# 6 The Abstract Format

This document describes the standard representation of parse trees for Erlang
programs as Erlang terms. This representation is known as the **abstract format**.
Functions dealing with such parse trees are compile:forms/[1,2]
and functions in the modules
epp,
erl_eval,
erl_lint,
erl_pp,
erl_parse,
and
io.
They are also used as input and output for parse transforms (see the module
compile).

We use the function Rep to denote the mapping from an Erlang source construct C to its abstract format representation R, and write R = Rep(C).

The word LINE below represents an integer, and denotes the number of the line in the source file where the construction occurred. Several instances of LINE in the same construction may denote different lines.

Since operators are not terms in their own right, when operators are mentioned below, the representation of an operator should be taken to be the atom with a printname consisting of the same characters as the operator.

### 6.1 Module Declarations and Forms

A module declaration consists of a sequence of forms that are either function declarations or attributes.

- If D is a module declaration consisting of the forms F_1, ..., F_k, then Rep(D) = [Rep(F_1), ..., Rep(F_k)].
- If F is an attribute -export([Fun_1/A_1, ..., Fun_k/A_k]), then Rep(F) = {attribute,LINE,export,[{Fun_1,A_1}, ..., {Fun_k,A_k}]}.
- If F is an attribute -import(Mod,[Fun_1/A_1, ..., Fun_k/A_k]), then Rep(F) = {attribute,LINE,import,{Mod,[{Fun_1,A_1}, ..., {Fun_k,A_k}]}}.
- If F is an attribute -module(Mod), then Rep(F) = {attribute,LINE,module,Mod}.
- If F is an attribute -file(File,Line), then Rep(F) = {attribute,LINE,file,{File,Line}}.
- If F is a function declaration Name Fc_1 ; ... ; Name Fc_k, where each Fc_i is a function clause with a pattern sequence of the same length Arity, then Rep(F) = {function,LINE,Name,Arity,[Rep(Fc_1), ...,Rep(Fc_k)]}.
- If F is a function specification -Spec Name Ft_1; ...; Ft_k, where Spec is either the atom spec or the atom callback, and each Ft_i is a possibly constrained function type with an argument sequence of the same length Arity, then Rep(F) = {attribute,Line,Spec,{{Name,Arity},[Rep(Ft_1), ..., Rep(Ft_k)]}}.
- If F is a function specification -spec Mod:Name Ft_1; ...; Ft_k, where each Ft_i is a possibly constrained function type with an argument sequence of the same length Arity, then Rep(F) = {attribute,Line,spec,{{Mod,Name,Arity},[Rep(Ft_1), ..., Rep(Ft_k)]}}.
- If F is a record declaration -record(Name,{V_1, ..., V_k}), where each V_i is a record field, then Rep(F) = {attribute,LINE,record,{Name,[Rep(V_1), ..., Rep(V_k)]}}. For Rep(V), see below.
- If F is a type declaration -Type Name(V_1, ..., V_k) :: T, where Type is either the atom type or the atom opaque, each V_i is a variable, and T is a type, then Rep(F) = {attribute,LINE,Type,{Name,Rep(T),[Rep(V_1), ..., Rep(V_k)]}}.
- If F is a wild attribute -A(T), then
Rep(F) = {attribute,LINE,A,T}.

#### Record Fields

Each field in a record declaration may have an optional explicit default initializer expression, as well as an optional type.

- If V is A, then Rep(V) = {record_field,LINE,Rep(A)}.
- If V is A = E, where E is an expression, then Rep(V) = {record_field,LINE,Rep(A),Rep(E)}.
- If V is A :: T, where T is a type, then Rep(V) = {typed_record_field,{record_field,LINE,Rep(A)},Rep(T)}.
- If V is A = E :: T, where E is an expression and T is a type, then Rep(V) = {typed_record_field,{record_field,LINE,Rep(A),Rep(E)},Rep(T)}.

#### Representation of Parse Errors and End-of-file

In addition to the representations of forms, the list that represents a module declaration (as returned by functions in erl_parse and epp) may contain tuples {error,E} and {warning,W}, denoting syntactically incorrect forms and warnings, and {eof,LINE}, denoting an end-of-stream encountered before a complete form had been parsed.

### 6.2 Atomic Literals

There are five kinds of atomic literals, which are represented in the same way in patterns, expressions and guards:

- If L is an atom literal, then Rep(L) = {atom,LINE,L}.
- If L is a character literal, then Rep(L) = {char,LINE,L}.
- If L is a float literal, then Rep(L) = {float,LINE,L}.
- If L is an integer literal, then Rep(L) = {integer,LINE,L}.
- If L is a string literal consisting of the characters C_1, ..., C_k, then Rep(L) = {string,LINE,[C_1, ..., C_k]}.

Note that negative integer and float literals do not occur as such; they are parsed as an application of the unary negation operator.

### 6.3 Patterns

If Ps is a sequence of patterns P_1, ..., P_k, then Rep(Ps) = [Rep(P_1), ..., Rep(P_k)]. Such sequences occur as the list of arguments to a function or fun.

Individual patterns are represented as follows:

- If P is an atomic literal L, then Rep(P) = Rep(L).
- If P is a bit string pattern <<P_1:Size_1/TSL_1, ..., P_k:Size_k/TSL_k>>, where each Size_i is an expression that can be evaluated to an integer and each TSL_i is a type specificer list, then Rep(P) = {bin,LINE,[{bin_element,LINE,Rep(P_1),Rep(Size_1),Rep(TSL_1)}, ..., {bin_element,LINE,Rep(P_k),Rep(Size_k),Rep(TSL_k)}]}. For Rep(TSL), see below. An omitted Size_i is represented by default. An omitted TSL_i is represented by default.
- If P is a compound pattern P_1 = P_2, then Rep(P) = {match,LINE,Rep(P_1),Rep(P_2)}.
- If P is a cons pattern [P_h | P_t], then Rep(P) = {cons,LINE,Rep(P_h),Rep(P_t)}.
- If P is a map pattern #{A_1, ..., A_k}, where each A_i is an association P_i_1 := P_i_2, then Rep(P) = {map,LINE,[Rep(A_1), ..., Rep(A_k)]}. For Rep(A), see below.
- If P is a nil pattern [], then Rep(P) = {nil,LINE}.
- If P is an operator pattern P_1 Op P_2, where Op is a binary operator (this is either an occurrence of ++ applied to a literal string or character list, or an occurrence of an expression that can be evaluated to a number at compile time), then Rep(P) = {op,LINE,Op,Rep(P_1),Rep(P_2)}.
- If P is an operator pattern Op P_0, where Op is a unary operator (this is an occurrence of an expression that can be evaluated to a number at compile time), then Rep(P) = {op,LINE,Op,Rep(P_0)}.
- If P is a parenthesized pattern ( P_0 ), then Rep(P) = Rep(P_0), that is, parenthesized patterns cannot be distinguished from their bodies.
- If P is a record field index pattern #Name.Field, where Field is an atom, then Rep(P) = {record_index,LINE,Name,Rep(Field)}.
- If P is a record pattern #Name{Field_1=P_1, ..., Field_k=P_k}, where each Field_i is an atom or _, then Rep(P) = {record,LINE,Name,[{record_field,LINE,Rep(Field_1),Rep(P_1)}, ..., {record_field,LINE,Rep(Field_k),Rep(P_k)}]}.
- If P is a tuple pattern {P_1, ..., P_k}, then Rep(P) = {tuple,LINE,[Rep(P_1), ..., Rep(P_k)]}.
- If P is a universal pattern _, then Rep(P) = {var,LINE,'_'}.
- If P is a variable pattern V, then Rep(P) = {var,LINE,A}, where A is an atom with a printname consisting of the same characters as V.

Note that every pattern has the same source form as some expression, and is represented the same way as the corresponding expression.

### 6.4 Expressions

A body B is a nonempty sequence of expressions E_1, ..., E_k, and Rep(B) = [Rep(E_1), ..., Rep(E_k)].

An expression E is one of the following alternatives:

- If E is an atomic literal L, then Rep(E) = Rep(L).
- If E is a bit string comprehension <<E_0 || Q_1, ..., Q_k>>, where each Q_i is a qualifier, then Rep(E) = {bc,LINE,Rep(E_0),[Rep(Q_1), ..., Rep(Q_k)]}. For Rep(Q), see below.
- If E is a bit string constructor <<E_1:Size_1/TSL_1, ..., E_k:Size_k/TSL_k>>, where each Size_i is an expression and each TSL_i is a type specificer list, then Rep(E) = {bin,LINE,[{bin_element,LINE,Rep(E_1),Rep(Size_1),Rep(TSL_1)}, ..., {bin_element,LINE,Rep(E_k),Rep(Size_k),Rep(TSL_k)}]}. For Rep(TSL), see below. An omitted Size_i is represented by default. An omitted TSL_i is represented by default.
- If E is a block expression begin B end, where B is a body, then Rep(E) = {block,LINE,Rep(B)}.
- If E is a case expression case E_0 of Cc_1 ; ... ; Cc_k end, where E_0 is an expression and each Cc_i is a case clause then Rep(E) = {'case',LINE,Rep(E_0),[Rep(Cc_1), ..., Rep(Cc_k)]}.
- If E is a catch expression catch E_0, then Rep(E) = {'catch',LINE,Rep(E_0)}.
- If E is a cons skeleton [E_h | E_t], then Rep(E) = {cons,LINE,Rep(E_h),Rep(E_t)}.
- If E is a fun expression fun Name/Arity, then Rep(E) = {'fun',LINE,{function,Name,Arity}}.
- If E is a fun expression fun Module:Name/Arity, then Rep(E) = {'fun',LINE,{function,Rep(Module),Rep(Name),Rep(Arity)}}. (Before the R15 release: Rep(E) = {'fun',LINE,{function,Module,Name,Arity}}.)
- If E is a fun expression fun Fc_1 ; ... ; Fc_k end, where each Fc_i is a function clause then Rep(E) = {'fun',LINE,{clauses,[Rep(Fc_1), ..., Rep(Fc_k)]}}.
- If E is a fun expression fun Name Fc_1 ; ... ; Name Fc_k end, where Name is a variable and each Fc_i is a function clause then Rep(E) = {named_fun,LINE,Name,[Rep(Fc_1), ..., Rep(Fc_k)]}.
- If E is a function call E_0(E_1, ..., E_k), then Rep(E) = {call,LINE,Rep(E_0),[Rep(E_1), ..., Rep(E_k)]}.
- If E is a function call E_m:E_0(E_1, ..., E_k), then Rep(E) = {call,LINE,{remote,LINE,Rep(E_m),Rep(E_0)},[Rep(E_1), ..., Rep(E_k)]}.
- If E is an if expression if Ic_1 ; ... ; Ic_k end, where each Ic_i is an if clause then Rep(E) = {'if',LINE,[Rep(Ic_1), ..., Rep(Ic_k)]}.
- If E is a list comprehension [E_0 || Q_1, ..., Q_k], where each Q_i is a qualifier, then Rep(E) = {lc,LINE,Rep(E_0),[Rep(Q_1), ..., Rep(Q_k)]}. For Rep(Q), see below.
- If E is a map creation #{A_1, ..., A_k}, where each A_i is an association E_i_1 => E_i_2 or E_i_1 := E_i_2, then Rep(E) = {map,LINE,[Rep(A_1), ..., Rep(A_k)]}. For Rep(A), see below.
- If E is a map update E_0#{A_1, ..., A_k}, where each A_i is an association E_i_1 => E_i_2 or E_i_1 := E_i_2, then Rep(E) = {map,LINE,Rep(E_0),[Rep(A_1), ..., Rep(A_k)]}. For Rep(A), see below.
- If E is a match operator expression P = E_0, where P is a pattern, then Rep(E) = {match,LINE,Rep(P),Rep(E_0)}.
- If E is nil, [], then Rep(E) = {nil,LINE}.
- If E is an operator expression E_1 Op E_2, where Op is a binary operator other than the match operator =, then Rep(E) = {op,LINE,Op,Rep(E_1),Rep(E_2)}.
- If E is an operator expression Op E_0, where Op is a unary operator, then Rep(E) = {op,LINE,Op,Rep(E_0)}.
- If E is a parenthesized expression ( E_0 ), then Rep(E) = Rep(E_0), that is, parenthesized expressions cannot be distinguished from their bodies.
- If E is a receive expression receive Cc_1 ; ... ; Cc_k end, where each Cc_i is a case clause then Rep(E) = {'receive',LINE,[Rep(Cc_1), ..., Rep(Cc_k)]}.
- If E is a receive expression receive Cc_1 ; ... ; Cc_k after E_0 -> B_t end, where each Cc_i is a case clause, E_0 is an expression and B_t is a body, then Rep(E) = {'receive',LINE,[Rep(Cc_1), ..., Rep(Cc_k)],Rep(E_0),Rep(B_t)}.
- If E is a record creation #Name{Field_1=E_1, ..., Field_k=E_k}, where each Field_i is an atom or _, then Rep(E) = {record,LINE,Name,[{record_field,LINE,Rep(Field_1),Rep(E_1)}, ..., {record_field,LINE,Rep(Field_k),Rep(E_k)}]}.
- If E is a record field access E_0#Name.Field, where Field is an atom, then Rep(E) = {record_field,LINE,Rep(E_0),Name,Rep(Field)}.
- If E is a record field index #Name.Field, where Field is an atom, then Rep(E) = {record_index,LINE,Name,Rep(Field)}.
- If E is a record update E_0#Name{Field_1=E_1, ..., Field_k=E_k}, where each Field_i is an atom, then Rep(E) = {record,LINE,Rep(E_0),Name,[{record_field,LINE,Rep(Field_1),Rep(E_1)}, ..., {record_field,LINE,Rep(Field_k),Rep(E_k)}]}.
- If E is a tuple skeleton {E_1, ..., E_k}, then Rep(E) = {tuple,LINE,[Rep(E_1), ..., Rep(E_k)]}.
- If E is a try expression try B catch Tc_1 ; ... ; Tc_k end, where B is a body and each Tc_i is a catch clause then Rep(E) = {'try',LINE,Rep(B),[],[Rep(Tc_1), ..., Rep(Tc_k)],[]}.
- If E is a try expression try B of Cc_1 ; ... ; Cc_k catch Tc_1 ; ... ; Tc_n end, where B is a body, each Cc_i is a case clause and each Tc_j is a catch clause then Rep(E) = {'try',LINE,Rep(B),[Rep(Cc_1), ..., Rep(Cc_k)],[Rep(Tc_1), ..., Rep(Tc_n)],[]}.
- If E is a try expression try B after A end, where B and A are bodies then Rep(E) = {'try',LINE,Rep(B),[],[],Rep(A)}.
- If E is a try expression try B of Cc_1 ; ... ; Cc_k after A end, where B and A are a bodies and each Cc_i is a case clause then Rep(E) = {'try',LINE,Rep(B),[Rep(Cc_1), ..., Rep(Cc_k)],[],Rep(A)}.
- If E is a try expression try B catch Tc_1 ; ... ; Tc_k after A end, where B and A are bodies and each Tc_i is a catch clause then Rep(E) = {'try',LINE,Rep(B),[],[Rep(Tc_1), ..., Rep(Tc_k)],Rep(A)}.
- If E is a try expression try B of Cc_1 ; ... ; Cc_k catch Tc_1 ; ... ; Tc_n after A end, where B and A are a bodies, each Cc_i is a case clause, and each Tc_j is a catch clause then Rep(E) = {'try',LINE,Rep(B),[Rep(Cc_1), ..., Rep(Cc_k)],[Rep(Tc_1), ..., Rep(Tc_n)],Rep(A)}.
- If E is a variable V, then Rep(E) = {var,LINE,A}, where A is an atom with a printname consisting of the same characters as V.

#### Qualifiers

A qualifier Q is one of the following alternatives:

- If Q is a filter E, where E is an expression, then Rep(Q) = Rep(E).
- If Q is a generator P <- E, where P is a pattern and E is an expression, then Rep(Q) = {generate,LINE,Rep(P),Rep(E)}.
- If Q is a bit string generator P <= E, where P is a pattern and E is an expression, then Rep(Q) = {b_generate,LINE,Rep(P),Rep(E)}.

#### Bit String Element Type Specifiers

A type specifier list TSL for a bit string element is a sequence of type specifiers TS_1 - ... - TS_k, and Rep(TSL) = [Rep(TS_1), ..., Rep(TS_k)].

- If TS is a type specifier A, where A is an atom, then Rep(TS) = A.
- If TS is a type specifier A:Value, where A is an atom and Value is an integer, then Rep(TS) = {A,Value}.

#### Associations

An association A is one of the following alternatives:

- If A is an association K => V, then Rep(A) = {map_field_assoc,LINE,Rep(K),Rep(V)}.
- If A is an association K := V, then Rep(A) = {map_field_exact,LINE,Rep(K),Rep(V)}.

### 6.5 Clauses

There are function clauses, if clauses, case clauses and catch clauses.

A clause C is one of the following alternatives:

- If C is a case clause P -> B, where P is a pattern and B is a body, then Rep(C) = {clause,LINE,[Rep(P)],[],Rep(B)}.
- If C is a case clause P when Gs -> B, where P is a pattern, Gs is a guard sequence and B is a body, then Rep(C) = {clause,LINE,[Rep(P)],Rep(Gs),Rep(B)}.
- If C is a catch clause P -> B, where P is a pattern and B is a body, then Rep(C) = {clause,LINE,[Rep({throw,P,_})],[],Rep(B)}.
- If C is a catch clause X : P -> B, where X is an atomic literal or a variable pattern, P is a pattern, and B is a body, then Rep(C) = {clause,LINE,[Rep({X,P,_})],[],Rep(B)}.
- If C is a catch clause P when Gs -> B, where P is a pattern, Gs is a guard sequence, and B is a body, then Rep(C) = {clause,LINE,[Rep({throw,P,_})],Rep(Gs),Rep(B)}.
- If C is a catch clause X : P when Gs -> B, where X is an atomic literal or a variable pattern, P is a pattern, Gs is a guard sequence, and B is a body, then Rep(C) = {clause,LINE,[Rep({X,P,_})],Rep(Gs),Rep(B)}.
- If C is a function clause ( Ps ) -> B, where Ps is a pattern sequence and B is a body, then Rep(C) = {clause,LINE,Rep(Ps),[],Rep(B)}.
- If C is a function clause ( Ps ) when Gs -> B, where Ps is a pattern sequence, Gs is a guard sequence and B is a body, then Rep(C) = {clause,LINE,Rep(Ps),Rep(Gs),Rep(B)}.
- If C is an if clause Gs -> B, where Gs is a guard sequence and B is a body, then Rep(C) = {clause,LINE,[],Rep(Gs),Rep(B)}.

### 6.6 Guards

A guard sequence Gs is a sequence of guards G_1; ...; G_k, and Rep(Gs) = [Rep(G_1), ..., Rep(G_k)]. If the guard sequence is empty, Rep(Gs) = [].

A guard G is a nonempty sequence of guard tests Gt_1, ..., Gt_k, and Rep(G) = [Rep(Gt_1), ..., Rep(Gt_k)].

A guard test Gt is one of the following alternatives:

- If Gt is an atomic literal L, then Rep(Gt) = Rep(L).
- If Gt is a bit string constructor <<Gt_1:Size_1/TSL_1, ..., Gt_k:Size_k/TSL_k>>, where each Size_i is a guard test and each TSL_i is a type specificer list, then Rep(Gt) = {bin,LINE,[{bin_element,LINE,Rep(Gt_1),Rep(Size_1),Rep(TSL_1)}, ..., {bin_element,LINE,Rep(Gt_k),Rep(Size_k),Rep(TSL_k)}]}. For Rep(TSL), see above. An omitted Size_i is represented by default. An omitted TSL_i is represented by default.
- If Gt is a cons skeleton [Gt_h | Gt_t], then Rep(Gt) = {cons,LINE,Rep(Gt_h),Rep(Gt_t)}.
- If Gt is a function call A(Gt_1, ..., Gt_k), where A is an atom, then Rep(Gt) = {call,LINE,Rep(A),[Rep(Gt_1), ..., Rep(Gt_k)]}.
- If Gt is a function call A_m:A(Gt_1, ..., Gt_k), where A_m is the atom erlang and A is an atom or an operator, then Rep(Gt) = {call,LINE,{remote,LINE,Rep(A_m),Rep(A)},[Rep(Gt_1), ..., Rep(Gt_k)]}.
- If Gt is a map creation #{A_1, ..., A_k}, where each A_i is an association Gt_i_1 => Gt_i_2 or Gt_i_1 := Gt_i_2, then Rep(Gt) = {map,LINE,[Rep(A_1), ..., Rep(A_k)]}. For Rep(A), see above.
- If Gt is a map update Gt_0#{A_1, ..., A_k}, where each A_i is an association Gt_i_1 => Gt_i_2 or Gt_i_1 := Gt_i_2, then Rep(Gt) = {map,LINE,Rep(Gt_0),[Rep(A_1), ..., Rep(A_k)]}. For Rep(A), see above.
- If Gt is nil, [], then Rep(Gt) = {nil,LINE}.
- If Gt is an operator guard test Gt_1 Op Gt_2, where Op is a binary operator other than the match operator =, then Rep(Gt) = {op,LINE,Op,Rep(Gt_1),Rep(Gt_2)}.
- If Gt is an operator guard test Op Gt_0, where Op is a unary operator, then Rep(Gt) = {op,LINE,Op,Rep(Gt_0)}.
- If Gt is a parenthesized guard test ( Gt_0 ), then Rep(Gt) = Rep(Gt_0), that is, parenthesized guard tests cannot be distinguished from their bodies.
- If Gt is a record creation #Name{Field_1=Gt_1, ..., Field_k=Gt_k}, where each Field_i is an atom or _, then Rep(Gt) = {record,LINE,Name,[{record_field,LINE,Rep(Field_1),Rep(Gt_1)}, ..., {record_field,LINE,Rep(Field_k),Rep(Gt_k)}]}.
- If Gt is a record field access Gt_0#Name.Field, where Field is an atom, then Rep(Gt) = {record_field,LINE,Rep(Gt_0),Name,Rep(Field)}.
- If Gt is a record field index #Name.Field, where Field is an atom, then Rep(Gt) = {record_index,LINE,Name,Rep(Field)}.
- If Gt is a tuple skeleton {Gt_1, ..., Gt_k}, then Rep(Gt) = {tuple,LINE,[Rep(Gt_1), ..., Rep(Gt_k)]}.
- If Gt is a variable pattern V, then Rep(Gt) = {var,LINE,A}, where A is an atom with a printname consisting of the same characters as V.

Note that every guard test has the same source form as some expression, and is represented the same way as the corresponding expression.

### 6.7 Types

- If T is an annotated type A :: T_0, where A is a variable, then Rep(T) = {ann_type,LINE,[Rep(A),Rep(T_0)]}.
- If T is an atom or integer literal L, then Rep(T) = Rep(L).
- If T is a bit string type <<_:M,_:_*N>>, where M and N are singleton integer types, then Rep(T) = {type,LINE,binary,[Rep(M),Rep(N)]}.
- If T is the empty list type [], then Rep(T) = {type,Line,nil,[]}.
- If T is a fun type fun(), then Rep(T) = {type,LINE,'fun',[]}.
- If T is a fun type fun((...) -> T_0), then Rep(T) = {type,LINE,'fun',[{type,LINE,any},Rep(T_0)]}.
- If T is a fun type fun(Ft), where Ft is a function type, then Rep(T) = Rep(Ft). For Rep(Ft), see below.
- If T is an integer range type L .. H, where L and H are singleton integer types, then Rep(T) = {type,LINE,range,[Rep(L),Rep(H)]}.
- If T is a map type map(), then Rep(T) = {type,LINE,map,any}.
- If T is a map type #{A_1, ..., A_k}, where each A_i is an association type, then Rep(T) = {type,LINE,map,[Rep(A_1), ..., Rep(A_k)]}. For Rep(A), see below.
- If T is an operator type T_1 Op T_2, where Op is a binary operator (this is an occurrence of an expression that can be evaluated to an integer at compile time), then Rep(T) = {op,LINE,Op,Rep(T_1),Rep(T_2)}.
- If T is an operator type Op T_0, where Op is a unary operator (this is an occurrence of an expression that can be evaluated to an integer at compile time), then Rep(T) = {op,LINE,Op,Rep(T_0)}.
- If T is ( T_0 ), then Rep(T) = Rep(T_0), that is, parenthesized types cannot be distinguished from their bodies.
- If T is a predefined (or built-in) type N(T_1, ..., T_k), then Rep(T) = {type,LINE,N,[Rep(T_1), ..., Rep(T_k)]}.
- If T is a record type #Name{F_1, ..., F_k}, where each F_i is a record field type, then Rep(T) = {type,LINE,record,[Rep(Name),Rep(F_1), ..., Rep(F_k)]}. For Rep(F), see below.
- If T is a remote type M:N(T_1, ..., T_k), then Rep(T) = {remote_type,LINE,[Rep(M),Rep(N),[Rep(T_1), ..., Rep(T_k)]]}.
- If T is a tuple type tuple(), then Rep(T) = {type,LINE,tuple,any}.
- If T is a tuple type {T_1, ..., T_k}, then Rep(T) = {type,LINE,tuple,[Rep(T_1), ..., Rep(T_k)]}.
- If T is a type union T_1 | ... | T_k, then Rep(T) = {type,LINE,union,[Rep(T_1), ..., Rep(T_k)]}.
- If T is a type variable V, then Rep(T) = {var,LINE,A}, where A is an atom with a printname consisting of the same characters as V. A type variable is any variable except underscore (_).
- If T is a user-defined type N(T_1, ..., T_k), then Rep(T) = {user_type,LINE,N,[Rep(T_1), ..., Rep(T_k)]}.

#### Function Types

A function type Ft is one of the following alternatives:

- If Ft is a constrained function type Ft_1 when Fc, where Ft_1 is a function type and Fc is a function constraint, then Rep(T) = {type,LINE,bounded_fun,[Rep(Ft_1),Rep(Fc)]}. For Rep(Fc), see below.
- If Ft is a function type (T_1, ..., T_n) -> T_0, where each T_i is a type, then Rep(Ft) = {type,LINE,'fun',[{type,LINE,product,[Rep(T_1), ..., Rep(T_n)]},Rep(T_0)]}.

#### Function Constraints

A function constraint Fc is a nonempty sequence of constraints C_1, ..., C_k, and Rep(Fc) = [Rep(C_1), ..., Rep(C_k)].

- If C is a constraint is_subtype(V, T) or V :: T, where V is a type variable and T is a type, then Rep(C) = {type,LINE,constraint,[{atom,LINE,is_subtype},[Rep(V),Rep(T)]]}.

#### Association Types

- If A is an association type K => V, where K and V are types, then Rep(A) = {type,LINE,map_field_assoc,[Rep(K),Rep(V)]}.
- If A is an association type K := V, where K and V are types, then Rep(A) = {type,LINE,map_field_exact,[Rep(K),Rep(V)]}.

#### Record Field Types

- If F is a record field type Name :: Type, where Type is a type, then Rep(F) = {type,LINE,field_type,[Rep(Name),Rep(Type)]}.

### 6.8 The Abstract Format After Preprocessing

The compilation option debug_info can be given to the compiler to have the abstract code stored in the abstract_code chunk in the BEAM file (for debugging purposes).

In OTP R9C and later, the abstract_code chunk will contain

{raw_abstract_v1,AbstractCode}

where AbstractCode is the abstract code as described in this document.

In releases of OTP prior to R9C, the abstract code after some more processing was stored in the BEAM file. The first element of the tuple would be either abstract_v1 (R7B) or abstract_v2 (R8B).