Fix bug with UTF-8 pattern group name; and document the rules

pod/perldiag.pod

@@ -2866,9 +2866,19 @@ has since been undefined.
 =item Group name must start with a non-digit word character in regex; marked by 
 S<<-- HERE> in m/%s/
 
-(F) Group names must follow the rules for perl identifiers, meaning
-they must start with a non-digit word character.  A common cause of
-this error is using (?&0) instead of (?0).  See L<perlre>.
+(F) Group names must follow the rules for Perl identifiers, meaning
+they must start with a character that matches C<\p{XID_Start}> plus the
+underscore.  This means the first character may not be a digit.
+Subsequent characters must match C<\p{XID_Continue}>.
+
+A common cause of this error is using (?&0) instead of (?0).
+
+This message was formulated before Perl supported Unicode; so it is
+not accurate for Unicode characters outside the ASCII-range.  There are
+many word characters in Unicode that may not start a group name, and a
+few that may not be a continuation character.
+
+See L<perlre>.
 
 =item ()-group starts with a count
 

regcomp.c

@@ -2533,8 +2533,8 @@ S_reg_scan_name(pTHX_ RExC_state_t *pRExC_state, U32 flags)
                  && (advance = isWORDCHAR_utf8_safe( (U8 *) RExC_parse,
                                                      (U8 *) RExC_end)));
     } else {
-        RExC_parse_inc_by(1); /* so the <- from the vFAIL is after the offending
-                         character */
+        /* so the <- from the vFAIL is after the offending character */
+        RExC_parse_inc_safe();
         vFAIL("Group name must start with a non-digit word character");
     }
     sv_name = newSVpvn_flags(name_start, (int)(RExC_parse - name_start),

t/re/reg_mesg.t

@@ -176,7 +176,7 @@ my @death =
  '/(?/' => 'Sequence (? incomplete {#} m/(?{#}/',
 
  '/(?;x/' => 'Sequence (?;...) not recognized {#} m/(?;{#}x/',
- '/(?<;x/' => 'Group name must start with a non-digit word character {#} m/(?<;{#}x/',
+ '/(?<;name>match)/' => 'Group name must start with a non-digit word character {#} m/(?<;{#}name>match)/',
  '/(?\ix/' => 'Sequence (?\...) not recognized {#} m/(?\{#}ix/',
  '/(?\mx/' => 'Sequence (?\...) not recognized {#} m/(?\{#}mx/',
  '/(?\:x/' => 'Sequence (?\...) not recognized {#} m/(?\{#}:x/',
@@ -222,7 +222,7 @@ my @death =
  '/\g{-abc}/' => 'Group name must start with a non-digit word character {#} m/\g{-{#}abc}/',
  '/\g{1-1}/'  => 'Sequence \g{... not terminated {#} m/\g{1{#}-1}/',
  '/\g{ -1 foo }/'  => 'Sequence \g{... not terminated {#} m/\g{ -1 {#}foo }/',
- '/(?<;x/' => 'Group name must start with a non-digit word character {#} m/(?<;{#}x/',
+ '/(?<;name>match)/' => 'Group name must start with a non-digit word character {#} m/(?<;{#}name>match)/',
 
  'my $m = "\\\"; $m =~ $m', => 'Trailing \ in regex m/\/',
 
@@ -310,7 +310,7 @@ my @death =
  'm/(?&a/'  => 'Sequence (?&... not terminated {#} m/(?&a{#}/',
  'm/(?P=/' => 'Sequence ?P=... not terminated {#} m/(?P={#}/',
  "m/(?'/"  => "Sequence (?'... not terminated {#} m/(?'{#}/",
- "m/(?</"  => "Sequence (?<... not terminated {#} m/(?<{#}/",
+ "m/(?<name)/"  => "Sequence (?<... not terminated {#} m/(?<name{#})/",
  'm/(?&/'  => 'Sequence (?&... not terminated {#} m/(?&{#}/',
  'm/(?(</' => 'Sequence (?(<... not terminated {#} m/(?(<{#}/',
  "m/(?('/" => "Sequence (?('... not terminated {#} m/(?('{#}/",
@@ -485,7 +485,7 @@ my @death_utf8 = mark_as_utf8(
  '/ネ(?/' => 'Sequence (? incomplete {#} m/ネ(?{#}/',
 
  '/ネ(?;ネ/' => 'Sequence (?;...) not recognized {#} m/ネ(?;{#}ネ/',
- '/ネ(?<;ネ/' => 'Group name must start with a non-digit word character {#} m/ネ(?<;{#}ネ/',
+ '/ネ(?<;name>match)ネ/' => 'Group name must start with a non-digit word character {#} m/ネ(?<;{#}name>match)ネ/',
  '/ネ(?\ixネ/' => 'Sequence (?\...) not recognized {#} m/ネ(?\{#}ixネ/',
  '/ネ(?^lu:ネ)/' => 'Regexp modifiers "l" and "u" are mutually exclusive {#} m/ネ(?^lu{#}:ネ)/',
 '/ネ(?lil:ネ)/' => 'Regexp modifier "l" may not appear twice {#} m/ネ(?lil{#}:ネ)/',
@@ -547,6 +547,7 @@ my @death_utf8 = mark_as_utf8(
  '/[\cネ]/' => "Character following \"\\c\" must be printable ASCII {#} m/[\\cネ{#}]/",
  '/\b{ネ}/' => "'ネ' is an unknown bound type {#} m/\\b{ネ{#}}/",
  '/\B{ネ}/' => "'ネ' is an unknown bound type {#} m/\\B{ネ{#}}/",
+ '/ネ(?<‿name>match)ネ/; #no latin1' => 'Group name must start with a non-digit word character {#} m/ネ(?<‿{#}name>match)ネ/',
 );
 push @death, @death_utf8;