Site Navigation

Friday, August 3, 2007

Submit a Bug!

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.


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. ;-)


Jeremy Straub said...


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

Xogede said...

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).

DOK said...

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.

Anonymous said...

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");

//works for everything but forms
function discardElement(element) {
var garbageBin = document.getElementById('IELeakGarbageBin');
if (!garbageBin) {
garbageBin = document.createElement('DIV'); = 'IELeakGarbageBin'; = 'none';
// move the element to the garbage bin
garbageBin.innerHTML = '';

Xogede said...

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)!

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "">
<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.

CloCkWeRX said...

IE8 Beta 1:

onchange() doesn't trigger correctly.

CloCkWeRX said...

Probably a dupe, but

textContent() is not implemented in IE 7/8

aer said...

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:

<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);

Nils T. Devine said...

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.

Marcus said...

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

Anonymous said...

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()"
thead thead
tbody id="tblbdy" name="tblbdy"
jsp:include page = "essay_table.jsp"/
tfoot tfoot
div name = "json" id = "json"

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

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 =

yields the correct value.

André said...

workaround to Mozila - firefox: flash and flex 100% height and width

Anonymous said...

In javascript you might be tempted to do this: = null;

and, you might expect this to "nullify" the value of foo.


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.

Kamil said...

On bug # 193: ( when using the onClick as a workaround, try clicking the already clicked radio button again: The event fires again.

Tested in IE7

Vincent Tollu said...

Found a bug when going forth and back in fullscreen mode via automation. When back, the "tab" toolbar is gone...

Sample code (c#):

SHDocVw.InternetExplorer ie = new SHDocVw.InternetExplorerClass();
object missing = new object();
ie.Navigate("", ref missing, ref missing, ref
missing, ref missing);
ie.Visible = true;
ie.FullScreen = true;
ie.FullScreen = false; //There, the tab bar has disappeared...

Arron said...

this is regarding bug 312 page title in IE7 re: named anchors. you say workarounds: none. here's one

(Include jQuery Source here)
$(document).ready(function () {
document.title = "Your Page Title";

note: jquery is only used because of the fact that many functions can be bound to the jquery $(document).ready() function but only one window.onload() function can execute. this way adding this code won't prevent other functions from firing.

Small Potato said...

This one is with Google Chrome. It doesn't fire onBeforeUnload event if it's inside an iframe
(I'm using Google Chrome but I suspect this bug affects all previous versions too)

let's say we have two pages test1.htm and test2.htm:

in test1.htm we have:
<body onbeforeunload="alert(1)">
<iframe width=100 height=100 src="test2.htm"/>

in test2.htm we have:
<html><body onbeforeunload="alert(2)"/></html>

if we visit test2.htm directly and navigate away, we can see it firing the alert(2).
but if we visit test1.htm (which has test2.htm included in the iframe) and navigate away, we only see alert(1) but no alert(2).

both IE and Firefox are able to give me two alerts in the latter case, but Google Chrome fails to fire the beforeunload event for the iframe.

no work arounds have been found so far, but you may use onunload instead of onbeforeunload if that is sufficient to fulfil your needs. onunload seems to behaviour properly on google chrome even inside an iframe

JeffC said...

I am struggling with memory leak/growth. I have a .js file that gets updated externally, and I want the .htm to file read it periodically and update items in the HTML. Every refresh, it adds to memory (seen via Task Manager)

Here is the HTML and below is the .js. This Submit page won't accept leading < so I changed them all to LESSTHAN. Just cut and paste back into Notepad and replace appropriately. Any help is appreciated.

LESSTHANtitle>Memory Leak TestLESSTHAN/title>
function ReLoadJS()
new_script = document.createElement('script');
new_script.setAttribute("language", "JavaScript");
new_script.setAttribute("type", "text/javascript");
new_script.setAttribute("src", "MemData2.js?" + (new Date()).getTime());
new_script = null;
LESSTHANbody onload="ReLoadJS();">
LESSTHANtable border="0" width="100%" id="table1">
LESSTHANtd width="100%" align="left" colspan="6">
LESSTHANdiv align="center" id="CALCdate">%Calculated Date%LESSTHAN/div>LESSTHAN/font>LESSTHAN/span>LESSTHAN/td>
LESSTHANtd width="17%" align="left" bgcolor="#000000" bordercolor="#FFFF00">
LESSTHANfont color="#FFFF00">Row1LESSTHAN/font>LESSTHAN/td>
LESSTHANtd id="^A01" >LESSTHAN/td>
LESSTHANtd id="^B01" >LESSTHAN/td>

This is file: MemData2.js

try {
document.getElementById("CALCdate").innerHTML="Saturday, April 25, 2009 10:41:55 AM"
document.getElementById('^A01').innerHTML = "LESSTHANfont size='4'>123456LESSTHAN/font>"
document.getElementById('^B01').innerHTML = "LESSTHANfont size='4'>98765LESSTHAN/font>"
} catch(e) {
} finally { }

Anonymous said...

Yet another IE bug...

The following secuence of event causes appendChild to missbehave:

1.- Execute the following code:

function f1(obj){

var ct = document.createElement("div")


ct.innerHTML = "< div >< /div >"
dt = ct.childNodes[0]
dt.contentEditable = true

2.- click on the editable object, type something and then select the text you just typed

3.- execute: document.execCommand("JustifyCenter")

The next time you call appendChild it won't do nothing, just return as if the append was successfull, but no object is appended to the DOM.

Hilbrand Edskes said...

I'm the author of Internet Explorer Collection.
Internet Explorer Collection contains multiple IE versions from 1.0 to 8.0, which are standalone so they can be used at the same time.'ve found a couple of very weird bugs in Internet Explorer 8.
One of these bugs causes the entire page to disappear.
Other bugs I discovered cause boxes to expand even when a (max-)width and/or (max-)height have been explicitly set. ie8overflowandexpandingboxbugs.htm

Anonymous said...


internet explorer 8.
Any 'object' always appears as the uppermost z-index.

This is sooooo annoying when you have to use svg or similar as a plugin as its not supported.

Hell who wanted scalable vector graphics since 1998 anyway?!?!??

steven said...

for bug 143 - createTextNode() doesn't work on the style tag in IE

there is an issue with the workaround. you have to append the style tag to the DOM before you can access the styleSheet property or IE will throw an error.

Rajeev said...

Sidebar hangs when reopens a pop up window
I have a sidebar application, I am using an IM functionality inside this sidebar. When a user is logging in, the connection with chat server will generate on this sidebar.The sidebar itself acts as a parent window. When a user click on the IM icon inside the sidebar, a child Pop up window will open. The child Pop up window uses the connection objects from the Parent sidebar. When closing this child Pop up window, the sidebar hangs. When I tried to repoen the child Popup window, the Pop up window also hangs for a long time. But I can not find the exact reason behind this hanging of sidebar, after closing the Pop up window.

Also this issue is only exists in main server. In my local server its working fine. Also I am using JavaScript caching for main server.

Thanks Raju

Rajeev said...

I have a sidebar application, When a user click on the IM icon inside the sidebar, a Pop up window will open. When closing this child Pop up window, the sidebar hangs. When I tried to repoen the Popup window again, the Pop up window also hangs for a long time. But I can not find the exact reason behind this hanging of sidebar, after closing the Pop up window.

Also this issue is only exists in main server. In my local server its working fine. Also I am using JavaScript caching for main server.

Thanks Raju

Anonymous said...

I have found a IE8 javascript array bug.
When you define an array this way:
note that I have one extra comma at the end; it is simple to generate it unconditionally in automatically generated array data. the property array.length will now be 4 although there are only 3 elements. Works fine in other browsers. Not tried in older IE versions.

Brad said...

Bug in current version of IE8 (8.0.6001.18702) concerning text-align:left in td elements.

Wish I could post the (simple) HTML here, but you can see it here.

Anonymous said...

Operating system: XP SP3.

After installing KB976325 Internet Explorer 8 throws an exception when opening certain web pages.

For example you can try opening and click "Платформы" link to the left, then wait until the page is fully loaded. Notice that IE8 constantly reloads the page.

The exception itself happens in mshtml.dll and looks like this:

Unhandled exception at 0x3fba6a3b in iexplore.exe: 0xC00000FD: Stack overflow.

Uninstalling KB976325 fixes the problem.

jfjauvin said...

setAttribute() on the "value" property of a button element is half working.

IE8 / HTML 4.01


<button id="B1">Click Me</button>


// Working
document.getElementById("B1").value = 1;

alert(document.getElementById("B1").value); // Works
alert(document.getElementById("B1").getAttribute("value")); // Works

// Half Working
document.getElementById("B1").setAttribute("value", 1);

alert(document.getElementById("B1").value); // Doesn't work
alert(document.getElementById("B1").getAttribute("value")); // Works


Educação XXI Lisboa said...

On the main page options I am not able to change or organize the items that appear in the end of each post.

Not sure what is happening, but I've tried everything I can do on the backend.

Victor said...

Opera 12 and Opera Mobile 11.50 have bug (or feature) when showing text on buttons with CSS property text-shadow (see full example at In this example text on two buttons looks similar in Chrome and Firefox, but different in Opera:

.shadow {
margin: 0;
padding: 5px 10px;
text-shadow: rgba(0, 0, 0, 0.4) 1px 2px 1px;
<button class="shadow">Text1</button>
<input type="button" value="Text2" class="shadow" />

Text-shadow properly applies to button tag, but not applies to input type="button". Bug is already reported to Opera (ID: DSK-371187).

Steven Joseph said...

This may be related to bug 411.

Seems to be a bug in IE10, this code causes IE10 to CRASH.

var elements = document.getElementsByName('someelementthatdoesnotexist');

will crash ie10

workaround (if u use jquery) is to use


But getElementsByName is extensively used and this is just one case.

Steven Joseph said...

This may be related to bug 411.

Seems to be a bug in IE10, this code causes IE10 to CRASH.

var elements = document.getElementsByName('someelementthatdoesnotexist');

will crash ie10

workaround (if u use jquery) is to use


But getElementsByName is extensively used and this is just one case.

Victor said...
Chromium and Chrome up to builds of version 26 unable to edit the content of the contentEditable element when it is placed inside of draggable element.

Anonymous said...


IE10 completely ignores table column widths in colgroup/col for "table-layout: fixed".

I have documented my research at

This CSS 2.1 feature works on all browsers (including IE 6...9) but not on IE 10...
MS says, it's by design (you can reed it at ms connect, follow the link in my stackoverflow article).