Windows Integration Guide - murphy78

Well, not if the August Rollup doesn't get fixed.
I worked on this for 4 days straight since I was including OEM indexes this month.

While I personally didn't notice any font problems or bsods, they could be related to hardware compatibility problems.
I have a rule that I never release anything that I wouldn't proudly use on my own system and any family member's system.

 

I will do win9 when the final version is released. I don't deal with previews or leaked non-rtm builds.
As far as whether or not I modify it with updates, it depends if MS releases a ton of updates the first month like they've done for win8 and win81.
I probably won't bother making an AIO until we have all the various versions and MS has released a certain amount of updates.

Also, there's a distinct possibility that they will severely change the kms activation to use white-listing and/or certs. So depending on how much they modify things,
it might be a better idea to watch for transition deals like they did with win8. Well, I consider it always a good idea to support MS, but their 100 and 200$ price tags are ridiculous.

 

I'm not entirely sure. That would be my speculation, but none of us have seen any actual MS blog post or actual MS-sourced information about that.
We've only heard a rumor about it from MJF.

 

Well... It depends on the OS and/or Office version.
The important bits are:

• Nobody has cracked Office 2010/2013 Retail activation
• KMS activation is the only way of activating without stealing someone else's key/activation charge
• Office 2010 activation is independent of the OS
• Office 2013 uses the host operating system's KMS protocol if the version is greater or equal to v5.

MTK works by setting up an emulated server in the system's memory. They've gotten it down to the point where an actual virtual machine is not necessary and
they can simply 'pretend' a server is on the other end of the connection. It's very complicated how they return the successful response. If you want to read more
about the actual guts of the process you can read the KMS source code thread. I think it should still be up.

Basically the activator sets itself up as if it were the KMS Server and receives the encrypted request.
It then decrypts the request and finds the data it needs to create a successful response.
It uses a SALT IV (basically a starting point for a hash), a few relevant data points, and puts things in the correct order with buffer data and such.
It also encrypts the response. Not all of that stuff happens in that order.

Although it uses TCP/IP, it doesn't actually need to transfer anything over the internet.
In MTK 2.4.0 and beyond, it actually doesn't even transfer to a virtual machine.
Depending on the Office version it will set the OSPP to activate to either the localhost (127.x.x.2-254) or some dummy network address in the case of the TAP/tunmirror method (10.3.x.x)
The 10.3.x.x range stuff is what intranets use and there are ways to copy things from localhost to the 10.3.x.x address (hence the mirror portion).

There is also a windivert option which does a very similar process but using a kernel driver instead of a TAP adapter, but does the same essential thing of bypassing Win8.1's localhost block.
There is secohook, which is the default option in 2.5.1 and 2.5.2. This option intercepts the sppcomobj iirc and basically cuts out the network stuff and jumps straight to the successful response processing.

In 2010, this stuff was only specific to the Office activation method OSPP.
In 2013, it depends on the OS. In Win8.0(2012) or greater, it uses the Operating System's KMS protocol.
In Win7/2008r2 or before, it uses KMS v4 I believe.

I could be mistaken on those details. Office is a little weird in that it doesn't just support a single version.
Office supports many different kms activation protocols. For instance, Office 2010 can use KMS 1.1 or greater.
There could be an upward limit. I'm not savvy with some of the more technical aspects of the details.
I've been using a lot of the stuff, but I didn't do any of the programming. Honestly, I barely understand a lot of that stuff.
I tend not to delve too much into programming because I have an eye problem that gives me headaches if I spend too much time reading.

It would simply take me too long to get anything done if I had to only program for 30 minutes at a time.
I used to do it when I was a teenager, but the headaches made me quit.

 

I don't believe you can boot from an esd. You could try changing the boot lines in the {default} part of the bcd, but I really doubt it would work.

 

You need to login to view this posts content.

 

Ahh yes, that's because they are permanent and won't get unlisted unless every single component of the update is replaced and the system has gone through a /resetbase.
At least I believe it will remove them...
You can try yourself by installing on a vm and doing a resetbase and see if they go away.

I wish we could resetbase offline images; but, alas, the pending issue won't let us.
I wish they'd change it so it was similar to the winpe ability to resetbase in waik5.1

 

Yah, I don't typically do syspreps because they don't play well with AIO versions.
Offline integration queues things up using the same files.
Sysprep likes to create unique identifiers or something for each update you install.

Offline AIO integration nets like 4gb for an x64 wim.
Online sysprep AIO ends up being like 6.5gb despite all indexes being /resetbased

Thinking of your integration issue with non-english versions... I believe you mentioned an update or two that couldn't integrate offline?
Perhaps if you mentioned the order things were integrated when you ran into this problem, we could troubleshoot.
Who knows? Maybe it just doesn't like to offline integrate a particular update if netfx3 is enabled or something weird.

Try using a script and pipe it to a txt file so we can see where the error occurs and suggest...
like:

Code:

call integratex64.cmd >x64.txt

 

You need to login to view this posts content.