Fix CPAN random tester regressions

src/main/java/org/perlonjava/frontend/parser/NumberParser.java

@@ -548,7 +548,10 @@ public static RuntimeScalar parseNumber(RuntimeScalar runtimeScalar) {
 
     // parseNumber(RuntimeScalar, String) method - with optional operation context for warnings
     public static RuntimeScalar parseNumber(RuntimeScalar runtimeScalar, String operation) {
-        String str = (String) runtimeScalar.value;
+        String str = runtimeScalar.toString();
+        if (str == null) {
+            str = "";
+        }
 
         RuntimeScalar result = numificationCache.get(str);
         if (result != null) {

src/main/java/org/perlonjava/runtime/perlmodule/XMLParserExpat.java

@@ -33,6 +33,8 @@
 public class XMLParserExpat extends PerlModuleBase {
 
     public static final String XS_VERSION = "2.56";
+    private static final Pattern SYSTEM_ID_PATTERN =
+            Pattern.compile("\\bSYSTEM\\s+(['\"])(.*?)\\1", Pattern.DOTALL);
 
     // Namespace separator character (same as expat's NSDELIM = 0xFC)
     private static final char NS_SEP = '\u00FC';
@@ -208,6 +210,9 @@ static class ParserState {
         // Base URI from InputSource for un-resolving SAX systemIds
         String parseBaseUri;
 
+        // Raw DTD system IDs keyed by SAX-resolved equivalents.
+        Map<String, String> rawSystemIds;
+
         // Protocol encoding (e.g. "ISO-8859-1") from ParserCreate
         String protocolEncoding;
 
@@ -715,9 +720,7 @@ public static RuntimeList ParseString(RuntimeArray args, int ctx) {
                     ? xmlString.getBytes(StandardCharsets.ISO_8859_1)
                     : xmlString.getBytes(StandardCharsets.UTF_8);
             xmlBytes = convertEncoding(xmlBytes);
-            state.bytesProcessed = 0;
-            state.inputBytes = xmlBytes;
-            state.inputScanPos = 0;
+            prepareInputBytes(state, xmlBytes);
             doParse(state, new ByteArrayInputStream(xmlBytes));
             return scalarTrue.getList();
         } catch (PerlDieException e) {
@@ -784,9 +787,7 @@ public static RuntimeList ParseStream(RuntimeArray args, int ctx) {
 
             byte[] xmlBytes = baos.toByteArray();
             xmlBytes = convertEncoding(xmlBytes);
-            state.bytesProcessed = 0;
-            state.inputBytes = xmlBytes;
-            state.inputScanPos = 0;
+            prepareInputBytes(state, xmlBytes);
             doParse(state, new ByteArrayInputStream(xmlBytes));
             return scalarTrue.getList();
         } catch (PerlDieException e) {
@@ -840,9 +841,7 @@ public static RuntimeList ParseDone(RuntimeArray args, int ctx) {
                     : xml.getBytes(StandardCharsets.UTF_8);
             xmlBytes = convertEncoding(xmlBytes);
             state.partialIsByteString = false;
-            state.bytesProcessed = 0;
-            state.inputBytes = xmlBytes;
-            state.inputScanPos = 0;
+            prepareInputBytes(state, xmlBytes);
             doParse(state, new ByteArrayInputStream(xmlBytes));
             return scalarTrue.getList();
         } catch (PerlDieException e) {
@@ -998,6 +997,7 @@ private static void doParse(ParserState state, InputStream input) throws Excepti
                         sb.insert(insertPos, doctypeDecl);
                         byte[] newBytes = sb.toString().getBytes(StandardCharsets.UTF_8);
                         state.inputBytes = newBytes;
+                        rememberRawSystemIdsFromInput(state);
                         input = new ByteArrayInputStream(newBytes);
                     }
                 }
@@ -1552,12 +1552,13 @@ public void ignorableWhitespace(char[] ch, int start, int length) throws SAXExce
         @Override
         public void unparsedEntityDecl(String name, String publicId, String systemId,
                                        String notationName) throws SAXException {
+            String rawSysId = unresolveSysId(systemId, state);
+            rememberRawSystemId(rawSysId, state);
             if (state.unparsedHandler != null) {
                 RuntimeArray callArgs = new RuntimeArray();
                 RuntimeArray.push(callArgs, state.selfRef);
                 RuntimeArray.push(callArgs, new RuntimeScalar(name));
                 RuntimeArray.push(callArgs, state.base != null ? new RuntimeScalar(state.base) : scalarUndef);
-                String rawSysId = unresolveSysId(systemId, state);
                 RuntimeArray.push(callArgs, new RuntimeScalar(rawSysId != null ? rawSysId : ""));
                 RuntimeArray.push(callArgs, publicId != null ? new RuntimeScalar(publicId) : scalarUndef);
                 RuntimeArray.push(callArgs, new RuntimeScalar(notationName));
@@ -1574,8 +1575,7 @@ public void unparsedEntityDecl(String name, String publicId, String systemId,
                 RuntimeArray.push(callArgs, state.selfRef);
                 RuntimeArray.push(callArgs, new RuntimeScalar(name));
                 RuntimeArray.push(callArgs, scalarUndef); // val (undef for external entities)
-                String rawSysId2 = unresolveSysId(systemId, state);
-                RuntimeArray.push(callArgs, rawSysId2 != null ? new RuntimeScalar(rawSysId2) : scalarUndef);
+                RuntimeArray.push(callArgs, rawSysId != null ? new RuntimeScalar(rawSysId) : scalarUndef);
                 RuntimeArray.push(callArgs, publicId != null ? new RuntimeScalar(publicId) : scalarUndef);
                 RuntimeArray.push(callArgs, new RuntimeScalar(notationName)); // ndata
                 RuntimeArray.push(callArgs, scalarZero); // is_param
@@ -1590,12 +1590,13 @@ public void unparsedEntityDecl(String name, String publicId, String systemId,
         @Override
         public void notationDecl(String name, String publicId, String systemId)
                 throws SAXException {
+            String rawNotSysId = unresolveSysId(systemId, state);
+            rememberRawSystemId(rawNotSysId, state);
             if (state.notationHandler != null) {
                 RuntimeArray callArgs = new RuntimeArray();
                 RuntimeArray.push(callArgs, state.selfRef);
                 RuntimeArray.push(callArgs, new RuntimeScalar(name));
                 RuntimeArray.push(callArgs, state.base != null ? new RuntimeScalar(state.base) : scalarUndef);
-                String rawNotSysId = unresolveSysId(systemId, state);
                 RuntimeArray.push(callArgs, rawNotSysId != null ? new RuntimeScalar(rawNotSysId) : scalarUndef);
                 RuntimeArray.push(callArgs, publicId != null ? new RuntimeScalar(publicId) : scalarUndef);
                 try {
@@ -1665,6 +1666,7 @@ public void endCDATA() throws SAXException {
 
         @Override
         public void startDTD(String name, String publicId, String systemId) throws SAXException {
+            rememberRawSystemId(systemId, state);
             if (state.doctypeHandler != null) {
                 RuntimeArray callArgs = new RuntimeArray();
                 RuntimeArray.push(callArgs, state.selfRef);
@@ -1732,12 +1734,13 @@ public void internalEntityDecl(String name, String value) throws SAXException {
         @Override
         public void externalEntityDecl(String name, String publicId, String systemId)
                 throws SAXException {
+            String rawExtSysId = unresolveSysId(systemId, state);
+            rememberRawSystemId(rawExtSysId, state);
             if (state.entityDeclHandler != null) {
                 RuntimeArray callArgs = new RuntimeArray();
                 RuntimeArray.push(callArgs, state.selfRef);
                 RuntimeArray.push(callArgs, new RuntimeScalar(name));
                 RuntimeArray.push(callArgs, scalarUndef); // value (external entities have no inline value)
-                String rawExtSysId = unresolveSysId(systemId, state);
                 RuntimeArray.push(callArgs, rawExtSysId != null ? new RuntimeScalar(rawExtSysId) : scalarUndef);
                 RuntimeArray.push(callArgs, publicId != null ? new RuntimeScalar(publicId) : scalarUndef);
                 RuntimeArray.push(callArgs, scalarUndef); // notation
@@ -2135,6 +2138,10 @@ private static String escapeXmlAttr(String value) {
      */
     private static String unresolveSysId(String systemId, ParserState state) {
         if (systemId == null) return null;
+        if (state.rawSystemIds != null) {
+            String raw = state.rawSystemIds.get(systemId);
+            if (raw != null) return raw;
+        }
         // Try to strip the parse base URI that we set on the InputSource
         if (state.parseBaseUri != null && systemId.startsWith(state.parseBaseUri)) {
             return systemId.substring(state.parseBaseUri.length());
@@ -2153,32 +2160,50 @@ private static String unresolveSysId(String systemId, ParserState state) {
                 return systemId.substring(base.length());
             }
         }
-        // Try to strip file:// + CWD prefix to recover relative or absolute file paths
+        // Preserve explicit file: URIs. Expat passes the lexical SYSTEM id through
+        // to Perl callbacks, so `file:///tmp/x` must not become `/tmp/x`.
         if (systemId.startsWith("file:")) {
-            try {
-                String cwd = System.getProperty("user.dir");
-                String filePath;
-                if (systemId.startsWith("file:///")) {
-                    filePath = systemId.substring(7); // file:///path -> /path
-                } else if (systemId.startsWith("file://")) {
-                    filePath = systemId.substring(7); // file://path -> path  
-                } else if (systemId.startsWith("file:/")) {
-                    filePath = systemId.substring(5); // file:/path -> /path
-                } else {
-                    filePath = systemId.substring(5); // file:path -> path
-                }
-                if (cwd != null) {
-                    String cwdWithSlash = cwd.endsWith("/") ? cwd : cwd + "/";
-                    if (filePath.startsWith(cwdWithSlash)) {
-                        return filePath.substring(cwdWithSlash.length());
-                    }
-                }
-                return filePath;
-            } catch (Exception ignored) {}
+            return systemId;
         }
         return systemId;
     }
 
+    private static void rememberRawSystemId(String rawSystemId, ParserState state) {
+        if (rawSystemId == null) return;
+        if (state.rawSystemIds == null) {
+            state.rawSystemIds = new HashMap<>();
+        }
+        state.rawSystemIds.put(rawSystemId, rawSystemId);
+        try {
+            java.net.URI rawUri = new java.net.URI(rawSystemId);
+            if (rawUri.isAbsolute()) {
+                state.rawSystemIds.put(rawUri.toString(), rawSystemId);
+            } else if (rawSystemId.startsWith("/")) {
+                state.rawSystemIds.put(new File(rawSystemId).toURI().toString(), rawSystemId);
+            } else if (state.parseBaseUri != null) {
+                state.rawSystemIds.put(new java.net.URI(state.parseBaseUri).resolve(rawUri).toString(), rawSystemId);
+            }
+        } catch (Exception ignored) {
+        }
+    }
+
+    private static void prepareInputBytes(ParserState state, byte[] xmlBytes) {
+        state.bytesProcessed = 0;
+        state.inputBytes = xmlBytes;
+        state.inputScanPos = 0;
+        state.rawSystemIds = null;
+        rememberRawSystemIdsFromInput(state);
+    }
+
+    private static void rememberRawSystemIdsFromInput(ParserState state) {
+        if (state.inputBytes == null) return;
+        String input = new String(state.inputBytes, StandardCharsets.ISO_8859_1);
+        Matcher matcher = SYSTEM_ID_PATTERN.matcher(input);
+        while (matcher.find()) {
+            rememberRawSystemId(matcher.group(2), state);
+        }
+    }
+
     /**
      * Format an error with line/column info, matching expat error format.
      * SAX error messages are wrapped with "not well-formed (invalid token)"

src/main/perl/lib/CPAN/Config.pm

@@ -42,6 +42,10 @@ sub _bootstrap_prefs {
         'IO-HTML.yml'                => 'PerlOnJava/CpanDistroprefs/IO-HTML.yml',
         'Image-BMP.yml'              => 'PerlOnJava/CpanDistroprefs/Image-BMP.yml',
         'Javascript-Menu-Full.yml'   => 'PerlOnJava/CpanDistroprefs/Javascript-Menu-Full.yml',
+        'CGI-Widget-Tabs.yml'        => 'PerlOnJava/CpanDistroprefs/CGI-Widget-Tabs.yml',
+        'Devel-Symdump.yml'          => 'PerlOnJava/CpanDistroprefs/Devel-Symdump.yml',
+        'Pod-Parser.yml'             => 'PerlOnJava/CpanDistroprefs/Pod-Parser.yml',
+        'Test-Class.yml'             => 'PerlOnJava/CpanDistroprefs/Test-Class.yml',
         'ExtUtils-CBuilder.yml'      => 'PerlOnJava/CpanDistroprefs/ExtUtils-CBuilder.yml',
         'ExtUtils-ParseXS.yml'       => 'PerlOnJava/CpanDistroprefs/ExtUtils-ParseXS.yml',
         'Module-Build.yml'           => 'PerlOnJava/CpanDistroprefs/Module-Build.yml',
@@ -82,6 +86,7 @@ sub _bootstrap_prefs {
         'HTTP-Daemon.yml'            => 'PerlOnJava/CpanDistroprefs/HTTP-Daemon.yml',
         'WWW-RobotRules.yml'         => 'PerlOnJava/CpanDistroprefs/WWW-RobotRules.yml',
         'libwww-perl.yml'            => 'PerlOnJava/CpanDistroprefs/libwww-perl.yml',
+        'LWP-Protocol-https.yml'     => 'PerlOnJava/CpanDistroprefs/LWP-Protocol-https.yml',
         'PerlIO-via-Timeout.yml'     => 'PerlOnJava/CpanDistroprefs/PerlIO-via-Timeout.yml',
     );
     $pref_install{'OpenAI-API.yml'} = $ENV{PERLONJAVA_OPENAI_LIVE_TESTING}
@@ -177,6 +182,8 @@ sub _bootstrap_patches {
           'PerlOnJava/CpanPatches/Image-BMP-1.26/BMP.pm.patch' ],
         [ 'Javascript-Menu-Full-2.02/NoCGIDependency.patch',
           'PerlOnJava/CpanPatches/Javascript-Menu-Full-2.02/NoCGIDependency.patch' ],
+        [ 'CGI-Widget-Tabs-1.14/OptionalAuthorAndCGITests.patch',
+          'PerlOnJava/CpanPatches/CGI-Widget-Tabs-1.14/OptionalAuthorAndCGITests.patch' ],
         [ 'Data-Dmp-0.242/PerlOnJava.patch',
           'PerlOnJava/CpanPatches/Data-Dmp-0.242/PerlOnJava.patch' ],
         [ 'Capture-Tiny-0.50/NoForkTeeCatchErrors.patch',
@@ -187,6 +194,8 @@ sub _bootstrap_patches {
           'PerlOnJava/CpanPatches/Error-Pure-0.34/PlainLexicalConstants.patch' ],
         [ 'String-ShellQuote-1.04/SkipForkScriptTests.patch',
           'PerlOnJava/CpanPatches/String-ShellQuote-1.04/SkipForkScriptTests.patch' ],
+        [ 'LWP-Protocol-https-6.15/SkipForkProxyTest.patch',
+          'PerlOnJava/CpanPatches/LWP-Protocol-https-6.15/SkipForkProxyTest.patch' ],
         [ 'Type-Tiny-2.010001/SkipRegexCallbackTests.patch',
           'PerlOnJava/CpanPatches/Type-Tiny-2.010001/SkipRegexCallbackTests.patch' ],
         [ 'PerlIO-via-Timeout-0.32/SkipViaRuntimeTest.patch',

src/main/perl/lib/PerlOnJava/CpanDistroprefs/CGI-Widget-Tabs.yml

@@ -0,0 +1,16 @@
+---
+comment: |
+  PerlOnJava distroprefs for CGI::Widget::Tabs.
+
+  CGI::Widget::Tabs 1.14 declares Test::Distribution as a build requirement
+  even though t/01_distribution.t already treats it as optional author-test
+  coverage.  Pulling it in adds Pod::Coverage dependencies that are unrelated
+  to installing CGI::Widget::Tabs.
+
+  Its main test also tries to skip itself when neither CGI nor CGI::Minimal is
+  installed, but it emits a fixed 11-test plan first.  Patch the distribution
+  so these optional test dependencies are genuinely optional.
+match:
+  distribution: "^SRSHAH/CGI-Widget-Tabs-"
+patches:
+  - "CGI-Widget-Tabs-1.14/OptionalAuthorAndCGITests.patch"

src/main/perl/lib/PerlOnJava/CpanDistroprefs/Devel-Symdump.yml

@@ -0,0 +1,12 @@
+---
+comment: |
+  PerlOnJava distroprefs for Devel::Symdump.
+
+  Devel::Symdump is a pure-Perl prerequisite of Pod::Coverage.  Its upstream
+  tests assert exact Perl symbol table details, including entries that differ
+  under PerlOnJava's runtime, but the built module is still useful to downstream
+  POD coverage tooling.
+match:
+  distribution: "^ANDK/Devel-Symdump-"
+test:
+  commandline: "PERLONJAVA_TEST_IGNORE_FAILURES"

src/main/perl/lib/PerlOnJava/CpanDistroprefs/Error-Pure.yml

@@ -9,7 +9,7 @@ comment: |
   Allow installation so downstream pure-Perl modules such as Class::Utils can
   test against the built module.
 match:
-  distribution: "^SKIM/Error-Pure-"
+  distribution: "^SKIM/Error-Pure-[0-9]"
 patches:
   - "Error-Pure-0.34/PlainLexicalConstants.patch"
 test:

src/main/perl/lib/PerlOnJava/CpanDistroprefs/LWP-Protocol-https.yml

@@ -0,0 +1,11 @@
+---
+comment: |
+  PerlOnJava distroprefs for LWP::Protocol::https.
+
+  t/https_proxy.t forks a child process to run local proxy/server sockets.
+  PerlOnJava does not implement POSIX fork(), so skip that one test on
+  no-fork platforms while leaving the rest of the upstream suite intact.
+match:
+  distribution: "^OALDERS/LWP-Protocol-https-"
+patches:
+  - "LWP-Protocol-https-6.15/SkipForkProxyTest.patch"

src/main/perl/lib/PerlOnJava/CpanDistroprefs/Pod-Parser.yml

@@ -0,0 +1,12 @@
+---
+comment: |
+  PerlOnJava distroprefs for Pod::Parser.
+
+  Pod::Parser supplies Pod::Find and Pod::Parser for Pod::Coverage.  The 1.67
+  test suite expects File::Find to resolve to a filesystem core-library path,
+  while PerlOnJava loads bundled core modules from jar:PERL5LIB.  Ignore that
+  packaging-path assertion so downstream CPAN distributions can install.
+match:
+  distribution: "^MAREKR/Pod-Parser-"
+test:
+  commandline: "PERLONJAVA_TEST_IGNORE_FAILURES"

src/main/perl/lib/PerlOnJava/CpanDistroprefs/Test-Class.yml

@@ -0,0 +1,13 @@
+---
+comment: |
+  PerlOnJava distroprefs for Test::Class.
+
+  Test::Class is a pure-Perl test helper used by downstream CPAN test-only
+  dependencies such as Test::MockTime::HiRes.  Its upstream suite exercises
+  Test::Builder diagnostic formatting and several attribute/CHECK edge cases
+  that differ under PerlOnJava, while the built module remains usable for the
+  downstream tests that require it.
+match:
+  distribution: "^SZABGAB/Test-Class-"
+test:
+  commandline: "PERLONJAVA_TEST_IGNORE_FAILURES"

src/main/perl/lib/PerlOnJava/CpanPatches/CGI-Widget-Tabs-1.14/OptionalAuthorAndCGITests.patch

@@ -0,0 +1,53 @@
+--- Build.PL.orig
++++ Build.PL
+@@ -5,7 +5,6 @@ Module::Build->new(
+ 		   license     => 'perl',
+ 		   module_name => 'CGI::Widget::Tabs',
+ 		   build_requires => {
+-				      'Test::Distribution'  => 1.14,
+ 				      'Test::More'	    => 0,
+ 				     },
+ 		   requires    => {
+--- META.yml.orig
++++ META.yml
+@@ -9,7 +9,6 @@ requires:
+   HTML::Entities: 0
+   URI::Escape: 0
+ build_requires:
+-  Test::Distribution: 1.14
+   Test::More: 0
+ provides:
+   CGI::Widget::Tabs:
+--- t/02_main.t.orig
++++ t/02_main.t
+@@ -3,11 +3,15 @@
+ use Test;
+ use CGI::Widget::Tabs;
+
+-BEGIN { plan tests => 11 };
+-
+-my $cgi;
+-if ( $cgi = cgi_available() ) {
+-    ok(1); # If we made it this far, we're ok.
++my $cgi = cgi_available();
++unless ($cgi) {
++    print "1..0 # Skip CGI or CGI::Minimal not installed\n";
++    exit 0;
++}
++
++plan tests => 11;
++
++ok(1); # If we made it this far, we're ok.
+
+     # --- First the simple headings
+
+@@ -83,9 +87,7 @@
+                      query    => "t3",
+                      cgi      => $cgi } ),
+         "t3" );
+
+-}
+-
+
+ ############################################################
+