UCyborg

For anyone that might be interested in the possibility to continue using partially defective RAM, I have one such module in my laptop. I figured from Memtest86 results (addresses) that the faulty parts are somewhere near the end, somewhere in its last 40 megabytes. There's just one 2 GB stick. So one can tell the OS to not use the last 40 MB by running this in the Command Prompt: bcdedit /set {current} removememory 40 If there's 2x 1 GB sticks for instance and those 40 MB would be in the one in the first slot, it would be a good idea to swap their places as cutting whole 1064 MB isn't desirable. Those addresses in hexadecimal notation address the bytes, so one has to convert to decimal and do some math or use some (online) conversion tool to get the megabytes and calculate the difference from that point to the end of the module. If I remember correctly, I actually put the result in the {globalsettings}, not {current}, so that way, it works for the recovery environment and other Windows versions booted through that Windows Boot Manager instance, not just the running OS from which bcdedit was run. There's also truncatememory that can be used instead of removememory, which accepts address, but I've never used that option (https://docs.microsoft.com/en-us/windows-hardware/drivers/devtest/boot-parameters-to-manipulate-memory). In case someone has very minimal faulty bits, they can be more precisely mapped out using: bcdedit /set {badmemory} badmemorylist PFN1 PFN2 PFN3 ... Individual 4 KB blocks (pages) can be mapped out of action this way, so whatever addresses Memtest86 gives, they have to be rounded down to the nearest page (so one gets the multiple of 0x1000) and then the last 3 zeros have to be cut from the result to get the Page Frame Number that Windows accepts in hexadecimal notation with a 0x prefix. RAMMap can be used to verify that bad pages are no longer on the list. Got this from Super User. Also, badmemoryaccess doesn't have to be explicitly set anywhere, it already defaults to no if not present; with default configuration, it's only explicitly added and set to yes on the entry for launching Windows Memory Diagnostic, so parts marked as bad are accessible and can be tested.

It might, but definitely not on vanilla Windows 98. No need to (at least not for me), it's more like a passing curiosity. Haven't had a Win9x compatible computer in over a decade.

Is there anything like f.lux for 98? Supposedly some old f.lux build from 2016 works on it with KernelEx.

What about web browsers? I settled on Via.

If your motherboard's BIOS actually supports ACPI, Windows should have been installed by invoking the setup program in win98/win9x folder on the installation CD from the command line like: setup /p j You can't avoid the message to turn off the computer if your hardware lacks ACPI support because such computer can't be turned off by software.

Aye, but I haven't found the archive of RTSS. Maybe I didn't look hard enough. Does it try to connect every time even if you set the longest allowed checking interval? If not: I noticed Config file under Profiles sub-folder has a timestamp under LastUpdateCheck, unix epoch time in seconds in hexadecimal format with a 'h' suffix. Maybe setting it to the distant future date would help, eg. FFFFFFFFh or at least 7FFFFFFFh.

MSI Afterburner displaying some stats via RTSS on a D3D11 game and a D3D8 game on Windows 8.1. RTSS settings were left at defaults. Windows 8.1 is at the December 2017 patch level + .NET Framework 4.8 (KB4486105). .NET likely doesn't matter in this case unless .NET based game is involved. More information would be needed to solve the OP's problem. Sometimes the module name where it crashed gives a hint, but it's not necessarily where the main issue resides. The crash log can be found in Event Viewer (Windows Logs->System) or in Action Center (Maintenance->View reliability history). The developers of RTSS would be the best equipped to diagnose the issue if they're provided with the crash dump. Crash dumping is enabled by creating a registry key described here and adding described settings. Default values are probably OK for the start, the person analyzing the dump should be able to tell if it should be configured differently to give better results. Obviously you can link it here and hope someone that comes by can make a sense out of it, but it's easiest for the person that has the source code of the crashing program since then Visual Studio will be able show where it crashed at the source level, variables will be identifiable by name etc.

MSI Afterburner 4.6.3.16053 with RivaTuner Statistics Server 7.3.1.24485 works fine on my Windows 8.1 installation, so it seems OP's issue can't be Windows version alone. It is yet to be tried, but an older version of RivaTuner Statistics Server's RTSSHooks.dll bundled with MSI Afterburner 4.6.2 doesn't depend on K32GetProcessMemoryInfo function. I only dug through to the DLL with 7-Zip to check its dependencies with PE Viewer. Found the installer on Filehorse. My theory is that they wanted the world to forget about the old systems at some point, so "supported" would take into the account whether an OS is still supported (at the time of writing). Also, you were looking at the documentation for driver developers. Average Joe would go here: https://docs.microsoft.com/en-us/windows/win32/api/winnt/nf-winnt-interlockedcompareexchange (this one says XP is the minimum).

From what I could gather, CsEnabled setting doesn't work since 2004, but some users reported success with PlatformAoAcOverride on 20H2. But what do I know, my HW only knows real S3 sleep mode. Obviously tinkering with Windows settings won't help if the issue is at the lower level.

Windows NT 4.0 and Windows 98 both have all 4 functions mentioned in the post above.

Would any of these help? If not for the laptop in question, maybe it's still of use on another piece of hardware. New registry setting to disable modern standby: Windows Registry Editor Version 5.00 [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Power] "PlatformAoAcOverride"=dword:00000000 EFI program that patches ACPI tables to disable modern standby: https://github.com/ElectronicElephant/Modern-Standby-Byby

https://microg.org/ And what were the results? I figured I have 3 apps that depend on those services and could potentially continue to use them without Google's stuff, assuming microG satisfies their dependencies. There are lots of issues opened on its GitHub page.

Indifferent when it comes to code names. I don't mind if it has them, but so far, version number tells me more than the codename...though for software updated very frequently, those lose meaning as well over time, at least for me, I don't have the mental capacity to keep up.

Assuming there aren't any other changes that would require modifying extension's code, it should be enough.

My DVD drive tray stopped opening some time ago. It would get stuck in initial position after pushing the button. I opened it manually by pushing a needle in the designated hole, now it opens normally most of the time, sometimes it still gets stuck when trying to open it for the first time in the day. So far, it always opened on the second try. In 12 years I've had this computer, DVD drive was used very rarely.

A desktop and a laptop.

Most likely reason is that you have N edition of Windows 10 and have to install Media Feature Pack to get missing components that certain software depends on.

With default settings, DEP is already effective for all applications that have a flag in its executable header that it's compatible with DEP. Your typical internet facing applications most likely have it set already. If you check the .exe with Process Hacker's PE Viewer, it should have NX compatible listed under Characteristics. CPU must support the feature and I've read that the motherboard/BIOS settings may prevent it from being usable, even when supported by the CPU. First Pentium 4 CPUs based on Northwood core don't have it, later Prescott based ones do. On AMD side, Wikipedia suggests support started with Athlon 64.

Reading about CRTs and such here in 2021 while "Weird Al" Yankovic sang "hundred gigabytes of RAM" and "flat screen monitor forty inches wide" in 1999! https://www.youtube.com/watch?v=qpMvS1Q1sos

There are multiple drivers for that GPU on Microsoft Update Catalog/Windows Update, so unless your Windows installation is fubared, one of them should have been installed automatically. The update procedure can be triggered manually from Device Manager->your graphics card's (display adapter) properties window->Driver tab.

I'm not involving older OS in this case. At the first glance, deleting reparse points is not a problem as long as you get the updated OS to do it, which is better familiar with its specific reparse points, though copying them to another file system is still another story. Using Ln for the first time, I didn't turn on logging back then, so I didn't get the entire picture of what else failed to copy, besides 8 reparse points, the summary noted 5 folders as well. I'm going to take alternative route to do what I'm trying to do - merge data of 2 partitions onto a single partition - so instead of copying Windows to another partition, I'll copy other stuff to Windows partition.

There's still a bit of blue in there, it was solid white on my end. I don't know if anything can be done about builtin start menu tiles. In older builds, it's possible to color individual program's tiles by placing a special XML file in a program's directory that specifies color and what image to use as a tile icon, but the color parameter is ignored now.

Apologies, read that too fast. Still, looking at the actual content from the official Win10 20H2 Slovenian ISO's install.wim file, it's not there. Haven't seen it in english ISOs neither. I've read about people using renamed copy of official aero.msstyles as that changes behavior when it comes to coloring windows' title bar. I get similar appearance as on your screenshot. Inactive windows' title bar is white then, which I believe is connected to aero.msstyles being loaded with its original name.

Aye, its health is at 100%. Yes, but I only tried Windows 10 and its PE environment. I imagine it would be the same with Windows 7. I once tried deleting all Windows 10's folders from its root partition with good old rd /S /Q command using Windows 8.1 PE, it was unable to delete such reparse points, but it was possible with Windows 10 PE. Definitely seems to be the case here. Interesting, it seems Set-ExecutionAlias could be used to replicate them on another partition. But why there isn't a generic way to replicate arbitrary reparse point on another NTFS file system since you don't really care about the content.

You must have taken someone else's theme file then - it refers to %SystemRoot%\resources\Themes\colored\colored.msstyles Windows loads the default one if it doesn't find it. @Tonny52 is right though and it doesn't look like there's a setting to set the old start menu style.