jaclaz

Isn't this covered by the first 4 FAQ's here? http://www.techspot.com/community/topics/windows-2000-oldcigarettes-windows-2000-xp-api-wrapper-pack.167843/ jaclaz

You mean the Registry as a filesystem driver? http://reboot.pro/topic/7681-the-registry-as-a-filesystem/ jaclaz

@Dencorso NO. The issue was expected and it should be connected to the writability of disk sectors without LOCKing the disk (or the drive/volume/partition) Trustedinstaller should have no connection with this, and however the idea is again to make things as simple as possible, and using TrustedInstaller would be an overcomplex addition, we'd better shift back to "wasting" a few Mb and make a (slightly) more standard partitioning scheme @Dave-H Your last report makes no sense to me *Something* must have changed between your post #180 and #182. in post #180 you posted: The bolded part comes from the SUB :doswitchDA, which means that the CALL to it, which is near the end of this snippet: The post #182 is instead the message that comes from the check in red above. This can only mean that somehow the previous diskpart command failed, which makes no sense as it worked before. However, try running manually. Make sure you are in the right condition, i.e. PC1814096. Make sure the disk is \\.\Physicaldrive5. Run in diskpart: select disk 5 list partition you should get the same output as in post #176 The Partition 2 has offset 32 Mb, (as expected because we know that you selected 32 Mb size in the batches when you created the disk) so it's bootsector is at absolute offset 32*1024*1024=33554432 and you want to run: lockdismount -lock 5 dsfi \\.\PhysicalDrive5 33554432 4096 as4kbNTFS.bss and then a diskpart RESCAN. jaclaz

You mean these?: ZIP installer (32/64 bit) | 16 MB Viewer Manual (PDF) | 7 MB <- seemingly the manual, NOT the program EXE installer (32/64 bit) | 16 MB Portable version (ZIP) | 8 MB Portable Version (OCR Lang Files) | 8 MB<- seemingly the OCR Lang files, NOT the program MSI Network Installer (32 bit) | 27 MB MSI Network Installer (64 bit) | 32 MB Language Localization Tool | 13 M<- seemingly the Langiage Localization tool, NOT the programI believe it depends on your exact location and on which time of the day you plan to install the program, Zip installer will work almost always, but reportedly it may have issues every first friday of the month or when there is a full moon (or if you are below the tropic of Capricorn ), the portable version has more or less the same limitations but allows to install on portable computers such as notebooks or netbooks , the MSI network is the more reliable one - as long as you choose the right bitness, either 32 or 64 bit - because being connected to the MSI network will use Greenwich time as reference, though you will have to manually convert to local time. Come on , WHICH one among them is the default selected one? Get that one (the Zip installer), unless you know better. jaclaz

Yeah, sure, guess WHERE WinSetupfromUSB was developed? jaclaz

Maybe "bear" would do http://thesz.diecru.eu/content/bear.php jaclaz

Yep , good (which again means bad ) That is where the difference between what is expected: and what you actually get: Becomes relevant. There are no issues of course to avoid the repetitive loop (that is actually my bad , forgot to put a check about it). The issue is probably about the mechanism that the stupid Vista or later implements to protect parts of the hard disks (but not all of them) unless they are part of something it "recognizes". It is a CATCH22, since it doesn't recognize offset 33554432 as being part of the volume, it doesn't allow to write the (correct) bootsector to it, BUT when it finds already the "correct" bootsector, since it recognizes that offset as being part of a volume, THEN allows potentially to write to it a "wrong" one, it sounds crazy but that is seemingly what happens: http://www.msfn.org/board/topic/173265-formatting-an-external-drive-using-different-interfaces/?p=1093173 To be fair, the stupid Windows 8.1 is actually "right" (hard as it is to say this ) as in the specific prilog scheme the beginning of the NTFS volume is in the address actually assigned to the FAT12 partition (not by the actual volume, only by the MBR partition table entry), i.e. the two partitions (but not the volumes) are actually overlapping. There are four ways out as I see it, all to be tested, but basically: 1) change the setup from "prilog" to "twolog" 2) keep the FAT12 volume and partition entries "as they are now" BUT move the NTFS partition and volume further so that they don't overlap 3) keep the FAT 12 Volume and NTFS partition and volume "as they are" but find a way to also switch the MBR making the FAT12 partition entry "tight" on the space occupied by the volume. 4) find a way, like a MBR write or putting the disk offline to be able to get access to the bootsector of the NTFS volume bootsector. I don't like very much #1 as it would make (even on the 512 bytes interface) the disk impossible to be made bootable (unless a MBR+hidden sectors bootmanager like grub4dos is used, though this might be needed anyway, to pre-switch to right geometry to access the NTFS volume) and as well I am not really sure that #3 switching the MBR also, is a good idea (doing it from a batch running from a volume residing on that same disk) or #4 (that implies to either do another kingd of MBR trick or on 7 and later put the disk offline). #2 still seems to me the best one (actually more lke the "simpler" one), though it has an attached string, you will lose overall capacity of the disk (roughly 7 times the size of the FAT12 partition). In practice, this has no real effect/weight until the size of the FAT12 volume is kept small enough, like: FAT12:1 Mb in size=7 Mb unusable area of disk FAT12:2 Mb in size=14 Mb unusable area of disk ... FAT12:32 Mb in size=224 Mb unusable area of disk and it is still to be seen if the non-standard 1 sector Extended partition is an issue, but since it works on the 512 interface, I don't think that it is part of the issue. If you are OK with this latter approach, and willing to try it with either a smaller partition (the 1 Mb one is more than enough to hold just the "switcher related files") or you are fine with having the 1 Tb reduces by at most some 224 Mb, I would recheck/correct the previously posted batches so that you can recreate the "prilog" disk from scratch. To be exact there is even a further approach, making use of a "temp" directory on another disk, let's say your "normal" %TEMP%, by copying to it the Switcher.cmd and thus being surely able (on 7 and later) to quickly put the disk offline in order to write the bootsector. Though 3/4 to 4/5 of all programs you run normally do this (beside senselessly writing to your Registry and what not) I somehow feel like this approach to be "inappropriate". A further approach would be to add (yet another) little excutable to the FAT12 partition. The _K over at reboot.pro put together a little tool that should be doing exactly what needed, Lock the disk and Unlock it after having performed another task (dsfi in our case), it has to be anyway tested on 8.x, as it was tested on Vista and 7 at the time, though nothing should have changed. Get the tool here: http://reboot.pro/topic/12413-lockdismount-v0300-update/ http://www.mediafire.com/download/5q6795h7pp32623/LockDismount0300.zip and add it in the root of the FAT12 volume. Edit the switcherDA.cmd changing this: to: jaclaz

But still there is no real need for an embedded system, among the new features of "normal" Windows 8.x is it's "Kiosk mode" or "assigned access": http://blogs.technet.com/b/canitpro/archive/2013/12/17/step-by-step-enabling-kiosk-mode-in-windows-8-1-via-assigned-access.aspx i.e. something that has been done for years before by everyone else by adding or removing a couple things . Windows 7 (i.e. the "normal", licensed OS corresponding to the OP's PE3) example: http://sysadministrivia.blogspot.it/2012/05/creating-kiosk-machine-with-windows-7.html jaclaz

Well, to be fair, in the good ol' times you payed for the use of a computer mainframe on a hourly base (and/or when computers needed to be booked in advance of weeks, say in universities) you had: more time to think how to write the program, and at least wrote it in a way that was readable you made d@mn sure that the whatever you wrote did work/had not stupid bugs and only a few years before you made ALSO d@mn sure to punch the stupid cards correctlyAnd yes, before someone comments, this is intentionally a WHY, in MY day ... https://tinyapps.org/blog/misc/200702250700_why_in_my_day.html type of post. jaclaz

Sure Here is a good start: http://bit.ly/1w0PUSU jaclaz

A .inf file in the c:\windows\inf\ folder is compressed to it's correspondent .pnf file when installed, but usually there is not very much to gain from removing the "original" .inf files, and you have the nuisance (if you disable/uninstall a device for - say - troubleshooting an issue) that you will miss the driver and have to supply it from CD/DVD or however install media". The "modem" files are of course OK to delete (but they amount to bunch of used bytes). Most probably the "right" approach is once the system has been in use for some time, so that all devices have been used/installed to delete ONLY the .inf files that DO NOT HAVE a corresponding .pnf, as those will be, just like the mdm*.inf "unused" INF files. All in all a SAFER way would be to compress in a - say - .7z file ALL the .inf files (no matter if they HAVE OR DO NOT HAVE a corresponding .pnf), since .inf files are very compressible being basically "plain text" files this won't be much different from deleting them, and this way you can always extract the .inf from the .7z archive when needed. jaclaz

Hmmm. jaclaz

Good. Try the attached, Switcher009, renamed to SwitcherDA (stands for Direct Access). If it works as intended, on PCn814096 it should have an output similar to: "DUAL" disk found as \\.\PHYSICALDRIVE6Disk is connected as 4096 bytes/sectorThe NTFS volume has not been found.This may seemingly happen on 8.x, and we can then try with direct access.The current NTFS bootsector is NOT the 4kb bytes oneswitching it ...dsfi \\.\PHYSICALDRIVE6 33554432 4096 as4kbNTFS.bssOK, written 4096 bytes at offset 33554432attempting to run a rescanDal computer CINQUEAttendere. Rilevamento della configurazione corrente in corso.............DiskPart ha completato l'analisi della configurazione.This is the most we can do, if after next iterationthe NTFS volume is still not found but the bootsectoris detected as the "right" one, you will need todisconnect the disk through "Safely Remove Hardware" andreconnect it ...Premere un tasto per continuare . . .whether the Diskpart rescan will be enough to "create" the volume or if a physical disconnection and re-connection will be needed is to be seen ... jaclaz Switcher009.zip

Which is actually a good thing , as you won't be able to measure TTL levels with a multimeter (you would need an oscilloscope or similar). The 7200.11 should communicate normally. If you have already tried exchanging the Tx and Rx wires, then likely the adapter is a dod or however is not properly working, as from the link you provided it does seem like the "right" one . jaclaz

... which is what I call - maybe improperly - freedom. jaclaz

Well, some oddities are as said expected , particularly in Disk Manager. The BSOD is queer, however it may well be a "glitch in the matrix". I need the "full" set of reports for the PC181512, just like the ones you posted on #172 for the PC1814096. In those the "anomaly" was that: I expect that on the PC181512, like on all the XP based tests, a volume is actually associated to the partition (which is what allows the SwitcherNG to work to switch from 4kb to 512). However the good news about: I:\>switcherngVista , or later, ...OK, I am a local admin ..."DUAL" disk found as \\.\PHYSICALDRIVE5Disk is connected as 4096 bytes/sectorTHe NTFS volume has not been found.are that the batch works correctly, i.e. it cannot found a volume because the volume is not actually there, but the detection of both the Physicaldrive and of the sector size is correct. :thumbsup: I can now add to the batch (in the case the volume has not been found) a "more direct access" routine. jaclaz

Yep .Though I am sorry to create thus the need for frequent reboots . At next occasion, however do post please the PC181512bs4kb, I need it to understand (hopefully) if the 8.1 on 512 interface behaves like XP or in a "third way" different from both XP and 8.1 4096. jaclaz

And now, for NO apparent reason: http://www.msfn.org/board/topic/140891-paging-file-set-at-installation-and-contiguous/ http://www.msfn.org/board/topic/154667-the-most-stupidawkward-things-you-sawheard-in-the-it-environment/ (and no, let's not start the usual discussion on pagefile size and *need*) jaclaz

Good . (which means bad ) The issue is definitely here: On PC1XP4096 you should have (please check, but I have no doubts): Can you please try the same (and post results) on PC181512bs4096? Would it be possible that the good MS guys have removed altogether the concept of "Logical Volume" BUT only for 4096 byte/sector devices? BTW this will probably not prevent us from "switching", I think I can deploy the bootsector correctly by using the \\.\Physicaldrive (with the right offset), what has to be seen if the "blank" MBR trick (or the "offline disk" one) is needed or not, i.e. if the bootsector is write protected or not. jaclaz

What exactly provides voltage in a PC (AND BTW contains capacitors as well)? Let me think what I would try next .... jaclaz

Add - just like it is rumoured about car names - some research should be done on what happens in other languages... ... and no, I will not post what Cortana rhymes with in Italian , but let's say that it has some common points with a Mazda model name: http://en.wikipedia.org/wiki/Mazda_Laputa jaclaz

I am curious to see how the dual mode disk is seen in 8.1. It seems like the volume is not detected at all. , it is possible that it belongs to one of the "safety measures" (or whatever) in Vista or later. Forget about Disk Manager. When you are in that situation (possibly Windows 8.1, connected as 4kb, BUT actual bootsector 512 i.e. PC1814096bs512) run in diskpart: and post output. (of course disk 5 is the output you posted, and partition 2 is the actual partition number as it should be, if you don't have the partition 2 in the output of list partition, go directly to the list volume command). Please also run the "usual": MOUNTVOL>mountvol.log REG QUERY HKLM\SYSTEM\MountedDevices>regexe.log and attach them. Now, I may seem grumpy - mainly because I am grumpy, but can you ALWAYS from now on, use the "code" that we have seemingly agreed to : as reports like this: are (to me) confusing and I spend some time to try understanding them and I am not even sure if I get them fully right jaclaz

Then there has been a misunderstanding. What I would like to have is freedom (for everyone) to choose either the desktop or the NCI (let's call it "modern" since it cannot be called "metro" anymore) AND to be able to customize the one or the other according to personal habits, likes and likes not, whatever. You prefer the "search" paradigm? Good, then you are free to use it on BOTH your office PC and on the tablet you carry with you. I prefer the Start button cascading links? Good, then I am free to use it on BOTH my office PC and on the tablet I carry with me. Someone else wants to use the one on the desktop PC and the other on the tablet (or - maybe better - viceversa)? Good as well. Someone (like myself) cannot really distinguish the scrolling bars? Then he/she is free to change their colour or give them a more visible/contrasting look, different form what someone else in a not-so-far-away galaxy defined as JFYI : https://social.msdn.microsoft.com:443/Forums/en-US/ed1de1dc-1389-4980-acf2-aefc95947ac1/changing-to-a-theme-with-color-in-office-2013? jaclaz

Well, when the good MS guys are involved everything (and the contrary of it) is possible , but I still believe that the issue is *somewhere* in the batch(es) , I mean, the second NTFS volume is as "logical" when mounted as 512 as it is when mounted as 4kb, as a matter of fact the only things that change (besides the bootsector) is the address of the EPBR. However, try the attached, SwitcherNG.cmd (the NG stands for Next Generation , should anyone be led to think it could stand for No Good ). Hopefully it is a bit more reliable in performing the switching and the re-detection/re-scan. It's not yet as good as I would have expected it to be, it seems that I can still manage (and surely you will be able to replicate it) to find "sequences" of connections/disconnections that "hang" or that however seem somehow "sensitive" to opened windows/tools and/or being sometimes much slower than others, and I still have to understand why in some cases drive letters remain "sticky". A good part of the issues are surely given by the "false" drive letter assignment caused by the network drive assigned to an "early" drive letter (actually the otherwise "first free" one), but I believe that this new approach should work if not "better", "less worse" than the early one. Try this new one, and report. jaclaz Switcher008.zip

That is a "green" LP drive? Those usually need the read channel to be shorted, but you should have anyway a LED: kind of error. Are you sure-sure that your adapter is TTL 3.3V? (the loopback only can tell if the adapter works, but not the TTL level it uses). Why don't you try on another Seagate drive, just to check it works properly? jaclaz