So, you found a bug you'd like to share? Great! Add a comment below with the details, and I'll be sure to include it and all applicable links.
If you have a comment about this Blog though, please add it on the Site Feedback page, or for any other questions, try checking out the FAQ page.
Thanks.
Oh, one last thing! If we verify that the bug is truly an issue (e.g. not just a misconfiguration), then we'll also submit a bug to the applicable Browser maker, and update this blog with the tracker ID and a direct link. ;-)
Site Navigation
13 comments:
HI,
This actually relates to bug #153 -- when you use the Microsoft System.Xml.XPath.XPathDocument and
System.Xml.Xsl.XslCompiledTransformwith an XML file that contains a script tag in it (even if it is formatted as <script ...></script> the transform consolidates it into a <script ... /> that IE won't parse -- an interesting quirk!
- Jeremy Straub
When using screen.height and screen.width, IE returns the size of the main screen, instead of the screen it is currently on (or even the screen it was started on).
Also re bug #153: I have found through very painful experience that div and textarea tags must have separate closing tags.
I am updating a large website that has been running successfully for several years, but suddenly these unclosed tags are a problem in IE7.
there is a psuedo leak with forms where you are unable to remove them completely from the page and they dont garbage collected inline. They will get cleaned up on a page refresh. Other elements have the same issue but if you set the parents innerhtml='' then it gets cleaned up. With forms it just doesnt work.
function leakforms(){
for(var i=0;i<10000;i++){
var form=document.createElement("FORM");
document.body.appendChild(form);
document.body.removeChild(form);
}
}
//works for everything but forms
function discardElement(element) {
var garbageBin = document.getElementById('IELeakGarbageBin');
if (!garbageBin) {
garbageBin = document.createElement('DIV');
garbageBin.id = 'IELeakGarbageBin';
garbageBin.style.display = 'none';
document.body.appendChild(garbageBin);
}
// move the element to the garbage bin
garbageBin.appendChild(element);
garbageBin.innerHTML = '';
}
This must really be the strangest bug ever. It seems to affect both IE6 and IE7.
An absolutely positioned div with a set width (either with width: or a combination of left: and right:) does not recieve any events outside of it's contents (e.g. text).
The strange thing is that this only occurs in "standards" mode, which was supposed to reduce the number of bugs.
But wait, there's more! The bug disappears if you give the div a background-color (even if it's white)!
Example:
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<div style="position:absolute;top:0;left:0;width:200px;border:1px solid black" onclick="alert('Clicked!')">Testing</div>
First, try clicking on "Testing", then click anywhere else inside the border. Then, add background-color:white to the div, and try again.
IE8 Beta 1:
http://clockwerx.blogspot.com/2008/03/ie8-beta-1-onchange-for-checkboxes.html
onchange() doesn't trigger correctly.
Probably a dupe, but http://clockwerx.blogspot.com/2008/03/unsupported-by-ie78-textcontent.html
textContent() is not implemented in IE 7/8
Downloading via iframe doesn't work in ie7, ie8
In our AJAX application we try to download a file in the end of the callback. To accomplish this task, we assign a location to an iframe object. If “Downloads/Automatic prompting for the downloads” security option is disabled (by default it is disabled in the Internet profile), the appearing information bar prompts for downloading a file, but when it have been committed, the downloading process doesn't start. So, an end-user must click again to download this file.
We cannot start the downloading process immediately on an end-user's click, because the callback request sends additional information from the aspx page, which was used to generate the target file on the fly. The simplest example that reproduces this problem uses the setTimeout function instead of sending a callback. fake.rtf is the arbitrary rtf file that resides in the same folder in a web site:
<html>
<head>
<title></title>
</head>
<body>
<a onclick="runDownload();">Click here to start download</a>
<iframe id="fr1" style="width:10px;height:10px;"></iframe>
<script type="text/javascript">
function foo() {
window.frames["fr1"].location = "fake.rtf";
}
function runDownload() {
setTimeout("foo()", 0);
}
</script>
</body>
</html>
I've got a nasty one. In firefox (prior to version 3) the cursor sometimes fails to appear in input text fields and textareas. It's an official one (Bugzilla@Mozilla – Bug 167801), with a long discussion and several workarounds, but I'm not having much luck finding a clean solution that doesn't muck up other things.
Internet Explorer 8 [beta 1] has a bug supporting the valign attribute on table cells. It doesn't matter what you set the value to - all cells will apply "top" as the vertical alignment.
On a related note it appears that something is messed up with cellspacing and cellpadding too. If you apply a value of "1" for each, and set the background color for the table, and td node types, you will get a 2-3 pixel border at the bottom of the table, not the requested "1" pixel value.
Marcus T
given: a html page with 4 frames: 2 visible, 2 invisible.
situation: in an ajax response handler one of the invisible frames has been loaded with a jsp page.
the jsp page has an onload event handler.
the body of the jsp page:
body id="respbdy" name = "respbdy" onload="showAndSortResponse()"
table
thead thead
tbody id="tblbdy" name="tblbdy"
jsp:include page = "essay_table.jsp"/
tbody
tfoot tfoot
table
div name = "json" id = "json"
${jsonResponse}
div
body
task: in the aforementioned onload event handler access the rows in the table of the jsp page above.
// document holding the jsp page
var respDoc = parent.frames[3].document;
// essay body
var newEssayBdy = respDoc.getElementById("tblbdy")
var cnt = 0;
while (true) {
var newRow = respDoc.getElementById(cnt +"_tr");
// do something with the new row
}
in FF and IE 7 newRow is indeed a
row in the table in the jsp page.
in IE 6 newRow is null.
further: change the div in the jsp page to:
div name = "json" id = "json"
span test span
${jsonResponse}
div
var json = respDoc.getElementById("json").innerHTML;
in FF and IE 7:
json includes the contents of ${jsonResponse}
in IE 6:
json includes span test span
but not the contents of ${jsonResponse}
also: in all 3 browsers
var x = respDoc.body.id
yields the correct value.
workaround to Mozila - firefox: flash and flex 100% height and width
http://mfolio.wordpress.com/2008/06/27/mozila-firefox-flash-and-flex-100-height-and-width/
In javascript you might be tempted to do this:
document.foo.value = null;
and, you might expect this to "nullify" the value of foo.
BUT NO!
In Internet Explorer this sets the value of foo to the String "null".
That's not null. That's a string.
So, tests for null will fail.
Post a Comment