Internet Explorer 7 Beta 2 goes public!

[mark87]

Does anyone know how I can install IE7 Beta 2 yet keep IE6? I can't seem to find any info on it.

A post today on the IE blog says we should test our sites in IE7 (http://blogs.msdn.com/ie/archive/2006/01/31/520883.aspx) so I suppose I better but I don't want it to overwrite IE6.

[liorean]

Title says it. Release notes (http://msdn.microsoft.com/ie/releasenotes/default.aspx). What's new (http://msdn.microsoft.com/workshop/essentials/whatsnew/whatsnew_70_sdk.asp). Information Index for Internet Explorer 7 Beta 2 Preview (http://msdn.microsoft.com/ie/infoindex/default.aspx)

Download ie7wb2 (http://go.microsoft.com/fwlink/?LinkId=58957)

[mark87]

Don't worry found out how from an article (http://weblogs.asp.net/jgalloway/archive/2005/08/12/422335.aspx).

Download IE7.

Open the install file with WinRAR (yep, it's an EXE, but it's a self extracting EXE) and extract them to a folder. Alternatively, you can just run the EXE, make a copy of the files which are extracted in the first step, and cancel the install. I was too chicken to do this on my work computer.

Look in the folder you extracted to for the file SHLWAPI.DLL. Delete this file.
Create a new text file in the folder, then rename it to IEXPLORE.exe.local.

[cjwsb]

Very cool, I gotta do this... Do you still need an invite to DL IE7B2 or is it public?

[mark87]

Nope it's public - http://go.microsoft.com/fwlink/?LinkId=58957

Now I'm getting error messages in IE6 though! Even though I did the stand-alone thing. :(

Most of my sites look OK, one I'm having a bit of a problem with because it's rendering more like Firefox so I took out the IE hacks and placed them in a seperate stylesheet, then used a conditional comment (<!--[if lt IE 7]>) to suposedly only serve the hacks to IE6 and below, but it doesn't seem to be working (IE6 displays it how it used to, so does FF, but IE7 is still looking a bit weird)!

[cjwsb]

No offense, man, but I think I'll let you work out the kinks before I dive into IE7. Hardy har har... *smacks myself on your behalf*

Chris

[liorean]

Ie comes with a rather small browser part, and a very large system part. The system part will replace the older system part if you install it. Then all trident users on the system, ie5.0 ie5.5, ie6.0, ie7 included, will use the system parts for some things, even if they have been localised. The way the program is structured today doesn't allow Microsoft to create an entirely self contained version, though they are working on providing a stand alone rendering engine part (hopefull we'll see this come in fruitation for the version after ie7).

One thing that will always use the system parts is the conditional comments syntax, including version detection. So all of them will behave as if they were the system installed variant when it comes to conditional comments.

[mark87]

I just had a *big* problem with using IE6, after running IE7! Everytime I clicked on a favourite, opened a new window, typed in a new address, went Home etc, muliple windows and errors were popping up! I finally found a solution so this should help the guys who choose to run IE7 alongside IE6.

http://windowsxp.mvps.org/ie/905915.htm

All it is is a simple registry key deletion, all working now.

I'd advise at least trying IE7 (of course it's up to you though), I've noticed it renders very differently from IE6, and even differently from Firefox too; so you may want to check out any sites you have before the full IE7 comes out. I know it's a while but I like to be prepared.

We now basically have a new browser to work with!

[drhowarddrfine]

It's running different because every site that has an IE hack now fails the hack.

[drhowarddrfine]

they are working on providing a stand alone rendering engine part (hopefull we'll see this come in fruitation for the version after ie7).Where did you hear about this? It's new to me.

[mark87]

It's running different because every site that has an IE hack now fails the hack.

Exactly. Which I why I tried using conditional comments like so -

<!--[if lt IE 7]>
<link rel="stylesheet" type="text/css" href="ie.css" />
<![endif]-->

And then placed all IE hacks in ie.css.

But even though IE6 still displays with the hacks, IE7 looked no different with the IE6 hacks applied or not. :confused:

Ah well, too tired to work on it now.

[liorean]

Where did you hear about this? It's new to me.It was to me too, when I saw it an hour ago.
# re: IE7 Beta 2 Preview Available (http://blogs.msdn.com/ie/archive/2006/01/31/520812.aspx#520894)
Tuesday, January 31, 2006 1:58 PM by cwilso
Noah,
we understand and have heard this feedback. "IE" is actually a collection of system components - networking, browser hosting, core HTML rendering, printing, etc. When we install a new version of IE, we're installing it for all applications that use these system components - including the tiny iexplore.exe itself. That's why it's hard to have multiple side-by-side IEs. We are working on a solution for just HTML rendering, but we can't have multiple system components installed with the same name.

-Chris Wilson [MS]

[brothercake]

It may not support the old hacks, but that doesn't mean it fixes the problems the old hacks were used to prevent, or that it doesn't introduce new ones. I shall reserve judgement ;)

[marcus1060]

I had to increase my text-indent from 1000em to 2000 to make the text for the header invisable, even though it worked in IE6, FF, and Opera the other way round.

[Masterslave]

Mmm interesting. IE 6 and 7 next to each other.:rolleyes:
Hope I'll get it fixed. :D

[mark87]

Now after I used that registry fix to run IE6 again without all of those error messages etc, IE7 just opens for less than a second and closes without error! :(

*Searches to find a solution.*

[brothercake]

It's good to see that their own pre-testing has been so comprehensive :rolleyes:

http://www.brothercake.com/cf/ie7-startpage.gif

[brothercake]

I've run it through the filter tests at dithered.com to produce these results (it's a spreadsheet)

Haven't analyzed them yet, but a quick glance confirms what we knew already: "* html" is gone, attribute selectors work, so do + and > selectors,

and a couple of interesting nuggets: does still support the nonvalidating "html*element" hack, doesn't chew over XHTML mode pages (but that doesn't mean it supports them - haven't checked - it just doesn't prompt for download)

[Joey123321]

I can't wait for it's release.

[brothercake]

Well we'd better hope this turns out good, because further analysis reveals no pure CSS filter - the only non-scripted way to identify IE7 uniquely is to use conditional comments, which is bad in my book because conditional comments are HTML hacks, and therefore, not as clean as CSS hacks - changing a site to support IE7 could mean editing hundreds of individual pages, rather than a single stylesheet.

I'm forced to conclude that the IE7 team have done us a disservice by removing * html, and we should pressure them to put it back (http://blogs.msdn.com/ie/archive/2006/01/31/520883.aspx#522663), because with * html in combination with a CSS2 attribute selector, it would be possible to identify IE7 uniquely

[brothercake]

In fact the more I think about it, the more annoyed I become - the removal of * html seems very much like Microsoft putting their own commercial interests above the interests of the web in general.

Now it's true that with * html still there, sites which are using it to feed CSS to IE6 might break in IE7, thus forcing basically everybody who uses it to update their site. Except ... we've got to do that anway, the stagnation of IE6 forced us into using these hacks, and the popularity of Windows forces us to consider IE7 - Microsoft leave us no choice but to update our sites.

Yet, they've taken away one of the key tools that could do the job.

[liorean]

James: I don't agree. If they had it left in the engine, all pages that exploited the "* html" bug for any particular rendering bug which they fixed in ie7 would essentially be broken and need to be fixed. As the exploit would still be targetting a live user agent all uses of it would have to be checked over for all rendering bugs fixed. By comparison, as it looks right now, that exploit now targets dead user agents - ie5/6w and ie5m. That means no page that uses it to target mentioned dead browsers will need to change their current hacks. They just need to add ie7 filtering for all remaining or emerging rendering bugs in ie7.

As I see it, not breaking the current web is more important than providing a hack to filter the new user agent.

[drhowarddrfine]

Coding hacks to work for IE5/5.5: $500
Coding hacks to work for IE6: $500
Recoding sites to work with IE7: $500
Coding once for Firefox and all other browsers: priceless

[drhowarddrfine]

Besides the now known security DoS issue, CSS errors are popping up. Here is a serious one. (http://www.brunildo.org/test/fenc7.html)

[brothercake]

Here's a simpler test case of what I think is the same problem - http://www.brothercake.com/reference/ie7/float-overflow.html

[brothercake]

Found another CSS bug - this one is box-model related: http://www.brothercake.com/reference/ie7/position-boxmodel.html

[drhowarddrfine]

IE7 Bugs Abound (http://news.zdnet.com/2100-1009_22-6034054.html)

[drhowarddrfine]

On another board one guy said he got an email from Juno saying IE7 does not work with their software and you may not be able to connect.