Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.javascript > #32363 > unrolled thread

understanding the expression "ws && ~[2, 3].indexOf(ws.readyState)"

Back to article view | Back to comp.lang.javascript

Contents

  understanding the expression "ws && ~[2, 3].indexOf(ws.readyState)" Une Bévue <unbewusst.sein@fai.invalid> - 2017-01-27 07:06 +0100
    Re: understanding the expression "ws && ~[2, 3].indexOf(ws.readyState)" Ben Bacarisse <ben.usenet@bsb.me.uk> - 2017-01-27 11:36 +0000
      Re: understanding the expression "ws && ~[2, 3].indexOf(ws.readyState)" Thomas 'PointedEars' Lahn <PointedEars@web.de> - 2017-01-27 13:04 +0100
      Re: understanding the expression "ws && ~[2, 3].indexOf(ws.readyState)" Une Bévue <unbewusst.sein@fai.invalid> - 2017-01-27 18:00 +0100
        Re: understanding the expression "ws && ~[2, 3].indexOf(ws.readyState)" "Evertjan." <exxjxw.hannivoort@inter.nl.net> - 2017-01-28 10:41 +0100
          Re: understanding the expression "ws && ~[2, 3].indexOf(ws.readyState)" Ben Bacarisse <ben.usenet@bsb.me.uk> - 2017-01-28 15:33 +0000
      Re: understanding the expression "ws && ~[2, 3].indexOf(ws.readyState)" Une Bévue <unbewusst.sein@fai.invalid> - 2017-01-29 09:12 +0100

#32363 — understanding the expression "ws && ~[2, 3].indexOf(ws.readyState)"

The expression :
ws && ~[2, 3].indexOf(ws.readyState)

is within a test :

     if (ws && ~[2, 3].indexOf(ws.readyState)) {
       connected = false;
       init(openCallback, closeCallback);
     }

ws being an instance of WebSocket created by :
ws = new WebSocket(url);

when WebSocket is ready the value of ws.readyState is 1.

the test "if (ws && ~[2, 3].indexOf(ws.readyState)) {"

re-init a connection to the websocket when ws.readyState is not equal to 1

(as far as i understand well)

i did some tests :

~[2, 3].indexOf(1); // gives 0
~[2, 3].indexOf(2); // gives -1
~[2, 3].indexOf(3); // gives -2

~ is a bitwise NOT operator, for example :
~ 5 being ~0101 gives 1010

(in fact the numbers aren't 4 bits wide rather 32 bits in JavaScript)

I'd like to decompose the above expression :

ws && ~[2, 3].indexOf(ws.readyState)

ws is a kind of existance test "true" when ws is not null nor undefined ;

~[2, 3].indexOf(ws.readyState) equals 0 (logical false) when WebSocket 
is in ready state.

then clearly the test is false when ws.readyState = 1, no attempt to 
reconnect using init(...).

however when ws.readyState is of 2 or 3 what is the logical value of -1 
or -2 ?

does negative numbers have an equivalent logical value of false ?

[toc] | [next] | [standalone]

#32374

Une Bévue <unbewusst.sein@fai.invalid> writes:

> The expression :
> ws && ~[2, 3].indexOf(ws.readyState)
<snip>
> i did some tests :
>
> ~[2, 3].indexOf(1); // gives 0
> ~[2, 3].indexOf(2); // gives -1
> ~[2, 3].indexOf(3); // gives -2
<snip>
> ~[2, 3].indexOf(ws.readyState) equals 0 (logical false) when WebSocket
> is in ready state.
>
> then clearly the test is false when ws.readyState = 1, no attempt to
> reconnect using init(...).
>
> however when ws.readyState is of 2 or 3 what is the logical value of
> -1 
> or -2 ?
>
> does negative numbers have an equivalent logical value of false ?

Any non-zero number (other than NaN) is considered to be true in a
Boolean context.

The test ~[2, 3].indexOf(n) is a tricky way of testing if n is 2 or 3.
All the valid indexes (0 and 1) have bit-negations that are non-zero and
hence "true".  The return that indicates that n is not present (-1)
bit-negates to 0 which is taken to be false.

I'd write n == 2 || n == 3 but if there were lots of values to consider
I'd probably write [2, 3, 8, 9, 77, 102].indexOf(n) >= 0 instead.

-- 
Ben.

[toc] | [prev] | [next] | [standalone]

#32376

Ben Bacarisse wrote:

> The test ~[2, 3].indexOf(n) is a tricky way of testing if n is 2 or 3.
> All the valid indexes (0 and 1) have bit-negations that are non-zero and
> hence "true".  The return that indicates that n is not present (-1)
> bit-negates to 0 which is taken to be false.
> 
> I'd write n == 2 || n == 3

Keep in mind that you need to write “===” for equivalence to .indexOf().

> but if there were lots of values to consider I'd probably write [2, 3, 8,
> 9, 77, 102].indexOf(n) >= 0 instead.

For few values, and if they are Number values with proper string 
representation, you can also write

  n in {2: 1, 3: 1}

instead (“1” serving as the shortest character sequence for a property 
value; any value works).

“||” is most backwards-compatible, but requires the most space.

With String values, you have to create the object with Object.create(null) 
(or equivalent) unless you can be sure that names of built-in properties of 
Object instances and their prototype could never be matched.  However, both 
Object.create() and Array.prototype.indexOf() were introduced in ECMAScript 
Ed 5., and you can emulate both to support the desired behavior.  As of 
ECMAScript 2016, you can also use Map instances to avoid name collisions and 
support non-string values (as keys).

Either approach is useful if you need to test repeatedly whether a value is 
contained in a set of non-unique values, where repeated linear search is 
comparably inefficient.

-- 
PointedEars
FAQ: <http://PointedEars.de/faq> | <http://PointedEars.de/es-matrix>
<https://github.com/PointedEars> | <http://PointedEars.de/wsvn/>
Twitter: @PointedEars2 | Please do not cc me./Bitte keine Kopien per E-Mail.

[toc] | [prev] | [next] | [standalone]

#32386

Le 27/01/2017 à 12:36, Ben Bacarisse a écrit :
> Any non-zero number (other than NaN) is considered to be true in a
> Boolean context.

ok, fine, it's what I expected.

> The test ~[2, 3].indexOf(n) is a tricky way of testing if n is 2 or 3.
> All the valid indexes (0 and 1) have bit-negations that are non-zero and
> hence "true".  The return that indicates that n is not present (-1)
> bit-negates to 0 which is taken to be false.

ok, thanks clear enough.

> I'd write n == 2 || n == 3 but if there were lots of values to consider
> I'd probably write [2, 3, 8, 9, 77, 102].indexOf(n) >= 0 instead.

Yes, sure, then I've to investigate what could be the different values 
of #readyState.

[toc] | [prev] | [next] | [standalone]

#32401

=?UTF-8?Q?Une_B=c3=a9vue?= <unbewusst.sein@fai.invalid> wrote on 27 Jan 2017 
in comp.lang.javascript:

> Le 27/01/2017 … 12:36, Ben Bacarisse a ‚crit :
>> Any non-zero number (other than NaN) is considered to be true in a
>> Boolean context.

Please reconsider:

alert(0 == false) // true
alert(0 === false) // false

alert(17 == true) // false ???????????
alert(true == 17) // false ???????????
alert(!!17) // true
 
> ok, fine, it's what I expected.

-- 
Evertjan.
The Netherlands.
(Please change the x'es to dots in my emailaddress)

[toc] | [prev] | [next] | [standalone]

#32407

"Evertjan." <exxjxw.hannivoort@inter.nl.net> writes:
<snip>
>> Le 27/01/2017 … 12:36, Ben Bacarisse a ‚crit :
>>> Any non-zero number (other than NaN) is considered to be true in a
>>> Boolean context.
>
> Please reconsider:

What should I be re-considering?

> alert(0 == false) // true
> alert(0 === false) // false
>
> alert(17 == true) // false ???????????
> alert(true == 17) // false ???????????
> alert(!!17) // true

There is only one example of a non-zero number in a Boolean context
there and it behaves as I described.

Maybe I should have been explicit about what I meant by a Boolean
context.  It's where the abstract toBoolean operation is used in the
language specification.

<snip>
-- 
Ben.

[toc] | [prev] | [next] | [standalone]

#32411

Le 27/01/2017 à 12:36, Ben Bacarisse a écrit :
> I'd write n == 2 || n == 3 but if there were lots of values to consider
> I'd probably write [2, 3, 8, 9, 77, 102].indexOf(n) >= 0 instead.

According to 
<https://developer.mozilla.org/en-US/docs/Web/API/WebSocket> the values 
of readyState are :

Constant 	Value 	Description
CONNECTING 	0 	The connection is not yet open.
OPEN 		1 	The connection is open and ready to communicate.
CLOSING 	2 	The connection is in the process of closing.
CLOSED 		3 	The connection is closed or couldn't be opened.

Then I only test #readyState verssus 1...

[toc] | [prev] | [standalone]

Back to top | Article view | comp.lang.javascript

csiph-web