist,
      * upon an unescaped ']' that isn't one ending a regclass.  To do both
      * these things, we need to realize that something preceded by a backslash
      * is escaped, so we have to keep track of backslashes */
-    if (SIZE_ONLY) {
-        UV depth = 0; /* how many nested (?[...]) constructs */
-
+    if (PASS2) {
         Perl_ck_warner_d(aTHX_
             packWARN(WARN_EXPERIMENTAL__REGEX_SETS),
             "The regex_sets feature is experimental" REPORT_LOCATION,
@@ -12782,6 +12785,9 @@ S_handle_regex_sets(pTHX_ RExC_state_t *pRExC_state, SV** return_invlist,
                 UTF8fARG(UTF,
                          RExC_end - RExC_start - (RExC_parse - RExC_precomp),
                          RExC_precomp + (RExC_parse - RExC_precomp)));
+    }
+    else {
+        UV depth = 0; /* how many nested (?[...]) constructs */
 
         while (RExC_parse < RExC_end) {
             SV* current = NULL;
@@ -13282,7 +13288,9 @@ S_add_above_Latin1_folds(pTHX_ RExC_state_t *pRExC_state, const U8 cp, SV** invl
         default:
             /* Use deprecated warning to increase the chances of this being
              * output */
-            ckWARN2reg_d(RExC_parse, "Perl folding rules are not up-to-date for 0x%02X; please use the perlbug utility to report;", cp);
+            if (PASS2) {
+                ckWARN2reg_d(RExC_parse, "Perl folding rules are not up-to-date for 0x%02X; please use the perlbug utility to report;", cp);
+            }
             break;
     }
 }
@@ -13750,8 +13758,8 @@ parseit:
 		    bool valid = grok_bslash_o(&RExC_parse,
 					       &value,
 					       &error_msg,
-                                               SIZE_ONLY,   /* warnings in pass
-                                                               1 only */
+                                               PASS2,   /* warnings only in
+                                                           pass 2 */
                                                strict,
                                                silence_non_portable,
                                                UTF);
@@ -13770,7 +13778,7 @@ parseit:
 		    bool valid = grok_bslash_x(&RExC_parse,
 					       &value,
 					       &error_msg,
-					       TRUE, /* Output warnings */
+					       PASS2, /* Output warnings */
                                                strict,
                                                silence_non_portable,
                                                UTF);
@@ -13782,7 +13790,7 @@ parseit:
 		    goto recode_encoding;
 		break;
 	    case 'c':
-		value = grok_bslash_c(*RExC_parse++, SIZE_ONLY);
+		value = grok_bslash_c(*RExC_parse++, PASS2);
 		break;
 	    case '0': case '1': case '2': case '3': case '4':
 	    case '5': case '6': case '7':
@@ -13822,7 +13830,7 @@ parseit:
                         if (strict) {
                             vFAIL("Invalid escape in the specified encoding");
                         }
-                        else if (SIZE_ONLY) {
+                        else if (PASS2) {
                             ckWARNreg(RExC_parse,
 				  "Invalid escape in the specified encoding");
                         }
diff --git a/t/lib/warnings/regcomp b/t/lib/warnings/regcomp
index b55959e070..f62f5f19bf 100644
--- a/t/lib/warnings/regcomp
+++ b/t/lib/warnings/regcomp
@@ -1,5 +1,6 @@
   regcomp.c	These tests have been moved to t/re/reg_mesg.t
-		except for those that explicitly test line numbers.
+		except for those that explicitly test line numbers
+                and those that don't have a <-- HERE in them.
 
 __END__
 use warnings 'regexp';
@@ -7,3 +8,25 @@ $r=qr/(??{ q"\\b+" })/;
 "a" =~ /a$r/; # warning should come from this line
 EXPECT
 \b+ matches null string many times in regex; marked by <-- HERE in m/\b+ <-- HERE / at - line 3.
+########
+# regcomp.c
+use warnings 'digit' ;
+my $a = qr/\o{1238456}\x{100}/;
+my $a = qr/[\o{6548321}]\x{100}/;
+no warnings 'digit' ;
+my $a = qr/\o{1238456}\x{100}/;
+my $a = qr/[\o{6548321}]\x{100}/;
+EXPECT
+Non-octal character '8'.  Resolved as "\o{123}" at - line 3.
+Non-octal character '8'.  Resolved as "\o{654}" at - line 4.
+########
+# regcomp.c.c
+use warnings;
+$a = qr/\c,/;
+$a = qr/[\c,]/;
+no warnings 'syntax';
+$a = qr/\c,/;
+$a = qr/[\c,]/;
+EXPECT
+"\c," is more clearly written simply as "l" at - line 3.
+"\c," is more clearly written simply as "l" at - line 4.
diff --git a/t/re/reg_mesg.t b/t/re/reg_mesg.t
index 6c8566d879..c03964bb50 100644
--- a/t/re/reg_mesg.t
+++ b/t/re/reg_mesg.t
@@ -320,21 +320,23 @@ push @death, @death_utf8;
 # In the following arrays of warnings, the value can be an array of things to
 # expect.  If the empty string, it means no warning should be raised.
 
-##
-## Key-value pairs of code/error of code that should have non-fatal regexp warnings.
-##
-my @warning = (
-    'm/\b*/' => '\b* matches null string many times {#} m/\b*{#}/',
-    'm/[:blank:]/' => 'POSIX syntax [: :] belongs inside character classes {#} m/[:blank:]{#}/',
 
-    "m'[\\y]'"     => 'Unrecognized escape \y in character class passed through {#} m/[\y{#}]/',
+# Key-value pairs of code/error of code that should have non-fatal regexp
+# warnings.  Most currently have \x{100} appended to them to force them to be
+# upgraded to UTF-8, and the first pass restarted.  Previously this would
+# cause some warnings to be output twice.  This tests that that behavior has
+# been fixed.
 
-    'm/[a-\d]/' => 'False [] range "a-\d" {#} m/[a-\d{#}]/',
-    'm/[\w-x]/' => 'False [] range "\w-" {#} m/[\w-{#}x]/',
-    'm/[a-\pM]/' => 'False [] range "a-\pM" {#} m/[a-\pM{#}]/',
-    'm/[\pM-x]/' => 'False [] range "\pM-" {#} m/[\pM-{#}x]/',
+my @warning = (
+    'm/\b*\x{100}/' => '\b* matches null string many times {#} m/\b*{#}\x{100}/',
+    'm/[:blank:]\x{100}/' => 'POSIX syntax [: :] belongs inside character classes {#} m/[:blank:]{#}\x{100}/',
+    "m'[\\y]\\x{100}'"     => 'Unrecognized escape \y in character class passed through {#} m/[\y{#}]\x{100}/',
+    'm/[a-\d]\x{100}/' => 'False [] range "a-\d" {#} m/[a-\d{#}]\x{100}/',
+    'm/[\w-x]\x{100}/' => 'False [] range "\w-" {#} m/[\w-{#}x]\x{100}/',
+    'm/[a-\pM]\x{100}/' => 'False [] range "a-\pM" {#} m/[a-\pM{#}]\x{100}/',
+    'm/[\pM-x]\x{100}/' => 'False [] range "\pM-" {#} m/[\pM-{#}x]\x{100}/',
     'm/[\N{LATIN CAPITAL LETTER A WITH MACRON AND GRAVE}]/' => 'Using just the first character returned by \N{} in character class {#} m/[\N{U+100{#}.300}]/',
-    "m'\\y'"     => 'Unrecognized escape \y passed through {#} m/\y{#}/',
+    "m'\\y\\x{100}'"     => 'Unrecognized escape \y passed through {#} m/\y{#}\x{100}/',
     '/x{3,1}/'   => 'Quantifier {n,m} with n > m can\'t match {#} m/x{3,1}{#}/',
     '/\08/' => '\'\08\' resolved to \'\o{0}8\' {#} m/\08{#}/',
     '/\018/' => '\'\018\' resolved to \'\o{1}8\' {#} m/\018{#}/',
@@ -343,52 +345,59 @@ my @warning = (
     '/(?=a)*/' => '(?=a)* matches null string many times {#} m/(?=a)*{#}/',
     'my $x = \'\m\'; qr/a$x/' => 'Unrecognized escape \m passed through {#} m/a\m{#}/',
     '/\q/' => 'Unrecognized escape \q passed through {#} m/\q{#}/',
-    '/(?=a){1,3}/' => 'Quantifier unexpected on zero-length expression {#} m/(?=a){1,3}{#}/',
-    '/(a|b)(?=a){3}/' => 'Quantifier unexpected on zero-length expression {#} m/(a|b)(?=a){3}{#}/',
+
+    # Feel free to modify these 2 tests, should they start failing because the
+    # marker of where the problem is becomes wrong.  The current behavior is
+    # bad, always marking at the very end of the regex instead of where the
+    # problem is.  See [perl #122680] regcomp warning gives wrong position of
+    # problem.
+    '/(?=a){1,3}\x{100}/' => 'Quantifier unexpected on zero-length expression {#} m/(?=a){1,3}\x{100}{#}/',
+    '/(a|b)(?=a){3}\x{100}/' => 'Quantifier unexpected on zero-length expression {#} m/(a|b)(?=a){3}\x{100}{#}/',
+
     '/\_/' => "",
     '/[\_\0]/' => "",
     '/[\07]/' => "",
     '/[\006]/' => "",
     '/[\0005]/' => "",
-    '/[\8\9]/' => ['Unrecognized escape \8 in character class passed through {#} m/[\8{#}\9]/',
-                   'Unrecognized escape \9 in character class passed through {#} m/[\8\9{#}]/',
+    '/[\8\9]\x{100}/' => ['Unrecognized escape \8 in character class passed through {#} m/[\8{#}\9]\x{100}/',
+                   'Unrecognized escape \9 in character class passed through {#} m/[\8\9{#}]\x{100}/',
                   ],
-    '/[:alpha:]/' => 'POSIX syntax [: :] belongs inside character classes {#} m/[:alpha:]{#}/',
-    '/[:zog:]/' => 'POSIX syntax [: :] belongs inside character classes {#} m/[:zog:]{#}/',
-    '/[.zog.]/' => 'POSIX syntax [. .] belongs inside character classes {#} m/[.zog.]{#}/',
+    '/[:alpha:]\x{100}/' => 'POSIX syntax [: :] belongs inside character classes {#} m/[:alpha:]{#}\x{100}/',
+    '/[:zog:]\x{100}/' => 'POSIX syntax [: :] belongs inside character classes {#} m/[:zog:]{#}\x{100}/',
+    '/[.zog.]\x{100}/' => 'POSIX syntax [. .] belongs inside character classes {#} m/[.zog.]{#}\x{100}/',
     '/[a-b]/' => "",
-    '/[a-\d]/' => 'False [] range "a-\d" {#} m/[a-\d{#}]/',
-    '/[\d-b]/' => 'False [] range "\d-" {#} m/[\d-{#}b]/',
-    '/[\s-\d]/' => 'False [] range "\s-" {#} m/[\s-{#}\d]/',
-    '/[\d-\s]/' => 'False [] range "\d-" {#} m/[\d-{#}\s]/',
-    '/[a-[:digit:]]/' => 'False [] range "a-[:digit:]" {#} m/[a-[:digit:]{#}]/',
-    '/[[:digit:]-b]/' => 'False [] range "[:digit:]-" {#} m/[[:digit:]-{#}b]/',
-    '/[[:alpha:]-[:digit:]]/' => 'False [] range "[:alpha:]-" {#} m/[[:alpha:]-{#}[:digit:]]/',
-    '/[[:digit:]-[:alpha:]]/' => 'False [] range "[:digit:]-" {#} m/[[:digit:]-{#}[:alpha:]]/',
-    '/[a\zb]/' => 'Unrecognized escape \z in character class passed through {#} m/[a\z{#}b]/',
-    '/(?c)/' => 'Useless (?c) - use /gc modifier {#} m/(?c{#})/',
-    '/(?-c)/' => 'Useless (?-c) - don\'t use /gc modifier {#} m/(?-c{#})/',
-    '/(?g)/' => 'Useless (?g) - use /g modifier {#} m/(?g{#})/',
-    '/(?-g)/' => 'Useless (?-g) - don\'t use /g modifier {#} m/(?-g{#})/',
-    '/(?o)/' => 'Useless (?o) - use /o modifier {#} m/(?o{#})/',
-    '/(?-o)/' => 'Useless (?-o) - don\'t use /o modifier {#} m/(?-o{#})/',
-    '/(?g-o)/' => [ 'Useless (?g) - use /g modifier {#} m/(?g{#}-o)/',
-                    'Useless (?-o) - don\'t use /o modifier {#} m/(?g-o{#})/',
+    '/[a-\d]\x{100}/' => 'False [] range "a-\d" {#} m/[a-\d{#}]\x{100}/',
+    '/[\d-b]\x{100}/' => 'False [] range "\d-" {#} m/[\d-{#}b]\x{100}/',
+    '/[\s-\d]\x{100}/' => 'False [] range "\s-" {#} m/[\s-{#}\d]\x{100}/',
+    '/[\d-\s]\x{100}/' => 'False [] range "\d-" {#} m/[\d-{#}\s]\x{100}/',
+    '/[a-[:digit:]]\x{100}/' => 'False [] range "a-[:digit:]" {#} m/[a-[:digit:]{#}]\x{100}/',
+    '/[[:digit:]-b]\x{100}/' => 'False [] range "[:digit:]-" {#} m/[[:digit:]-{#}b]\x{100}/',
+    '/[[:alpha:]-[:digit:]]\x{100}/' => 'False [] range "[:alpha:]-" {#} m/[[:alpha:]-{#}[:digit:]]\x{100}/',
+    '/[[:digit:]-[:alpha:]]\x{100}/' => 'False [] range "[:digit:]-" {#} m/[[:digit:]-{#}[:alpha:]]\x{100}/',
+    '/[a\zb]\x{100}/' => 'Unrecognized escape \z in character class passed through {#} m/[a\z{#}b]\x{100}/',
+    '/(?c)\x{100}/' => 'Useless (?c) - use /gc modifier {#} m/(?c{#})\x{100}/',
+    '/(?-c)\x{100}/' => 'Useless (?-c) - don\'t use /gc modifier {#} m/(?-c{#})\x{100}/',
+    '/(?g)\x{100}/' => 'Useless (?g) - use /g modifier {#} m/(?g{#})\x{100}/',
+    '/(?-g)\x{100}/' => 'Useless (?-g) - don\'t use /g modifier {#} m/(?-g{#})\x{100}/',
+    '/(?o)\x{100}/' => 'Useless (?o) - use /o modifier {#} m/(?o{#})\x{100}/',
+    '/(?-o)\x{100}/' => 'Useless (?-o) - don\'t use /o modifier {#} m/(?-o{#})\x{100}/',
+    '/(?g-o)\x{100}/' => [ 'Useless (?g) - use /g modifier {#} m/(?g{#}-o)\x{100}/',
+                    'Useless (?-o) - don\'t use /o modifier {#} m/(?g-o{#})\x{100}/',
                   ],
-    '/(?g-c)/' => [ 'Useless (?g) - use /g modifier {#} m/(?g{#}-c)/',
-                    'Useless (?-c) - don\'t use /gc modifier {#} m/(?g-c{#})/',
+    '/(?g-c)\x{100}/' => [ 'Useless (?g) - use /g modifier {#} m/(?g{#}-c)\x{100}/',
+                    'Useless (?-c) - don\'t use /gc modifier {#} m/(?g-c{#})\x{100}/',
                   ],
       # (?c) means (?g) error won't be thrown
-     '/(?o-cg)/' => [ 'Useless (?o) - use /o modifier {#} m/(?o{#}-cg)/',
-                      'Useless (?-c) - don\'t use /gc modifier {#} m/(?o-c{#}g)/',
+     '/(?o-cg)\x{100}/' => [ 'Useless (?o) - use /o modifier {#} m/(?o{#}-cg)\x{100}/',
+                      'Useless (?-c) - don\'t use /gc modifier {#} m/(?o-c{#}g)\x{100}/',
                     ],
-    '/(?ogc)/' => [ 'Useless (?o) - use /o modifier {#} m/(?o{#}gc)/',
-                    'Useless (?g) - use /g modifier {#} m/(?og{#}c)/',
-                    'Useless (?c) - use /gc modifier {#} m/(?ogc{#})/',
+    '/(?ogc)\x{100}/' => [ 'Useless (?o) - use /o modifier {#} m/(?o{#}gc)\x{100}/',
+                    'Useless (?g) - use /g modifier {#} m/(?og{#}c)\x{100}/',
+                    'Useless (?c) - use /gc modifier {#} m/(?ogc{#})\x{100}/',
                   ],
-    '/a{1,1}?/' => 'Useless use of greediness modifier \'?\' {#} m/a{1,1}?{#}/',
-    '/b{3}  +/x' => 'Useless use of greediness modifier \'+\' {#} m/b{3}  +{#}/',
-);
+    '/a{1,1}?\x{100}/' => 'Useless use of greediness modifier \'?\' {#} m/a{1,1}?{#}\x{100}/',
+    '/b{3}  +\x{100}/x' => 'Useless use of greediness modifier \'+\' {#} m/b{3}  +{#}\x{100}/',
+); # See comments before this for why '\x{100}' is generally needed
 
 my @warnings_utf8 = mark_as_utf8(
     'm/ネ\b*ネ/' => '\b* matches null string many times {#} m/ネ\b*{#}ネ/',