Your assumption isn't really valid.

Perhaps this kind of array should be printed like a dictionary, but it can be misleading as well.

> a = [1,2,3]; delete a[1]; a.toString()
'1,,3'

What do you prefer?

I can see why the output would be misleading.

[ 1, 3 ] suggests that the index of 3 is 1, but arr[1] is undefined. arr[2] is 3 even after deleting a preceding value.

And this is why we don't use delete for arrays, we us Array.slice()

deleting items from array lead to misleading results ;-)

That's because deleting is improper on an array. :P

Sent from my shiny glass rectangle

On Sep 5, 2011, at 12:43 AM, [email protected] wrote:

deleting items from array lead to misleading results ;-)

Reply to this email directly or view it on GitHub:
#1651 (comment)

True, but there should be some way of printing this kinds of arrays.
I'm +1 on @bnoordhuis' solution. We could make it look like [ 1, , 3 ] after util.format.

@mmalecki this would be misleading as well, I would say it'll be best to display it with all properties named then:
[ '0': 1, '2': 3 ]

@medikoo These are not string indexes, so it should be: [ 0: 1, 2: 3 ], I think. But yes, it looks good (pretty hard to mislead with dictionary).
I will try to provide patch for this one.

I think a dictionary misleads people into thinking it's a dictionary.

Also +1 on @bnoordhuis solution

In both cases I've originally mentioned, it holds that:

• list.length == 3
• list[1] === undefined

For me, this looks like a sufficient reason to consider [ 1, undefined, 3 ] the valid and most intuitive output in both originally mentioned cases.
Is there ANY practical reason to distinguish between them?

[ 1, undefined, 3 ] would be inconsistent, as it may show same for two different arrays:

var a = [1, undefined, 3];
console.log(1 in a); // true

var b = [1, 2, 3];
delete b[1];
console.log(1 in b); // false

Currently when we declare named property on array, we have output like following:

var a = [1, 2, 3];
a.foo = 'bar';
console.log(a); // [ 1, 2, 3, foo: 'bar' ]

That is great.

So I guess best would be to do same for "disconnected" indexes:

var a = [1, 2, 3, 4];
delete a[2];
console.log(a); // [1, 2, 4: 4]

Either way it's icky.

On Sep 6, 2011, at 1:14 AM, [email protected] wrote:

[ 1, undefined, 3 ] would be inconsistent, as it may show same for two different arrays:

var a = [1, undefined, 3];
console.log(1 in a); // true

var b = [1, 2, 3];
delete b[1];
console.log(1 in b); // false

Currently when we declare named property on array, we have output like following:

var a = [1, 2, 3];
a.foo = 'bar';
console.log(a); // [ 1, 2, 3, foo: 'bar' ]

That is great.

So I guess best would be to do same for "disconnected" indexes:

var a = [1, 2, 3, 4];
delete a[2];
console.log(a); // [1, 2, 4: 4]

Reply to this email directly or view it on GitHub:
#1651 (comment)

How about c6fe2b3? (This is a workaround for a workaround...)

> a = [1, 2, 3]
[ 1, 2, 3 ]
> delete a[1]
true
> a
[ 1, , 3 ]
> require('util').inspect(a, true)
'[ 1, , 3, [length]: 3 ]'

But... unfortunately:

> new Array(10)
[ , , , , , , , , ,  ]

(´・ω・`)

I approve of that behavior.

On Sep 6, 2011, at 8:31 AM, [email protected] wrote:

How about c6fe2b3? (This is a workaround for a workaround...)

> a = [1, 2, 3]
[ 1, 2, 3 ]
> delete a[1]
true
> a
[ 1, , 3 ]
> require('util').inspect(a, true)
'[ 1, , 3, [length]: 3 ]'

But... unfortunately:

> new Array(10)
[ , , , , , , , , ,  ]

(´・ω・`)

Reply to this email directly or view it on GitHub:
#1651 (comment)

@koichik +1

@NuckChorris, @TooTallNate - Thanks.
@bnoordhuis - Can you review c6fe2b3?

@koichik - LGTM but I think it's time we start breaking up that 200 line inspect function, it's getting unwieldy.

@bnoordhuis - I agree. I will extract the anonymous function that is parameter of keys.map().

@bnoordhuis - I have just divided util.inspect() into some subroutines.
Please review b86210f and a39c639.

@koichik - Love it. There should be two newlines between functions but otherwise LGTM.

There should be two newlines between functions

Done. Thanks!