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

Groups > comp.lang.javascript > #32524 > unrolled thread

removeEventListener why

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

Contents

  removeEventListener why Andrew Poulos <ap_prog@hotmail.com> - 2017-02-07 11:46 +1100
    Re: removeEventListener why "Julio P. Di Egidio" <julio@diegidio.name> - 2017-02-07 05:51 +0100
      Re: removeEventListener why "Evertjan." <exxjxw.hannivoort@inter.nl.net> - 2017-02-07 11:41 +0100
        Re: removeEventListener why Julio Di Egidio <julio@diegidio.name> - 2017-02-07 03:00 -0800
          Re: removeEventListener why "Evertjan." <exxjxw.hannivoort@inter.nl.net> - 2017-02-07 12:44 +0100
        Re: removeEventListener why JJ <jj4public@vfemail.net> - 2017-02-07 22:50 +0700
          Re: removeEventListener why "Evertjan." <exxjxw.hannivoort@inter.nl.net> - 2017-02-07 18:17 +0100
            Re: removeEventListener why JJ <jj4public@vfemail.net> - 2017-02-08 23:11 +0700
              Re: removeEventListener why "Evertjan." <exxjxw.hannivoort@inter.nl.net> - 2017-02-08 19:39 +0100
          Re: removeEventListener why Thomas 'PointedEars' Lahn <PointedEars@web.de> - 2017-02-09 02:02 +0100
            Re: removeEventListener why Julio Di Egidio <julio@diegidio.name> - 2017-02-08 20:33 -0800
              Re: removeEventListener why "Evertjan." <exxjxw.hannivoort@inter.nl.net> - 2017-02-09 11:51 +0100
              Re: removeEventListener why Thomas 'PointedEars' Lahn <PointedEars@web.de> - 2017-02-09 12:01 +0100
            Re: removeEventListener why JJ <jj4public@vfemail.net> - 2017-02-09 18:36 +0700
              Re: removeEventListener why Thomas 'PointedEars' Lahn <PointedEars@web.de> - 2017-02-09 13:09 +0100
    Re: removeEventListener why "Michael Haufe (TNO)" <tno@thenewobjective.com> - 2017-02-07 06:32 -0800
      Re: removeEventListener why Thomas 'PointedEars' Lahn <PointedEars@web.de> - 2017-02-07 17:31 +0100
        Re: removeEventListener why Julio Di Egidio <julio@diegidio.name> - 2017-02-07 11:59 -0800
          Re: removeEventListener why Thomas 'PointedEars' Lahn <PointedEars@web.de> - 2017-02-07 21:55 +0100
            Re: removeEventListener why Julio Di Egidio <julio@diegidio.name> - 2017-02-07 13:23 -0800
              Re: removeEventListener why "Michael Haufe (TNO)" <tno@thenewobjective.com> - 2017-02-08 06:03 -0800
        Re: removeEventListener why "Michael Haufe (TNO)" <tno@thenewobjective.com> - 2017-02-08 05:54 -0800

Page 1 of 2  [1] 2  Next page →

#32524 — removeEventListener why

If I use addEventListener on an element in the page should I then use 
removeEventListener before the page is unloaded?

Is it good practice to remove all listeners I add or does it make no 
difference? I'm usually careful about removing all listeners I add but I 
don't know if today it serves any real purpose.

In the past I used to be worried about memory leaks. Is it still an issue?

Andrew Poulos

[toc] | [next] | [standalone]

#32526

"Andrew Poulos"  wrote in message 
news:hbCdnUYE2rVRhQTFnZ2dnUU7-dvNnZ2d@westnet.com.au...
>
> If I use addEventListener on an element in the page should I then use 
> removeEventListener before the page is unloaded?
>
> Is it good practice to remove all listeners I add or does it make no 
> difference? I'm usually careful about removing all listeners I add but I 
> don't know if today it serves any real purpose.
>
> In the past I used to be worried about memory leaks. Is it still an issue?

Yes, you should destroy all event handlers before the page is unloaded, 
browsers may, and AFAICT still do, leak memory otherwise.  There was an 
article explaining the details which I cannot find now: it was about 
orphaned references not getting automatically released... or something along 
that line.

HTH,

Julio

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

#32539

"Julio P. Di Egidio" <julio@diegidio.name> wrote on 07 Feb 2017 in
comp.lang.javascript: 

> "Andrew Poulos"  wrote in message 
> news:hbCdnUYE2rVRhQTFnZ2dnUU7-dvNnZ2d@westnet.com.au...
>>
>> If I use addEventListener on an element in the page should I then use 
>> removeEventListener before the page is unloaded?
>>
>> Is it good practice to remove all listeners I add or does it make no 
>> difference? I'm usually careful about removing all listeners I add but
>> I don't know if today it serves any real purpose.
>>
>> In the past I used to be worried about memory leaks. Is it still an
>> issue? 
> 
> Yes, you should destroy all event handlers before the page is unloaded, 
> browsers may, and AFAICT still do, leak memory otherwise.  There was an 
> article explaining the details which I cannot find now: it was about 
> orphaned references not getting automatically released... or something
> along that line.

No, you should not.

Any 'page' has so many listenters by default that you trust are destroyed 
when you leave the page, you could as well suspect a modern browser not 
keeping good care for that. btw, should you reset all variables before you 
exit a page too?

Memory-leakage is a task for the [individual browser]-maker to worry about, 
not for the page programmer.

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

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

#32540

On Tuesday, February 7, 2017 at 11:41:34 AM UTC+1, Evertjan. wrote:
> "Julio P. Di Egidio" <julio@diegidio.name> wrote on 07 Feb 2017 in
> comp.lang.javascript: 
> 
> > "Andrew Poulos"  wrote in message 
> > news:hbCdnUYE2rVRhQTFnZ2dnUU7-dvNnZ2d@westnet.com.au...
> >>
> >> If I use addEventListener on an element in the page should I then use 
> >> removeEventListener before the page is unloaded?
> >>
> >> Is it good practice to remove all listeners I add or does it make no 
> >> difference? I'm usually careful about removing all listeners I add but
> >> I don't know if today it serves any real purpose.
> >>
> >> In the past I used to be worried about memory leaks. Is it still an
> >> issue? 
> > 
> > Yes, you should destroy all event handlers before the page is unloaded, 
> > browsers may, and AFAICT still do, leak memory otherwise.  There was an 
> > article explaining the details which I cannot find now: it was about 
> > orphaned references not getting automatically released... or something
> > along that line.
> 
> No, you should not.

A professional programmer definitely should.

> Any 'page' has so many listenters by default

And you should rightly expect that *those* are automatically released.
As I have hinted at above, there are reasons why to today browsers cannot
do better than that, i.e. cannot do better than leave the programmer to care
for his or her own code.

That may be unfortunate, I do not really disagree, but in order to write pages
that do not leak it is a matter of fact that one *has to* destroy event-related
resources explicitly.   Note that this is not just about preventing the browser
overall from licking, so a matter of overall user experience, which you might
even make a point not to care about, it is also about writing heavily scripted
groups of apps and pages and similar, i.e. sometimes not caring is not even an
option, depending on the complexity of an application.

HTH,

Julio

> that you trust are destroyed 
> when you leave the page, you could as well suspect a modern browser not 
> keeping good care for that. btw, should you reset all variables before you 
> exit a page too?
> 
> Memory-leakage is a task for the [individual browser]-maker to worry about, 
> not for the page programmer.

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

#32542

Julio Di Egidio <julio@diegidio.name> wrote on 07 Feb 2017 in
comp.lang.javascript: 

> On Tuesday, February 7, 2017 at 11:41:34 AM UTC+1, Evertjan. wrote:
>> "Julio P. Di Egidio" <julio@diegidio.name> wrote on 07 Feb 2017 in
>> comp.lang.javascript: 
>> 
>> > "Andrew Poulos"  wrote in message 
>> > news:hbCdnUYE2rVRhQTFnZ2dnUU7-dvNnZ2d@westnet.com.au...
>> >>
>> >> If I use addEventListener on an element in the page should I then
>> >> use removeEventListener before the page is unloaded?
>> >>
>> >> Is it good practice to remove all listeners I add or does it make no
>> >> difference? I'm usually careful about removing all listeners I add
>> >> but I don't know if today it serves any real purpose.
>> >>
>> >> In the past I used to be worried about memory leaks. Is it still an
>> >> issue? 
>> > 
>> > Yes, you should destroy all event handlers before the page is
>> > unloaded, browsers may, and AFAICT still do, leak memory otherwise. 
>> > There was an article explaining the details which I cannot find now:
>> > it was about orphaned references not getting automatically
>> > released... or something along that line.
>> 
>> No, you should not.
> 
> A professional programmer definitely should.

This is nonsense.

As a professional programmer, 
I should not go on hearsay,
but go for facts and numbers.

Do you have facts to which browsers are so disfunctional,
and how much memory is lost for each leaking listener in such browser?

Would it take 10 minutes, 10 years or 1000 years for such a defunctional 
browser regularily accessing such web-pages to loose an important part of 
its memory-pool?

>> Any 'page' has so many listenters by default
> 
> And you should rightly expect that *those* are automatically released.

Again, just hearsay, why are you so trusting here?

> As I have hinted at above, there are reasons why to today browsers
> cannot do better than that, i.e. cannot do better than leave the
> programmer to care for his or her own code.

Oh yes, what are those reasons? 

Please elaborate.

> That may be unfortunate, I do not really disagree, but in order to write
> pages that do not leak it is a matter of fact that one *has to* destroy
> event-related resources explicitly. 

How do you know? 

I don't say you don't know, but you should elaborate instead of 
having us [sorry: me] believe you you at unseeen face-value.

> Note that this is not just about
> preventing the browser overall from licking, so a matter of overall user
> experience, 

Why? The only limiting factor in user-experience would be craching due to 
memory pool loss. So please elaborate.

> which you might even make a point not to care about, it is
> also about writing heavily scripted groups of apps and pages and
> similar, i.e. sometimes not caring is not even an option, depending on
> the complexity of an application. 

Ah, authority experience.

PS: I did not say above I was a professional programmer,
nor that I was not.

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

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

#32550

On Tue, 07 Feb 2017 11:41:34 +0100, Evertjan. wrote:
> 
> No, you should not.
> 
> Any 'page' has so many listenters by default that you trust are destroyed 
> when you leave the page, you could as well suspect a modern browser not 
> keeping good care for that. btw, should you reset all variables before you 
> exit a page too?
> 
> Memory-leakage is a task for the [individual browser]-maker to worry about, 
> not for the page programmer.

This won't work when developing some kind of HTML demo which can be
stressful such as HTML based particle demo.

e.g. the script creates many elements and listen to them. Before presenting
the next set of demo, if it simply remove the elements without removing the
listeners, the garbage will fill up eventually. When the browser finally
collect the garbage, it may interfere the progress of the current demo.

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

#32553

JJ <jj4public@vfemail.net> wrote on 07 Feb 2017 in comp.lang.javascript:

> On Tue, 07 Feb 2017 11:41:34 +0100, Evertjan. wrote:
>> 
>> No, you should not.
>> 
>> Any 'page' has so many listenters by default that you trust are
>> destroyed when you leave the page, you could as well suspect a modern
>> browser not keeping good care for that. btw, should you reset all
>> variables before you exit a page too?
>> 
>> Memory-leakage is a task for the [individual browser]-maker to worry
>> about, not for the page programmer.
> 
> This won't work when developing some kind of HTML demo which can be
> stressful such as HTML based particle demo.
> 
> e.g. the script creates many elements and listen to them. Before
> presenting the next set of demo, if it simply remove the elements
> without removing the listeners, the garbage will fill up eventually.
> When the browser finally collect the garbage, it may interfere the
> progress of the current demo. 

Different problem.

We were discussing removing the listeners when the page is unloaded.

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

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

#32565

On Tue, 07 Feb 2017 18:17:56 +0100, Evertjan. wrote:
> 
> Different problem.
> 
> We were discussing removing the listeners when the page is unloaded.

Anything which aren't cleaned up, are garbage.

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

#32566

JJ <jj4public@vfemail.net> wrote on 08 Feb 2017 in comp.lang.javascript:

> On Tue, 07 Feb 2017 18:17:56 +0100, Evertjan. wrote:
>> 
>> Different problem.
>> 
>> We were discussing removing the listeners when the page is unloaded.
> 
> Anything which aren't cleaned up, are garbage.

Wrong grammar, still just a tautology.

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

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

#32568

JJ wrote:

> On Tue, 07 Feb 2017 11:41:34 +0100, Evertjan. wrote:
>> No, you should not.
>> 
>> Any 'page' has so many listenters by default that you trust are destroyed
>> when you leave the page, you could as well suspect a modern browser not
>> keeping good care for that. btw, should you reset all variables before
>> you exit a page too?
>> 
>> Memory-leakage is a task for the [individual browser]-maker to worry
>> about, not for the page programmer.
> 
> This won't work when developing some kind of HTML demo which can be
> stressful such as HTML based particle demo.
> 
> e.g. the script creates many elements and listen to them. Before
> presenting the next set of demo, if it simply remove the elements without
> removing the listeners, the garbage will fill up eventually. When the
> browser finally collect the garbage, it may interfere the progress of the
> current demo.

You are grasping at straws here.

1. The HTML demo, if professionally done, would not attach event listeners 
   to each element, but delegate them to an upper level, probably the 
   document, based on the properties of the element.

2. The HTML demo would not add elements and event listeners on the fly,
   because modifying the document tree extensively after the document
   has been loaded incurs a performance penalty and a risk of failure
   that a presenter cannot afford.

3. The HTML demo would be for a sufficiently recent user agent in which
   this bug has been fixed.

4. There is no good reason to assume that an event listener for a hidden
   element would be interfering with one for a displayed one.

Your example is unrealistic, to say the least.

-- 
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]

#32570

On Thursday, February 9, 2017 at 2:02:56 AM UTC+1, Thomas 'PointedEars' Lahn wrote:
> JJ wrote:
> > On Tue, 07 Feb 2017 11:41:34 +0100, Evertjan. wrote:
> >> No, you should not.
> >> 
> >> Any 'page' has so many listenters by default that you trust are destroyed
> >> when you leave the page, you could as well suspect a modern browser not
> >> keeping good care for that. btw, should you reset all variables before
> >> you exit a page too?
> >> 
> >> Memory-leakage is a task for the [individual browser]-maker to worry
> >> about, not for the page programmer.
> > 
> > This won't work when developing some kind of HTML demo which can be
> > stressful such as HTML based particle demo.
> > 
> > e.g. the script creates many elements and listen to them. Before
> > presenting the next set of demo, if it simply remove the elements without
> > removing the listeners, the garbage will fill up eventually. When the
> > browser finally collect the garbage, it may interfere the progress of the
> > current demo.
> 
> You are grasping at straws here.

I can confirm he isn't: apparently, you are simply yet another dick head...

*Plonk*

Julio

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

#32573

Julio Di Egidio <julio@diegidio.name> wrote on 09 Feb 2017 in 
comp.lang.javascript:

>> You are grasping at straws here.
> 
> I can confirm he isn't: apparently, 

Sorry, you cannot reliably confirm something by just saying so.

> you are simply yet another dick head...

... even if you resort to ad hominem attacks.



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

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

#32574

Julio Di Egidio trolled *via Google Groups*:

> *Plonk*

Funny, I was going to post the same, but in my case it actually *works*.

F'up2 poster

-- 
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]

#32576

On Thu, 09 Feb 2017 02:02:50 +0100, Thomas 'PointedEars' Lahn wrote:

> JJ wrote:
> 
>> On Tue, 07 Feb 2017 11:41:34 +0100, Evertjan. wrote:
>>> No, you should not.
>>> 
>>> Any 'page' has so many listenters by default that you trust are destroyed
>>> when you leave the page, you could as well suspect a modern browser not
>>> keeping good care for that. btw, should you reset all variables before
>>> you exit a page too?
>>> 
>>> Memory-leakage is a task for the [individual browser]-maker to worry
>>> about, not for the page programmer.
>> 
>> This won't work when developing some kind of HTML demo which can be
>> stressful such as HTML based particle demo.
>> 
>> e.g. the script creates many elements and listen to them. Before
>> presenting the next set of demo, if it simply remove the elements without
>> removing the listeners, the garbage will fill up eventually. When the
>> browser finally collect the garbage, it may interfere the progress of the
>> current demo.
> 
> You are grasping at straws here.
> 
> 1. The HTML demo, if professionally done, would not attach event listeners 
>    to each element, but delegate them to an upper level, probably the 
>    document, based on the properties of the element.
> 
> 2. The HTML demo would not add elements and event listeners on the fly,
>    because modifying the document tree extensively after the document
>    has been loaded incurs a performance penalty and a risk of failure
>    that a presenter cannot afford.
> 
> 3. The HTML demo would be for a sufficiently recent user agent in which
>    this bug has been fixed.
> 
> 4. There is no good reason to assume that an event listener for a hidden
>    element would be interfering with one for a displayed one.
> 
> Your example is unrealistic, to say the least.

No. You simply have fast system with abundant resources, and not aware of
the problem.

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

#32578

JJ wrote:

> On Thu, 09 Feb 2017 02:02:50 +0100, Thomas 'PointedEars' Lahn wrote:
>> Your example is unrealistic, to say the least.
> 
> No. You simply have fast system with abundant resources, and not aware of
> the problem.

Your reply has nothing to do with what I wrote.  I did not say that there 
could not be a problem, that there could not be memory leaks.

I said that the situation that could produce them that *you* describe (an 
HTML demo working *this* way) is unrealistic.  IOW, if  you would not work 
so hard to *produce* the problem in your example due to *incompetence*, 
there would not be one in the first place.  

And I hope that you will agree that if one does something stupid which 
causes bad things to happen, it is not the bad things that are to blame for 
this but the stupid actions that lead to them, and the person doing them.

-- 
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]

#32547

On Monday, February 6, 2017 at 6:46:14 PM UTC-6, Andrew Poulos wrote:
> If I use addEventListener on an element in the page should I then use 
> removeEventListener before the page is unloaded?
> 
> Is it good practice to remove all listeners I add or does it make no 
> difference? I'm usually careful about removing all listeners I add but I 
> don't know if today it serves any real purpose.
> 
> In the past I used to be worried about memory leaks. Is it still an issue?

Doesn't matter anymore on modern browsers. Beware of doing things with an <object/> <embed /> or <applet /> tag though if you have them. God knows what happens when you interact with browser plugins with a bridge.

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

#32552

Michael Haufe (TNO) wrote:

> On Monday, February 6, 2017 at 6:46:14 PM UTC-6, Andrew Poulos wrote:
>> Is it good practice to remove all listeners I add or does it make no
>> difference? I'm usually careful about removing all listeners I add but I
>> don't know if today it serves any real purpose.
>> 
>> In the past I used to be worried about memory leaks. Is it still an
>> issue?
> 
> Doesn't matter anymore on modern browsers.

True; however, it is a good rule of thumb.  Backwards-compatibility can be 
important; I would think that it is in Andrew’s case (IIRC, he is developing 
e-learning applications based on Moodle).

In which browser/layout engine versions is this issue resolved?

> Beware of doing things with an <object/> <embed /> or <applet /> tag
> though if you have them. God knows what happens when you interact with
> browser plugins with a bridge.

ACK.


PointedEars
-- 
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]

#32555

On Tuesday, February 7, 2017 at 5:31:07 PM UTC+1, Thomas 'PointedEars' Lahn wrote:
> Michael Haufe (TNO) wrote:
> > On Monday, February 6, 2017 at 6:46:14 PM UTC-6, Andrew Poulos wrote:
> >> Is it good practice to remove all listeners I add or does it make no
> >> difference? I'm usually careful about removing all listeners I add but I
> >> don't know if today it serves any real purpose.
> >> 
> >> In the past I used to be worried about memory leaks. Is it still an
> >> issue?
> > 
> > Doesn't matter anymore on modern browsers.
> 
> True;

It is not true, even on the latest browsers...

Julio

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

#32557

Julio Di Egidio wrote:

> […] Thomas 'PointedEars' […] Lahn wrote:
>> Michael Haufe (TNO) wrote:
>> > On Monday, February 6, 2017 at 6:46:14 PM UTC-6, Andrew Poulos wrote:
>> >> Is it good practice to remove all listeners I add or does it make no
>> >> difference? I'm usually careful about removing all listeners I add but
>> >> I don't know if today it serves any real purpose.
>> >> 
>> >> In the past I used to be worried about memory leaks. Is it still an
>> >> issue?
>> > Doesn't matter anymore on modern browsers.
>> True;
> 
> It is not true, even on the latest browsers...

Interesting.  Cite evidence.

-- 
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]

#32558

On Tuesday, February 7, 2017 at 9:55:32 PM UTC+1, Thomas 'PointedEars' Lahn wrote:
> Julio Di Egidio wrote:
> > […] Thomas 'PointedEars' […] Lahn wrote:
> >> Michael Haufe (TNO) wrote:
> >> > On Monday, February 6, 2017 at 6:46:14 PM UTC-6, Andrew Poulos wrote:
> >> >> Is it good practice to remove all listeners I add or does it make no
> >> >> difference? I'm usually careful about removing all listeners I add but
> >> >> I don't know if today it serves any real purpose.
> >> >> 
> >> >> In the past I used to be worried about memory leaks. Is it still an
> >> >> issue?
> >> > Doesn't matter anymore on modern browsers.
> >> True;
> > 
> > It is not true, even on the latest browsers...
> 
> Interesting.  Cite evidence.

If your daily browsing experience isn't enough to convince you, this needs a little test bench, which I might put together as soon as I manage, this may indeed be of general interest.  Of course, feel free to beat me to it.

Julio

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

Page 1 of 2  [1] 2  Next page →

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

csiph-web