Jump to content
Strawberry Orange Banana Lime Leaf Slate Sky Blueberry Grape Watermelon Chocolate Marble
Strawberry Orange Banana Lime Leaf Slate Sky Blueberry Grape Watermelon Chocolate Marble

MSFN is made available via donations, subscriptions and advertising revenue. The use of ad-blocking software hurts the site. Please disable ad-blocking software or set an exception for MSFN. Alternatively, register and become a site sponsor/subscriber and ads will be disabled automatically. 


GreenMachine

Developer
  • Content Count

    3,070
  • Donations

    $0.00 
  • Joined

  • Last visited

Everything posted by GreenMachine

  1. RHSM V2, Cool! Glad to see you left the fixes in the I386\SVCPACK directory, I feel safer that way. I should have waited, 'cause I guess now I'll have to go modify my own version of the code I stole from you ... As you said, gotta be careful with the unsupported stuff. Nice to make it easy, but I don't have time on the clock to do MS's debugging. So perhaps those that know can post a list of good and bad?Thanks for sharing.
  2. I have seen that others here install and then kill the process once the actual install is over. I prefer the simple method: I just copy all the winamp files and shortcuts to the target machine, then run a .reg file for creating the file associations. All winamp settings are in it's .ini file. (I wish they were all like this ...)
  3. Windows XP was supposed to have QCHAIN functionality built into the hotfixes, but it turns out that some of the earlier patches do, in fact, still need it. The exact date is somewhere on MS's site. For info: Hotfixes create a list of system files to replace on the next boot. This list is created in the order the hotfixes are run. If run out of order, an older file version may overwrite a newer one. QCHAIN sorts the list of files to be replaced, putting the files in a oldest to newest order to prevent this problem. I used this method before integrating the hotfixes into the install, and even if run when not needed, it will not harm the instalation / fixes.
  4. I believe the settings are not saved until you logoff. Compare a registy export from before the changes, and after the changes + logoff / logon.
  5. NTUSER.DAT contains the HOT_KEY_CURRENT_USER keys, not the HOT_KEY_LOCAL_MACHINE key.
  6. Oh, Man, I should never have let you look at my code. In less than 24 hours, you busted my program execution order in svcpack.inf, and now my files are no longer correctly digitally signed! Of the four type 2 fixes I do, I think two are IE and OE updates. I renamed them to the corrosponding KB article because ... because I'm that way. I do almost all the critical and recommended upates. When I go to Windows Update, there are only 4 (was three until a couple of days ago) recommended updates: (I just checked: now there is a new 5th recommended). These are: Journal Viewer, Windows Messenger, Advanced Networking Pack (Q817778), and now Windows Right Management and KB822603. I wonder about the sp1.cat file. It would help to know how the installer works: if t finds a newer version in I386 (newer than that inside the cab file), could it use it instead? Even if it does not like the versioning, there is only one single version in the possible multiple copies. so what's it going to do about it? The newer versions do have a signature somewhere, in some .cat file. Could that be found? On a side note: We build the svcpack.inf file. What about adding everything else there as well: Java, WMP/MM, IE6.1, MDAC, RegEdit /S? Also, perhaps when building the svcpack.inf, list the directory by date in the loop to create the file names, after having merged the two types of fixes.
  7. @RyanVM: For Q823980, the original Blaster Patch, this was the case. (Thread Here) It's only one example, but one is enough for me. I, too, remove all remaining uninstall keys. Note that the above example has been superceeded with another, so it is no longer "active", but I am no longer silly enough to believe I won't see this again. Burn me once, shame on you. Burn me twice, shame on me.
  8. Frustrating, isn't it? I would say from the looks of things, you have us all stumped. The syntax looks fine to me. I can only offer you some hints for further debugging: Seperate the commands, perhaps with a pause in between, to determine which of the two is guilty. I have it working on my system with this code: NET USER SysAdmin SysAdmin /ADD /COMMENT:"System Administrator Account" /PASSWORDREQ:YES NET USER User User /ADD /COMMENT:"Local User Account" /PASSWORDREQ:YES NET LOCALGROUP Administrators SysAdmin /ADD NET USER Administrator /ACTIVE:NOReplace yours with this, and see if that works (Perhaps Sky and Windows do not jive - anything is possible at this point). (Pay attention, my code disactivates the administrator account, and creates the users SysAdmin and User, both with passwords. But DON'T modify my code, as it works, and in doing so you are adding one more unknown to the debugging. After you get it working as I have it, makes changes one at a time, until you find the error.) Open a command prompt in the setup, just before the NET commands: START CMDTry your commands, either by stepping through the batch file from one pause to the next, and see if anything is happening in the Users List - You can get to the Local Users and Groups Snap-In as such: From the command prompt, type "mmc" (without the quote - same applies for following examples) This will open a Microsoft Management Console. Select File - Add/Remove Snap In. Click "Add" in the new window. Select "Local users and Groups", then click "Add". Select "Local Computer" and click "Finish". Click"Close". Click "OK". You should now have the familiar Local users and Group Snap-In running in the console.While you have the command prompt, try anything else you can think of... Now you've got me curious. Give some or all of that a try, and let us know what you discover.
  9. It's the "pretty sure" and "should" that worry me ... If a problem should arise, you are at a handicap from the start in determining the cause, and I am not a fan of debugging Windows. Windows Update has also been known to rely on the HKLM\SOFTWARE\Microsoft\Windows\CurrentVersion\Uninstall key do determine if a certain hotfix is present. I am not trying to belittle anyone's ideas, but as neither Windows nor their HotFixes are known to be the most stable, and I have no desire to increase that instability.
  10. I believe I mentioned in one or two threads that the variable PROFILESDIR was available for use, and I would like to tell those whom I may have mis-lea, that I myself was mis-led. You are able to set the location of the users profiles with this entry in WINNT.SIF, but the variable itself is not availabe. I interupted setup at the cdlines.txt point, and discovered this list of available variables: ALLUSERSPROFILE=C:\Documents and Settings\All Users CommonProgramFiles=C:\Program Files\Common Files\ COMPUTERNAME=XXXXXX ComSpec=C:\WINDOWS\system32\cmd.exe Path=C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\system32\Wbem PATHEXT=.COM;.EXE;.BAT;.CMD;.VBS;.WS ProgramFiles=C:\Program Files PROMPT=$P$G SystemDrive=C: SystemRoot=C:\WINDOWS Upgrade=False USERPROFILE=C:\Documents and Settings\Default User windir=C:\WINDOWS __PROCESS_HISTORY=C:\WINDOWS\system32\setup.exe Barring any typos, this list should be correct. The actual values of the variable, of course, may vary.
  11. I noticed the risk because the file advpack.dll was the wrong version. It was listed in the setuperr.log (or .txt) file, and that started me looking. Even then, it did not seem to harm anything. Still, better safe than unsure. I had not really given a thought to the fact that they were run the second time out of order, in svcpack.inf. Oooops. I think I will be renaming them so that alphabetical and chronological order are the same. I do not care about automatically renaming them, because I like to scrutinize the hotfix before deciding to use it, and the renaming is not a big issue - I've already got my hands in there. The Q12345~1.exe method works fine I think, but I never quite swallowed how poorly windows went from 8.3 to long files, so I (almost) exclusivly use 8.3 from the get-go. If I'd've known about %~dp0 I would have used it, but I am probably better at assembler than DOS batch commands. You said you had a good look at my code. Did you notice, like I did, that is becomes much clearer between the first and second pint? Once you get the kinks out of your new script, please do share (again)
  12. The uninstall switch when installing the hotfix avoids the need for this (in ALMOST all hotfixes). (No offense, Doug...)
  13. RoyalBox! My base version of the RHSM (Royal Hotfix Scripting Method) is your original post, plus changes added by others to include HotFix Type 2. I ran into issues with all KB hotfixes running first, and then all Q hotfixes, and then the same thing with type 2 hotfixes. They were not in chronological order, and I did see some incorrect overwrites. I thought about nameing them all Q######.exe, but I remember reading on MS's site that the hotfix number itself could not insure correct component versioning. XCOPY was the best I could come up with, though it would be ideal if it could do a version check instead of a date check. I extract them each to a seperate directory (Q######) so that I can look at them afterwards, should the need arise. The PREPDIR thing, as I am sure you figured out, is just to alow runing the script from anywhere, as the type 2 hotfixes, unlike the type 1, require a complete extraction path. This enables me to get them into a subdirectory of my working directory. As regards the type 2 hotfixes working, it seems they do, but in any event ALL these hotfixes are run during setup, so the changes get there one way or another, and the updated files will be present on the CD, so I think it's OK. I have the most faith in the hotfix itself, and that is why I will never install without them, as suggested in another thread. I only do 4 this way: KB814078, Q327405, Q330994, Q822925, and have not had any problems (that I noticed ...) Bottom line is, I don't think it hurts. The other system updates: .Net, MDAC, JavaVM and combined WMP/MM, I consider special items, and do not attempt to slipstream them. I run those 4 items, plus the WMP update, from the batch file called in CMDLINES.TXT. On a side note, I was bothered that the Q817287 hotfix did not seem to accept the uninstall switch, and it would leave behind a C:\Windows\$NTUninstall.....$ folder. It was the only one I had, so the imperfection bugged me... I extracted the hotfix, and it seemed to me that the type 2 hotfix setup was just a wrapper for a type 1 hotfix. The setup basicaly stopped a service, called the type 1 fix, and restarted the service. I just use the extracted type 1 fix, and it works fine as such, as the service in question was never started anyway. No entry in the uninstall list, no hidden uninstall directory.
  14. Does this method prevent the error message in the setuperr.txt (or whatever it is called)? I did as many others, and just copied a version posted here into the I386 directory. Works fine, yet is listed as a problem in the error log, and causes a red flaged event in the event log. Does hacking the checksum overcome this problem as well, or does it just prevent setup from crashing?
  15. I will certainly take the exrta 10 minutes to start over: copy the files from CD, slipstream SP2, ... and wait for the next wave of hotfixes.
  16. I agree 100% with the snoozer, but then if you cannot do it with a batch file and regedit, it's not for me. If I were to use another method, I would simply log on to a new installation, make all my desired changes, save the user's registry hive file (NTUSER.DAT), and copy it to the Default User's profile during setup.
  17. Do you think that we should be using the "start" command in the generated .cmd files in that case? i.e. start /wait "HFTYPE1\Q282010.exe" /q /x:EXTRACT\Q282010instead of HFTYPE1\Q282010.exe /q /x:EXTRACT\Q282010
  18. For reasons stated HERE, my vote is for #2, the Standard and Documented Microsoft way, regardless of whether this is done manually, or in a batch/script routine.
  19. It is true that removing the actual hotfixes saves about 40 MBs, but that is off no matter to me: my windows installations are windows only installations. I need alot more that 40 MBs to install all the additional programs I will install, but I like to keep things modular - thus simple and easy to debug. Without really knowing what each of the hotfix setups does, I would not remove them. Do they do the exact same thing on all computers? If there is a single "IF" satement in the code... I will never try to understand that code so in depth: I will never remove them. At best, it will introduce instability questions. At worst, I smell a maintenance and weekly (wednesday) updating nightmare. My interest in slipstreaming the hotfixes is added stability. This is, to me, a step in the other direction. Windows update must also work correctly, or, in the words of another member, what's the point? The extra 15 or so minutes does not concern me, either: I am not there to see this, nor the fancy .inf / RunOnceEX installations. Now that the Project's Father is back on line, perhaps we should baptise this the Royal Hotfix Sipstreaming Method? God Bless the Queen.
  20. The SET command takes care of the variables - adds, deletes, modifies and lists 'em. You can do it from the command prompt or in a batch file, there is not really much to it. Type set /? for usage ... SET VARNAME=VARVALUE ; You have the variable name VARNAME assigned the value VARVALUE SET VARNAME= ; Variable VARNAME is gone SET X=%VARNAME% ; X now equals VARVALUE
  21. For what it's worth, I PMd you my copy of your work. I'm done, so the balls back in your court.
  22. Oh, man, you took all the mystic out of the zz###s. I was just starting to enjoy it ...
  23. Hey, waita minute, does't anyone read my posts? in response to your post in THIS THREAD. You certainly have a right to your pride, and I understand that it would have been better in that thread. I would have liked to have seen it from the begining as well. If it is of any consolation, I think what happened is that during the MSFN down time of a couple of days earlier this month, people lost their train of thought. So that it does not go unsaid: Jolly good show, RoyalBox, and thank you for sharing your work with us! And thank you for being the cause of these lively discussions! (What good is an ego, if you can't show it once in a while?)
  24. I think you are correct: the Default User\NTUSER.DAT hive is already in use, so there is not much sense in loading / unloading it. Snooz seems to concure, as he said he has tried them both, and they both work. I have not tried yet, so I cannot speak from experience. Default User\NTUSER.DAT contains the New User Template, and some things in there will only be used to initialize a new profile. You make the changes there, and when you next create a new user, the changes will be visible. My experience leads me to believe that many explorer settings are saved to the registry just before shutdown / logoff, such as default folder view options, taskbar settings, and who knows what else, thus will always overwrite your changes. For this reason you must load the default hive, make the changes there, and use it as a template. You can get away with making changes during the pre-gui-reboot because the shell (explorer) has not yet been loaded, thus will not save it's settings on exit.
  25. I'd like a look. 7z, eh. Sounds like a pretty efficient zipper. Do they have an SFX option?
×
×
  • Create New...