Proneax

yeah, type-o... use the batch file honestly tho, you didn't even try using the batch file?? you create a copy of the inf because if you don't copy the original file back, the installer can tell you modified it and won't install. I don't understand your first question. The installer comes as a compressed package... if you run it it just extracts all the stuff then runs the exes in order. We just do that manually so we can modify the install process.

So, I had gotten Beta 1 to work a while ago using a batch file and a sleep timer to copy the update.inf over, however when I applied this method to the final release it didn't work, if I set the timer to anything 1ms or greater it would give me the 'cannot validate blah inf', and without the timer it would give me wrong windows version errors. Anyway, I figured out a way to delay the inf detection and allow me to copy the inf. This involved renaming the security catalog file and replacing it with a dummy. When the cat file is missing, there is a long timeout, so I tell my batch file to start update.exe, wait 100ms, copy the inf, then copy the cat file over. Update will wait for the real cat file to be appear then validate, so it sees the original inf to compare with the security catalog after it approved the windows version with the edited ini file. I repeated this step with all the extractable exes inside the installer. Here is the batch file: @ECHO OFF SET COPYCMD=/Y echo hi xcopy Mupdate.inf update.inf start update.exe start /wait Sleeper.exe 100 xcopy Cupdate.inf update.inf xcopy wmp11.cat55 wmp11.cat pause So, you need to: 1.) extract the downloaded installer package to a folder 2.) extract each exe inside the folder from (1.) into its own folder (5 extractable exes) 3.) go to the update folder within each extracted folder from (2.) 4.) create a copy of the original inf and edit as below, then rename the copy Mupdate.inf (see below for values) 5.) create a copy of the original inf file as Cupdate.inf 6.) rename the appropriate cat file *.cat55 and create a new file (ie a new text file) named *.cat (purpose of creating the blank file is to suppress a copy confirmation in the batch file) 7.) Make sure the batch file has the correct "cat" names, they are different for each exe. 8.) Run the batch file for each update folder. MaxNtMinorVersionToUpdate=2 MaxNtMajorVersionToUpdate=5 NtBuildToUpdate=3790 NtMinorVersionToUpdate=2 NtMajorVersionToUpdate=5 MaxNtBuildToUpdate=3790 MinNtServicePackVersion=0 MaxNtServicePackVersion=2304 The format is not the same in all the infs but the values are. oh i attached sleeper.exe too in the batch above, it waits 100ms, which was good for me but you may need to adjust if you get errors. If you get an error you can just re-run the batch file it saves the infs but you ahve to delete the *.cat file and re-create the fake one manually. Have Fun! Sleeper.exe

There are several things here that can cause confusion. C:\XPCD\$OEM$\$1\install\Applications\OfficeXP\Unattended.MST is from the local directory c:\xpcd that this person is using to organize the cd before imaging and burning. When you do an unattended installation, all files are copied over to the hard drive and then everything is executed. So, in theory you could reference the file in terms of c:\ if you were certain your install drive is c:\ It would look like this: C:\install\Applications\OfficeXP\Unattended.MST because $OEM$\$1\ is copied to the root of the drive (ie c:\) However it is much easier to use a relative path (no drive letters) in executing the command. just using TRANSFORMS=Unattended.MST means the MST file is in the same directory as the setup file.

I've been trying to figure out the best way to have an unattended install that includes programs I want to have the latest versions of, like winamp and firefox, without re-burning the entire disk. Well, I finally figured out that you can use nero to delete or update single files on a cd or dvd rw without affecting other files. So, once a month you could update your cd/dvd with the latest program versions. The only drawback is that it doesn't actually 'erase' the files, so they still take up space, but you can't 'see' them on the disk. - meaning you can add updates until you run out of extra disk space. A couple of specific things you have to do for this to work the best: -first burn a session with all 'permanent' files (name it 'Permanent') -have a reference in cmdlines.txt (or something) to the batch file in your 'updates' session. - burn a session with all 'updatable' files, including the batch file to install them / add them to runonce. (name it 'update_1') The reason for multiple sessions/names is when you go to update, you have to pick a session to update from. Having separate sessions just gives more flexibility. - To update files you can do one of several things: 1.) keep an image of your 'updates' folder on disk, change anything necessary in that folder, then in the 'new compilation' window multisession tab in nero, under 'options' select 'delete files' 'add new' and 'replace file/date'. After you select the updates session, it should automatically update the cd with any changes in that directory (it has to be the same directory you originally burned the files from). It will give you some statistics so you can check, and also bold the files to be updated. 2.) You can manually select current files on the disk and just hit 'delete' or if you copy new files of the same name you can 'replace' the old files. Given the fact that these options are pretty apparent to anyone using nero I'm not sure if a lot of people know this already, but I figured I would throw it out there as I just figured it out. I can post more details if anyone wants.

the whole problem was that the drive letter weren't set up properly when windows was putting stuff in 'all users' during installation. What i finally ended up doing was using a boot program called Bootit-ng to set up an ntfs partition (not formated) Then when I booted into setup it recognized the ntfs disk as c: (it was device 0) and the fat32 as d: I also used the "slide" function of Bootit-ng to align the clusters on my fat32 drive so I could convert to NTFS without getting the 512byte clusters that you usually get. It worked great and my clusters are 4k.

Ok, tried installing with only 1 hdd (this time using VMWare trial, which has much better performance) So now my system disk is c: but the cdrom is still d: and when I add the other disk it's e: :-( What I did was boot first time (textmode) to set up partition without the 2nd disk. Then when it restarted I shutdown the machine and installed the second disk, then started the machine. I actually tried taking the cdrom offline but apparently setup still needs stuff from the cd during GUI Mode, so it restarted and I had to add the cdrom again. I'm thinking It would be easier to set the profile directory to the sytem drive (c:) then move the profile directory after I install. I'm not exactly sure how to do this. I know that you can change some settings in the [.../profilelist] key but I actually tried doing that with my previous tests (using e: for the 2nd hdd) and it didn't work. So how would I go about moving the profile directory after install?

Thanks Tribble, I guess I'll have to disconnect the HD during install then. I might have to disconnect the cdrom as well so it doesn't make that D: I wish there was some way you could force windows to change this, like set the cdroms to w: and x: so they're out of the way. It might seem that way but oddly e: is the boot drive. - the fat32 drive doesn't even have anything on it. (in my virtual pc test setup)

I'm having trouble getting windows to setup my drives with the correct letters. I've been testing this in MS VirtualPC 2004. What I have is: Device (0,0,0) A brand-new, blank drive, unformated. Device (0,1,0) A current, fat32 drive with some data on it. I want to be able to format the new drive during setup and have it set to drive C: while the older drive become drive D: So, in my winnt.sif file, i have these relevant settings: [Data] AutoPartition=0 MsDosInitiated="0" UnattendedInstall="Yes" [Unattended] UnattendMode=FullUnattended OemSkipEula=Yes OemPreinstall=Yes UnattendSwitch="yes" TargetPath=\WINDOWS Repartition="no" [GuiUnattended] OEMSkipRegional=1 TimeZone=35 OemSkipWelcome=1 ProfilesDir ="D:\Documents and Settings\" So anyway, when I boot from cd first time, i get the parition screen, and I choose to partition the new drive as NTFS, the old drive shows up as fat32 and because Its the only partitioned drive I says c:. After it restarts again I go back to the partition screen just to check and it says: c: ntfs d: fat32 with the correct drive sizes corresponding, I must be all set. Well, doesn't work out that way. When setup finishes and I go to login, I get something like "windows could not create a temporary profiles directory" and "windows could not load your user profile - Access Denied" Notice I have my profilesdir set to D:\Documents and Settings, and that should put the profiles on the fat32 partition. So, if I'm an administrator it lets me login with a system profile and it shows that: c: fat32 d: cdrom e: ntfs Any Ideas on what I can do here to get this to work? I will say that I'm using dynamically expanding virtual drives, would it work better If I used a fixed sized virtual drive? I thought it was transparent, the error messages say nothing about drive space.

let me also add that if you have a username that is the same as the computer name, you can't use the net command to add that user to a user group. It will say something like 'user does not exist' error 3317.

Been looking into this a little and I thought I'd share my findings. Most people know that you can convert fat32 volumes to ntfs using microsoft's "convert" program. However, unless that fat32 volume was created using their new "oformat" utility, the first several clusters are only 2k and throw alignment off for the rest of the drive. This means that convert.exe will make your clusters 512bytes, which slows performance. To remedy this you can use a utility (bootit ng) that shifts all of your clusters to the proper alignment. There is a page here that explains this process in more detail. The program is shareware, but the 30 day trial gives you plenty of time to make the boot disk that I assume doesn't expire.

HKEY_CURRENT_USER\Software\Logitech\MouseWare\CurrentVersion\Control Center\Schemes\.Default\Devices\CordlessOptical6\0001 In case anyone is interested, the above key is where I found the configuration data that the mouseware utility edits. Under the "Usability" key there is some of the common settings like mousetrails, double click time, acceleration. Under the "assignments" key there is a key for each mouse button. There is no setting to tell what each button does but instead a subkey whose name indicates the function of the button. looks like [PermanentEvent\UnivScrollPermanent] has some extra features Also, the silent method requires a restart before the install will take effect. You may run into problems if you attempt to configure the mouse before restarting. attached is the entire logitech registry entry after a default configuration with only 1 change: the middle mouse button click is changed to "explore my computer" the function subkeys don't correspond exactly with their displayed names in the mouseware utility so the easiest thing to do is configure in the utility then export the registry file. Tested this and the entire configuration is based in the registry - remove the reg file and it thinks it's a new install, add the reg file after a new install and it skips the new install config screens. log.reg

Haven't seen anything about this on these forums or anywhere on google, but in my preliminary testing i've used "setup.exe /s" to silently install logitech mouseware. Let me explain. I used winrar to extract the exe file from their site. Setup.exe is in the root folder. I have tested this by uninstalling mouseware, verifying that logitech's custom mouse properties page is gone, then running this command. Upon completion, clicking mouse properties I get "Setup has detected mx700 mouse, configure now" or something like that. then I click next a couple times and it's done. I haven't tested this on a real installation yet, will soon.

First of all, to answer Bâshrat's question: AFAIK, when you specify IEAccess etc, you are NOT removing the fundamental component of Internet Explorer and therefore you still need to apply the hotpatches because they apply mostly to the underlying framework that you aren't removing and not the superficial IE overlay. This is different for OE, and I'm not sure about WMP, never seen it in action. Anyway, gosh, do you have any examples of programs that don't work well when IEAccess=Off is set? Bâshrat mentioned diskkeeper but that appears to be an isolated incident.

Looking around, I just saw that you can use : /minpwlen[:length] on the end of the "net accounts" line to do the same thing. When you set up the profiles on a separate disk, do you just use: net user [...] /profilepath[:path] is that sufficient or do you do more?

I wanted a way to have more control over the user setup process. The problem is when you set up a system for more than one user, you either have to include each user's password on the cd or make sure everyone changes their password. To get around this, you can use a utility called "cusrmgr". Unfortunately, it's only available with the win2kpro resource kit. Luckily, you can download the file here. The way I use this utility is to use the "net user" command to give each user a generic password, such as 'newpass' then I use the following command switch to force each user to change thier password at the next (first) login: cusrmgr -u [Username] +s MustChangePassword so my batch file looks like this for each user: net user test newpassword /add net localgroup "Power Users" test /add net accounts /maxpwage:unlimited cusrmgr -u test +s MustChangePassword that, coupled with setting the binary value 'MinPwdLen' to the desired value in [HKEY_CURRENT_USER\SOFTWARE\Microsoft\Windows\CurrentVersion\Policies\] provides a pretty good setup. what do you think?