Internet Explorer Bugs

Summary: InternetExplorerFeedback about (non-programming-related) bugs

Other pages that might be a better place for your contributions:
* All programming-related bugs should go in InternetExplorerProgrammingBugs
* Talk about InternetExplorerStandardsSupport
* Add to the InternetExplorerFeatureRequests
* Discuss InternetExplorerSecurity

Please, folks: First choose a well worded, meaningful heading, then find the right alphabetical place to put your helpful heading.

Thank you for your cooperation and understanding here. :)

application/ecmascript media type not recognized

Internet Explorer 6 / Windows does not seem to support the application/ecmascript media type for ECMA-262 content, it only supports the obsolete text/ecmascript media type which is no longer appropriate for use. More information about the media type is available at https://datatracker.ietf.org/public/pidtracker.cgi?command=view_id&dTag=7686

application/xhtml+xml media type is not recognized

IE does not consistently recognize the xhtml mime type - which is required for true XHTML compliance. EVEN WORSE, ASP.NET 2.0 is completely switching over to XHTML, and IE doesn't even fully support it yet! Insanity... we really need the xhtml mime header fixed... :( http://www.w3.org/TR/xhtml-media-types/

Auto-Complete Issues

The Auto-Complete feature is useful but flakey. There is no way of predicting when it will take new additions or just ignore them. It would be also nice if it were more obvious how to delete autocomplete entries (press down to highlight, then delete).
The bug (#137) is documented here, including where/when it happens, and a temporary workaround for this bug.
http://webbugtrack.blogspot.com/2007/08/bug-137-ie-autocomplete-hardly-ever.html

Built-in Webpage Quality indicator icon

Implement a feature which will report back to the user if a page uses valid code, has markup errors and/or parsing CSS errors: some sort of a Webpage Quality indicator icon (smiley or green check for valid page, frown or red 'X' when invalid) on the statusbar (or somewhere else) which when clicked would report more info to the user and give him more options among which one would be to validate the page with the World Wide Web Consortium validator.

Implement something like HTML Tidy as an extension or an option into IE 7 and for IE 7 users. Amaya 9.2.1 reports parsing errors. Icab 2+ reports parsing errors and other errors (warnings about deprecated elements and/or attributes). Dillo browser reports common mistakesmeter.html (improper nesting, attribute format) in webpages. BBedit and BB Tidy also reports frequently encountered markup errorschecking/#toperrors like improper nesting, absence of doctype decl., usage of @<font>@, @<center>@, non-standard or deprecated tag attributes, etc.

W3C HTML 4.01 specs recommend browsers to notify users about markup/syntax errors in pages: 'We also recommend that user agents provide support for notifying the user of such errors.'#h-B.1 http://www.w3.org/TR/html401/appendix/notes.html#h-B.1

Internet Explorer Validation and QA Toolbar 'Silent recovery from error is harmful.' In particular for improper nesting of elements (malformed markup code), missing closing quotes, incorrect attribute values, etc..

"Agents that recover from error by making a choice without the user's consent are not acting on the user's behalf. (...)
Experience with the cost of building a user agent to handle the diverse forms of ill-formed HTML content convinced the designers of the XML specification to require that agents fail upon encountering ill-formed content. Because users are unlikely to tolerate such failures, this design choice has pressured all parties into respecting XML's constraints, to the benefit of all."
Coming from: Architecture of the World Wide Web: http://www.w3.org/TR/webarch/#no-silent-recovery

"The whole reason nearly all Web pages on the Internet are malformed is because browsers let Web page authors get away with it. As long as browsers are permissive in their error handling and recovery, Web authors will continue to produce invalid Web pages, because they won't even have any idea the pages they are authoring are invalid!" coming from http://weblogs.mozillazine.org/hyatt/archives/2004_01.html#004702

		 -- DU/GT

BHO exception crashes IE

A very popular Browser Helper Object, Quickfind BHO, is able to crash IE when it sees this javascript:

<code>var Event = new Object();</code>

It will cause both IE6 + IE7 to crash.

The javascript line comes from the Prototype toolkit, which is very popular these days, so anyone with that particular BHO will get crashes when visting a LOT of sites. I suspect that other BHO's throwing the same exception will also cause the same crash.

I've described the situation in a lot more detail here: http://roderick.dk/parasite/

It's my experience that IE generally handles exceptions from BHOs and plugins quite well, so this must be one of the more obscure ones.

Recently, I've come in contact with the company that did the Quickfind BHO, so if someone from Microsofts IE team could contact me, we should be able to wrap this one up quickly and improve the browsing experience for a lot of users.

Cache Issues

IE 6.0: if the cache fills up, the cache index file (which maps URLs to the cached item) seems to get corrupted. This behaviour can lead to erroneous Cannot Find Server messages, missing images, missing style sheets, missing downloads, etc. The data has been downloaded, but the item cannot be found in the cache. Sometimes, the display of the Temporary Internet Files folder will also be corrupted. Clearing Temporary Internet Files fixes the problem, until the next time it fills up. -- MikeDimmick

FYI, the missing-download problem seems to be caused by the cache-size bug; IE will store the wrong 64-bit cache size in INDEX.DAT and then downloads are missing until it's fixed. Anyone working on this should check out CacheSentry to see a list of known IE cache bugs. -- Jay Levitt

Cached HTC Files Generate Unnecessary GET Requests

DHTML Behaviors are just awesome. Unfortunately, this caching bug cripples the "reuse" feature of this component technology for the very situations in which it would be most useful - pages with the same control/behavior repeated many times. -- RobEberhardt, xDOM DHTML behavior suite

Cannot install on Dutch XPSP2

I cannot get IE7 installed on my Dutch XP box. It says the system language is different from the update language. Why is that a problem? I can install english Firefox or Opera perfectly fine. There's no Dutch version of IE7 now, but what if I just need an english browser?

Cascading Style Sheet(CSS) Memory Corruption (fixed)

There is something wrong in CSS handling. Certain elements make IE crash. As reported here
Fixed: in Windows XP SP2, great! --Mikko Salmi

Change colour of disabled 'select' does not work in IE

The color of the text of the selected option in a 'select' does not work in IE. The following works in Mozilla, but not in IE. The IE version that i am using is 6.0. Havent tried with other versions.

<html>
<head>
<style>
selectdisabled {color:red!important;}

</style>
<script type="text/javascript">
function changeMyCol()
{
var s = document.getElementById("s1")
s.disabled="true"
}

</script>

</head>
<body >

<select id="s1" >

		 <option id="o1">Colour1</option>
		 <option id="o2">Colour2</option>

</select>
<p>
<input type="button" id="b1" value="Click me to disable and change colour" onClick="changeMyCol()"/>

</body>
</html>

Ctrl + A (select all) in address bar

Doesn't work, it just beeps. Please just fix this annoying bug and we'll love you forever, just like when you fixed Notepad's keyboard shortcuts in Windows 2000!!! --infrared

This seems to work fine for me. Version: 7.0.5730.11. --erh

Clear Forms Button Doesn't Do Anything

When ever i click Tools >> Internet Options then the content tab, then auto complete button, then i click clear forms and click ok. then guess what. i go to google and all those searches of mine are still their when i click the search thing and press down

cookies bug

If i set my username to " "(ascii 160) in Windows XP sp2 japanese region setting, then IE will not remember any cookies

Cursor does not disappear when typing.

Test in outlook or VS!(both use IE in edit mode) - even in this wiki the mouse cursor is still showing as I type.

Default Home Page Not Loading When It's Set To a XHTML-Trans Page

Our company website was rewritten to use XHTML-Trans doctype with CSS (and validated by W3C's validator as XHTML-Trans).

Strange error: I have the company home page as my default Home Page in IE6 SP2 (running under XP SP2); ever since the site rewrite to XHTML-Trans, when opening new browser window IE6 locks up and does not display the page. (All I see is a white page, and a status indicator showing page load about half complete. Task Manager - Processes shows iexplorer using 50% of CPU and about 20M memory usage...)

When opening IE6 through 'Start-Favorites' to any other page, IE6 opens. When open, I can enter the company home page URI in the Address bar and the page displays properly. (This lockup does not occur w/ IE6 SP1 running under Win2K Pro or IE6 SP1 running under XP SP1.)

This lockup has been replicated on 2 laptops and 2 desktops also running XP SP2/IE6 SP2.

Could it be an issue with the way IE6 SP2 tries to read or render the doctype code when opening (since, when open, it does not lock up when displaying the page when the URI is typed into the address bar?

Page code initial section:

		 <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
		 <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
		 <head>

The rest of the page is copied from the old standard HTML 4 Trans page--the only changes were the doctype, adding the required end tag to empty elements (such as <br /> and <img.... />), and elimination of inline display code now listed in the associated external CSS file.

Weird one, huh--but most annoying as it affects every IE6 instance launched!

Discussions Bar (not an IE bug)

This thing should probably just be removed altogether. People turn it on accidentally, and it breaks all kinds of things.

This is not an Internet Explorer bug - it is a plugin that Microsoft Office installs. I believe you can uninstall it through the office installer in a custom install, but I'm not 100% sure. Same thing goes with the "research" feature.

DNS no response make IE stuck

GUI of IE is stuck when DNS server give no response, it just poll each server configured in system, again and again. This problem is really serious when it come to tabbed browser.

DOM generated form fields are _NOT_ accessible by name in the form or elements array/objects.

For example, a checkbox with the name 'foo'.
e.g. Using .createElement('input'), .setAttribute('type','checkbox'), and .setAttribute('name','foo') then .appendChild(...) to add it to the page.

e.g. Any of the following attempts to read the value will *FAIL*.

var val = document.formsyourIndex 'foo'.value;
var val = document.formsyourIndex.elements'foo'.value;
var val = document.formsyourIndex.foo.value;
var val = document.formsyourIndex.elements.foo.value;

And of course, it isn't just the value, the object just isn't added to the form's named array... so if you want the length of an element, or details about the options of a select list, or the class/style attributes, you're out of luck.

As a horribly ugly workaround, you can iterate over the entire contents of the form, checking the name of each element against your desired name.

PS This is broken in IE6 too.
PS setting obj.name = 'foo'; will not work either.

Addendum:

Reduced testcase showing the bug

Reference: "The NAME attribute cannot be set at run time on elements dynamically created with the createElement method." coming from NAME Attribute at MSDN_2.asp and from NAME Attribute at MSDN2

This bug has been reported by Aaron Gustafson in September 2005 in his Death to bad DOM Implementations. "I just encountered a DOM implementation issue in IE which took about three hours to solve (and like a year off my life)."

It has also been reported by Bennett McElwee in his Setting the 'name' attribute in Internet Explorer

DU/GT

Download File Dialog freezes

When the last location you saved to was a network path (eg. \\fileserver\myhomedir) the Download File dialog freezes (IE6SP2). This can be fixed by changing HKCU\Software\Microsoft\Internet Explorer\Download Directory to a local path. This happens to me often after working over a VPN.

Download Silently Fails

When downloading a file IE gives no error if the download is shorter than expected. Instead, it just saves what it did download and acts like the download in fact completed. IE should instead alert the user to the failure of the download (downloading only 800K of a 1.9MB file is an error) or restart/resume the download automatically.

Even more serious bug: sometimes after IE7 downloaded short file and you try to redownload it, IE7 just loads the file from cache. You need to clear cache to download the file. -- Fduch

Download to Desktop and Open Folder

If you download a file to the desktop, the download complete window does dumb things. If you click Open Folder, it will pop a message box stating 'The content is on the desktop'. Then, when you click OK the entire download complete window disappears (which mean that you have to dig under your existing windows, then search your desktop to find the file).

This User-Hostile behavior isn't acceptable. The quickest way to fix it would be to make the dialog not close the download window. The better way would be to disable the Open Folder button, if the file is on the desktop.

Download window goes away when you close IE

I really like the fact that I can start a download and close the IE window containing the download link without interfering with the download. I really don't like the fact that if I accidently close the last IE window on the screen, my download gets silently blown away. Let me tell you a true story...

I go to a site with some videos of an augmented reality demo; I start to download one. Its 80MB, so the download will take a while. I close the IE window, minimize the download window, and contine browsing in another IE window. The download chugs along in the background.

A few minutes later, I go to download the latest beta of RSS Bandit. Thats a lot smaller; it starts downloading pretty quick. My video download is about 70% done. I close the IE window with the RSS Bandit download link on it - whoops! That was the last IE window on the screen! Both downloads are gone. Now, that is irritating!

Downloading Ignored Content

IE downloads stuff it will not use, such as alternate style media types (handheld, braille, etc.) or script MIME types (application/x-notscript). This means users have to wait for unneccessary downloads and HTTP congestion. It also makes true browser stats checking (by technology support, rather than capricious, obfuscated user agent strings) more difficult than it needs to be.

Dynamically created binary data

It's possible to do myImageHandle.src="data:image/bmp;charset=iso-8859-7,"+BMPData in IE but it seems to fail whenever the added data contains the byte 0 (0x00). This would be really useful since it allows images and other binary data to be rendered and re-rendered/updated after the page is done loading.
This works in both FireFox and Opera and I'm pretty sure it's part of a standard somewhere, sorry I can't remember which one. -Henrik Danielsson

favicon.ico

I know it's really fussy <g>, but if you come up with an enhancement in the creation of favicon.ico it really shouldn't loose them or only display them when you have a site bookmarked. Mozilla's implementation seems a lot nicer, with it appearing when you browse to a site, and indeed icons appearing which seem to completly confuse IE.

I'd like the favicon.ico file to be used as the icon displayed in the Windows Taskbar instead of the generic IE logo. At work when I have dozens of apps open it is difficult to tell which IE window I need to click on in the task bar to get to the needed application. Displaying the favicon.ico in the taskbar would allow each web-based app to have their own "look" on the task bar.

The favicon.ico feature should be expanded to support favicon.jpg and favicon.gif. I believe firefox recognizes these files.

While "favicons" were introduced with IE, only modern browsers truly support it. Having the site icon consistently appear in the address bar, Favorites, and tabs is much more desirable that Internet Explorer's current random application of said icon. -- DalanGalma

""MyIE2"" (now Maxthon"")"" already does this - you get the icon in the address bar for every site you visit.

Cool - now let's see standard IE get it. -- DalanGalma

IE stores the icons in the Temporary Internet Files cache. As soon as you clear the cache (delete temporary files), you lose all of your icons. It would be nice to see a documented, "non-cache" storage location implemented for favicons. Currently, whenever I visit a site with a favicon, I copy it from the cache to a hidden "Shorcut Icons" folder in my Favorites folder and then edit the icon properties of the link to point to the new location. -- tthibodeau

IE often doen't show favicons. When I often 10 pages on a site with favicon, about 2 pages are displayed without it. -- Fduch

Favorites

Importation of Favorites simply doesn't work and never has since IE4. Everybody has been forced to manually copy the individual files to the appropriate folder. This is very rough on less adept users. The import function should be smart about where all versions of Windows back to Win98 store their Favorites and bring them to the proper place without user input. It should also be smart about not duplicating the Favorites that are pre-installed.

File Save Problems

I'm not sure if these problems affect everyone or just me: when a File Save (to local hard disk) is obeyed, IE usually re-downloads the entire web page and all its files, even if everything has already been fully loaded. Sometimes IE then refuses to save the page. Sometimes it appears to save the page but when you open it again later there's nothing there. This is annoying for most web sites and possibly catastrophic to the user for ecommerce and online banking operations. These problems, and the constant security scares, are what stop me using IE. John259.

Yeah, this is a nasty bug. It always happens to me after I have bought something on an e-commerce site and I want to save the receipt - I don't have a printer at home right now.

Imagine you opened a page with useful info. But it was soon replaced by "the information is gone" page. When you save the useful page, in reality the meaningless page is saved. -- Fduch

Find dialog Error 57

If a page contains images loaded from an invalid IP such as 0.0.0.0 or 255.255.255.255 (say, an ad site that has been blacklisted in the HOSTS file), trying to search the page too soon fails with "An error has occurred in this dialog. Error: 57. Unspecified error." Dismissing this dialog also closes the original Find dialog. - anfortas

Focusing a secondary window and the target attribute

If a hyperlink opens in a new named window, and that window already exists, IE does not switch focus

I've seen this lots of times with beginner users. If a hyperlink opens a page in a new named window, and that window already exists in the background, the page will load in that window but it is not activated. The user is sat there clicking the link repeatedly saying "It's not working". You may not think this is a major thing, but I've seen it destroy a user's task on countless occasions.

"The biggest fault with pop-ups is that it takes the focus away from the main browser window, and this can be disconcerting. It presents general usability issues aside from accessibility. How often have you seen someone launch a pop-up and then inadvertently click back on the launcher window and thinking that nothing's happened, click the link again with nothing happening? Of course the window has opened but is now under the launcher window, and only moving down to the task-bar and selecting the window from there will solve this." Ian Lloyd in http://www.accessify.com/tutorials/the-perfect-pop-up.asp

"Furthermore, Windows XP groups multiple windows of the same application in the taskbar, so there is virtually no visible indication that a new window has even been opened." Mark Pilgrim in Dive into Accessibility, Not opening new windows16notopeningnew_windows.html

"current operating systems have miserable window management" J. Nielsen regarding Opening New Browser Windows in The Top Ten New Mistakes of Web Design

Several other interesting quotes at: http://www.gtalbot.org/FirefoxSection/Popup/PopupAndFirefox.html#RaiseLowerSetting

"some people can use Windows applications for years without understanding the concept of task switching. (When I point to the task bar and ask them what it's for, they can't tell me.) (...) In another recent study, six out of 17 users had difficulty with multiple windows, and three of them required assistance to get back to the first window and continue the task." Carolyn Snyder, Seven tricks that Web users don't know, 7. Second browser windows_7tricks.htm#7 June 2001
--DU/GT

Frameset and IE Performance

IE Performance continously degrades when pushing html content across frames

When pushing html content from one frame to anothe frame IE perfromance continously degrades.
A simple test case hosted at http://users.rcn.com/ckhima/ieperf/frameset.htm
shows the problem.

The test case consists of a header frame, main frame, srcFrame (where all js are loaded and JS objects are created) and finally proxy frame. In the header frame you will see a bunch of links that loads pages in the proxy frame simulating server round trip. The html loaded in the proxy frame uses JS lib from the srcframe to create JS objects (in this example bunch of absolute positioned table control is created) and finally renders the generated html content on to the main frame. A debug window display how much time it took to actually render the content on the main frame. Cycle thru all the links from
page 1 to page 5 for several rounds. In each round you will see that the rendering time keeps on increasing. For example to render page 5 intially it took 2.6 sec but in the second round it takes 3 sec and in third round it takes around 3.5 sec and it keeps increasing

Now do the same test on a Firefox browser you will see the rendering time stays same no matter how many times we cycle through the pages. Why IE performance is degrading?

GIF (animated) performance

The longer gif animated, the higher cpu it takes. For looped gif, frame 0 takes lowest cpu, and last frame take most cpu.

FTP Data Loss

Try this... open the FTP client/browser to a site where you have read & write access. Cut a pile of files from the server, and paste them on your desktop. Press cancel once the transfer begins. The transfer will cancel, but IE will happily delete all of the files that were cancelled! YIKES! I'd classify this one as a serious bug.... --ShadowChaser

FTP username/password (not a bug)

when i have the username set to an email address the ftp url is not accepted and IE doesn't find the site. for example

		 ftp://user@site.com:password@site.com

doesn't work. --mwirth

Answer: This is not a bug. It was disabled (in XP SP2, I believe) because it was the source of too many customer security problems -- JonathanHardwick

Answer: Actually, no; username/password support was disabled in a patch to SP1+ for HTTP only. The reason his FTP URL doesn't work is that it's not legal; you cannot have an @ sign within the USERINFO Component of a URL. EricLaw MSFT

FTP Support

The FTP Browser in Internet Explorer at a glance looks like a very easy to use FTP Browser when compared to everything else out there. Unfortunately, it makes dozens if not hundreds of connections with FTP servers (which place a great deal of strain on FTP servers and can get users banned from FTP servers for making more than 1 simultaneous connection) and doesn't exactly work. If I recall the last time I used it, when I log on to an FTP server via it, I can't modify permissions or transfer files either way until it opens up a new window in which I can. In my book, this renders an FTP Browser which could arguably become the best FTP browser in the world unusable. -- Shining Arcanine

Also, I agree with Shining Arcinine that the FTP client in IE is seriously troubled. It often crashes the whole browser and sometimes the whole Explorer process when it loses its connection (ie walking out of wireless range).

Here's a more specific problem: when you drag/drop an already existing folder into an IE FTP connection, it creates multiple folders with the same name. If you refresh, IE reloads the directory list and everything's fine, but if you change directories and change back, you will see more than one of the same folder - happens with files too. -- evanbro

History Bug

If you set the amount of days for IE to maintain its history to 0, it defaults to 1 day while maintaining 0 in internet options. --Shining Arcanine

History Sometimes Skips Pages

Sometimes (seen roughly once every few thousand pageviews) the Back button will take a user back two pages prior in the browse history, skipping the prior page. Specifically, if I start out at page A, click a link to go to page B, then click a link to go to page C, and when I click Back I'll be delivered to page A, skipping over B. Clicking Forward will take me forward to C, again skipping B. It's as if B just never made it into the history list. This is semi-rare, but it can be very annoying. -- DavidWeekly

The "by order visited today" history bar option has similar problems. I'll KNOW I visited a site at a certain time in the day - yet it's not in the list. Sometimes revisiting the site will move it in the list, instead of putting two entries in. It's very strange. -- ShadowChaser

History wipe on blackout/crash/freeze.

Open IE7. Browse for a while. Look at the history. Browse more. Hit reset.
The session history is gone.

---

History list perfomance problem.

Looks like a typical beginner programmer mistake.
When you change modes in history (e.g. from by visits to by site) IE7 doesn't disable list layout while updating list contents. Looks funny, but Ridiculous. -- Fduch

History Drop Down (By the back/forward arrow)

The bottom option in this drop down is History Unreadable Char Ctrl+H

It looks messy.

Home page settings are not honored by IE

I set my home page to about:blank or http://www.google.com, and from time to time when I open IE, instead of my
home page, another MSN or windows update page hijacks it. This is really annoying.

HTTP Digest Access Authentication

HTTP Digest Access Authentication as defined in RFC 2617 is broken in all versions of Internet Explorer that support it.

When a server does not send a qop-value in the challenge (as is explicitly allowed by RFC 2617 for backwards compatibility with RFC 2069 ) Internet Explorer sends two consecutive requests without an 'Authorization:' header and then gives up. When it does so it shows the standard "the page can not be display" page, thereby placing the blame on the server.
Even when using RFC 2617 Internet Explorer gets it wrong when there is a query part in the URL. It then drops the question mark and everything behind it when computing the response. This leads to an incorrect response which will not be accepted by the server.

Feel free to test point 1 and point 2 -- Henryk Plötz

HTTP POST error with non-ISO-8859-1 chars and disabled controls

When entering text containing non-ISO-8859-1 chars into textarea controls in a HTML form declared as of method=POST, enctype=multipart/form-data and containing some "disabled" controls at its end, it is possible that IE6 does compose an illegal (malformed) HTTP POST request. The malformed request is not empty but misses the first multipart boundary string and the directly following static text. The page is eventually required to be set to "charset=iso-8859-1" to show the behaviour.

Maybe the problem can be worked around by declaring the page delivering utf-8 instead.

An example of a page showing the behaviour is available from the author of this paragraph ZipwiZ, but you need a network traffic sniffer to watch things happen.

HTTP POST: send html files with proper content-type

Since the release of XP SP2, IE sends html files (through HTTP Post) as text/plain. It should be text/html.
See e.g http://www.lachy.id.au/blogs/log/2004/09/validating-xhtml-with-ie-using-file

HTTP Redirect error

When having a HTTP/302, IE use the wrong port in the 'Host:' header value. Analyze with ethereal show:

		 GET [/AccountId?accountno=2038221] HTTP/1.1
		 Accept: */*
		 Accept-Language: fr
		 Accept-Encoding: gzip, deflate
		 User-Agent: Mozilla/4.0 (compatible; MSIE 5.01; Windows NT 5.0)
		 Host: pfq-1.somewhere:7132
		 Connection: Keep-Alive
		 Cookie: aaasessionid=NL5R; bbbsessid=5XBD; ccc_context=864a

		 HTTP/1.1 302 ok
		 Date: Tue, 03 Aug 2004 12:50:25 GMT
		 Server: Apache/1.3.12 (Unix)
		 Set-Cookie: ddddsessid=1CCT;Domain=.somewhere;Path=/
		 Cache-Control: no-cache="set-cookie,set-cookie2"
		 Expires: Thu, 01 Dec 1994 16:00:00 GMT
		 location: [http://pfq-1.somewhere:11733/aaa/AAAAuthPresEXT?appliName=DDDD&DNSName=http://pfq-1.somewhere:7132]
		 Set-Cookie: accountDDDD=2038221
		 Keep-Alive: timeout=15, max=50
		 Connection: Keep-Alive
		 Transfer-Encoding: chunked
		 Content-Type: text/html
		 Content-Language: en

		 GET [/aaa/AAAAuthPresEXT?appliName=DDDD&DNSName=http://pfq-1.somewhere:7132] HTTP/1.1
		 Accept: */*
		 Accept-Language: fr
		 Accept-Encoding: gzip, deflate
		 User-Agent: Mozilla/4.0 (compatible; MSIE 5.01; Windows NT 5.0)
		 Host: pfq-1.somewhere:7132
		 Connection: Keep-Alive
		 Cookie: aaasessionid=NL5R; bbbsessid=5XBD; ccc_context=864a; accountDDDD=2038221

		 HTTP/1.1 404 Not Found 
		 Date: Tue, 03 Aug 2004 12:50:31 GMT
		 Server: IBM_HTTP_SERVER/1.3.26.2  Apache/1.3.26 (Unix)
		 Last-Modified: Thu, 10 Jul 2003 09:22:44 GMT
		 [ETag:] "9f-72b-3f0d3064" 
		 Accept-Ranges: bytes
		 Content-Length: 1835
		 Keep-Alive: timeout=15, max=100
		 Connection: Keep-Alive
		 Content-Type: text/html

		 <!-- No Cache -->
		 <html>
		 (...)

The second time, host value have the old port number

Answer: Fixed in a patch.

HTTP/1.1 Caching and passing content to external applications

IE's handling of non-cacheable content is fundamentally broken. The "design" seems to be "if I'm not allowed to cache it I won't pass it to external applications". See Kb316421 for a problem description.

HTTP/1.1 Expires Response Header is broken

When IE receives a http response with an expires header - IE will use the Expires header value directly in determining if content needs to be refetched. If you actually READ the http standard (RFC2616) it says to subtract the Expiry header from the Date header and use the delta to determine cache lifetime. This is to avoid clock skew issues. Because if you implement it how IE does - the server and client have to have their clocks set identically for this mechanism to work!!

Answer: Expires isn't a HTTP1.1 header. Use Cache-Control: max-age instead.

-- Not Really an answer - you are going to want to have both Expires and Cache-Control: max-age in the headers in a lot of cases. E.g. the server returns a fixed set of headers for a given file i.e. doesn't return headers based on the http request version (you still need to handle http1.0 for old proxy servers).

-- The answer is completely incorrect. Expires is an HTTP 1.1 header, it's defined in RFC 2616, section 14.21. Whoever said it wasn't didn't read the HTTP specification as was suggested.

HTTP/1.1 Compression and Entity Tags

These are two HTTP/1.1 features which can significantly improve web performance. Both are supported by IE, but unfortunately not at the same time. If IE receives an HTTP response with both Content-Encoding: gzip and ETag: "blah", it will ignore the ETag and fail to send an If-None-Match: "blah" request header next time that entity is fetched. This results in needless refetching of cached content.

HTTP/1.1 Compression and plug-in content

If IE isn't going to handle compressed content properly - why does it make a request saying that it does? Specifically - most plug-in content like pdfs for example die badly because IE passes compressed content to the plug-in not the uncompressed content.

Answer: Is there something that leads you to believe it's not the plugin making a buggy request?

-- I assumed Adobe would have gotten Acrobat correct if they could have

<iframe> implementation bug

There seems to be a bug in the implementation of <iframe border="0" frameborder="0" ..> attribute (without the desired effects). Many wrapper applications using IE5/6 as their underlying browser object do not make such a mistake (e.g., Tencent Traveller).

workaround This is another one of IEs case-sensitive attribute issues. If you set frameBorder="0" (note the capital B) this will work in IE6 & 7

Image Loading

For some sites, when I type their domain without the trailing slash into my browser the images do not load. Some examples would be:
* http://www.pokemonfanuniverse.com
* http://www.pokecommunity.com

It seems to only occur for the domain's index page when there are relative srcs that lack the / before them. The Mozilla developers resolved this in Mozilla by auto inserting the trailing slash after the domain if it isn't there. Doing the same for IE would resolve this issue. -- Shining Arcanine

Hmmm, never seen that before. However most web servers these days are configured to redirect to the index page with the trailing slash out of the box, IIRC. -- Simon

This is not true. Most web servers are configured to redirect with subdirectories, not the domain's root directory. As far as the server is concerned, there is no difference between http://www.example.com and http://www.example.com/ and in fact servers cannot tell the difference between the two. -- Jim

I think I have one possible reason: this happens on sites which have enabled anti-image-hotlinking measures via Apache's .htaccess configuration file and Rewrite Engine. These sites add some Rewrite rules to redirect (or discard) image requests with a referer not within their website. The typical code that can be found on htaccess "trick" collections is:

		 [RewriteCond] %{HTTP_REFERER} !^$
		 [RewriteCond] %{HTTP_REFERER} !^http://(www\.)?mydomain\.com/.*$ [NC]
		 [RewriteRule] .*\.(jpg|gif)$ other_URL_image_address_to_redirect.jpg [R,NC,L]

I think this is also the code that cPanel generates automatically when enabling its built-in anti-hotlinkinig option.

This Rewrite rule allows HTTP requests of any image without referer (first line/rule, URLs writen directly in the browser's address bar), and allows any HTTP request whose referer is our own website (second line/rule) and finally redirects to a tiny placeholder image all other requests coming from 3rd party websites. Please note that the second line ends with "...\.com/.*$ NC". To match this regular expression, the trailing slash must be present even when requesting the root page. Otherwise it's considered to be an "alien" domain and therefore it's treated like a leecher external website: images aren't served properly.

To fix this issue from the webmaster point of view, the easiest way is to remove the slash after ".com" from that regular expression. However, this could allow a website named, say, "mydomain.come" to access our images, because it matches "mydomain\.com.*". Quite unlikely, uh? If we're perfectionists, we can leave the second line intact and add another rule before the last one:

		 [RewriteCond] %{HTTP_REFERER} !^http://(www\.)?mydomain\.com$ [NC]

This way we're allowing exactly the naked domain name as referer without trailing slash and nothing else following the ".com" (the $ is the end of string marker).

I discovered this issue while analyzing the effectivity of my anti-image-hotlinking measures, and I found some legitimate requests being redirected. The only reason I found was the missing trailing slash. The reported User-Agent header on those cases was:

		 Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1)

All the browsers I've used (as far as I can remember) added the slash when it was missing. I've reached this page searching why that particular browser didn't. Hope this is helpful to someone. -- Chungalin

The "Red X"

??IE Chat Transcript From July04?? *

		 John_MS : Q: Why do gifs show red X's in IE?
		 John_MS : A: If you can give me a consistently reproable example of this I will be happy to follow up on it. I haven't seen this...

This behavior started after ""IE6"" Service Pack 1. It seems that pages with numerous occurrences of the same .gif image intermittently cause the red "X" image to appear in place of the actual image when loaded. This can be reproduced by refreshing a page numerous times in a row.

Here are some links with in-depth explanation of when the problem occurs and the troubleshooting around trying to fix it:

http://www.experts-exchange.com/Web/BrowserIssues/Q20869911.html
http://www.experts-exchange.com/CommunitySupport/EEBugs/Q_20882550.html

Here is an example site where the problem is very reproducible (continue to refresh the page until you see, or don't see, little arrow images on the menu items): http://www.dmregister.com

I noticed this bug instantly after installing ""IE6"" Service Pack 1 and it has been annoying me ever since. -- tthibodeau

Some methods to solve this problem: http://www.dslreports.com/faq/6720 -- infrared

* I have tried all of the "solutions" from the link above without any results. I think people are grasping at straws for an answer to this bug. The simple fact is that pre-SP1, images loaded properly. Immediately after, not loading properly (thank for the link, though). -- tthibodeau

JPEG image crashes IE in XP SP2

Even after applying the latest patches for the latest JEPEG vulnerability, this JPEG will still crash IE6 in XP SP2:

		 http://sylvana.net/test/AP4.jpg

The Slashdot thread where I found this

Seems to be fixed in IE7 on Vista -- Fduch

Kerberos negotiation

IE6 is not able to negotiate Kerberos tickets when talking to a proxy server, even ISA, as per Q321728. It would be required in a modern secure environment.

Language issues

Links Toolbar

The Links toolbar is displayed with fixed-width tabs. I've always renamed everything in my LINKS folder to very short (2-3 char) abbreviations so I can see a whole bunch across the toolbar without opening a menu. In IE7 I can only see 5 or 6 even though their names are very short.

Loading non-western web pages

If you switch in IE to a non-western character encoding to display a web page in a foreign language, e.g Russian, Chinese, Japanese, it takes up to between one and two minutes until the page is displayed. Reproduce the bug by browsing such a page: http://www.baidu.com
When such a page is loaded, the bug does not occur again only if you open the links on the page or open a new window by going to File->New->Window
Otherwise, if a new IE window is opened by executing iexplore.exe, the bug occurs.

Logging into HP Government/Education Order Status

When logging on to HP GEM Order status page, http://gem.compaq.com the pop-up window that opens uses all "tiny" text. The formatting is correct, except the text size is tiny. This is not an issue using IE6SP1, or Firefox 1.0.6.

Character encoding doesn't adapt to page content

When a non-western web page has been viewed and after all IE windows have been closed, the character encoding in a new IE window is set to this non-western language and produces the loading delay bug described above.
See also the Wiki entry "Locale Issues".

Layers and forms

This a huge problem for developers as they have to write some hacks to hide form input tags to show layer above the form. This is a problem thay may be also seen with flash animation and DIV layer.

Locale Issues

When using English version of Windows and the non-unicode program language is set to Traditional Chinese, some web pages are displayed incorrectly. Many English web sites are loaded with Chinese encoding and require manual changes. Certain special characters in English pages (like the apostrophe, the temp degree symbol, the registered trademark symbol etc) are displayed incorrectly even when the encoding is set to English.

MSDN WebPage printing gets scaled down to barely readable

When printing an MSDN page (e.g. http://msdn.microsoft.com/msdnmag/issues/04/05/VisualC2005/default.aspx) using the print icon provided on the page the printout gets scaled down to two pages with a font size that makes the printout almost unreadable.

offsetLeft, offsetTop and offsetParent: incoherences in implementations.

Incoherences with offsetLeft, offsetTop and offsetParent implementations. If an object has no offsetParent, then the offsetLeft value and offsetTop value should be undefined. If an object's offsetLeft value and offsetTop value are integer values, then such object must have an offsetParent otherwise the definitions provided at MSDN do not make sense at all.
The offsetLeft, offsetTop and offsetParent definitions provided at MSDN are circularly dependent, mutually inter-dependent on each other. Therefore, the slightlest implementation bug on one of them impacts the understanding (and implementation) of the others.

Reduced testcase showing the bug

PDF viewing from an ASPX is broken in XP SP2

Under XP SP2, navigate to any .aspx page that generates a PDF and returns content type of "application/pdf". You just get a blank screen, and IE refuses to launch the viewer or acknowledge that anything is there or anything is wrong. Right-click on the blank page to view properties, and you see nothing. Do View Source, and you get "The XML source file is unavailable for viewing." Print, and the page is filled with an empty rectangle with a scroll bar. Close, and IE takes an access violation. - anfortas

I've tried to reproduce this with the sample code in KB 306654 and it doesn't reproduce for me with Acrobat 5.0.1 or 5.1 (I refuse to downgrade to 6.x). Probably the infernal cache bugs again. Clear your cache (Tools > Internet Options > click Delete Files) and see if it's reproducible. For me, fixing the cache bugs (listed above) should be one of the highest priorities (after security, of course) since it interferes badly with the browsing experience and causes highly non-reproducible conditions.

Answer: Yup, it definitely seems to be a cache problem. A few of my machines with clean caches do work fine. I still think it's strange, though, that downloading a .aspx causes the problem and downloading a .pdf doesn't. - anfortas

PDF viewing disables onclick (IMG only?)

If your Adobe settings are set to the default, which is to open the PDF in a separate window whenever */pdf content types are opened, the page you opened them from forgets onclick functions at least for IMG tags. This happens regardless of whether the onclick is statically defined as an attribute on the tag, or dynamically added via Javascript. Refreshing the page reinstates the onclicks. Fails in IE6 and IE7... and Firefox?

Printing pages re-requests images from webserver

Printing (or print-previewing) a webpage makes IE re-request the images on the page before printing, this gives some unexpected behaviour when using dynamically generated images that is based on data that change a lot. Suddenly the print doesn't match the page on the screen. The page contains no cache directives or anything, and the re-requesting have been seen by inserting a proxy between the browser and the server. -- ChrisPoulsen

Answer: Send correct headers on the images to explicitly permit clientside caching.

-- That's no help. Read the description - it's a page containing volatile data. Marking it as cachable will break things because then people won't see changes to the data.

Printing web pages often clips text on the right

Printing web pages from IE will often clip the text on the right, making the printout useless. Sometimes, printing in landscape helps. IE should be able to print any page it loads by resizing it to fit the currently selected printer settings.
If you visit this webpage: http://www.michaelkenna.net/html/iviews/phorevu.html and check out print preview some text on the right is missing.

That's because it's in a fixed-size frame - if you resize the IE window you'll see the same effect. People use fixed-size frames for all sorts of reasons, so if IE were to resize things to make it fit on a printed page it would inevitably "break" some layouts.

Printing web pages with blank input buttons prints submit query

When you try to print a web page that has input type button with no text inside and a specific width, the printed page show up a 'Submit Query' text instead of the blank text. --Solrac

Resizable windows at all times

Non-chrome windows, script initiated-window.open() windows should always be resizable. Always.
Give absolute veto power to the users when it comes to windowFeatures (like chrome toolbars presence and browser window functionalities like resizability and scrollability) of secondary window generated via window.open() calls. There is nothing more frustrating than a poorly coded link that creates a crippled window (non-resizable and without scrollbars). The window is frustratingly unusable and inaccessible in such cases and such crippled window can not serve the authors' goals nor give the user access to the content. Non-resizable windows can not serve the authors' best interests nor the users' best interests. All other browser manufacturers (Opera, Mozilla, Konqueror, Safari) out there have given the user entire veto power over web developers on window.open() windowFeatures. All non-chrome windows should always be resizable. Earlier versions of Internet Explorer ignored the resizable=no feature and allowed you to resize the new window. http://support.microsoft.com/default.aspx?scid=kb;en-us;211068
-- DU/GT

Sight impaired people usually need to disable stylesheets, that means that a lot of popup windows are too small to display the text, ie one can't read all of it - so ALL windows should be resizable, and the designer shouldn't be allowed to disable this. -- IceHunter

Save does not save Processiong Instructions

When a page containing a processing instruction is saved using IE's "Web page, complete" format using the Save As command, the processing instructions (PIs) are removed causing the saved page to render differently than the original page. Heresaveasbug.htm is a page that demonstrates the problem.

Scaling of graphics and fonts on High-Res and Widescreen Displays

When browsing on a laptop using a high resolution wide screen display (1920x1200) the scaling of graphics is off causing the images to be pixelated. If you choose to turn off the scaling, the text will not correctly size with images. I understand this has something to do with a non standard DPI setting on wide screens but these are now very popular and we need a fix for this.

Scripts loading issue, unable to pipeline/preload

When a page contains bunch of <script src=".."></script>, IE load it one by one, not using HTTP/1.1 pipeline. and even you preload it using var img; (img = new Image()).src = thescripturl; for each script url, IE load script using pipeline which triggered by Image(), but rerequest for <script src="..."> without pipeline.

Security Zones don't update "Custom Settings" when the slider moves

Go to Tools->Options->Security, change a zone's security level to medium, click Apply. Move the slider to High. Click Custom Settings and the settings are not updated to the High level, they are still medium. The danger here is that you may change one setting thinking you're in High security zone, and yet all the other settings are medium. (eg., "Run activex controls"-- medium: enabled || high: disabled)

Now you might call this expected behavior since it doesn't apply to IE yet anyway, but when you change somethig in "custom settings", IE calls the zone custom (which means your changes + settings of the last zone you pressed Apply for). I've learned over many months to isolate this problem & just press Apply after changing the slider of the zone, but this is not how it's supposed to work. --infrared

Select box width

When using a fixed width on a select box:

		<select style="width: 150px">
		    <option>text</option>
		    <option>long text will get cut off</option>
		</select>

The second option's text will be cut off as the dropdown list is only 150px wide.

Other browser (Firefox/Safari/Opera) handle this by extending the dropdown list to fit the content.

This is a known issue with IE5, IE6, IE7 and now IE8.

<span> inside an anchor (<a> tag) looses whitespace after

If you have some html that looks like this:

<a href="#">foo <span>lala</span></a> blue

It'll render like this:

foo lalablue

This is probably related to other issue people have had with spacing and overflow within anchor elements, also with img tags.

Standard Buttons Never Stay When You Open IE7

Every single time i open IE7, i have to right click and then click stardard buttons.
But sometimes when i open a html file from my hard drive it has the standard buttons already opened

Status bar disappearing (fixed)

IE6 is 99% fantastic but its frustrating habit of hiding the status bar drives me mad on a daily basis. Before I click on many links, I want to roll my mouse over them to find out where they will take me. Having to manually turn on the status bar every time is ridiculous (maybe a Tools -> Options option would be better?).

Fixed: in XP SP2. For earlier versions, see fixes here_explor.php or here -- JonathanHardwick

Submitting a Form Changes Security Zone

Why not Medium? I can't see what is different between that zone (medium) & the default, directly post-install zone (custom). It's confusing --infrared.

This is because the second you fill out a form, IE asks you if you want to allow the submission (ie/ google). Once you click yes, it changes your security policy and it displays as custom! The form-submission policy should be removed from the security area or at least reprogram the security zone code in IE to have policies that do not affect the current zone (ie/ form submissions).

Even worse, the Baseline Security Analyzer shows a problem whenever a user has submitted a form because their security zone has changed! :(

(I've added some detail to this and renamed it from it's previous title of "Zones are set to "custom" by default") -- ShadowChaser

Table with more than a 1000 rows

internet explorer hangs when displaying table with more than a 1000 rows

Has anyone seen this? I can post the HTML code if you would like to recreate the problem.
--WebZoner

TCP ACK Bug @ end of file transfers

As detailed in this USENET post:

A bug in the IE6 browser that manifests when more than one file is being downloaded simultaneously. (It is possible that the problem is more general and happens whenever more than one simultaneous TCP connection is established.)

When download of one file completes, download of the other file can halt until a TCP timeout restarts transmission. This is because no TCP ACK is sent for the first packet received on the uncompleted download after the other download completes. If the uncompleted connection has a high retransmission timeout value (due to the end-to-end flow being slow, congested, or maxi-hop), the uncompleted download can hang for a minute or more before a TCP timeout restarts the flow. In other circumstances the hang is short enough to be difficult to spot, but I've had hangs of two minutes or more.

Title bar errors

IE shows some kind of error messages in it's own title bar. For instance, when it tries to load from an unresponsive server or cannot resolve the domain name, it changes its title bar changes to something like "Cannot find server - Microsoft Internet Explorer". When you retry on the same window using the same URL and succeed downloading the page, IE won't replace the title for the actual title of the page. This affects usability in a few ways. For instance when you later want to create a shortcut to the URL to add it to Desktop or Favorites, you will get a file with the name "Cannot find server.url".

Toolbar bugs

When having additional toolbars installed in IE, the names of the toolbars are incorrect (clicking toolbar A shows toolbar B and vice versa). Also, after arranging the toolbars and locking them, the next time the browser is opened, the positions are different, but unlocking and locking them again would restore the position.

This problem seems to be caused by a corruption of the data in the HKEYCURRENTUSER\Software\Microsoft\Internet Explorer\Toolbar\WebBrowser registry key. Deleting it seems to force Internet Explorer to rebuild the toolbar list and corrects the bug. It also seems to affect the Windows Explorer (HKEYCURRENTUSER\Software\Microsoft\Internet Explorer\Toolbar\Explorer).

Transfer-Encoding: chunked not working on first request

IE does not seem to load the chunked response the first time a page is rendered. This applied for both in-site links and open URLs in a new instance/window. After the initial load, refreshing same page will make IE load the chunked response as it's suppose to be (one chunk of response at a time). More details with demo movie and HTTP header log is available here: http://blogs.fanms.com/sondre/PermaLink.aspx?guid=1e89cc07-0af3-4f71-a638-c8a7f17d7571

Unicode

Handling of multiple BOMs inside a file

The current treat of multiple occurrences of U+FEFF (UTF-16BE BOM) inside a HTML page is quite bad. To me it ends up showing lots of blank space on the top of a table (the occurrence of the additional BOMs was inside the table between the rows). Of course it's not recommended to have multiple BOMs inside a HTML page, but since it could happen it should be dealt with IMHO.

The unicode.org FAQ kindly covers this issue: What should I do with U+FEFF in the middle of a file?_bom.html#38 . In the last sentence it states: In that case, any U+FEFF occurring in the middle of the file can be ignored, or treated as an error.

Thus i'd like to suggest doing as this FAQ suggests, by ignoring the additional BOMs completely. Treating them as an error would be not very tolerant, and that's what the IE is supposed to be in similiar error conditions, isn't it :-) -- Clemens

User Interface issues

I know many of these are because of the IE/Explorer duality. Eg the status bar needs to be on in both (open just one window in each, turn it on in each, close both and it's set on for both). There are too many places where IE's interface isn't like anything else and hasn't been touched in years. Take Favourites - which everyone uses a lot, many developers want to be able to use/manage/integrate with. Please take the Favourites dialog box!! For one it should have the same behaviour as the Favourites pane in the tree view but add extra buttons and features. It should be like every other 'save' dialog and remember where I was when I last saved things - because a bookmark is a shortcut so it's a Save dialog. and it should have the same behaviour as the menu - update the menu and the task pane to have the same order as I've chosen in the dialog rather than not alphetising them until I do it manually! - Mary Branscombe

Here's one that worked in IE4 and broken in later versions. Let the bloody program remember if it was maximized!!!! -- Scott
Neugroschl

Something like Firefox's or Opera's bookmarks manager would be nice. The current bookmark manager (when you click Organize Favorites) is useless in MANY ways and I'd much rather use Explorer (the shell) to organize them. At least then I can select and copy/paste multiple bookmarks. ORganize Favorites is single click everywhere and therefore you can't select multiple bookmarks. --infrared

View-Source protocol is broken in XPSP2

Typing "view-source:http://microsoft.com/" in the address bar should bring up the source in notepad. It just doesn't work anymore in IE with XPSP2, instead giving a "page cannot be displayed" error.
* MSDN docs on the view-source protocol
* A forum thread for reference (What's up with view-source:sURL run in IE 6 SP 2?)

Perhaps this was intentional -- if so, why isn't it documented here (Changes to Functionality in Microsoft Windows XP Service Pack 2) or here (Fine-Tune Your Web Site for Windows XP Service Pack 2) ?

This is very bad. Please fix it.

Answer: This was removed. Need to update docs.

Windows integration in IE7

Playing with both IE7 Beta 1 & 2 in Windows XP, I'm a bit annoyed between the way the windows integration changed between IE6 & IE7 & Reklama & Serwery & Hosting .

In IE6, if I was in Explorer navigating my HDD and then suddenly I decided to jump online, I simply typed a Url in the address bar and my window switched from Explorer to Internet Explorer (and vice versa).

In IE7 though, a new IE window will open rather than reusing the Explorer window (same thing happens for the reverse. If I'm online and want to navigate my HDD (something I often do after downloading files) it won't reuse the window, but rather launch a new explorer window.) Please fix this.

While we're at it, tabbed regular explorer (for Windows not the Internet) would be cool too...

Window size

The user should have more control over the default window size, the size set by Ctrl-Closing the last window often is lost again. It should also be possible to set it to "maximized".

(In other words, when clicking a link that opens a new window, the size is not the one specified by the last window closed. Resizing, closing and reopening does not update IE's window size setting.)

Don't add stuff here - scroll up and find the right alphabetical place to put your heading

kooziesgirl

Internet explorer is the nigghtmare of any front end developer, granted IE 7 and IE 8 are much better then IE6 but sadly there still are some idiots out there who use IE 6 and for them we need to speand a crap load of time optimizing. Koozies

internetexplorerbugs