The set of tokens is $\{$IDENT, expr, decl, if, else, switch, while, do, for, goto, continue, break, return, case, default, :, ;, {, }, (, ) $\}$. The token IDENT refers to an identifier (such as a label name), that is, a non-empty sequence of alphanumeric characters and underscore, not starting by a digit. We also dispose of tokens expr and decl, which denote expressions and declarations respectively. Thus, instead of parsing those, we use these ‘tokens’. Note that the expressions and declarations represented by them are not empty. The most elemental statements are: nothing, an expression, or a list of declarations and statements enclosed in curly brackets ({ and }). All statements must end in a semicolon (;), except for the aforementioned list in curly brackets, which does not need a semicolon after the closing bracket. A statement does not need to have its own ending semicolon if it ends in another statement, because that statement will have its own semicolon or be a block {...}, which does not require one. For instance, the statement “expr;” has an ending semicolon, so in the if statement “if(expr) expr;”, it is not required to add another semicolon. A statement can be labeled, in which case it is preceded by a label identifier (IDENT) and the token “:”. Alternatively, instead of a label identifier we can have the keyword case followed by an expression, or the keyword default (these options are only semantically valid in the context of a switch statement, but this doesn’t have to be considered by the parser). There are also the following types of statements:

• An if statement starts with the keyword if, followed by the condition (i.e., an expression) in parentheses, and a statement. Optionally, this statement can be followed by the keyword else and another statement.
• A switch statement starts with the keyword switch, followed by an expression in paretheses and then a statement.
• A while statement starts with the keyword while, followed by the condition (an expression) in paretheses and then a statement.
• A do-while statement starts with the keyword do, followed by a statement, the keyword while and a condition (an expression) in parentheses.
• A for statement starts with the keyword for, followed by a list of three expressions enclosed in parentheses and separated by semicolons, and then a statement. Any of the expressions can be missing, but the separating semicolons must remain.
• A goto statement starts with the keyword goto, followed by a label identifier.
• A return statement starts with the keyword return followed by an optional expression.

Moreover, the statements continue and break consist of just the keywords continue and break, respectively. Remarks about AST construction:

• An empty statement must be represented by a special node named nop (no operation).
• A statement consisting of an expression must be represented simply by the expr literal token.
• A list of statements and declarations enclosed in curly brackets must be represented by a tree with a special node named block as root, and the element of the list as children.
• A labeled statement must be represented by a tree with a special node named label as root, the label identifier as first child and the statement itself as second child.
• A case statement (a statement starting with case) must be represented by a tree with the token case as root and the expression and the statement as children.
• A default statement (a statement starting with default) must be represented by a tree with the token default as root and the statement as only child.
• An if statement must be represented by a tree with the token if as root and the condition and statement as children. In case an else clause is present, the corresponding statement must appear as third child.
• A switch statement must be represented by a tree with the token switch as root and the expression and statement as children.
• A while statement must be represented by a tree with the token while as root and the condition and statement as children.
• A do-while statement must be represented by a tree with the token do as root and the statement and condition as children.
• A for statement must be represented by a tree with the token for as root and four children, one for each expression of the list and one for the statement. In case any of the expressions are missing, the corresponding son must consist of a special node named nop.
• A goto statement must be represented by a tree with the token goto as root and the label identifier as only child.
• A return statement must be represented by a tree with the token return as root. If an expression is given, then it must appear as child. Otherwise, a special node named nop must appear as child.
• A continue or break statement must be represented by a node with just the corresponding keyword.