Closed Bug 354123 Opened 18 years ago Closed 18 years ago

Setting of "dependent" parameter of window.open has no effect

Categories

(Firefox :: General, defect)

x86
Windows XP
defect
Not set
normal

Tracking

()

RESOLVED INVALID

People

(Reporter: mjuhos, Unassigned)

References

Details

(Keywords: testcase)

Attachments

(1 file)

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; cs; rv:1.8.0.7) Gecko/20060909 Firefox/1.5.0.7
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; cs; rv:1.8.1b2) Gecko/20060821 Firefox/2.0b2

Setting of "dependent" parameter (to value "yes") of window.open has no effect.

Reproducible: Always

Steps to Reproduce:
1. Open test case
2. Click on button "Open window"
3. Click into main window

Actual Results:  
Main window is focused and moved ahead (opened window isn't dependent).

Expected Results:  
Main window will be focused but still on background, opened window will be ahead.

Test case:
<html>
  <head>
    <script>
      function run() {
        window.open('','','width=100px,height=100px,dependent=yes');
      }
    </script>
  </head>
  <body>
    <button onclick="run()">Open window</button>
  </body>
</html>
Attached file Testcase
we only support dependent for trusted apps (generally chrome).
Status: UNCONFIRMED → RESOLVED
Closed: 18 years ago
Resolution: --- → INVALID
Comment 2:
Are you sure? In http://developer.mozilla.org/en/docs/window.open is noted for trusted apps is allowed parameter "modal" but there is no mention that parameter "dependent" is allowed to use only in trusted apps.
Keywords: testcase
note sure, no. i'm pretty sure i wouldn't want it supported.

but personally, i'd rather not support any flags.
But this problem is new in Firefox 2.0. It looks like a regression. I didn't find any bug that would show that this is an intentional change.

IMHO this bug should not be resolved as INVALID.
(In reply to comment #3)
> In http://developer.mozilla.org/en/docs/window.open is noted for
> trusted apps is allowed parameter "modal" but there is no mention that
> parameter "dependent" is allowed to use only in trusted apps.

Hello Miroslav,

I am the original author of that MDC window.open webpage, because of bug 195867. I will update accordingly that document to reflect on *_recent_* changes: it's now on a long "to-do" list. I'm quite busy these days. I need to update that document with other changes too and some corrections/inaccuracies since the last time I edited that document.

I keep a more up-to-date version at:

http://www.gtalbot.org/BugzillaSection/Bug195867GDR_WindowOpen.html

I do tests from time to time with

http://www.gtalbot.org/BugzillaSection/MozPopupHelpTestcaseDebugger.html

and I need to also update:

http://www.gtalbot.org/FirefoxSection/Popup/PopupAndFirefox.html
Miroslav, 
this bug has been resolved as INVALID possibly because (just an hypothesis) bug 214867 has been fixed by someone some time ago.
(In reply to comment #7)
> Miroslav, 
> this bug has been resolved as INVALID possibly because (just an hypothesis) bug
> 214867 has been fixed by someone some time ago.
> 

Gérard, see your bug 214867 is in "reopened" status. So :
1°) I have the problem with firefox 2.0 and 1.5.0.7 too (so it not a regression but just a bug)
2°) your "updated" doc at http://www.gtalbot.org/BugzillaSection/Bug195867GDR_WindowOpen.html doesn't not tell that a specific config is needed for "dependent" feature to be effective, so it should work at any time.
3°)your test page at http://www.gtalbot.org/BugzillaSection/MozPopupHelpTestcaseDebugger.html shows that dependent window is not closed nor reduced seamingly with the parent opener. So the window.open with dependent=yes is breaking its own definition.

conclusion : 
-> It seems you draw your "INVALID" joker to fast and Miroslav is right !
-> You must status on "is Firefox supporting popup", according to the definition of popups given in Gecko DOM reference (as dependent windows)

If Firefox does support dependent windows, the "dependent" feature is not working as expected

If Firefox doesn't support dependent windows, the "dependent" feature must be marked as deprecated.
*** Bug 363937 has been marked as a duplicate of this bug. ***
Is anything being done about this bug. It is serially impacting several site that i know of and others that i am on the development team of. I agree that this is indeed a bug and one that is causing lots of headaches in the communality. Can we please soon action taken here as soon as possible to rectify this bug. 
Also another related problem is the fact that alwaysRaised no longer works as the documentation says it should. Previously one could use alwaysRaised and dependent in conjunction to create a modeless window for various applications. So it should be noted that its not just dependent that has been broken by a bug alwaysRaised has also been broken.

Here is the documentation that says FireFox should support both attributes:
http://developer.mozilla.org/en/docs/DOM:window.open 
> Is anything being done about this bug. 

Bug 214867 is the bug related to dependent windowFeature.

> It is serially impacting several site
> that i know of and others that i am on the development team of. 

Can you please list those websites in bug 214867? and maybe explain, elaborate on why you need dependent windows, why you can't use something else, etc. in bug 214867?


> alwaysRaised no longer works as
> the documentation says it should. 
> Previously one could use alwaysRaised and
> dependent in conjunction to create a modeless window for various applications.
> So it should be noted that its not just dependent that has been broken by a bug
> alwaysRaised has also been broken.


One issue per bug. 
Can you verify if what you're referring to is already addressed/tackled in bug 274088?

 
> Here is the documentation that says FireFox should support both attributes:
> http://developer.mozilla.org/en/docs/DOM:window.open 

There is a serious misunderstanding regarding the MDC DOM:window.open documentation. Such documentation is not a list of specification requirements but rather the result of a long, thorough empirical research done by me on window.open. As bugs surfaced and as changes in features happen, that documentation gets outdated.

Since neither Opera, IE5+, Safari, Icab, etc.. support dependent and since there is no standard whatsoever regarding the window features to be supported by a non-standard method like open(), then people should examine their webpage design and relativize the need to rely on window.open() calls.

Mozilla DOM 0 module owner can decide whatever he wants regarding any/all window features of the open() method.

The MDC DOM:window.open document clearly recommends  "Avoid resorting to window.open() Generally speaking, it is preferable to avoid resorting to window.open() (...)"

> If Firefox doesn't support dependent windows, the "dependent" feature 
> must be marked as deprecated.

That's what bug 214867 is about. As soon as bug 214867 is fixed (or bug 243388), I'll update accordingly the MDC DOM:window.open document.
(In reply to comment #13)
> Bug 214867 is the bug related to dependent windowFeature.
Yes this seems to be...

> Can you please list those websites in bug 214867? and maybe explain, elaborate
> on why you need dependent windows, why you can't use something else, etc. in
> bug 214867?
In short, most of the sites that i know of that use these features are either sites that require a login or are a part of an intranet. So even if i listed the sites here i don’t think that it would help very much. As far as the situations where i have seen it used is in places were the developer is trying to produce a similar look and feel to that of a thick client. For example, if a user wants to see the details of a particular record, instead of redirecting the user to a details page and losing the context of the page they were on, they can view and manipulate both pages at the same time whilst the details page that they are working on, always remains on top (at least till they close it). Also by having dependent set it gives a much cleaner overall look and feel. 
 
> One issue per bug. 
> Can you verify if what you're referring to is already addressed/tackled in bug
> 274088?
That’s fine, but these problems are all related to window.open

> There is a serious misunderstanding regarding the MDC DOM:window.open
> documentation. Such documentation is not a list of specification requirements
> but rather the result of a long, thorough empirical research done by me on
> window.open. As bugs surfaced and as changes in features happen, that
> documentation gets outdated.
> Since neither Opera, IE5+, Safari, Icab, etc.. support dependent and since
> there is no standard whatsoever regarding the window features to be supported
> by a non-standard method like open(), then people should examine their webpage
> design and relativize the need to rely on window.open() calls.
> Mozilla DOM 0 module owner can decide whatever he wants regarding any/all
> window features of the open() method.
> The MDC DOM:window.open document clearly recommends  "Avoid resorting to
> window.open() Generally speaking, it is preferable to avoid resorting to
> window.open() (...)"
That’s fine but i believe that window.open and the options that we have a developer are critical. If you ruled out using this feature you are basically placing the development of web based applications back 10 years, when the only reason people popped a new page was so that they could have multiple windows open. Back then it didn't serve any functional purpose. But now with you preventing us use features like window.open you are taking away our ability as developers to guide the user and provide a better overall experience. 


You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: