A Proc object is an encapsulation of a block of code, which can be stored in a local variable, passed to a method or another Proc, and can be called. Proc is an essential concept in Ruby and a core of its functional programming features.
square = Proc.new {|x| x**2 } square.call(3) #=> 9 # shorthands: square.(3) #=> 9 square[3] #=> 9
Proc objects are closures, meaning they remember and can use the entire context in which they were created.
def gen_times(factor) Proc.new {|n| n*factor } # remembers the value of factor at the moment of creation end times3 = gen_times(3) times5 = gen_times(5) times3.call(12) #=> 36 times5.call(5) #=> 25 times3.call(times5.call(4)) #=> 60
Creation
There are several methods to create a Proc
-
Use the
Procclass constructor:proc1 = Proc.new {|x| x**2 }
-
Use the
Kernel#procmethod as a shorthand ofProc.new:proc2 = proc {|x| x**2 }
-
Receiving a block of code into proc argument (note the
&):def make_proc(&block) block end proc3 = make_proc {|x| x**2 }
-
Construct a proc with lambda semantics using the
Kernel#lambdamethod (see below for explanations about lambdas):lambda1 = lambda {|x| x**2 }
-
Use the Lambda literal syntax (also constructs a proc with lambda semantics):
lambda2 = ->(x) { x**2 }
Lambda and non-lambda semantics
Procs are coming in two flavors: lambda and non-lambda (regular procs). Differences are:
-
In lambdas,
returnmeans exit from this lambda; -
In regular procs,
returnmeans exit from embracing method (and will throwLocalJumpErrorif invoked outside the method); -
In lambdas, arguments are treated in the same way as in methods: strict, with
ArgumentErrorfor mismatching argument number, and no additional argument processing; -
Regular procs accept arguments more generously: missing arguments are filled with
nil, singleArrayarguments are deconstructed if the proc has multiple arguments, and there is no error raised on extra arguments.
Examples:
p = proc {|x, y| "x=#{x}, y=#{y}" } p.call(1, 2) #=> "x=1, y=2" p.call([1, 2]) #=> "x=1, y=2", array deconstructed p.call(1, 2, 8) #=> "x=1, y=2", extra argument discarded p.call(1) #=> "x=1, y=", nil substituted instead of error l = lambda {|x, y| "x=#{x}, y=#{y}" } l.call(1, 2) #=> "x=1, y=2" l.call([1, 2]) # ArgumentError: wrong number of arguments (given 1, expected 2) l.call(1, 2, 8) # ArgumentError: wrong number of arguments (given 3, expected 2) l.call(1) # ArgumentError: wrong number of arguments (given 1, expected 2) def test_return -> { return 3 }.call # just returns from lambda into method body proc { return 4 }.call # returns from method return 5 end test_return # => 4, return from proc
Lambdas are useful as self-sufficient functions, in particular useful as arguments to higher-order functions, behaving exactly like Ruby methods.
Procs are useful for implementing iterators:
def test [[1, 2], [3, 4], [5, 6]].map {|a, b| return a if a + b > 10 } # ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ end
Inside map, the block of code is treated as a regular (non-lambda) proc, which means that the internal arrays will be deconstructed to pairs of arguments, and return will exit from the method test. That would not be possible with a stricter lambda.
You can tell a lambda from a regular proc by using the lambda? instance method.
Lambda semantics is typically preserved during the proc lifetime, including &-deconstruction to a block of code:
p = proc {|x, y| x } l = lambda {|x, y| x } [[1, 2], [3, 4]].map(&p) #=> [1, 2] [[1, 2], [3, 4]].map(&l) # ArgumentError: wrong number of arguments (given 1, expected 2)
The only exception is dynamic method definition: even if defined by passing a non-lambda proc, methods still have normal semantics of argument checking.
class C define_method(:e, &proc {}) end C.new.e(1,2) #=> ArgumentError C.new.method(:e).to_proc.lambda? #=> true
This exception ensures that methods never have unusual argument passing conventions, and makes it easy to have wrappers defining methods that behave as usual.
class C def self.def2(name, &body) define_method(name, &body) end def2(:f) {} end C.new.f(1,2) #=> ArgumentError
The wrapper def2 receives body as a non-lambda proc, yet defines a method which has normal semantics.
Conversion of other objects to procs
Any object that implements the to_proc method can be converted into a proc by the & operator, and therefore con be consumed by iterators.
class Greater def initialize(greating) @greating = greating end def to_proc proc {|name| "#{@greating}, #{name}!" } end end hi = Greater.new("Hi") hey = Greater.new("Hey") ["Bob", "Jane"].map(&hi) #=> ["Hi, Bob!", "Hi, Jane!"] ["Bob", "Jane"].map(&hey) #=> ["Hey, Bob!", "Hey, Jane!"]
Of the Ruby core classes, this method is implemented by Symbol, Method, and Hash.
:to_s.to_proc.call(1) #=> "1" [1, 2].map(&:to_s) #=> ["1", "2"] method(:puts).to_proc.call(1) # prints 1 [1, 2].each(&method(:puts)) # prints 1, 2 {test: 1}.to_proc.call(:test) #=> 1 %i[test many keys].map(&{test: 1}) #=> [1, nil, nil]
static VALUE
rb_proc_s_new(int argc, VALUE *argv, VALUE klass)
{
VALUE block = proc_new(klass, FALSE);
rb_obj_call_init(block, argc, argv);
return block;
}
Creates a new Proc object, bound to the current context. Proc::new may be called without a block only within a method with an attached block, in which case that block is converted to the Proc object.
def proc_from Proc.new end proc = proc_from { "hello" } proc.call #=> "hello"
static VALUE
proc_compose_to_left(VALUE self, VALUE g)
{
VALUE proc, args, procs[2];
rb_proc_t *procp;
int is_lambda;
procs[0] = self;
procs[1] = g;
args = rb_ary_tmp_new_from_values(0, 2, procs);
GetProcPtr(self, procp);
is_lambda = procp->is_lambda;
proc = rb_proc_new(compose, args);
GetProcPtr(proc, procp);
procp->is_lambda = is_lambda;
return proc;
}
Returns a proc that is the composition of this proc and the given g. The returned proc takes a variable number of arguments, calls g with them then calls this proc with the result.
f = proc {|x| x * x } g = proc {|x| x + x } p (f << g).call(2) #=> 16
static VALUE
proc_compose_to_right(VALUE self, VALUE g)
{
VALUE proc, args, procs[2];
rb_proc_t *procp;
int is_lambda;
procs[0] = g;
procs[1] = self;
args = rb_ary_tmp_new_from_values(0, 2, procs);
GetProcPtr(self, procp);
is_lambda = procp->is_lambda;
proc = rb_proc_new(compose, args);
GetProcPtr(proc, procp);
procp->is_lambda = is_lambda;
return proc;
}
Returns a proc that is the composition of this proc and the given g. The returned proc takes a variable number of arguments, calls g with them then calls this proc with the result.
f = proc {|x| x * x } g = proc {|x| x + x } p (f >> g).call(2) #=> 8
static VALUE
proc_arity(VALUE self)
{
int arity = rb_proc_arity(self);
return INT2FIX(arity);
}
Returns the number of mandatory arguments. If the block is declared to take no arguments, returns 0. If the block is known to take exactly n arguments, returns n. If the block has optional arguments, returns -n-1, where n is the number of mandatory arguments, with the exception for blocks that are not lambdas and have only a finite number of optional arguments; in this latter case, returns n. Keyword arguments will be considered as a single additional argument, that argument being mandatory if any keyword argument is mandatory. A proc with no argument declarations is the same as a block declaring || as its arguments.
proc {}.arity #=> 0 proc { || }.arity #=> 0 proc { |a| }.arity #=> 1 proc { |a, b| }.arity #=> 2 proc { |a, b, c| }.arity #=> 3 proc { |*a| }.arity #=> -1 proc { |a, *b| }.arity #=> -2 proc { |a, *b, c| }.arity #=> -3 proc { |x:, y:, z:0| }.arity #=> 1 proc { |*a, x:, y:0| }.arity #=> -2 proc { |a=0| }.arity #=> 0 lambda { |a=0| }.arity #=> -1 proc { |a=0, b| }.arity #=> 1 lambda { |a=0, b| }.arity #=> -2 proc { |a=0, b=0| }.arity #=> 0 lambda { |a=0, b=0| }.arity #=> -1 proc { |a, b=0| }.arity #=> 1 lambda { |a, b=0| }.arity #=> -2 proc { |(a, b), c=0| }.arity #=> 1 lambda { |(a, b), c=0| }.arity #=> -2 proc { |a, x:0, y:0| }.arity #=> 1 lambda { |a, x:0, y:0| }.arity #=> -2
static VALUE
proc_binding(VALUE self)
{
VALUE bindval, binding_self = Qundef;
rb_binding_t *bind;
const rb_proc_t *proc;
const rb_iseq_t *iseq = NULL;
const struct rb_block *block;
const rb_env_t *env = NULL;
GetProcPtr(self, proc);
block = &proc->block;
again:
switch (vm_block_type(block)) {
case block_type_iseq:
iseq = block->as.captured.code.iseq;
binding_self = block->as.captured.self;
env = VM_ENV_ENVVAL_PTR(block->as.captured.ep);
break;
case block_type_proc:
GetProcPtr(block->as.proc, proc);
block = &proc->block;
goto again;
case block_type_symbol:
goto error;
case block_type_ifunc:
{
const struct vm_ifunc *ifunc = block->as.captured.code.ifunc;
if (IS_METHOD_PROC_IFUNC(ifunc)) {
VALUE method = (VALUE)ifunc->data;
VALUE name = rb_fstring_lit("<empty_iseq>");
rb_iseq_t *empty;
binding_self = method_receiver(method);
iseq = rb_method_iseq(method);
env = VM_ENV_ENVVAL_PTR(block->as.captured.ep);
env = env_clone(env, method_cref(method));
/* set empty iseq */
empty = rb_iseq_new(NULL, name, name, Qnil, 0, ISEQ_TYPE_TOP);
RB_OBJ_WRITE(env, &env->iseq, empty);
break;
}
else {
error:
rb_raise(rb_eArgError, "Can't create Binding from C level Proc");
return Qnil;
}
}
}
bindval = rb_binding_alloc(rb_cBinding);
GetBindingPtr(bindval, bind);
RB_OBJ_WRITE(bindval, &bind->block.as.captured.self, binding_self);
RB_OBJ_WRITE(bindval, &bind->block.as.captured.code.iseq, env->iseq);
rb_vm_block_ep_update(bindval, &bind->block, env->ep);
RB_OBJ_WRITTEN(bindval, Qundef, VM_ENV_ENVVAL(env->ep));
if (iseq) {
rb_iseq_check(iseq);
RB_OBJ_WRITE(bindval, &bind->pathobj, iseq->body->location.pathobj);
bind->first_lineno = FIX2INT(rb_iseq_first_lineno(iseq));
}
else {
RB_OBJ_WRITE(bindval, &bind->pathobj,
rb_iseq_pathobj_new(rb_fstring_lit("(binding)"), Qnil));
bind->first_lineno = 1;
}
return bindval;
}
Returns the binding associated with prc.
def fred(param) proc {} end b = fred(99) eval("param", b.binding) #=> 99
0
static VALUE
proc_call(int argc, VALUE *argv, VALUE procval)
{
/* removed */
}
Invokes the block, setting the block’s parameters to the values in params using something close to method calling semantics. Returns the value of the last expression evaluated in the block.
a_proc = Proc.new {|scalar, *values| values.map {|value| value*scalar } } a_proc.call(9, 1, 2, 3) #=> [9, 18, 27] a_proc[9, 1, 2, 3] #=> [9, 18, 27] a_proc.(9, 1, 2, 3) #=> [9, 18, 27] a_proc.yield(9, 1, 2, 3) #=> [9, 18, 27]
Note that prc.() invokes prc.call() with the parameters given. It’s syntactic sugar to hide “call”.
For procs created using lambda or ->() an error is generated if the wrong number of parameters are passed to the proc. For procs created using Proc.new or Kernel.proc, extra parameters are silently discarded and missing parameters are set to nil.
a_proc = proc {|a,b| [a,b] } a_proc.call(1) #=> [1, nil] a_proc = lambda {|a,b| [a,b] } a_proc.call(1) # ArgumentError: wrong number of arguments (given 1, expected 2)
See also Proc#lambda?.
static VALUE
proc_curry(int argc, const VALUE *argv, VALUE self)
{
int sarity, max_arity, min_arity = rb_proc_min_max_arity(self, &max_arity);
VALUE arity;
if (rb_check_arity(argc, 0, 1) == 0 || NIL_P(arity = argv[0])) {
arity = INT2FIX(min_arity);
}
else {
sarity = FIX2INT(arity);
if (rb_proc_lambda_p(self)) {
rb_check_arity(sarity, min_arity, max_arity);
}
}
return make_curry_proc(self, rb_ary_new(), arity);
}
Returns a curried proc. If the optional arity argument is given, it determines the number of arguments. A curried proc receives some arguments. If a sufficient number of arguments are supplied, it passes the supplied arguments to the original proc and returns the result. Otherwise, returns another curried proc that takes the rest of arguments.
b = proc {|x, y, z| (x||0) + (y||0) + (z||0) } p b.curry[1][2][3] #=> 6 p b.curry[1, 2][3, 4] #=> 6 p b.curry(5)[1][2][3][4][5] #=> 6 p b.curry(5)[1, 2][3, 4][5] #=> 6 p b.curry(1)[1] #=> 1 b = proc {|x, y, z, *w| (x||0) + (y||0) + (z||0) + w.inject(0, &:+) } p b.curry[1][2][3] #=> 6 p b.curry[1, 2][3, 4] #=> 10 p b.curry(5)[1][2][3][4][5] #=> 15 p b.curry(5)[1, 2][3, 4][5] #=> 15 p b.curry(1)[1] #=> 1 b = lambda {|x, y, z| (x||0) + (y||0) + (z||0) } p b.curry[1][2][3] #=> 6 p b.curry[1, 2][3, 4] #=> wrong number of arguments (given 4, expected 3) p b.curry(5) #=> wrong number of arguments (given 5, expected 3) p b.curry(1) #=> wrong number of arguments (given 1, expected 3) b = lambda {|x, y, z, *w| (x||0) + (y||0) + (z||0) + w.inject(0, &:+) } p b.curry[1][2][3] #=> 6 p b.curry[1, 2][3, 4] #=> 10 p b.curry(5)[1][2][3][4][5] #=> 15 p b.curry(5)[1, 2][3, 4][5] #=> 15 p b.curry(1) #=> wrong number of arguments (given 1, expected 3) b = proc { :foo } p b.curry[] #=> :foo
static VALUE
proc_hash(VALUE self)
{
st_index_t hash;
hash = rb_hash_start(0);
hash = rb_hash_proc(hash, self);
hash = rb_hash_end(hash);
return ST2FIX(hash);
}
Returns a hash value corresponding to proc body.
See also Object#hash.
VALUE
rb_proc_lambda_p(VALUE procval)
{
rb_proc_t *proc;
GetProcPtr(procval, proc);
return proc->is_lambda ? Qtrue : Qfalse;
}
Returns true for a Proc object for which argument handling is rigid. Such procs are typically generated by lambda.
A Proc object generated by proc ignores extra arguments.
proc {|a,b| [a,b] }.call(1,2,3) #=> [1,2]
It provides nil for missing arguments.
proc {|a,b| [a,b] }.call(1) #=> [1,nil]
It expands a single array argument.
proc {|a,b| [a,b] }.call([1,2]) #=> [1,2]
A Proc object generated by lambda doesn’t have such tricks.
lambda {|a,b| [a,b] }.call(1,2,3) #=> ArgumentError lambda {|a,b| [a,b] }.call(1) #=> ArgumentError lambda {|a,b| [a,b] }.call([1,2]) #=> ArgumentError
Proc#lambda? is a predicate for the tricks. It returns true if no tricks apply.
lambda {}.lambda? #=> true proc {}.lambda? #=> false
Proc.new is the same as proc.
Proc.new {}.lambda? #=> false
lambda, proc and Proc.new preserve the tricks of a Proc object given by & argument.
lambda(&lambda {}).lambda? #=> true proc(&lambda {}).lambda? #=> true Proc.new(&lambda {}).lambda? #=> true lambda(&proc {}).lambda? #=> false proc(&proc {}).lambda? #=> false Proc.new(&proc {}).lambda? #=> false
A Proc object generated by & argument has the tricks
def n(&b) b.lambda? end n {} #=> false
The & argument preserves the tricks if a Proc object is given by & argument.
n(&lambda {}) #=> true n(&proc {}) #=> false n(&Proc.new {}) #=> false
A Proc object converted from a method has no tricks.
def m() end method(:m).to_proc.lambda? #=> true n(&method(:m)) #=> true n(&method(:m).to_proc) #=> true
define_method is treated the same as method definition. The defined method has no tricks.
class C define_method(:d) {} end C.new.d(1,2) #=> ArgumentError C.new.method(:d).to_proc.lambda? #=> true
define_method always defines a method without the tricks, even if a non-lambda Proc object is given. This is the only exception for which the tricks are not preserved.
class C define_method(:e, &proc {}) end C.new.e(1,2) #=> ArgumentError C.new.method(:e).to_proc.lambda? #=> true
This exception ensures that methods never have tricks and makes it easy to have wrappers to define methods that behave as usual.
class C def self.def2(name, &body) define_method(name, &body) end def2(:f) {} end C.new.f(1,2) #=> ArgumentError
The wrapper def2 defines a method which has no tricks.
static VALUE
rb_proc_parameters(VALUE self)
{
int is_proc;
const rb_iseq_t *iseq = rb_proc_get_iseq(self, &is_proc);
if (!iseq) {
return rb_unnamed_parameters(rb_proc_arity(self));
}
return rb_iseq_parameters(iseq, is_proc);
}
Returns the parameter information of this proc.
prc = lambda{|x, y=42, *other|} prc.parameters #=> [[:req, :x], [:opt, :y], [:rest, :other]]
VALUE
rb_proc_location(VALUE self)
{
return iseq_location(rb_proc_get_iseq(self, 0));
}
Returns the Ruby source filename and line number containing this proc or nil if this proc was not defined in Ruby (i.e. native).
static VALUE
proc_to_proc(VALUE self)
{
return self;
}
Part of the protocol for converting objects to Proc objects. Instances of class Proc simply return themselves.
static VALUE
proc_to_s(VALUE self)
{
const rb_proc_t *proc;
GetProcPtr(self, proc);
return rb_block_to_s(self, &proc->block, proc->is_lambda ? " (lambda)" : NULL);
}
Returns the unique identifier for this proc, along with an indication of where the proc was defined.