Fixed MethodCallExprContext generic parameter issue. [NullType must not fail matchTypeParameters]

[zcbbpo]

Hi ,

It's seem paser generic parameter will throws java.lang.UnsupportedOperationException: com.github.javaparser.symbolsolver.model.typesystem.NullType

code:

public class ISSUES_Generic_Parameter {

    public static Object test() {
        return Foo.error("code", null, "arg1");
    }

    public static class Foo<T> {
        public T data;
        public static <T> Foo<T> error(String resultCode, T data, Object... params) {
            return null;
        }
    }

}

I will modified com.github.javaparser.symbolsolver.javaparsermodel.contexts.MethodCallExprContext#matchTypeParameters

if (type.isNull()) {
    type = MyObjectProvider.INSTANCE.object();
}

And add com.github.javaparser.symbolsolver.resolution.javaparser.contexts.MethodCallExprContextResolutionTest#testGenericParameter for test.

Please check it.

[jlerbsc]

I don't understand what this PR is trying to solve. The test case below does not throw an exception. So what are you doing to meet your problem?

void test() {
    String code = 
            "public class ISSUES_Generic_Parameter {\n" + 
            "\n" + 
            "    public static Object test() {\n" + 
            "        return Foo.error(\"code\", null, \"arg1\");\n" + 
            "    }\n" + 
            "\n" + 
            "    public static class Foo<T> {\n" + 
            "        public T data;\n" + 
            "        public static <T> Foo<T> error(String resultCode, T data, Object... params) {\n" + 
            "            return null;\n" + 
            "        }\n" + 
            "    }\n" + 
            "\n" + 
            "}";
    
    ParserConfiguration config = new ParserConfiguration();
    config.setSymbolResolver(new JavaSymbolSolver(new ReflectionTypeSolver(false)));
    StaticJavaParser.setConfiguration(config);

    CompilationUnit cu = StaticJavaParser.parse(code);
    Set<MethodCallExpr> methodCallExpr = new HashSet<>(cu.findAll(MethodCallExpr.class));
    for (MethodCallExpr expr : methodCallExpr) {
        ResolvedMethodDeclaration rd = expr.resolve();
        System.out.println("\t Solved : " + rd.getQualifiedSignature());
    }
}

result is

Solved : ISSUES_Generic_Parameter.Foo.error(java.lang.String, T, java.lang.Object...)

[zcbbpo]

Hi @jlerbsc
Please try this:

@Test
    public void test() throws Exception {
        String code =
                "public class ISSUES_Generic_Parameter {\n" +
                        "\n" +
                        "    public static Object test() {\n" +
                        "        return Foo.error(\"code\", null, \"arg1\");\n" +
                        "    }\n" +
                        "\n" +
                        "    public static class Foo<T> {\n" +
                        "        public T data;\n" +
                        "        public static <T> Foo<T> error(String resultCode, T data, Object... params) {\n" +
                        "            return null;\n" +
                        "        }\n" +
                        "    }\n" +
                        "\n" +
                        "}";

        ParserConfiguration config = new ParserConfiguration();
        config.setSymbolResolver(new JavaSymbolSolver(new ReflectionTypeSolver(false)));
        StaticJavaParser.setConfiguration(config);

        CompilationUnit cu = StaticJavaParser.parse(code);
        cu.findAll(ReturnStmt.class).forEach(returnStmt -> {
            returnStmt.getExpression().ifPresent(expression -> {
                ResolvedType resolvedType = expression.calculateResolvedType();
                System.out.println(expression);
                System.out.println(resolvedType);
            });
        });

    }

[jlerbsc]

Hi @MysterAitch I think it's a correct patch in my opinion it can be merged. In this case null type could be assigned to Object type. JLS 5.2 assignment contexts says "a value of the null type (the null reference is the only such value) may be assigned to any reference type, resulting in a null reference of that type."
Perhaps the PR can be modified to remove the changes that are not needed and the test case can be better integrated with existing tests or it would be necessary to create a specific one for this issue.

[MysterAitch]

@zcbbpo -- thank you for the PR!

Hi @MysterAitch I think it's a correct patch in my opinion it can be merged.

I agree! :)

In this case null type could be assigned to Object type. JLS 5.2 assignment contexts says "a value of the null type (the null reference is the only such value) may be assigned to any reference type, resulting in a null reference of that type."

Good spot! 👍

I've just added a comment to directly cross-reference to the JLS.

Perhaps the PR can be modified to remove the changes that are not needed

@jlerbsc what did you have in mind? I've no objections to merging 👍

[jlerbsc]

@jlerbsc what did you have in mind?

Nothing too bad there was just a commit that did not do what we expected from the patch. commit 3c5ba3c on MethodCallExprContext which is corrected in commit b4e3731

[MysterAitch]

Perhaps [...] the test case can be better integrated with existing tests or it would be necessary to create a specific one for this issue.

I'll merge this now and it can be improved upon separately