fix: preserve Unicode regex interpolation provenance

docs/about/changelog.md

@@ -6,6 +6,7 @@ Release history of PerlOnJava. See [Roadmap](roadmap.md) for future plans.
 
 - Fix localization of numbered regex captures.
 - Fix IO-handle type checks and uninitialized-value warning locations.
+- Fix Unicode regex interpolation in substitutions.
 - Fix numeric-zero results from failed `s///` substitutions.
 - Bundle the complete CPAN `File::Path` 2.18 implementation, including modern
   `rmtree`/`remove_tree` options such as `keep_root`, `error`, `result`,

src/main/java/org/perlonjava/runtime/regex/RuntimeRegexTemplate.java

@@ -136,7 +136,12 @@ private static boolean isByteCompatiblePatternPart(RuntimeScalar scalar) {
     }
 
     static RuntimeScalar patternScalar(String pattern, boolean byteBackedPattern) {
-        return byteBackedPattern
+        // A byte-backed scalar cannot represent characters outside the Latin-1
+        // byte range. This matters when a Unicode qr// is stringified and
+        // embedded in an interpolated substitution.
+        boolean byteCompatible = byteBackedPattern
+                && pattern.chars().allMatch(ch -> ch <= 0xFF);
+        return byteCompatible
                 ? new RuntimeScalar(pattern.getBytes(StandardCharsets.ISO_8859_1))
                 : new RuntimeScalar(pattern);
     }

src/test/resources/unit/regex/unicode_qr_substitution_template.t

@@ -0,0 +1,15 @@
+use strict;
+use warnings;
+use utf8;
+use Test::More tests => 2;
+
+my $tld = qr/(?:com|みんな|中国|테스트|भारत|москва)/i;
+
+my $text = "Visit site.みんな today";
+$text =~ s{ (.*?) ($tld) | (.+?)$ }{defined $1 ? $1 : $3}gsex;
+is($text, 'Visit site. today', 'substitution compiles an interpolated Unicode qr');
+
+my $ascii = qr/(?:com|org)/i;
+my $ascii_text = 'Visit site.com today';
+$ascii_text =~ s{ (.*?) ($ascii) | (.+?)$ }{defined $1 ? $1 : $3}gsex;
+is($ascii_text, 'Visit site. today', 'ASCII qr interpolation remains compatible');