I've got a page with an iframe in it that shows the output of a CGI program (netsaint). When I use some of the features of that CGI, it will present a form to fill out, wich shows up inside the same iframe. One part of the form is a selection drop down... when I click on the little arrow, the list pops up but not where it should. It's up the screen farther, as though calculating its screen coordinates based on the main frame rather than the scrolled iframe.
Could you possibly post a testcase or better yet attach one to this bug using http://bugzilla.mozilla.org/createattachment.cgi?id=86909 ? Over to HTML Form Controls.
17 years ago
A test case is simple to create. Make a page with an iframe inside it that pulls in another page that has a form on it. In that form, create a dropdown selection list with a few items on it. Make it so the iframe is larger than the main page. +-------------------------------+ | Banner | | | +-------------------------------+ | menu | | | | iframe here | | | | ... ... ... ... make the iframe long enough that the main window must scroll to show all of it, and the contents of the ifram big enough so that you must scroll the iframe in order to see everything inside of it. Then click the drop down, and see that depending on where the scrollbars are, the popup menu part of the drop down is drawn in the wrong place. I think it's getting drawn relative to the main document, rather than relative to the iframe or something. You might just check there first rather than spend time creating the test case.
Possibly related to bug 53318
Created attachment 73660 [details] 2 iframe for ads and a login form I've got an similar weird thing with this. When typing (windows 2k mozilla 0.9.8 or 0.9.9), the second iframe "jumps" !
The example above run perfectly well with MacOs X. The weird thing happened with Windows2000.
These 5 bugs: bug#53318 bug#64902 bug#67661 bug#86909 bug#129782 all seem very similar to me. Are there any duplicates here? Or any useful information from any of the others that may help any of them get fixed?
This is almost certainly fixed on trunk now that bug 129782 is fixed. Please retest with a current build....
Bug#147806 is very similar, and has a patch
GavinS, I'm assuming that since think this might be a dup of #147806, this is still happening? (even allowing for bzbarbsky's comment 8?)
James, Yes, my local test case still shows this problem, as does the test case I attached to bug#64902 (you have to download and run locally), as does the latest testcase on bug#64902 (runnable directly). These are all as described on this bug - an IFRAME containing a <select> whose popup gets confused when the page is scrolled. Also, the issues with the EBAY UK SIdebar, and the Netscape Timezone sidebar from bug#67661 and bug#105035 still happen. As does the issue on bug#53318. All in all, I see lots of these issues -- I hope your patch fixes them all :) I haven't got a recent source tree, and a build takes about 2 days for me, but I'd be interested if you can reproduce any of these other issues, with and without your patch...
I also havce a simliar problem in Linux/X11. Here is a screenshot: http://caffeine.contactdesigns.com/~jolt/mozilla.png Hope that helps.
This needs an actual usable testcase (one along the lines described in comment 3, I guess, but actually attached to the bug so that it's possible to test this).
14 years ago
Created attachment 161223 [details] supporting file for testcase
The testcase I attached works for me on Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.3) Gecko/20040925 Firefox/0.10.
Jason, I see the bug on that testcase if I open the select when the page finishes loading, then scroll down about a line (after that the select starts going out of the viewport), and open it again. The second time, it's off by the line I scrolled by.
(In reply to comment #17) > Jason, I see the bug on that testcase if I open the select when the page > finishes loading, then scroll down about a line (after that the select starts > going out of the viewport), and open it again. The second time, it's off by the > line I scrolled by. Still WFM, but glad you could reproduce.
Note that I'm testing on Linux, not Windows.
Fixed by checkin for bug 64902