builtins.md

Builtin Functions

format

Returns a formatted string. The first argument must be a String object. See this for more details on formatting.

a := [1, 2, 3]
s := format("Foo: %v", a) // s == "Foo: [1, 2, 3]"

len

Returns the number of elements if the given variable is array, string, map, or module map.

v := [1, 2, 3]
l := len(v) // l == 3

copy

Creates a copy of the given variable. copy function calls Object.Copy interface method, which is expected to return a deep-copy of the value it holds.

v1 := [1, 2, 3]
v2 := v1
v3 := copy(v1)
v1[1] = 0
print(v2[1]) // "0"; 'v1' and 'v2' referencing the same array
print(v3[1]) // "2"; 'v3' not affected by 'v1'

append

Appends object(s) to an array (first argument) and returns a new array object. (Like Go's append builtin.) Currently, this function takes array type only.

v := [1]
v = append(v, 2, 3) // v == [1, 2, 3]

delete

Deletes the element with the specified key from the map type. First argument must be a map type and second argument must be a string type. (Like Go's delete builtin except keys are always string). delete returns undefined value if successful and it mutates given map.

v := {key: "value"}
delete(v, "key") // v == {}

v := {key: "value"}
delete(v, "missing") // v == {"key": "value"}

delete({}) // runtime error, second argument is missing
delete({}, 1) // runtime error, second argument must be a string type

splice

Deletes and/or changes the contents of a given array and returns deleted items as a new array. splice is similar to JS Array.prototype.splice() except splice is a builtin function and first argument must an array. First argument must be an array, and if second and third arguments are provided those must be integers otherwise runtime error is returned.

Usage:

deleted_items := splice(array[, start[, delete_count[, item1[, item2[, ...]]]])

v := [1, 2, 3]
items := splice(v, 0) // items == [1, 2, 3], v == []

v := [1, 2, 3]
items := splice(v, 1) // items == [2, 3], v == [1]

v := [1, 2, 3]
items := splice(v, 0, 1) // items == [1], v == [2, 3]

// deleting
v := ["a", "b", "c"]
items := splice(v, 1, 2) // items == ["b", "c"], v == ["a"]
// splice(v, 1, 3) or splice(v, 1, 99) has same effect for this example

// appending
v := ["a", "b", "c"]
items := splice(v, 3, 0, "d", "e") // items == [], v == ["a", "b", "c", "d", "e"]

// replacing
v := ["a", "b", "c"]
items := splice(v, 2, 1, "d") // items == ["c"], v == ["a", "b", "d"]

// inserting
v := ["a", "b", "c"]
items := splice(v, 0, 0, "d", "e") // items == [], v == ["d", "e", "a", "b", "c"]

// deleting and inserting
v := ["a", "b", "c"]
items := splice(v, 1, 1, "d", "e") // items == ["b"], v == ["a", "d", "e", "c"]

go

Starts an independent concurrent goroutine which runs fn(arg1, arg2, ...)

If fn is CompiledFunction, the current running VM will be cloned to create a new VM in which the CompiledFunction will be running. The fn can also be any object that has Call() method, such as BuiltinFunction, in which case no cloned VM will be created. Returns a goroutineVM object that has wait, result, abort methods.

The goroutineVM will not exit unless:

1. All its descendant goroutineVMs exit
2. It calls abort()
3. Its goroutineVM object abort() is called on behalf of its parent VM The latter 2 cases will trigger aborting procedure of all the descendant goroutineVMs, which will further result in #1 above.

var := 0

f1 := func(a,b) { var = 10; return a+b }
f2 := func(a,b,c) { var = 11; return a+b+c }

gvm1 := go(f1,1,2)
gvm2 := go(f2,1,2,5)

fmt.println(gvm1.result()) // 3
fmt.println(gvm2.result()) // 8
fmt.println(var) // 10 or 11

• wait() waits for the goroutineVM to complete up to timeout seconds and returns true if the goroutineVM exited(successfully or not) within the timeout. It waits forever if the optional timeout not specified, or timeout < 0.
• abort() triggers the termination process of the goroutineVM and all its descendant VMs.
• result() waits the goroutineVM to complete, returns Error object if any runtime error occurred during the execution, otherwise returns the result value of fn(arg1, arg2, ...)

1 client 1 server

Below is a simple client server example:

reqChan := makechan(8)
repChan := makechan(8)

client := func(interval) {
	reqChan.send("hello")
	for i := 0; true; i++ {
		fmt.println(repChan.recv())
		times.sleep(interval*times.second)
		reqChan.send(i)
	}
}

server := func() {
	for {
		req := reqChan.recv()
		if req == "hello" {
			fmt.println(req)
			repChan.send("world")
		} else {
			repChan.send(req+100)
		}
	}
}

gClient := go(client, 2)
gServer := go(server)

if ok := gClient.wait(5); !ok {
	gClient.abort()
}
gServer.abort()

//output:
//hello
//world
//100
//101

n client n server, channel in channel

sharedReqChan := makechan(128)

client = func(name, interval, timeout) {
	print := func(s) {
		fmt.println(name, s)
	}
	print("started")

	repChan := makechan(1)
	msg := {chan:repChan}

	msg.data = "hello"
	sharedReqChan.send(msg)
	print(repChan.recv())

	for i := 0; i * interval < timeout; i++ {
		msg.data = i
		sharedReqChan.send(msg)
		print(repChan.recv())
		times.sleep(interval*times.second)
	}
}

server = func(name) {
	print := func(s) {
		fmt.println(name, s)
	}
	print("started")

	for {
		req := sharedReqChan.recv()
		if req.data == "hello" {
			req.chan.send("world")
		} else {
			req.chan.send(req.data+100)
		}
	}
}

clients := func() {
	for i :=0; i < 5; i++ {
		go(client, format("client %d: ", i), 1, 4)
	}
}

servers := func() {
	for i :=0; i < 2; i++ {
		go(server, format("server %d: ", i))
	}
}

// After 4 seconds, all clients should have exited normally
gclts := go(clients)
// If servers exit earlier than clients, then clients may be
// blocked forever waiting for the reply chan, because servers
// were aborted with the req fetched from sharedReqChan before
// sending back the reply.
// In such case, do below to abort() the clients manually
//go(func(){times.sleep(6*times.second); gclts.abort()})

// Servers are infinite loop, abort() them after 5 seconds
gsrvs := go(servers)
if ok := gsrvs.wait(5); !ok {
	gsrvs.abort()
}

// Main VM waits here until all the child "go" finish

// If somehow the main VM is stuck, that is because there is
// at least one child VM that has not exited as expected, we
// can do abort() to force exit.
abort()

//output:
//3
//8
//hello
//world
//100
//101

//unordered output:
//client 4: started
//server 0: started
//client 4: world
//client 4: 100
//client 3: started
//client 3: world
//client 3: 100
//client 2: started
//client 2: world
//client 2: 100
//client 0: started
//client 0: world
//client 0: 100
//client 1: started
//client 1: world
//client 1: 100
//server 1: started
//client 1: 101
//client 2: 101
//client 4: 101
//client 0: 101
//client 3: 101
//client 3: 102
//client 0: 102
//client 2: 102
//client 1: 102
//client 4: 102
//client 0: 103
//client 3: 103
//client 2: 103
//client 1: 103
//client 4: 103

abort

Triggers the termination process of the current VM and all its descendant VMs.

makechan

Makes a channel to send/receive object and returns a chan object that has send, recv, close methods.

unbufferedChan := makechan()
bufferedChan := makechan(128)

// Send will block if the channel is full.
bufferedChan.send("hello") // send string
bufferedChan.send(55) // send int
bufferedChan.send([66, makechan(1)]) // channel in channel

// Receive will block if the channel is empty.
obj := bufferedChan.recv()

// Send to a closed channel causes panic.
// Receive from a closed channel returns undefined value.
unbufferedChan.close()
bufferedChan.close()

On the time the VM that the chan is running in is aborted, the sending or receiving call returns immediately.

type_name

Returns the type_name of an object.

type_name(1) // int
type_name("str") // string
type_name([1, 2, 3]) // array

string

Tries to convert an object to string object. See Runtime Types for more details on type conversion.

x := string(123) //  x == "123"

Optionally it can take the second argument, which will be returned if the first argument cannot be converted to string. Note that the second argument does not have to be string.

v = string(undefined, "foo")  // v == "foo"
v = string(undefined, false)  // v == false

int

Tries to convert an object to int object. See this for more details on type conversion.

v := int("123") //  v == 123

Optionally it can take the second argument, which will be returned if the first argument cannot be converted to int. Note that the second argument does not have to be int.

v = int(undefined, 10)    // v == 10
v = int(undefined, false) // v == false

bool

Tries to convert an object to bool object. See this for more details on type conversion.

v := bool(1) //  v == true

float

Tries to convert an object to float object. See this for more details on type conversion.

v := float("19.84") //  v == 19.84

Optionally it can take the second argument, which will be returned if the first argument cannot be converted to float. Note that the second argument does not have to be float.

v = float(undefined, 19.84)    // v == 19.84
v = float(undefined, false)    // v == false

char

Tries to convert an object to char object. See this for more details on type conversion.

v := char(89) //  v == 'Y'

Optionally it can take the second argument, which will be returned if the first argument cannot be converted to float. Note that the second argument does not have to be float.

v = char(undefined, 'X')    // v == 'X'
v = char(undefined, false)  // v == false

bytes

Tries to convert an object to bytes object. See this for more details on type conversion.

v := bytes("foo") //  v == [102 111 111]

Optionally it can take the second argument, which will be returned if the first argument cannot be converted to float. Note that the second argument does not have to be float.

v = bytes(undefined, bytes("foo"))    // v == bytes("foo")
v = bytes(undefined, false)           // v == false

If you pass an int to bytes() function, it will create a new byte object with the given size.

v := bytes(100)

time

Tries to convert an object to time value.

v := time(1257894000) // 2009-11-10 23:00:00 +0000 UTC

is_string

Returns true if the object's type is string. Or it returns false.

is_int

Returns true if the object's type is int. Or it returns false.

is_bool

Returns true if the object's type is bool. Or it returns false.

is_float

Returns true if the object's type is float. Or it returns false.

is_char

Returns true if the object's type is char. Or it returns false.

is_bytes

Returns true if the object's type is bytes. Or it returns false.

is_error

Returns true if the object's type is error. Or it returns false.

is_undefined

Returns true if the object's type is undefined. Or it returns false.

is_function

Returns true if the object's type is function or closure. Or it returns false. Note that is_function returns false for builtin functions and user-provided callable objects.

is_callable

Returns true if the object is callable (e.g. function, closure, builtin function, or user-provided callable objects). Or it returns false.

is_array

Returns true if the object's type is array. Or it returns false.

is_immutable_array

Returns true if the object's type is immutable array. Or it returns false.

is_map

Returns true if the object's type is map. Or it returns false.

is_immutable_map

Returns true if the object's type is immutable map. Or it returns false.

is_iterable

Returns true if the object's type is iterable: array, immutable array, map, immutable map, string, and bytes are iterable types in Tengo.

is_time

Returns true if the object's type is time. Or it returns false.