thr3ads.net - Rails spinoffs - Weird Prototype behavior [Dec 2007]

If this information is useful, please help other people find it:
Share via:

Matt Alexander

2007-Dec-06 18:12 UTC

Weird Prototype behavior

In this code:

<html>
<head>
<title>PT test</title>
<script type=''text/javascript''
src=''/js/prototype/1.6.0/
prototype.js''></script>
<script type=''text/javascript''>
//<![CDATA[
Event.observe(window, ''load'', function () {
   alert($(''login_username''));
});
//]]>
</script>
</script>
</head>
<body>
<input type=''text'' name=''login_username''
id=''login_username_index'' />
</body>
</html>

IE6 alerts "[object]".
IE7 alerts "[object]".
Opera 9 alerts "[object HTMLInputElement]".
FF2 alerts "null".

If I change the name of the input element from "login_username" to
anything else, all browsers alert "null". Why is the dollar function
reading the name attribute instead of the id? The way I''ve always
understood it, the dollar function ignores the name attribute.
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group.
To post to this group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs?hl=en
-~----------~----~----~----~------~----~------~--~---

Dan Dorman

2007-Dec-06 18:27 UTC

head link

Re: Weird Prototype behavior

On Dec 6, 2007 11:12 AM, Matt Alexander
<mattalexxpub-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:> In this code:
> [SNIP]
> <input type=''text''
name=''login_username''
id=''login_username_index'' />
> [SNIP]
> If I change the name of the input element from "login_username"
to
> anything else, all browsers alert "null". Why is the dollar
function
> reading the name attribute instead of the id? The way I''ve always
> understood it, the dollar function ignores the name attribute.
Well, the id of that input is actually "login_username_index" so
unless there''s another element with an id of
"login_username", $ is
going to be unable to find any matching elements, and thus returns
null.

:Dan Dorman

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group.
To post to this group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs?hl=en
-~----------~----~----~----~------~----~------~--~---

Matt Alexander

2007-Dec-06 18:43 UTC

head link

Re: Weird Prototype behavior

On Dec 6, 11:27 am, "Dan Dorman"
<dan.dor...-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:> Well, the id of that input is actually "login_username_index" so
> unless there''s another element with an id of
"login_username", $ is
> going to be unable to find any matching elements, and thus returns
> null.
>
> :Dan Dorman
Yes, in Firefox. The results of my tests reflect that Prototype 1.6.0
is reading the name attribute in some other browsers. If you run my
test and find something different, I''d love to hear about it.

== TEST 1 ========================
<input type=''text'' name=''login_username''
/>

IE6 alerts "[object]".
IE7 alerts "[object]".
Opera 9 alerts "[object HTMLInputElement]".
FF2 alerts "null".

== TEST 2 ========================
<input type=''text'' name=''blablabla'' />

IE6 alerts "null".
IE7 alerts "null".
Opera 9 alerts "null".
FF2 alerts "null".

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group.
To post to this group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs?hl=en
-~----------~----~----~----~------~----~------~--~---

Fabio

2007-Dec-06 19:09 UTC

head link

Re: Weird Prototype behavior

It''s not a Prototype issue.
document.getElementById in Opera and IE (incorrectly) returns elements
"identified" by the name.

Matt Alexander escreveu:
> On Dec 6, 11:27 am, "Dan Dorman"
<dan.dor...-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> > Well, the id of that input is actually
"login_username_index" so
> > unless there''s another element with an id of
"login_username", $ is
> > going to be unable to find any matching elements, and thus returns
> > null.
> >
> > :Dan Dorman
>
> Yes, in Firefox. The results of my tests reflect that Prototype 1.6.0
> is reading the name attribute in some other browsers. If you run my
> test and find something different, I''d love to hear about it.
>
> == TEST 1 ========================>
> <input type=''text''
name=''login_username'' />
>
> IE6 alerts "[object]".
> IE7 alerts "[object]".
> Opera 9 alerts "[object HTMLInputElement]".
> FF2 alerts "null".
>
> == TEST 2 ========================>
> <input type=''text'' name=''blablabla''
/>
>
> IE6 alerts "null".
> IE7 alerts "null".
> Opera 9 alerts "null".
> FF2 alerts "null".--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group.
To post to this group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs?hl=en
-~----------~----~----~----~------~----~------~--~---

Matt Alexander

2007-Dec-06 20:21 UTC

head link

Re: Weird Prototype behavior

I wonder if there are plans with the Prototype guys to take this loose
functionality out of the dollar function. Leave it in
document.getElementById where it belongs.

On Dec 6, 12:09 pm, Fabio
<fabiow...-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:> It''s not a Prototype issue.
> document.getElementById in Opera and IE (incorrectly) returns elements
> "identified" by the name.
>
> Matt Alexander escreveu:
>
> > On Dec 6, 11:27 am, "Dan Dorman"
<dan.dor...-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> > > Well, the id of that input is actually
"login_username_index" so
> > > unless there''s another element with an id of
"login_username", $ is
> > > going to be unable to find any matching elements, and thus
returns
> > > null.
>
> > > :Dan Dorman
>
> > Yes, in Firefox. The results of my tests reflect that Prototype 1.6.0
> > is reading the name attribute in some other browsers. If you run my
> > test and find something different, I''d love to hear about it.
>
> > == TEST 1 ========================>
> > <input type=''text''
name=''login_username'' />
>
> > IE6 alerts "[object]".
> > IE7 alerts "[object]".
> > Opera 9 alerts "[object HTMLInputElement]".
> > FF2 alerts "null".
>
> > == TEST 2 ========================>
> > <input type=''text''
name=''blablabla'' />
>
> > IE6 alerts "null".
> > IE7 alerts "null".
> > Opera 9 alerts "null".
> > FF2 alerts "null".--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group.
To post to this group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs?hl=en
-~----------~----~----~----~------~----~------~--~---

Tobie Langel

2007-Dec-06 20:41 UTC

head link

Re: Weird Prototype behavior

Matt,

What "loose" functionality are you talking about ?

Best,

Tobie


On Dec 6, 9:21 pm, Matt Alexander
<mattalexx...-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:> I wonder if there are plans with the Prototype guys to take this loose
> functionality out of the dollar function. Leave it in
> document.getElementById where it belongs.
>
> On Dec 6, 12:09 pm, Fabio
<fabiow...-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:>
It''s not a Prototype issue.
> > document.getElementById in Opera and IE (incorrectly) returns elements
> > "identified" by the name.
>
> > Matt Alexander escreveu:
>
> > > On Dec 6, 11:27 am, "Dan Dorman"
<dan.dor...-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> > > > Well, the id of that input is actually
"login_username_index" so
> > > > unless there''s another element with an id of
"login_username", $ is
> > > > going to be unable to find any matching elements, and thus
returns
> > > > null.
>
> > > > :Dan Dorman
>
> > > Yes, in Firefox. The results of my tests reflect that Prototype
1.6.0
> > > is reading the name attribute in some other browsers. If you run
my
> > > test and find something different, I''d love to hear
about it.
>
> > > == TEST 1 ========================>
> > > <input type=''text''
name=''login_username'' />
>
> > > IE6 alerts "[object]".
> > > IE7 alerts "[object]".
> > > Opera 9 alerts "[object HTMLInputElement]".
> > > FF2 alerts "null".
>
> > > == TEST 2 ========================>
> > > <input type=''text''
name=''blablabla'' />
>
> > > IE6 alerts "null".
> > > IE7 alerts "null".
> > > Opera 9 alerts "null".
> > > FF2 alerts "null".--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group.
To post to this group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs?hl=en
-~----------~----~----~----~------~----~------~--~---

Dan Dorman

2007-Dec-06 20:46 UTC

head link

Re: Weird Prototype behavior

On Dec 6, 2007 1:21 PM, Matt Alexander
<mattalexxpub-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:> I wonder if there are plans with the Prototype guys to take this loose
> functionality out of the dollar function. Leave it in
> document.getElementById where it belongs.
Why? Any potential problems come from sloppy coding. If all the needed
ids are in the proper place, everything''s fine. If a developer is
_relying_ upon the funky name-based behavior, then the onus is on the
developer to deal with the issues raised by coding in a non-standard
way.

:Dan Dorman

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group.
To post to this group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs?hl=en
-~----------~----~----~----~------~----~------~--~---

Matt Alexander

2007-Dec-06 22:00 UTC

head link

Re: Weird Prototype behavior

Tobie Langel wrote:> Matt,
>
> What "loose" functionality are you talking about ?
>
> Best,
>
> Tobie
document.getElementById should totally ignore the name attribute.
That''s loose functionality in JavaScript.

On Dec 6, 1:46 pm, "Dan Dorman"
<dan.dor...-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:> Why? Any potential problems come from sloppy coding. If all the needed
> ids are in the proper place, everything''s fine. If a developer is
> _relying_ upon the funky name-based behavior, then the onus is on the
> developer to deal with the issues raised by coding in a non-standard
> way.
>
> :Dan Dorman
I had to tweak my standard, not-sloppy code in order to sidestep this
sloppy JavaScript functionality: I was trying to detect whether a
certain element exists on the page:

<script>
if ($(''login_username'')) {
   // ...
}
</script>

The input that had the ID of "login_username" did *not* exist on the
page so the ifthen should have evaluated false. It didn''t though,
because there was another input on the page with a "name" attribute
set to "login_username". This is an example of a problem arising from
loose JavaScript functionality.
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group.
To post to this group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs?hl=en
-~----------~----~----~----~------~----~------~--~---

Dan Dorman

2007-Dec-06 23:15 UTC

head link

Re: Weird Prototype behavior

On Dec 6, 2007 3:00 PM, Matt Alexander
<mattalexxpub-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:> document.getElementById should totally ignore the name attribute.
> That''s loose functionality in JavaScript.
>
> The input that had the ID of "login_username" did *not* exist on
the
> page so the ifthen should have evaluated false. It didn''t though,
> because there was another input on the page with a "name"
attribute
> set to "login_username". This is an example of a problem arising
from
> loose JavaScript functionality.
As Fabio mentioned, it''s not really a Prototype problem, it''s
a
problem with how the browser vendors implemented the
document.getElementById method. You can make the argument that
Prototype is there to clean up these sorts of cross-browser issues,
since Prototype does just that very well in so many areas. That''s
valid, but the question is if it''s worth taking a performance hit to
one of Prototype''s core functions to satisfy an edge case that can be
easily remedied. After all, there will probably always be
cross-browser issues in Web development.

And I''m sorry if you thought the "sloppy coding" comment was
directed
at you, Matt, ''cause it wasn''t--I was just trying to make a
general
statement. Obviously, that wasn''t clear enough, and I apologize: no
slight was intended.

:Dan Dorman

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group.
To post to this group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs?hl=en
-~----------~----~----~----~------~----~------~--~---

Matt Alexander

2007-Dec-06 23:19 UTC

head link

Re: Weird Prototype behavior

No problem. :-)

On Dec 6, 4:15 pm, "Dan Dorman"
<dan.dor...-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:> On Dec 6, 2007 3:00 PM, Matt Alexander
<mattalexx...-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> And I''m sorry if you thought the "sloppy coding" comment
was directed
> at you, Matt, ''cause it wasn''t--I was just trying to make
a general
> statement. Obviously, that wasn''t clear enough, and I apologize:
no
> slight was intended.
>
> :Dan Dorman--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group.
To post to this group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs?hl=en
-~----------~----~----~----~------~----~------~--~---

LexNonScripta

2007-Dec-07 12:05 UTC

head link

Re: Weird Prototype behavior

Just a quick, non-standard hack for your problem, you could try:

if( $(''login_username'') &&
$(''login_username'').id==''login_username'')

Much faster than debugging wrongities.. :D

But Matt, this is really valuable information. I do my best to always
couple each form control with ... <input name="same_as_id"
id="same_as_id" val... but never actually figured out why... now I
know..

Cheers to cross browser issues!!

On Dec 7, 12:19 am, Matt Alexander
<mattalexx...-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:> No problem. :-)
>
> On Dec 6, 4:15 pm, "Dan Dorman"
<dan.dor...-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
>
> > On Dec 6, 2007 3:00 PM, Matt Alexander
<mattalexx...-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> > And I''m sorry if you thought the "sloppy coding"
comment was directed
> > at you, Matt, ''cause it wasn''t--I was just trying to
make a general
> > statement. Obviously, that wasn''t clear enough, and I
apologize: no
> > slight was intended.
>
> > :Dan Dorman--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group.
To post to this group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs?hl=en
-~----------~----~----~----~------~----~------~--~---

Dan Dorman

2007-Dec-14 19:44 UTC

head link

Re: Weird Prototype behavior

On Dec 6, 2007 12:09 PM, Fabio
<fabiowgus-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
wrote:>
> It''s not a Prototype issue.
> document.getElementById in Opera and IE (incorrectly) returns elements
> "identified" by the name.
Naturally, since I was so vehement that this issue could be so easily
worked around, I got bitten by a variant of the same problem when I
whipped up some sample code to fool around with Ajax.Updater.

The problem occurred when I was trying to load some scripts with
Ajax.Updater (thus evalScripts is set to true). If I declared a
function in the script being loaded, and gave that function a name
which matched any DOM element''s id, then I got an error in IE (Firefox
and Opera handled the situation without complaint).

Some code demonstrating the issue:

// index.html (simplified to show the relevant stuff)
<script>
document.observe(''dom:loaded'', function() {
  new Ajax.Updater(''foo'', ''ajax.html'' {
    evalScripts: true,
    onComplete: function() {
      $(''whee'').observe(''click'', function()
{ whee(); });
    }
  });
});
</script>

<div id="foo"></div>

// ajax.html
<input id="whee" value="whee" />
<script>
whee = function() {
  alert(''whee'');
}
</script>

However, if the conflictingly named function is declared in the same
script, IE can handle it without a problem.

Obviously, this is a somewhat contrived example with an easy
workaround (or two), but I thought it was funny and apropos that I ran
into the same problem I claimed was a no-brainer to avoid.
Nevertheless, it shows the issues with IE and DOM naming conflicts run
a bit deeper than just getElementById.

:Dan Dorman

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group.
To post to this group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs?hl=en
-~----------~----~----~----~------~----~------~--~---

Reasonably Related Threads

Search for more seemingly similar threads

Rails spinoffs - Dec 2007 - Weird Prototype behavior

Weird Prototype behavior

Re: Weird Prototype behavior

Re: Weird Prototype behavior

Re: Weird Prototype behavior

Re: Weird Prototype behavior

Re: Weird Prototype behavior

Re: Weird Prototype behavior

Re: Weird Prototype behavior

Re: Weird Prototype behavior

Re: Weird Prototype behavior

Re: Weird Prototype behavior

Re: Weird Prototype behavior

Reasonably Related Threads