Multiple versions of IE

Hi,

My understanding is that if I update to IE8 I won’t be able to retain IE7 and the IE7 rendering mode in IE8 isn’t exact. I want my sites to appear as I intend in both. Anyone know how I can run both IE7 and IE8 on the same machine? Do I need a virtual machine? If so, where can I get one (preferably for free) or at least what terms should I google?

(If there’s a better section of these forums for posting this question, just lemme know!)

Thanks so much!
James

Yes there is a better section and possibly: Computer Hardware and Software

Though a virtual machine should work I suggest you call a MOD and they’ll move this post.

Hi,

I moved the thread to web design as we often get browser questions here.:slight_smile:

The IE7 rendering on IE8 is not exact and there are a few bugs and differences but most of the time it will give you a rough idea.

You can use ietester to test multiple IE’s if you don’t want to run a virtual machine. I use it all the time. It’s a little unstable but works very well most of the time and is good if you want to check sites locally. I wouldn’t use t for on-line browsing unless it’s a site that you know.

There is also another collection here although I haven’t tried it.

IETester doesn’t work well for me. It crashes about every 5 minutes. But then again, I’m using it in a virtual environment on the Mac. From what I’ve heard, it’s much more robust an application running outside virtual environments.

This is a nice article that summarises most of the options you have. (many of them free)

It crashes quite a lot since I started using IE8 for real but when IE7 was the real version it hardly ever crashed.

I have parallels on my mac but haven’t tested ietester on there (not that I want to anyway). I have windows 7 running under parallels with IE8 and XP with IE7. (I have vista on my laptop so manage to be able to test most things locally :))

I have Windows XP and am running IE6 natively. All the other versions I’m testing via IETester or Spoon. The latter doesn’t even let me test anymore. Something with NET application needing to be installed. Weirdness. I haven’t had the nerve or time to sort that out, so I’m stuck with IETester for the moment.

At some point it might really make sense to buy a cheap, used laptop just for testing purposes that runs either XP or Vista.

I have two virtual machines on my Mac, one for IE8 and one for IE6. MultipleIE is also installed on the IE8 machine but it isn’t 100% reliable. If I find it necessary, I may get a third with IE7 but until now, compatibility mode has worked fine. Litmus.com can also be an idea for final compatibility testing.

I use Parallels for the VMs. Also tried VMware which was excellent. VirtualBox(.org) is free and so is VMware Player.

FWIW, here are some links I’ve collected over time, some of which are mentioned above:

BrowserShots

http://justfreetemplates.com/blog/2009/09/18/7-ways-to-check-cross-browser-compatibility.html

http://my-debugbar.com/wiki/IETester/HomePage

BrowserCam
http://www.sitepoint.com/forums/showthread.php?t=345568

VirtualPC for Windows (free):
http://www.microsoft.com/downloads/details.aspx?FamilyID=04d26402-3199-48a3-afa2-2dc0b40a73b6&displaylang=en

http://www.sitepoint.com/blogs/2006/10/27/browsershotsorg-test-nine-browser-for-free/

http://spoon.net/browsers/

There is also a plug-in for Firefox which allows you to render in ie5.5, ie6, ie7 and ie8. It passes the site over to http://ipinfo.info/netrenderer/index.php with the render type and url passed. The only downside with it is it isn’t able to handle local development, so it has to be in a publicly available site. But it is an option.

I would not bother with virtual machines - it is definitely an overkill as there are better solutions.

I was told multiple times that setting “browser mode” to IE7 in IE8 developer tools is 100% equivalent to running IE7 natively.

Well they were wrong I’m afraid and there are some small differences as have been noted in a number of posts on these forums.

Hers is an example of an IE7 bug that only shows in IE7.


<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
               "http://www.w3.org/TR/html4/strict.dtd">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<title>IE7 general sibling selector (~) bug 1</title>
<style type="text/css">
/* The ruleset below (empty or otherwise) is required for the bug! */
.holder b {
}
 .holder ~ span b {
 background: red;
}
</style>
</head>
<body>
<div class="holder"> <i>
    <!-- HTML COMMENT OR ANY ELEMENT -->
    <span> <span><b>My background is ony red in IE7</b></span> </span> </i> </div>
</body>
</html>

IE8 in any mode fails to find it or with any ie meta tag.

Therefore it is unreliable to use IE8 for testing IE7 because specific bugs are not found which would be the reason for checking in the first place :slight_smile:

Really? I took it to mean that IE7 in browser mode (Developer tools) was identical to IE7 and it was only IE8 (Compatibility mode) which was IE7.5 (the IE7 version with differences). If that is the case where IE7 doesn’t do the job within developer tools, I use the Spoon browser emulators and IETester with good success. :slight_smile:

No it’s not I’m afraid and there are differences that have been noted.

Just try my code above and you will see that IE7 and IE8 emulating IE7 (or IE7 browser mode in the developers tools) are all completely different. I would never use IE8 to test IE7 for anything important.

Ahh well, I use IETester to perform my cross browser checks (generally) anyway, it may not be entirely perfect and has the odd glitch here and there but IETester is the only non-virtualized product (along with Spoon’s testers) which genuinely use the EXACT same JavaScript + runtimes + engines that each version should use :slight_smile:

Interesting.

Well, this is not entirely true :slight_smile:

Wow, these links and resources are nice to put together as a sticky (or added to the one in Computer Hardware with a shortcut here).

I run Ubuntu (which is also nice to test in, as the Gnome windowing system DOES affect Firefox re fonts etc) with VirtualBox running two copies of Windows XP: one with native IE6 and Safari4, the other with IE7 and Tredosoft Multiple IE’s for quick IE6 testing (extensive testing is done in the native) and Safari 3. For Firefox 2 I use K-Meleon which still uses the FF2 version of Gecko. This was a hell of a lot easier than playing with profiles (seriously, that’s just silly).
What’s also good about having either Linux or a Mac for testing is for stupid stuff like how Mozilla decided to destroy all PNGs on those OS’s (I don’t notice a problem on FF3.5 on Windows) to “enhance” pngs. What a terrible idea… my own site looks like some retard built it (ok, some retard did build it, but the png thing is Mozilla’s fault, honest).

I’d love to know some cheap way of testing real Safari-on-a-Mac… without shelling out for a Mac.
I’ve resorted to waiting an hour for Browsershots to show me what Zapfino looked like on a site. Terrible.

Actually, you can reduce this to one WinXP installation - it is ease to install IE6, IE7, Safari 3 and Safari 4 on the same PC with BrowserSeal.BrowserPack - http://browserseal.com/index.php?option=com_content&view=article&id=7&Itemid=16

Thanks, I did see your sig earlier however I don’t want to add Java (no, I don’t have it that I know of) and it says “screenshots”. How does this test links, Javascript, browser resizing, Flash, and user interaction (highlighting text etc)? I also use the dev tools that come with IE. And, I already have the other browsers… I could have Saffy 3 and 4 on one VB OS by putting them in different folders but I was lazy and just didn’t : ) IE is pretty much the main issue due to its use of the DLL files.

Sounds useful as a one-time download for someone who doesn’t already have most of those browsers, with them being all tabbed.

You are right, BrowserSeal cannot test user interactions, however it does support Flash and JavaScript.

Ideally, one would want to combine something like Selenium and BrowserSeal, but unfortunately such a product does not exist…