thaisamirely / javaparser Goto Github PK

What steps will reproduce the problem?

1. The release is two year old
2. Don't answer the issues
3. Not enough documentacion

At the moment, AST nodes that have a list of children return null when
there were no child nodes in the input.  This makes processing the tree
awkward.  Processing code is full of pointless null checks.

Instead, the nodes should return an empty list when there are no child nodes.

I think this can be easily implemented by initialising list instance
variables to Collections.emptyList().

My open-source project java2objc http://code.google.com/p/java2objc/ depends on 
javaparser and is currently stuck at javaparser 1.0.1 which is available in 
Maven central repository. Please update it there so that I can upgrade to the 
latest one.

Japa does not seem to parse arrays of the form

Type id[]

correctly. It seems to discard the array brackets and just returns the 
AST  for "Type id".

E.g. in class java.lang.StringBuffer there is the following method:

public synchronized StringBuffer insert(int index, char str[], int 
offset, int len)

The "char str[]" is parsed as a simple PrimitiveType(type=char), not as a 
ReferenceType(arraycount=1,type=PrimitiveType(type=char)), as it should 
be.

I guess this case (array brackets after, not in front of the identifier) 
is just missing from java_1_5.jj and I would appreciate it if you could 
add support for these arrays as they are even - as you can see above - 
used in the standard Java libraries...

Thank you very much!
Martin

Hello. I hope I am posting this in the right place. :)

I was unable to find a method to retrieve the fully qualified name of a type. 
As an example, if I look at a field (for instance, String gF), I am unable
to retrieve 'java.lang.String' as a type-name for the field. Only the
type-name 'String'.

Does this feature exist? If not, is it possible to add this feature?

Thank you.

What steps will reproduce the problem?
1. parse the following code:
fooObject.
doStuff(v1, 
v2)
2. parse the code
3. try to determine the line and char position of the method (not the scope)

What is the expected output? What do you see instead?
I need to be able to tell the exact position of the method name, which is 
possible by adding the length of the scope to the position of the scope (plus 
one for the dot) as long as they are both on the same line.
If they aren't there is no way of telling where it is because it is only stored 
as a string

What version of the product are you using? On what operating system?
version 1.0.8 of the JavaParser
on different unix based systems (ubuntu 11.04 and ArchLinux)

Please provide any additional information below.
-

ModifierSet looks like this:

    public static final int PUBLIC = 1;
    public static final int PRIVATE = 2;
    public static final int PROTECTED = 4;
    public static final int STATIC = 8;
    public static final int FINAL = 16;
    public static final int SYNCHRONIZED = 32;
    public static final int VOLATILE = 64;
    public static final int TRANSIENT = 128;
    public static final int NATIVE = 256;
    public static final int ABSTRACT = 1024;
    public static final int STRICTFP = 2048;.

Ordinal of those values was different: public, protected, private...
Also why is suddenly ABSTRACT = 1024 and not 512?

I WONT TO KNOW HOW TO BUILD AST FORM A SOURCECODE JAVA WITH JAVAPARSER

Hi, I'm trying to use this grammar to develop some code completion
functionalities and I have run into a problem that I think you already know.

To obtain the completion options I need to know exactly the position of the
cursor by relation to the node I want to complete. I have tried with the
n.getEndLine() and n.getEndColumn() methods but they return the same value
than n.getBeginLine() and n.getBeginColumn(). I think you already know the
problem since this constructor:
    @Deprecated
    public Node(int line, int column) {

is deprecated.

Also, I have been using javacc for a while but I'm still a super newbie. I
haven't been able to generate the code from the grammar. If you help me to
do this I could change the references to the @Deprecated constructor and
send you a "diff"...

Thanks for you grammar, anyway.
Fernando.

for the declaration

static class myclass {  /**/}
the modifiers can't detect if the class is private or public (this apply 
also for any other TypeDeclaration)

so, the following statements always returns false:

ModifierSet.isPrivate(classNode.getModifiers())

been classNode the node representing the class definition above.
it also returns false for any other test (isPublic, isProtected, etc.)
Regards

Javadoc entire source code.

I faced some problems when performing simple semantic analysis.
Obviously I need to indicate line and column position of error
in parsed source file. But it seems it’s impossible to determine
position of class or method modifiers : "public", "private"...

Ideally, I need to know position of all tokens inside all AST nodes.
For example it would be good to be able to write analog of DumpVisitor -
ExactDumpVisitor that prints source that is not different from original
(in the sense that all tokens have exactly same positions as in original 
file).

Hi!

What steps will reproduce the problem?

1. I use javacc 4.1 and java_1_5.jj for creating parser.
2. then compile it.

What is the expected output?

All files of the source code must be compiled.

What do you see instead?

Not all...
For instance, Token.java does not contain now "public static class GTToken".
From for this ASTParserTokenManager.java is compiled with error.

What version of the product are you using?

1.0.3

Please provide any additional information below.

I think that it is necessary to correct java_1_5.jj.

I've created a new ConstructorDeclaration with the constructor
ConstructorDeclaration(int modifiers, String name)

I think it should create an empty constructor. But it's block field is not
initialized so it crashes with NullPointerException in VoidVisitorAdapter.

My workaround for the problem:
 public ConstructorDeclaration(int modifiers, String name) {
        this.modifiers = modifiers;
        this.name = name;
        block=new BlockStmt(null);//Added line
    }

I think the same should be done with MethodDeclaration

Would you please compile Java Parser using Java 5.

We are using Java 5 for Spring Roo (http://www.springsource.org/roo) and
would appreciate it if Java Parser could be released with Java 5 compiled
bytecode. We have needed to recompile Java Parser 1.0.7 with Java 5 so that
we can provide our users with Java 5 support, so we can confirm there are
no compilation errors against a Java 5 compilation.

Thank you very much for your assistance on this.

When recognizing an expression like the following the AST contains a
ClassCastExpr of type "foo".

int a = (foo)+bar;

I don't know if this is fixable in the AST parser because it needs semantic
knowledge. And its not that easy to check if "foo" is REALLY no type in
current scope. This behavior should at least be documented.

This is probably too much work, but it would be grand if it were possible to 
visit an AST and write a class file. I know one can use ASM to achieve a 
similar result but it works on much lower constructs namely byte codes, where 
as JP is one layer above working on java source constructs. 

This would make it possible to generate class files from source in text form 
which would be really cool :)

What steps will reproduce the problem?
For "this":
run this code:
    FieldDeclaration field = new FieldDeclaration(
                                      ModifierSet.PRIVATE, 
                                      new
PrimitiveType(PrimitiveType.Primitive.Boolean), 
                                      new VariableDeclarator(new
VariableDeclaratorId("isWorking")));

    MethodDeclaration getter = new MethodDeclaration(
                                     ModifierSet.PUBLIC, 
                                     field.getType(), 
                                     "get" + /*StringUtils.capitalize(*/
field.getVariables().get(0).getId().getName()/*)*/);

    BlockStmt block = new BlockStmt();
    ReturnStmt ret = new ReturnStmt(new ThisExpr(
             new NameExpr(field.getVariables().get(0).getId().getName())));
//ReturnStmt ret = new ReturnStmt(new SuperExpr(new
NameExpr(field.getVariables().get(0).getId().getName())));

    List<Statement> statements = new ArrayList<Statement>();
    statements.add(ret);
    block.setStmts(statements);
    getter.setBody(block);
    System.out.println(getter);

What is the expected output? What do you see instead?
Expected output:
public boolean getIsWorking() {
    return this.isWorking;
}

What's produced is:
public boolean getIsWorking() {
    return isWorking.this;
}
What version of the product are you using? On what operating system?
Latest. Mac OSX


Please provide any additional information below.
I played with the DumpVisitor class to get it fixed.. but I dont know if I
broke something else.. the unit tests failed for some reason..

I just found it very annoying when I am constructing a new method that I
have to create a list of parameters and then pass it the constructor 

Here is an example: 
I want to create a setter method so I go like this:
Parameter p = new Parameter(field.getType(), new VariableDeclaratorId("id"));

List<Parameter> ps = new ArrayList<Parameter>();
ps.add(p);

MethodDeclaration setter = new MethodDeclaration(ModifierSet.PUBLIC, 
                                field.getType(), "setId", ps);


Instead I want to be able to do all that one line:
MethodDeclaration setter = new MethodDeclaration(ModifierSet.PUBLIC, 
                                field.getType(), "setId", new
Parameter(field.getType(), new VariableDeclaratorId("id"));


it simplify things a lot ..
So the signature of the constructor I am suggesting is this:
public MethodDeclaration(int modifiers, Type type, String name,
Parameter... parameters);

and the complete implementation would be something like this in the
MethodDeclaration class

public MethodDeclaration(int modifiers, Type type, String name,
Parameter... parameters) {
        this.modifiers = modifiers;
        this.type = type;
        this.name = name;
        this.parameters = Arrays.asList(parameters);
    }

 Thanks

Is it possible to setup a git mirror of this project either on google code or 
preferrably on github?

Hi All,

What is the expected output? What do you see instead?
output :- /** This is Block Comments Print */

Sample Code which i have added in my code:- 

comments.add(new BlockComment("I Have added this program to print 
comment")); 
cu.setComments(comments);
System.out.println(cu.toString);         

This is printing empty.
What version of the product are you using? On what operating system?
windows

Please provide any additional information below.

I want to replace ReferenceTypes of a given type. If there would be an
equal method I could easily find out if the type matches.

non of the classes overrides the equal method

When using the JavaParser.parse() method to get a single CompilationUnit it 
works fine.  The problem arises when I attempt to have three separate 
CompilationUnits all from the same InputStream (that do not share a object 
reference). 

Only the first CompilationUnit has any data, the others are just empty 
CompilationUnits (not null though), it took me awhile to figure out, but 
what I think what is happening is that the parse method is not resetting 
back to the start of the file again at the end of its parsing, this meant 
that subsequent uses or JavaParser.parse() saw that they where already at 
the end of the file and didn't actually parse anything.

I am using a ugly solution where I close and re-construct my 
FileInputStream after every call to JavaParser.parse().  It would be nice 
if the parse method just returns to the start of the file after finishing 
its work, or even if there was a JavaParser.reParse() method or something.

Cheers, Bryce
P.S. Awesome project by the way. 

This class should extends Exception and not the Error class.

>public class TokenMgrError extends Error

I was trying with a different charset in a loop with exceptions. But don't 
work, since the TokenMgrError class extends from Error.

I donwload the sources changed and now works fine.

I'm trying to conduct some static analysis on top of javaparser, and find out 
that javaparser cannot distinguish long qualified name and long field access 
expressions.
 for example,
 for two expressions,

 Pa.Pb.Pc.D.fun();  //where Pa,Pb and Pc are package names, D is a class name 

 a.b.c.fun();//where a,b,c are variable names.

javaparser recognizes these two expressions' prefixes(Pa.Pb.Pc.D and a.b.c.) as 
field access expressions. When I get the semantic tree, I cannot make sure 
whether fun() is invoked on an object or just a static method from some class. 
I am wondering maybe this issue arises because the fact javaparser does not 
analysis symbols in a code snippet, so that it cannot distinguish package names 
from variable names?

  thank you for reading!

sample class:
public class FOO {
        // Line 1
        // Line 2
}       

cu.getComments() only contains '// Line 2'

Dear Team,

I am developing codesearch tool. In this tool, I am using your
javaparser1.0.5 for parsing java file and indexing using lucene. Without
using javaparser, i could index huge data like 500mb. But after i integrate
javaparser1.0.5, it will break while  the index size is near to 25mb. 

 Is there any issue in javaparser1.0.5 while parsing huge data. Is any
streams are not closing, which is open to read the data???????

Kindly reply me with the proper solution.

Thanks & Regards,
Muralidharan P.

japa.parser.ast.body.ModifierSet and java.lang.reflect.Modifier use
different constants for the same modifiers.
Just for matter of convenience please use the following constants:

    public static final int PUBLIC = Modifier.PUBLIC;

    public static final int PROTECTED = Modifier.PROTECTED;

    public static final int PRIVATE = Modifier.PRIVATE;

    public static final int ABSTRACT = Modifier.ABSTRACT;

    public static final int STATIC = Modifier.STATIC;

    public static final int FINAL = Modifier.FINAL;

    public static final int SYNCHRONIZED = Modifier.SYNCHRONIZED;

    public static final int NATIVE = Modifier.NATIVE;

    public static final int TRANSIENT = Modifier.TRANSIENT;

    public static final int VOLATILE = Modifier.VOLATILE;

    public static final int STRICTFP = Modifier.STRICT;


Thanks for the great job on the AST and parser!

What steps will reproduce the problem?
1. Compile DumperTestClass with any Java 6 compiler (Eclipse 3.6 JDT will do)
2. Observe the 'cyclic annotation' error

Eclipse JDT compiler error message:
Cycle detected: the annotation type DumperTestClass<T,X>.AnnotationTest cannot 
contain attributes of the annotation type itself

Cyclic annotations are indeed illegal (the new JDT has a testcase for that), 
but older compilers did not catch that problem. This prevents compiling and 
deploying the latest trunk version with Maven.

This is not a bug - but a request.

It would be nice, if parser also parsed javadoc into structured data.
The extension could be as simple as extending the JavadocComment class with
two methods:
1) public String getDescription()
   returns the description part of the javadoc
2) public Map<String, String> getTags()
   returns a LinkedHashMap (so that the order is preserved) containing a 
   attribute <-> value correlation.
   This map would could contain:
      key = "return"
      value = "The description of what is returned"

The grammar is not following the Java Language Specification entirely.
Some constructions must generate error, like:

{
  this.this.x = 1;
}

The getter and setter for the package of class CompilationUnit are called 
'getPakage' and 'setPakage' respectively. They should be called 
'getPackage' and 'setPackage' instead.

Using v1.0.5.

Hi,

I'm using the version 1.0.7 of Javaparser for one on mine Project.

It runs pretty good and I had no Problem.

But today I tryied to parse a German Project. In one Class there is : 
private int [] mÖpBär;

Something with accent seams to be not good for Javaparser. I take this 
exception :
Exception in thread "main" japa.parser.TokenMgrError: Lexical error at line 
1, column 1.  Encountered: "\ufeff" (65279), after : ""

After that the Program stops. Is it normal ? There is something that i can 
do to avoid that ?

In any case, thanks for this project it's really helpfully for me.

Bye

The AST nodes are all read only. Will be very useful if them can be edited
to made changes in a parsed CompilationUnit and then save it again to the
.java file, this is useful for refactoring changes.
Create a new CompilationUnit without parse a java file can be useful too.

I'm use 1.0.8 version.

When source contains mixed 4chars indent and tabs, expr.getBeginColumn()
returns incorrect value. 

JavaCharStream already has a setTabSize() method, but there is no way to
access it by JavaParser or ASTParser.

What steps will reproduce the problem?
1. input source code contains two consecutive commenst,no other expresses 
between them. the pre comment can not find in CompilationUnit. 
2.
3.

What is the expected output? What do you see instead?
if there are two consecutive  comments, the pre item can't find in 
CompilationUnit. 

What version of the product are you using? On what operating system?
1.0.8

Please provide any additional information below.

Perhaps this is already possible, but I was not able to find how.

It seems that currently, javaparser dumps all comments into 
CompilationUnit with only JavadocComments receiving special treatment and 
being attached to their respective ClassOrInterfaceDeclarations, 
FieldDeclarations and MethodDeclarations.

If javaparser is used, for example, to write a pretty-printer, it would be 
useful if the comments were kept as part of the AST. A possible way to 
achieve this could be having a:

List<Comment> getTrailingComments()

method in every Node, which would return the list of comments following 
that node.

Once this is implemented, DumpVisitor should be updated (perhaps via an 
optional flag in the constructor) to print comments as well as other nodes.

Currently, a workaround is possible by looking at endLine and endColumn of 
the node you're at and then looking for a Comment whose startLine and 
startColumn precede the next node you would've printed. This is fairly 
ugly and inefficient.

The issue why I personally need it for is GWT (http://code.google.com/
webtoolkit/) native methods whose syntax is:

private static native void foo()/*-{
    // javascript code goes here
}-*/;

In the meanwhile I've solved it with a new token type and adding 
pushNativeMethodBodyBlock() (similarly to how pushJavadoc()) at the 
MethodDeclaration production rule.

Javadoc entire source code.

[deleted issue]

This is an RFE.

It would be nice if there was also code to build the symbol table (for 
each identifier its scope, where it was declared etc.).

Each 3 subclasses of AnnotationExpr has the following code:
    private NameExpr name;
    public NameExpr getName() {
        return name;
    }
    public void setName(NameExpr name) {
        this.name = name;
    }

It might be better to pull them up.

I would like to use the library in a Maven project.
I have to upload it into my own repository.
It would be nice if the project would be also in a public repo.

Java 7 has some syntactical changes:

1. JSR 308: type annotations on Java
2. Project Coin changes, as described here 
(http://wikis.sun.com/display/ProjectCoin/2009ProposalsTOC).
3. Jigsaw module syntax
4. Potentially lambda/closure expressions

Attached is my code to get javaparser to support jsr 308 (not quite ready for 
integration yet).

What steps will reproduce the problem?
1. Use the following input file:

@Deprecated()
public class Test
{
}

2. Use the following test parser:
JavaParser.parse(new File(filename)).accept(new VoidVisitorAdapter(), 
null);

What is the expected output? What do you see instead?

Expected that code would not throw any exceptions.
Observed the following exception:
Exception in thread "main" java.lang.NullPointerException
    at 
japa.parser.ast.visitor.VoidVisitorAdapter.visit(VoidVisitorAdapter.java:570)
    at 
japa.parser.ast.expr.NormalAnnotationExpr.accept(NormalAnnotationExpr.java:54)
    at 
japa.parser.ast.visitor.VoidVisitorAdapter.visit(VoidVisitorAdapter.java:234)
    at 
japa.parser.ast.body.ClassOrInterfaceDeclaration.accept(ClassOrInterfaceDeclarat
ion.java:78)
    at 
japa.parser.ast.visitor.VoidVisitorAdapter.visit(VoidVisitorAdapter.java:282)
    at japa.parser.ast.CompilationUnit.accept(CompilationUnit.java:69)


What version of the product are you using? On what operating system?

Java1.5_parser_and_AST_2008-06-09.zip 

Please provide any additional information below.

Workaround: remove () from @Deprecated() in the input file

What steps will reproduce the problem?
1. Use a block statement visitor to traverse a piece of Java source code

What is the expected output? What do you see instead?
The java file containing an enum keyword is parsed as normal

What version of the product are you using? On what operating system?
Product - 1.0.8
Operating System - Windows 7

Please provide any additional information below.
I am using Java 1.4.2 where enum is not a keyword and could be used in the 
source.

Is there any resolution to this?

The declaration was:
            Enumeration enum = req.getParameterNames();

The exception thrown was:
japa.parser.ParseException: Encountered " "enum" "enum "" at line 2685, column 
25.
Was expecting one of:
    "(" ...
    ";" ...
    "=" ...
    "++" ...
    "--" ...
    "+=" ...
    "-=" ...
    "*=" ...
    "/=" ...
    "&=" ...
    "|=" ...
    "^=" ...
    "%=" ...
    "<<=" ...
    ">>=" ...
    ">>>=" ...

What steps will reproduce the problem?
1. Create a Java like this:
/*
 * Comment 1
 */

/*
 * Comment 2
 */

/*
 * Comment 3
 */


package net.perfectbug.test;

public class Test {

}
2. Parse it with JavaParser.
3. Use CompilationUnit.toString() to get the parsed source code.

What is the expected output? What do you see instead?
Expected:
All comments should be print.
Observation:
Not comments are print.
And CompilationUnit.comments only contains comment 3. It doesn't contains 
neither comment 1 nor comment 2.

What version of the product are you using? On what operating system?
1.0.8 on Windows XP.

Please provide any additional information below.

I believe that the parser handles some array declarations incorrect.

What steps will reproduce the problem?
1. Create the following method:
public void test(int argument[]) {}
2. parse it
3.

What is the expected output? What do you see instead?
Expected:
The paramater#type = int[]
The paramater#id = argument

Actual:
The paramater#type = int
The paramater#id = argument[]

What version of the product are you using? On what operating system?
1.05

Please provide any additional information below.

Create helper class to improve the node creation and reading.
The helper may have methods such:

getConstructors()
getFields()
getMethods()

newName() //accepts dotted names
newPackageDeclaration() // //accepts dotted names

etc...

What steps will reproduce the problem?
1. Recompile java_1_5.jj from javaparser-1.0.7 using JavaCC 4.1
2. See that there is a warning from JavaCC

What is the expected output? What do you see instead?

-> Expected output is no warning from JavaCC
-> Here's the output I see:

Java Compiler Compiler Version 4.1d1 (Parser Generator)
(type "javacc" with no arguments for help)
Reading from file java_1_5.jj . . .
Warning: Line 27, Column 3: Bad option name "TOKEN_FACTORY".  Option
setting will be ignored.
Parser generated with 0 errors and 1 warnings.


What version of the product are you using? On what operating system?

I use javaparser-1.0.7 on Windows XP SP3 with Java 6 Update 10.

Please provide any additional information below.

n.a.