Jump to content

Quisquose

Member
  • Posts

    59
  • Joined

  • Last visited

  • Donations

    0.00 USD 
  • Country

    United Kingdom

Everything posted by Quisquose

  1. Quisquose

    DX9

    Hi! Thanks for the very informative and helpful comments Speed. You've really helped clarify the various options available (as well as helping me understand the pros and cons of each). The version of RyanVM's DX pack that I was using was 1.0, and not the new 1.1 that he has out now (which is dated December - so I'm guessing that it's based on the Microsoft December release that you mentioned). I only used the 1.0 version because it was the one that I'd already downloaded from a few weeks ago when I'd just got my new PC up and running. When I encountered the wrong DX version being displayed, I thought I'd check for an update and that's when I saw that the 1.1 version had not long been released. So I downloaded that one as well. I did wonder why there was such a difference in file size between the 2 versions. So basically, (just to confirm I've understood this correctly) are you saying that I can just use RyanVM's DX pack 1.1 with NLite (as I did before) and because the new DX 1.1 pack is for all versions of Windows, it should be fine on Windows 2000?
  2. Ah, ok. I'm not a programmer (so I don't really understand these things). I thought that it would be quite trivial to implement. Sorry, I don't understand that bit. NLite does know if errors have been encountered. It's NLite that tells me (it pops up a dialog saying so). If the session ini file has already been created at the beginning of the process (at time that the hotfixes are added) then the file already exists and so it can have text added to it by the 'process' that generates the error pop up dialog. (Excuse my inaccurate programming terminology). There must be code somewhere that performs the function "if file is not of correct type, then pop up dialog to warn user". So all that needs to be added is "and then add a comment to the existing session ini file". NLite would not need to know of these errors before hand, it would just write them into the pre-existing session ini as each inividual error was encountered (during the NLiting process). But maybe I am over simplifying matters.
  3. Just thought I'd post an update with regards to the problem outlined above (in other words Windows 2000 failing to complete an installation and hanging at 55% on the 'Installing Devices' setup page). After spending inordinate amounts of time re-installing different Windows 2000 versions, my conclusions were as follows: Any NLited version of Windows 2000 based on USP5.1 plus added hotfixes would not complete. (Could be any one of the above software causing the problem, or it could be a hardware detection issue). Any NLited version of Windows 2000 with just USP5.1 (and no hotfixes) would not complete. (Hotfixes probably not causing the problem. I could be removing too much with NLite). USP5.1 (on its own) manually slipstreamed into Windows 2000 (no NLite at all) would not complete. (NLite ruled out as culprit. Maybe USP5.1 is to blame). Using Official SP4 (instead of Gurgelmeyer's USP5.1) + Microsoft hotfixes would not complete. (I might have incorrectly included a hotfix that clashes with an update already included in SP4 and USP5.1). Repeat the above installation but leave out all hotfixes (therefore using SP4 only). This does not complete. (Therefore it can't be a Hotfix issue, or Hotfix + Service pack clash. Also it can't be USP5.1 to blame). OK, Back to basics. Remove surplus hardware. Install original Windows 2000 (no service packs). This does complete the install. (Possibly a hardware issue?). Re-do the above installation, but using Windows 2000 with SP2. Leave hardware installed. This does complete the install. (Hardware is ok. Looking more like it's an SP4 issue - which would explain why USP5.1 also won't work). Re-do the above installation but this time using Windows 2000 SP3. This does complete the install. (Maybe something has been fixed by uninstalling then reinstalling the hardware). Try again using Windows 2000 SP4, to see if something has inadvertently been fixed. This does not complete. (Hmm.... problem still exists. Re-installing hardware hasn't changed anything. Maybe it's something in SP4). Install SP3 again (as this version is known to work) and compare the installation logs with the SP4 logs. (No significant errors found in SP4's logs, but the last entry before hanging at 55% referred to 'EHCI'. Doing a web search for some more information on what EHCI was (seeing as I'd never heard of it) pulled up one forum particular thread: http://www.techspot.com/vb/post359468-8.html This post mentioned Asus motherboards (which I am using) and issues when installing Windows (which I am experiencing). I went to my BIOS and looked for the EHCI handoff setting. It was set to disabled. I flipped it to enabled (just to see what would happen) and tried to install SP4 again. This time it worked! So then I tried to install USP5.1 with Hotfixes and NLite tweaks, and this also worked! So the issue of the service pack discrepancy between SP3 and SP4 was a red herring. For some reason you can install on SP3 (and earlier versions) with EHCI disabled in the BIOS, but for SP4 and later you must enable EHCI for Windows to be able to complete an install. (At least that's the case on my Asus motherboard). I can now understand why ISOs that I had made previously (which worked fine) suddenly stopped working. When I first built my new system I did a quick rough test to see if Windows would install OK (and to check that none of my new hardware was faulty). I then set about properly configuring my system. This started by me updating the BIOS (which was very out of date) on my motherboard. The Asus forums give strict instructions to clear the CMOS both before and after flashing the BIOS. Failure to do so has resulted in a trashed BIOS and a totally dead motherboard. So that's what I did, and in the process all of my BIOS settings got returned to their defaults. Asus (in their wisdom) have chosen 'disabled' as the default setting for EHCI. As I didn't really understand what that setting was, I left it at its default value after performing my BIOS upgrade. I still don't really understand what EHCI is, but it's something to do with the handling of USB and how it's passed from BIOS control to OS control (depending on how the system is being used - i.e. hardware POST or OS Loading). This kind of makes sense, as devices such as USB mice are available way before Windows drivers for them are loaded (and so it's probably being controlled by the BIOS at that stage). At some point this has to be handed over to the OS, and I think that this might be occurring at the 55% point of the 'Installing Devices' Windows Setup page. On subsequent inspection, you do see the mouse pointer disappear at this point and then re-appear about 15 seconds later (provided it hasn't hung), so maybe that's what's going on at that point. Apologies for this ridiculously long post, but hopefully people experiencing similar problems will stumble across this thread when Googling, and it will save them from tearing their hair out for days on end. If I had not found that TechSpot post by Sharkfood, I'd probably still be thinking that the problem was in SP4.
  4. Quisquose

    DX9

    Wow, well spotted Ponch! We have a sleuth in our midst. Even I hadn't noticed that particular clue (and I was the one who wrote it!) OK folks, thanks to every one for their comments. Here is an update: I have now managed (for the very first time) to install DX9 unattended Yes, after spending more than 75 hours over the last week trying to sort this problem out, I am finally getting somewhere. I used RyanVM's DX pack with a manually edited txtsetup.sif file (with Ver Greeneyes' corrections added) and not only did this install DX9, but it also allowed me to slipstream my ATI Radeon drivers for my PowerColor X1950 Pro. However, none of this would have been verifiable or testable had I not also solved the 55% hang issue because I was totally unable to complete any installation of Windows 2000. But now this problem has also been resolved (yes, I'm on a roll!) @ Speeddymon - Thanks for the kind offer to help me find a fix. A link to the 55% Hang bug was posted in my earlier post (message number 17 of this thread). I'll include another link here (see the green "55%" text above). I've coloured the link this time, to make the it a bit more noticeable. I'll update that 55% thread with my solution after I've finished writing this post. There were a couple of minor issues though (but nothing in comparison to the grief I've been enduring over the last week). 1. DX9 appears to be fully installed, allowing ATI Radeon drivers to work and all tests in DXDiag to complete successfully. However, the version number on the first tab of DXDiag is still showing as version 7 (i.e. the original Windows 2000 version of DirectX). 2. Drivers for the RealTek onboard HD audio (Azalia) were also integrated. These did get copied to the system, but the drivers don't actually get successfully installed. Closer investigation showed that a requirement for the installation of these sound drivers is that a minimum of DX8.1 be installed. So I am wondering if these 2 issues are related (perhaps the RealTek drivers just look at the DX version number and then refuse to install). I was going to post back here to seek assistance with these two issues, but having read the more recent comments in this thread, it seems that I was lucky to get this working at all! (because it seems the DX update packs are really not intended for Windows 2000 use). So, I think I should make another install with a more appropriate DX pack and see how that works instead. I'm just wondering which method (and which actual DX pack) would be the best to use in this circumstance. Although I really like the idea of the code being slipstreamed, I wonder what the downside would be to having it done as a silent install instead of a slipstream? Any comments or advice on this would be very much appreciated. Thanks again everyone
  5. Quisquose

    DX9

    Actually that's incorrect. If you re-read the thread you will see that I said it before you even posted. That's how Ver Greeneyes spotted the problem (he recognised that the XP style install commands I'd quoted in my message were incorrect for use on Windows 2000). Anyhow, it's sorted now. Onwards and upwards to try to defeat the 55% hang bug!
  6. Quisquose

    DX9

    I'm sorry, I can't do that ......... the information is classified TOP SECRET. KiKi's 'people' would hunt me down like dog if I talked. Well this is the update on the situation: Thanks to Mr. Greeneyes, I've managed to get rid of the text mode setup errors. (I changed all the 100, values in that section of the TXTSETUP.SIF to 1,) There was a second section with similar values (much further down the file) that I hadn't noticed. This caught me out, but the Windows error message that was generated gave a different file name to the one it was giving before, so I knew that the first lot of fixes must have worked because Windows stopped complaining about those. I also knew where to look to find the second lot of values that needed fixing because I could search for the new filename being stated in this latest error message. This time I went through the whole .sif file, making sure there were no other entries with the same starting value. I then saved the .sif and burned a new CD and it worked! Well, ... when I say "worked" I mean I am now back to my original problem of the setup locking up in GUI install mode at 55%. But still, at least the text mode section is completing without complaining now. The 55% GUI mode hang problem is really winding me up though. This has been going on for days now, and I think I'm slowly losing the will to live! I have a sneaking suspicion that it might be down to the USP5.1 that I've been using, so I'm gonna have to go through yet more tedious testing to narrow it down (which is not easy as I've been integrating quite a few post SP5.1 hotfixes too, and the problem could possibly be down to any one of those files or a particular combination of them). I've got a feeling that it's going to be days more testing before I get anywhere near having a usable custom setup CD Just wish I had a hint to help me narrow things down.
  7. Quisquose

    DX9

    Thanks for your message Kiki, I'll look into it
  8. Quisquose

    DX9

    I'm not trying to integrate a "later" distribution, I'm using Windows 2000 (which has no DirectX 9 in it all) and so I'm just trying to get any version of DirectX on to the system. I don't know about you, but I'm not prepared to run Windows in 4bit (16 colours) @ 60Hz, and if you try to install ATI drivers without DirectX 9 being present on the system, then you will get an instant BSOD crash. I don't even play any games (I don't have the slightest interest in gaming), I just want the damned thing to work at a decent colour depth and refresh rate. @ Ver Greeneyes now that you mention it, that section does stand out from the rest of the sif file because of the 100s (which none of the other entries seem to have). I'll edit the txtsetup.sif and see what happens. If it doesn't work, then at least I know it is more likely that Ryan's pack is to blame (as it not taking into account installation on Windows 2000). It gives me a direction to investigate further. Thanks for the heads up.
  9. Quisquose

    DX9

    Hi Kiki, I did think about posting to that forum, but I'm not really an HFSLIP user (I don't find it very intuitive) and I prefer NLite. I only resorted to trying HFSLIP after repeated attempts to integrate DirectX using NLite had failed (and I was getting so fed up that ANY method that worked would have been acceptable to me). The issue that I was having occurred when running HFSLIP for the second time. Basically nothing would happen at all. No error messages, no pop ups or anything. I could try running the file as many times as I wanted but it just did nothing. At first I left it for ages (thinking it might be doing something in the background) but it wasn't. I went back and followed the HFSLIP guidelines really carefully, step by step. The program always worked on the first launch (it created all the correct folders as expected) but then when it came to the second time to run it, nothing at all would happen. I tried uninstalling and starting again, but the same problem persisted. So I gave up and came back to NLite (which I am more comfortable with). Thanks for enquiring into my problem. The error happens during the text mode part of Windows 2000 setup (after the partioning and formatting have completed). The error message says: The following file in SIF file used by Setup is corrupted or missing value 0 on the line in section [sourcediskfiles.x86] with key "d3dx9_24.dll" I checked my files and the file mentioned is actually present (along with other DX files). I also checked my txtsetup.sif and the line referred to by the error message is: d3dx9_24.dll = 100,,,,,,,2,0,0 but I'm afraid that I don't really understand the contents of the txtsetup.sif file, so I don't know if that value is correct or not. These are all the DX related files that appear in that particular section of the txtsetup.sif: d3dx9_24.dll = 100,,,,,,,2,0,0 d3dx9_24.inf = 100,,,,,,,20,0,0,d3dx9_24_x86.inf d3dx9_25.dll = 100,,,,,,,2,0,0 d3dx9_25.inf = 100,,,,,,,20,0,0,d3dx9_25_x86.inf d3dx9_26.dll = 100,,,,,,,2,0,0 d3dx9_26.inf = 100,,,,,,,20,0,0,d3dx9_26_x86.inf d3dx9_27.dll = 100,,,,,,,2,0,0 d3dx9_27.inf = 100,,,,,,,20,0,0,d3dx9_27_x86.inf d3dx9_28.dll = 100,,,,,,,2,0,0 d3dx9_28.inf = 100,,,,,,,20,0,0,d3dx9_28_x86.inf d3dx9_29.dll = 100,,,,,,,2,0,0 d3dx9_29.inf = 100,,,,,,,20,0,0,d3dx9_29_x86.inf d3dx9_30.dll = 100,,,,,,,2,0,0 d3dx9_30.inf = 100,,,,,,,20,0,0,d3dx9_30_x86.inf d3dx9_31.dll = 100,,,,,,,2,0,0 d3dx9_31.inf = 100,,,,,,,20,0,0,d3dx9_31_x86.inf Originally I was working with NLite v1.2 Final, but thought I'd try 1.3 beta as well (just in case) but it made no difference. I still get the same error message even when integrating again from scratch with clean source files. My first reaction was that it was a problem with the DX pack itself, but it's difficult to raise issues with the pack when you are using NLite to integrate it (because you can't really ask for help on Ryan's forum). My NLite session ini file is attached. Thanks for your help.
  10. Quisquose

    DX9

    Well it wasn't for me. HFSlip refused to work. Is NLite incapable of integrating DirectX, because whenever someone asks about this people never give a straight answer, they just start talking about other programs or complicated manual methods. I tried using Ryan DX pack but I keep getting an error during Windows setup. HFSlip won't work either. As a nice simple solution like USP5.2 will probably never be released, I'm left wondering if I will ever get DX slipstreamed.
  11. Since when? Are you talking about XP? cos it's certainly not a part of Windows 2000 (which is what I use). In fact i STILL have not found a way to get DX slipstreamed into Windows without problems.
  12. Can anyone tell me how I go about specifying the name that Windows should use for the System32 folder? I'm soon to start working on my XP64 install and I would like to sort out the idiotic naming scheme that Microsoft has adopted for the System folders in 64bit XP. I've already fixed the stupid Program folder names (which is easy because they are defined in winnt.sif) but I imagine that renaming system folders is more challenging (possibly requiring editing of hive files or something like that). Any help very gratefully received. Thanks!
  13. What does it mean by new: 'DirectX' on the NLite history page for version 1.3 beta? Yeah, I know what Direct X is, but why is NLite mentioning it? There are no release notes with version 1.3, so I can't find any explanation of this new feature.
  14. I tried to slipstream some hotfixes into an install and during the process got warned by NLite that one of the hotfixes was not of the expected type (it was therefore skipped). However, when I checked my NLite session ini, the file was still listed in among all the other hotfixes with nothing to distinguish it and comment about the fact that it had not been added. Might it not be a good idea to have this line REMd out with a comment, so that when you go back to look at the session ini later, it more accurately reflects what has been actually been done rather than what was attempted? Just an idea. Thanks
  15. Jtbalogh, I think nuhi meant that shormanm should attach his NLite session.ini to his forum message, so that nuhi could examine the settings to see if there was any obvious cause for the problem. It wasn't a method for fixing the issue itself. I may be wrong, but that's my understanding of the comment.
  16. Hi nuhi, Thanks for taking the time to reply (and for confirming that my setup works for you). At least that is one thing that I can rule out. I was planning to use VMware to test out all my registry tweaks and silent installs, but I have not got to that point yet (seeing as I can't even install). There are a couple of other reasons why I haven't got round to using VMware yet; the first is that the problem I'm seeing occurs during hardware detection and device install, and a virtual environment won't be able to replicate / emulate the particular hardware combination that I've got in my new PC. Secondly, the PC that I am replacing is really very old (PII 233MHz) and running VMWare on that would be a real stretch. I think I'll try a 2 pronged attack; first install XP on the new PC so that I can test the Windows 2000 hotfixes and Service pack in VMWare. Then, if there are no conflicts or problems, I'll go back to real hardware boots to test the device detection and install procedure. Hopefully the problem will disappear now that I have totally started again from scratch, but if it hasn't, then the testing above should help me narrow down where to look.
  17. I'm being driven nuts by a bastard Windows 2000 install that will not complete on my new PC under any circumstances. It all started out fine, I did a test install of Windows 2000 slistreamed with SP4 and that worked. Then I NLited the install (nothing was added at all, I just removed useless components). That worked fine as well. Then I was collecting drivers for my new hardware so that I could add those via NLite. I was checking out options in the BIOS and managed to hose my perfectly fine NLited SP4 install, because I stated booting with AHCI enabled (where as the install had been done as standard IDE). Needless to say, switching back to IDE mode in the BIOS made no difference and the install had been trashed. No worries I thought, I'll just re-install ........ That was the beginning of my 48 hours of hell. No matter what I did, I just could not get Windows to complete an install. I made sure that I completely deleted any partitions on the disk and the re-created and formatted them (just to be sure). It all goes fine up to the GUI install part where it says "Installing Devices". The progress bar moves very quickly and smoothly up to 50% then it slows down, and then at approximately 55% the CD-ROM stops spinning, all hard disk activity stops and the PC locks up (can't move mouse pointer). I tried leaving it for up to 20 minutes to see if it would start up again but it didn't. I also tried several re-boots to see if the install could pick up where it left off previously (and hopefully skip the troublesome issue) but again that did not work either. So then I went back to the pre-NLited version of the ISO and burned that to CD again, but that hung at exactly the same place as the other installs. I then re-created a new ISO from fresh files but that didn't work. I tried re-setting my BIOS to defaults (not that I had altered that much anyway) and tried installing on to 2 different hard disks, but all the time it was the same 55% hang issue at "installing devices". I was just about to start yanking out hardware to see if that would help (the only thing I could remove is the sound card as that is the only add on apart from the graphic card) when I decided to try my XP install instead of Windows 2000. This worked, so the issue is Windows 2000 specific. I'm really hoping someone can help, as I am so p****d off right now. I did have a look through some of the log files from the install (not that any of it makes the slightest bit of sense to me) and DTCSetup.cat was the last file to be mentioned in a few of the logs, so I did a search on that file and found some people who'd had issue with it. In one thread (on MSFN) it said that the file is not really neeeded and that it could be rem'd out of Syssetup.in_, Txtsetup.sif and Layout.inf. I did this and remade the ISO and it still didn't work. However, I've done so many re-installs now that the log that had the most errors in it has been overwritten. Strangely, subsequent installs did not generate hardly any error messages (despite hanging at exactly the same place as before) I'm at a total loss as what to do next. Can anyone offer some advice please, to at least point me in a direction so that I can narrow down the investigation. I wonder if it might be graphic card related (Powercolour ATI Radeon X1950 Pro) as the problem occurs around the time when the screen flickers (on a normal install) although I don't actually ever get to that point. I've attached a log file which may (or may not) be of some relevance. My NLite settings attached to next post. Thanks! ..... and here are my NLite settings (it wouldn't fit on my previous post). These settings actually worked fine the first time I tried it, but all subsequent attempts (including total remakes) have not worked since.
  18. Does anyone know how I can get the standard property sheet "Summary" tab (and tooltip) to work on .url Internet Shortcut files? I can understand why Microsoft would add a shell extension for the Web Document tab (for access to the 'Make document available offline' checkbox and to edit URL path) but what's the point of removing the Summary tab that is present on nearly every other file type? I use Metadata a LOT, and it would be so much more useful to be able to add notes (Subject, Keywords, Comments etc) to an internet shortcut (and have those pop up in a tooltip) rather than being forced to guess what kind of web site it is from some cryptic url. I have tried all the obvious methods; such as adding the shell extensions under HKCR\* to .url, doing reg searches on all the GUIDs that have any asscoiation with the .url or internet shortcut keys in the Registry and removing them one by one to see what was making .url a special case, but none of it this has worked. The only thing that I have managed to do is to make the Web Document property sheet disappear, but even that did not result in the restoration of the Summary tab. I'm sure there must be a way to do this. Does anyone have any ideas? Thanks in advance for any help from the Registry gurus out there.
  19. I am quite taken aback by some of the responses in this thread (given the nature of this site). I too have been looking to find a way to do this. In fact a few years ago I asked an MVP on some Microsoft Newsgroup how to do it. He tried and tried, and after several days posted back a message ranting that he had wasted 3 days and was getting no where and so was going to abandon trying - lol! Why Microsoft insist on pi**ing people of by doing stuff like this is beyond me. The Win2k Send To folder was much better, as it was just a standard folder with shortcuts in it. You could even flip the hidden attribute on the shortcut to make it disappear from the menu (without having to permanently delete it). I'm a noob, so I'm not really in a position to speculate, but I would say there are 2 possible solutions (both of which I have yet to investigate myself). 1. Nuke the entire Send To menu, and re-create it with a 3rd party tool (so that you can control exactly what goes into it). 2. The "will not die under any circumstances" nature of these menu entries is behaviour that is typical of Shell Objects. This suggest it might well be a GUID key under CLSID with attributes set to prevent editing (although how you'd go about finding the relevant GUID is another matter). Sorry to not be of any help. I wish I knew the answer myself.
  20. Thanks everyone for the replies! Because that's not a variable. Hard-coding half of the path provides no benefit over hard-coding the entire path (if the path gets changed). Whoa! I never even knew things like this were possible! I will look into this. Thank you very much for the heads up. I found my own semi-solution in the end. I used SetX with -k switch to assign variables to values read from UserShellFolders. I even managed to then use those same variables back within the User Shell Folders key itself! (you just need to ensure that the variable is set early enough, so that when the reg key is read it knows what to expand). It's a bit kludgy, but it works reliably and does do what I wanted to achieve. Ricktendo64's suggestion is much more elegant though, and would be my preferred method, provided that I can get it to work - (I'm still a noob). Thanks again everyone.
  21. Ah, that makes sense. It all seems so obvious (now that you have fully expained how and why they both behave the way they do). Thanks again
  22. Yeah, me too. Those icons are totally minging. Fortunately they provided a way to turn off the tray icon (I hate it when stupid developers use the 'Notifcation Area' as a 'Quick Launch' area) so for me that just leaves the control panel applet (which I assume would have the icon located in the .cpl file) but I've yet to look.
  23. Does anybody know if there is a System variable value for Local Settings? I've been driving myself crazy trying to find a solution to getting an environmental variable value for this particular folder. After many, many hours of searching, I found some web sites that listed a more comprehensive collection of variables (rather than just the usual suspects that most sites trot out) but many of these lesser known variables (such as %recent% and %cookies%) simply did not work. Annoyingly, there is a variable for the Application Data folder (%AppData%) and this works fine, so I don't know why there isn't one for Local Data. I thought about creating my own system variable to achive the same end, however I can't figure out how to define the variable without hard-coding the path to decribe it (which defeats the purpose of the reson why I want to use a variable in the first place). Anyone have any ideas?
  24. Thanks very much for the detailed response Plamdi. So would you say that Winzip is doing something to not add a context menu to the Start button, or is WinRAR specifically doing something to make sure that it does add a context menu to the Start button? I think Winzip's way makes far more sense, and I can't believe that RARLab would have spent time hard-coding this non-sensical addition into a DLL.
  25. Geting to grips with the stuff is confusing enough as it is without having to contend with mistakes like this in an official instruction guide ..... But thanks very much for the clarification IcemanND, at least it shows that I was paying attention - lol
×
×
  • Create New...