Specifications
Getting Started
The files are:
Here is p4.zip file. It is recommended to start your project with it.
Name Analysis
The name analyzer will perform the following tasks:
Build symbol tables.
You will use the "list of hashtables" approach (using the
SymTable class from program 1).
Find multiply declared names, uses of undeclared names,
bad struct accesses, and bad declarations.
Like C, the b language allows the same name to be declared
in non-overlapping or nested scopes.
The formal parameters of a function
are considered to be in the same scope as the function body.
All names must be declared before they are used.
A bad struct access happens when either the left-hand
side of the dot-access is not a name already declared to be of a
struct type or the right-hand side of the dot-access is
not the name of a field for the appropriate type of struct.
A bad declaration is a declaration of anything other
than a function to be of type void as well as the
declaration of a variable to be of a bad struct type
(the name of the struct type doesn't exist or is not a struct type).
Add IdNode links:
For each IdNode in the abstract-syntax tree that
represents a use of a name (not a declaration)
add a "link" to the corresponding symbol-table entry.
(As stated above, you will need to modify the IdNode
class in ast.java to have a new field of type Sym.
That is the field that your name analyzer will fill in with
a link to the Sym returned by the symbol table's
globalLookup method.)
You must implement your name analyzer by writing appropriate methods
for the different subclasses of ASTnode.
Exactly what methods you write is up to you (as long as they do name
analysis as specified).
It may help to
start by writing the name analysis method for ProgramNode,
then work "top down", adding a method for DeclListNode
(the child of a ProgramNode), then for each kind of
DeclNode (except StructDeclNode), and so on
(and then handle StructDeclNode and perhaps other struct
related nodes at the end).
Be sure to think about which nodes' methods need to add a new hashtable
to the symbol table (i.e., when is a new scope being entered) and
which methods need to remove a hashtable from the symbol table
(i.e., when is a scope being exited).
Some of the methods will process the declarations in the program (checking
for bad declarations and checking
whether the names are multiply declared, and if not, adding appropriate
symbol-table entries) and some will process the statements in the program
(checking that every name used in a statement has been declared and
adding links).
Note that you should not add a link for an IdNode that represents
a use of an undeclared name.
struct Handling Issues
Name analysis issues surrounding structs come up in several situations:
Defining a struct type: for example
struct Point {
int x;
int y;
};
When defining a struct, the name of the struct type can't be a name that
has already been declared. The fields of a struct must be unique to that particular struct; however, they can be a name that has been declared outside of the struct definition. For this reason, a recommended approach is to have a separate symbol table associated with each struct definition and to store this symbol table in the symbol for the name of the struct type.
Declaring a variable to be of a struct type: for example
struct Point pt;
When declaring a variable of a struct type, in addition to determining if the variable
name has been previously declared (and issuing a "multiply declared" error if it is),
you should also check that the name of the struct type has been
previously declared and is actually the name of a struct type.
Accessing the fields of a struct: for example
pt.x = 7;
When doing name analysis on something like LHS.RHS, you will need to check
that LHS is the name of a variable that has previously been declared to be of a struct
type and that RHS is the name of a field in the struct type associated with
LHS.
Error Reporting
Your name analyzer should find all of the errors described
in the table given below;
it should report the specified position of the error,
and it should give exactly the specified error message
(each message should appear on a single line, rather than how it is
formatted in the following table).
Error messages should have the same format as in the scanner and parser
(i.e., they should be issued using a call to ErrMsg.fatal).
If a declaration is both "bad" (e.g., a non-function declared void)
and is a declaration of a name that has already been declared
in the same scope, you should give two error
messages (first the "bad" declaration error, then the
"multiply declared" error).
Type of Error |
Error Message |
Position to Report |
More than one declaration of an identifier in a given scope
(note: includes identifier associated with a struct definition) |
Multiply declared identifier |
The first character of the ID in the duplicate declaration |
Use of an undeclared identifier |
Undeclared identifier |
The first character of the undeclared identifier |
Bad struct access (LHS of dot-access is not of a struct type) |
Dot-access of non-struct type |
The first character of the ID corresponding to the LHS of the dot-access. |
Bad struct access (RHS of dot-access is not a field of the appropriate a struct) |
Invalid struct field name |
The first character of the ID corresponding to the RHS of the dot-access. |
Bad declaration (variable or parameter of type void) |
Non-function declared void |
The first character of the ID in the bad declaration. |
Bad declaration (attempt to declare variable of a bad struct type) |
Invalid name of struct type |
The first character of the ID corresponding to the struct type in the bad declaration. |
Note that the names themselves should not be printed as
part of the error messages.
During name analysis, if a function name is multiply declared
you should still process the formals and the body of the function;
don't add a new entry to the current symbol table for the function,
but do add a new
hashtable to the front of the SymTable's list for the names declared in
the body (i.e., the parameters and other local variables of the function).
If you find a bad variable declaration (a variable of type void
or of a bad struct type),
give an error message and add nothing to the symbol table.
Other Tasks
Extending the Sym Class
It is up to you how you store information in each symbol-table
entry (each Sym).
To implement the changes to the unparser described below you will need
to know each name's type.
For function names, this includes the return type and the number of
parameters and their types.
You can modify the Sym class by adding
some new fields (e.g., a kind field)
and/or by declaring some subclasses (e.g.,
a subclass for functions that has extra fields for the
return type and the list of parameter types).
You will probably also want to add new methods that return the values
of the new fields and it may be helpful to change the toString
method so that you can print the contents of a Sym for
debugging purposes.
Modifying the IdNode Class
Two changes to the IdNode class are needed:
Adding a new field of type Sym (to link the node with the
corresponding symbol-table entry), and
-
Changing the unparse method so that every use of an ID has
its type (in parentheses) after its name.
(The point of this is to help you to see whether your name
analyzer is working correctly; i.e., does it correctly match
each use of a name to the corresponding declaration, and does
it correctly set the link from the IdNode to the
information in the symbol table.)
For names of functions, the information should be of
the form:
param1Type, param2Type, ..., paramNType -> returnType.
For names of global variables, parameters, and local variables of a non-struct type , the
information should be int or bool.
For a global or local variable that is of a struct type,
the information should be the name of the struct type.
For example, given a program that contains this code:
struct Point {
int x;
int y;
};
int f(int x, bool b) { }
void g() {
int a;
bool b;
struct Point p;
p.x = a;
b = a == 3;
f(a + p.y*2, b);
g();
}
The unparser should print:
struct Point {
int x;
int y;
};
int f(int x, bool b) {
}
void g() {
int a;
bool b;
struct Point p;
p(Point).x(int) = a(int);
b(bool) = (a(int) == 3);
f(int,bool->int)((a(int) + (p(Point).y(int) * 2)), b(bool));
g(->void)();
}
Summary
struct
- If a variable or a function with the same name has been declared in the same scope before, then do not add a SymTable entry for the struct. You don't have to process the variables of the struct in this case.
- A variable inside a struct with the same name as a variable or a function outside the struct is legal.
- A variable x inside a struct with the same name as another variable inside the struct is illegal. In this case, create SymTable for the struct and add all variables up to but excluding the second occurrence of x and then continue with the rest of the members.
- If a struct is used without declaration like a.b, then you can report two errors (undeclared ID and dot access of non-struct type) or you can just report undeclared ID.
- The name of the struct is in a scope that is one level outside the scope of the struct itself. Thus, a struct and one of its members can have the same name.
function
- A function with the same name as another function in the same scope is illegal. You must not add a new SymTable entry in the outer scope for this second occurrence. You should process the formals and the local variables for both the functions.
- A function with the same name as another variable in the same scope is illegal. In this case, do not create a SymTable entry for the function. However, continue processing the body of the function.
- If a function with formal parameter a also has a variable declared as a, then create the SymTable for the function and add the formal parameter but not the local variable and then continue with processing.
- If a function has 2 formal parameters or 2 local variables with the same name, then create the SymTable, add the first parameter/local variable, report the error and then continue with processing.
- The name of the function is in a scope that is one level outside the scope of the function itself. Thus, a function and one of its formals/local variables can have the same name.
if/else/while
- if/else and while statements have their own scope. So, names can be reused inside these statements.
- The if part and the else part have different scopes. So, the same name can be declared in both of them.
P4.java
The main program, P4.java, will be similar to P3.java,
except that
- After parsing, if there are no syntax errors, it will call
the name analyzer.
- After that, if there are no errors so far (either scanning, parsing, or
name-analysis errors), it will call the unparser.
Calling the name analyzer means calling the
appropriate method of the ASTnode that is the root of the
tree built by the parser.
Modifying the ErrMsg Class
Your compiler should quit after the name analyzer has finished if any
errors have been detected so far (either by the scanner/parser or
the name analyzer).
To accomplish this, you can add a static boolean field to the ErrMsg
class that is initialized to false and is set
to true if the fatal method is ever called
(warnings should not change the value of this field).
Your main program can check the value of this field
and only call the unparser if it is false.
Writing Test Inputs
You will need to write two input files to test your code:
nameErrors.b should contain code with errors
detected by the name analyzer.
This means that it should include bad and multiply declared names
for all of the different kinds of names, and in all of the
different places that declarations can appear.
It should also include uses of undeclared names in all kinds
of statements and expressions as well as bad struct accesses.
-
test.b should
contain code with no errors that exercises all of the name-analysis
methods that you wrote for the different AST nodes.
This means that it should include (good) declarations of all of the
different kinds of names in all of the places that names can be
declared
and it should include (good) uses of names in all kinds
of statements and expressions.
Note that your nameErrors.b should cause error
messages to be output, so to know whether your name analyzer
behaves correctly, you will need to know what output to expect.
As usual, you will be graded in part
on how thoroughly your input files test your code.
Some Advice
Here are few words of advice about various issues that come up
in the assignment:
For this assignment you are free to make any changes you want
to the code in ast.java.
The tree-traversal code you wrote to perform unparsing provides
a good model for the traversal that you need to write to handle
name analysis.
However, you might not want to declare the name-analysis
methods to be abstract methods of class
ASTnode (as we did for unparse).
This is because you will not need those methods for all nodes; e.g.,
you probably won't want a name-analysis method for all of the sub-classes of the
TypeNode class.
However, you will need to declare the name-analysis methods
to be abstract methods of some of the
classes that are lower down in the inheritance hierarchy;
for example, you will need to declare an abstract name-analysis
method for the DeclNode class, because the method for
the DeclListNode class will call that method for each
node in the list.
-
If you are working with a partner, you will have to decide how to
divide up the work.
You might want to divide up some of the "incidental tasks"
(like modifying the ErrMsg, Sym, and IdNode classes),
then work together to get a small part of the name-analysis
phase working (e.g., finding multiply declared global variables).
Then you could split up the ASTnode subclasses and each
implement the name-analysis methods for your
subset of those classes (you might want to start by choosing just a
few each, until you have a better idea which ones will require the
most work).
Don't forget to test your work as you go along, rather than waiting
until everything is finished!
|