Bug Tracker

Ticket #1187 (closed bug: fixed)

Opened 1 year ago

Last modified 1 year ago

IE: removing of #__ie_init in a FRAMESET causes weird loading of another frame (SVN)

Reported by: blueyed Assigned to: brandon
Type: bug Priority: major
Milestone: 1.1.3 Component: event
Version: 1.1.2 Keywords:
Cc: Needs: Patch

Description

I've experienced the following problem using jQuery SVN and Internet Explorer 6.

If you have a frameset with two frames, the back/forward buttons behave weird: they cause a page to get loaded in a frame which does not belong there (and was not displayed there, before going to another link and pressing back).

I will attach files for testing this: 1. Extract the attachment to a directory 2. Put jquery.js into the directory (or adjust the path(s)) 3. Open frameset_1.html in Internet Explorer 4. Click on "2" in the top frame 5. This will open frameset_2.html 6. Click "Back" 7. You will see "frame_b_2.html" in the top frame, instead of frame_a.html

This is caused by the following jQuery code:

jQuery(window).load(function(){ jQuery("#__ie_init").remove(); });

which was added because of ticket #1061.

My hacky solution for this was adding

if( ! window.frames.length )

before it, so that the code won't get used for FRAMESETs.

The real fix probably is to not include jQuery in a FRAMESET page at all, but if it gets done it should not cause this weird behaviour.

Attachments

weird-ie-frame-test.zip (1.4 kB) - added by blueyed 1 year ago.
testcase

Change History

Changed 1 year ago by blueyed

testcase

Changed 1 year ago by blueyed

Like already mentioned in #1061, it would maybe be the best option to just always leave the ie_init-Hack-Script in the DOM, wouldn't it? At least it has caused two different problems until now.

Changed 1 year ago by john

  • need changed from Review to Patch
  • owner changed from john to brandon
  • component changed from core to event

I bet this could be fixed by doing jQuery(self) instead of jQuery(window).

Changed 1 year ago by blueyed

No, unfortunately not. At least not for the provided testcase.

Changed 1 year ago by brandon

Other than not cleaning up what are the problems (if any) of leaving the script tag in the DOM?

Changed 1 year ago by brandon

  • status changed from new to closed
  • resolution set to fixed

Fixed in Rev [2136].

Note: See TracTickets for help on using tickets.