jaclaz

Sorry, but I have to ask. What is the actual *need* to ever open more than one, two, maybe three, at the very most four DOS Boxes on a 9x system? I mean, if the hardcoded limit was - instead of 64 - 10 or 16, would anyone (outside of intentionally testing the limits of the OS) ever had noticed it in real life use of the OS? jaclaz

Well, a RAID 1 mirror is a mirror. The disks - until they failed - had exactly the same contents. So the Office install was on BOTH drives. If - after the "repair" you performed the booting drive has no Office install, it only means that *somehow* the repair procedure removed it or some parts of it. You can well (from another OS) copy the "working" Registry from first drive to second one, but the result will most probably be two separately working fine disk drives, BOTH without Office. One of the two (very likely) will have a modified Disk Signature (automatically recreated by the Windows 7 booting with "separate" drives due to the conflict it will have found, as such one of the two disks won't have proper drive letter assignments (and as such won't boot properly until you fix/delete the DosDevices letter assignments). As I see it your only way out (not easy at all, mind you) is to remove from the system (disconnect) the disk drive that is now booting and troubleshoot/fix (manually) the non-working one. If all that is needed is the Office key, usually it can be retrieved by using some script ot little thrtdpart utility, however this may depend on the version of Office you had installed. jaclaz

No, it's ok (meaning it isn't ). The disk partitions are intact as confirmed by the successful re-test on the windows 7 machine, the issue you are having is only the "generic" 764 Gb instead of 3 Tb seen that is connected with this (or that) XP driver and it is most probably not connected with the second partition. The check with Tiny Hexer would be only useful to understand if it actually can see/access the disk MBR (first sector of the physical drive) or if the 746 Gb (as I suspect) are just the result of some "queer" wrap around. To definitely check you can 0x00 the second partition table entry and try again, but I doubt that it will change anything, In any case the issue is solvable only if you can find and install fully working drivers (that may or may not exist for your specific board), or maybe - cannot say - it is some BIOS incompatibility, There is (or was) a DDO of sorts for Seagate drives, but using it would of course be another thing (and I believe it is hardcoded at the 2.2 Tb mark): http://knowledge.seagate.com/articles/en_US/FAQ/218619en? Easier would be - to see if you have any other motherboard (and XP install) that can see the full 3 Tb. jaclaz

Hmmm. Allow me to completely and utterly disagree. That is not how cases are closed, this is how cases are opened and later ignored, without investigating them. (possibly because of lazyness by the investigator ) jaclaz

That is exactly the thing that makes little sense (to me), it would be only an "added feature" to WinNTSetup without any particular advantage (saving one reboot maybe), I mean, cannot the autounattend.xml be created (and/or modified if needed) before and besides using WinNTSetup? jaclaz

@Atari800XL I am still perplexed/unsure. When does autounattend.xml run? After "apply" AND after "first reboot", right? And you want to try to change the name BEFORE that, i,e. after "apply" and before "first reboot", is this correct? So you can try, on a PE with the WMIC enabled if this can be done. And if yes, then try to replicate it manually without needing the WMI environment, by either mounting the offline registry and applying a .REG or by using the offline registry library. jaclaz

Well, it sounds just "normal" to me. At the time nlite was tested by n people and Nuhi did change nlite to be compatible with SP3, afterwards nlite devlopment was discontinued and it is to be expected that this or that glitch with unoffical SP4 happen. There is a Harkaz's Unofficial SP4 theead on MSFN also (just for the record): http://www.msfn.org/board/topic/171171-introducing-unofficial-windows-xp-sp4/ The Unofficial SP4 can be also integrated via nlite, maybe this makes a difference? jaclaz

Well, it is a "queer" question IMHO. You just asked to have this done by WinNTSetup, i.e. before the first reboot. If this is possible, it is possible, if it isn't the problem cannot be solved by WinNTSetup anyway. I mean, when does the autounatted,xml run? The doubt is (or should be) more about the potential need of changing some other keys (and/or setting a flag of some kind) besides changing the relevant keys values. If you prefer, does the WMIC command "WMIC ComputerSystem where Name="currentname" call Rename Name=newname" only chnage those four name instances or does it do *anything else*? In any case (for experimenting), there is normally no issue to add the WMI environment to a PE, I don't think that later PE's have changed since PE 2.x/3.x: http://www.msfn.org/board/topic/143622-how-to-install-wmic-in-winpe/ jaclaz

Computer name is (or should be) in 4 keys (on the running system): HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Tcpip\Parameters\Hostname HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Tcpip\Parameters\NV Hostname HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\ComputerName\ComputerName\ComputerName HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\ComputerName\ActiveComputerName\ComputerName Most probably you can change them while in the PE using the Offline Registry library (and tool). http://reboot.pro/topic/11212-offline-registry-library/ http://reboot.pro/topic/11312-offline-registry/ If you are doing it manually or with other tolls/methods, make sure to get the right "ControlSet". jaclaz

Also, check the actual SATA/IDE controller drivers on the XP, there are many reports of issues with drivers, often INTEL RST, causing that problem, it can happen also on Windows 7 machines, examples: https://www.servethehome.com/fix-746gb-3tb-hard-drive-issue/ https://www.chrisnewland.com/solved-3tb-hard-drive-only-shows-74652gb-in-windows-7-389 You can also try (since you are anyway in IDE mode) the "Standard Dual PCI IDE Controller": http://www.aomeitech.com/forum/discussion/682/what-to-do-when-new-3tb-hard-drive-only-showing-746-gb/p1 jaclaz

Let's see if I can sum it up. The issue here seems to be with the first partition, besides and before the second. When you were on 7 you had (initially): 1st partition: 1852.69 Gb Unused: 195.31 Unused (AND inaccessible in Disk Manager because past the 2.2 Tb) : 746.52 Later you had: 1st partition: 1852.69 Gb Unused: 195.31 2nd partition; 600 Gb Unused (and inaccessible in Disk Manager): 146.52 Now (on XP) you have only one partition (not formatted with a valid filesystem) with the same size of the extents beyond the 2.2 Tb. Queer, really queer. Can you check the actual contents of the MBR partition table? On XP you can use (say) TinyHexer with my MBR or PTN template: http://reboot.pro/topic/8734-tiny-hexer-scripts/ or any other suitable disk editor or MBR utility. Or simply reconnect the disk to the Windows 7 system and see what it finds now. It is well possible that Disk Manager mingled something (not unlike what it does on non cylinder aligned logical volumes inside extended): http://reboot.pro/topic/9897-vistawin7-versus-xp-partitioning-issue/ http://www.dcr.net/~w-clayton/Vista/DisappearingPartitions/DisappearingPartitions.htm but in that case some user action, besides merely opening the disk manager was needed, namely changing the active status of a primary partition. jaclaz

No, that wouldn't make any difference the extents are "reserved" to a partition/volume, no matter if it is filled up or not, that is the whole point about partitioning . Once the OS can access the volume, it doesn't matter, all the volume structures are relative to first sector, the NTFS filesystem has 64 bit fields, and - in any case - the "units addressed" in the filesystem are LCN (Logical Cluster Number) that is normally 8 kb in size, so it won't ever reach the 32 bit limit within a volume whose size in sectors, i.e. 512 bytes units, is below that. The potential issue is only if anything sums the LBA address and LBA size expressed in sectors. Booting would be probably tricky, not because of the NTFS or anything else connected to the volume, but only because of (maybe) some limitations in NTLDR (and/or NTDETECT.COM). There is/was a known issues with SETUPLDR.BIN on DVD's where if the (CDFS) LBA was too high it would not boot, I would be not surprised if something similar happened with NTLDR. But in any case there is no real *need* to boot from this "added" volume even in single disk systems, as the bootable volume can anyway be the "first" one (or however any volume entirely contained within the first 2.2 Tb. So, yes, attaching the disk as secondary on an XP system would be more than adequate for the test. jaclaz

Some files can be retrieved via Wayback Machine using their "office update" link: Sr1off97.exe http://officeupdate.microsoft.com/downloadDetails/sr1off97detail.htm Fm2paste.exe http://officeupdate.microsoft.com/downloadDetails/fm2paste.htm O97dtfix.exe http://officeupdate.microsoft.com/downloadDetails/o97dtfix.htm O97attch.exe http://officeupdate.microsoft.com/downloadDetails/O97attch.htm Sr2bof97.exe http://officeupdate.microsoft.com/downloadDetails/sr2off97detail.htm Jetcopkg.exe http://officeupdate.microsoft.com/downloadDetails/excel97odbc.htm And more generally: https://web.archive.org/web/*/http://officeupdate.microsoft.com:80/downloadDetails/* jaclaz

Try via Wayback Machine: https://archive.org/ Ideally, you could test each and every file, making a new post with the links working and another one with those not working. (so that someone can assist you with finding the missing ones). A lot of files won't be retrievable because of the "interstitial" pages but you can find often the filenames and then look for them, An issue may be with "localized" versions, most available files are only en-us or another specific language, it is rare to find "all" of them on the same mirror site, example: ftp://ftp.uni-rostock.de/pub/tools/microsoft/ServicePacks/Office8/us/ jaclaz

Good. The volume, even if it is partially "beyond the border" is to all effects a "normal" volume, the NTFS filesystem is exactly the same on MBR or GPT partitions/volumes, its size is well within NTFS capabilities (and actually it can be - at the very most and on 4 Tb disk - anyway smaller than the first volume which can be at the very most almost 2.2 Tb). The possible issues are only connected to testing if any of the various tools that are "normal use" may have an hiccup with the volume location, technically the issue is with *anything* that may treat the "volume" as a "partition" AND have "artificial" checks/limitation (be them either intentional or not), but not with "normal" operations at the volume level. I have very little doubts that Windows 7 may have any issues, since anyway it has GPT (and thus volume placed at large addresses) support. I would be more suspicious of NT/2K/XP. jaclaz

And AGAIN, there is NO PATCH whatever in Usher's method. IT DOES NOT INVOLVE ANY PATCH, only some minor modifications to SYSTEM.CB (only useful in Safe Mode). jaclaz

It opens just fine in XP Wordpad, just for the record. And - for example - this online converter seemingly wotks just fine with that RTF file: https://convertio.co/document-converter/ Attached result from above. jaclaz Hardware.pdf

Yes, it was only to show a possible issue in "inserting a batch tested on cmd,exe", if you (IMHO wrongly) implement the "autodetect policies" there would be no certainties about which command processor is used, and thus a same batch file may work fine in a machine and not work on another one because two different command processors are run depending on the autodetection. As said, no matter which small quirks this or that command processor .exe, they don't really matter, as long as the batch has been tested on that specific command processor. In this the "embedded" command processor is a very good idea. Putting it in a subdirectory of the 7-zip is (still IMHO) not only unneeded, but also an unneeded complication, however. What would be useful would be to be able to change the command processor, so that we are not limited to ROSCMD.EXE, as an example one might want to use 4NT/TCCLE: https://jpsoft.com/tccle-cmd-replacement.html http://www.portablefreeware.com/forums/viewtopic.php?t=3392 or Greg's Shell: https://www.gammadyne.com/cmdline.htm#gs or whatever other alternative command line processor, including (of course not for redistribution) a copy of a Windows CMD.EXE. Again, a simple .INI file with a number of simple choices possible would add greater flexibility without making anyway the thing too complex. Here is another quirk, this one serious, it seems like ROSCMD does not exits loops : @ECHO OFF SETLOCAL ENABLEDELAYEDEXPANSION SETLOCAL ENABLEEXTENSIONS FOR /L %%A IN (0,1,5) DO ( SET Number=%%A IF %%A==2 GOTO :out_of_loop IF %%A==2 ECHO %%A IS equal to 2 ECHO %%A IS NOT equal to 2 ) GOTO :EOF ECHO. ECHO You won't see this :out_of_loop ECHO Here is out of loop, number is %number% C:\>looptest.cmd 0 IS NOT equal to 2 1 IS NOT equal to 2 Here is out of loop, number is 2 C:\>roscmd ReactOS Operating System [Version 0.4.5 20170801-rUNKNOWN] (C) Copyright 1998-2017 ReactOS Team. C:\>looptest.cmd 0 IS NOT equal to 2 1 IS NOT equal to 2 2 IS equal to 2 2 IS NOT equal to 2 3 IS NOT equal to 2 4 IS NOT equal to 2 5 IS NOT equal to 2 Here is out of loop, number is 5 C:\>exit jaclaz

Sure , as often happens, cannons, flies .... jaclaz

No hurry whatever . Another test (when/if you have time) is to artificially fragment a few files and then try on them wincontig and contig (besides running defrag on the whole volume). To fragment, you can use myfragmenter, get it from mydefrag: https://web.archive.org/web/20150811002116/http://www.mydefrag.com/Downloads/Download.php?File=MyDefrag-v4.3.1.exe use UNIextract. jaclaz

The point is just that to write a batch file you (actually I) normally test the command in the exact same environment where the batch will be run, not really an issue (for me) but a potential user will (in this order): 1) add his/her batch file AND remove the roscmd.exe from the archive (FAIL) 2) add his/her batch file AND NOT remove the roscmd.exe BUT have no PAUSE in the batch (FAIL) 3) add his/her batch file AND NOT remove the roscmd.exe AND have PAUSE in the batch BUT the batch only works in CMD.EXE (FAIL) Personally I hate programs that "decide" what is better for me, where a simple file with a few settings would allow me to decide myself what is better on a case by case basis. Actually when/if the little ROSCMD.EXE incompatibilities will be "leveled out" it is actually "better" (IMHO) to use that command processor (and also "guaranteed" executables inside the 7-zip sfx), so that the batch (tested on that command processor and with that exact version of the other .exe's) surely works, avoiding possible incompatibilities between different CMD.EXE's on different OS's, including a few related to localization. A teeny-tiny issue with Roscmd.exe is the different behaviour of PUSHD. On Windows CMD.EXE once you have PUSHD a directory, running PUSHD without arguments will show the pushed directory, on ROSCMD, it doesn't, this may "break" some particular batches, this is a workaround for tilde assigning the variable %sfxextpath% so that it can be typed and the variable tilde: myhello.cmd: @echo off title %SFXPATH%%SFX% SET SFXEXTPATH=%~dp0 SET SFX SET TILDE=~ SET TILDE %~dp0\roscmd.exe then (output and commands): SFX=mytest.exe SFXEXTPATH=C:\DOCUME~1\jaclaz\IMPOST~1\temp\7z14024E48\ SFXPATH=C:\batches\Portabat\ ReactOS Operating System [Version 0.4.5 20170801-rUNKNOWN] (C) Copyright 1998-2017 ReactOS Team. C:\batches\Portabat>pushd %sfxextpath% C:\Documents and Settings\jaclaz\Impostazioni locali\temp\7z14024E48>pushd C:\Documents and Settings\jaclaz\Impostazioni locali\temp\7z14024E48>popd C:\Documents and Settings\jaclaz\Impostazioni locali\temp\7z14024E48>popd C:\batches\Portabat>cmd.exe Microsoft Windows XP [Versione 5.1.2600] (C) Copyright 1985-2001 Microsoft Corp. C:\batches\Portabat>set sfx SFX=mytest.exe SFXEXTPATH=C:\DOCUME~1\jaclaz\IMPOST~1\temp\7z14024E48\ SFXPATH=C:\batches\Portabat\ C:\batches\Portabat>pushd %sfxextpath% C:\DOCUME~1\jaclaz\IMPOST~1\temp\7z14024E48>pushd C:\batches\Portabat C:\DOCUME~1\jaclaz\IMPOST~1\temp\7z14024E48>popd C:\batches\Portabat> Another side-side note (very minor) is that traditionally .bat should be compatible with DOS/9x, while .cmd is the extension for batches to be executed by CMD.EXE (or ROSCMD.EXE). jaclaz

I will try again. What is the output on your setup of running this command: devcon listclass ports is it different from the one you obtain with devcon.exe /Find =Ports ? (actually "devcon find =Ports" should work just the same, no need for the slash) And what is the output of this one?: devcon listclass USB |FIND "(COM" The snippet Wireless posted is NOT to be issued on command line, but rather copied to a batch file that later is executed, *like*: @ECHO OFF SETLOCAL ENABLEEXTENSIONS (set s=) for /f "tokens=2 delims=()" %%A in ('findstr /i /r "^USB..*(COM[0-9][0-9]*)$" input.txt') do (call set "s=%%s%%,%%A" & call set "s=%%s:COM=%%") REM if defined s (echo "%s:~1%") > output.txt if defined s (echo "%s:~1%") If you want (why?) run it on command line you need to "unescape" the %%A variable, but most probably some more tricks will be needed for the expansion of the variables. Generally speaking it is not a good idea to save the output of console to a text file that then needs to be re-parsed, it makes a lot more sense to run the command in a batch and directly parse its output, without writing temporary files. jaclaz

That is the output of devcon listclass ports , right? Maybe you could do instead: devcon listclass USB |FIND "(COM" This way it is already a list of USB COM devices only. jaclaz

Maybe a link to the homepage of XHDD would be useful: http://optimizr.dyndns.org/dos/drivers.html jaclaz

echo ^G (that is CTRL+G). jaclaz