Home Creators Posts Import Register Favorites Logout

Downloads

Content

This release bundles the ICU DLLs and data files that are used by Qt 6.10 and up, and includes a fix to an issue affecting D3D11 and DXGI on Windows 8.1 that was introduced by the addition of a D3D11 feature that allowed Parsec to function on Windows 7 and up.Updated on 2026-06-18 at 2209 (UTC-4) to fix issues with FontForge and the latest Microsoft Edge installer on Windows 7.Updated on 2026-07-05 at 1126 (UTC-4) to add tbs.dll wrapper for Windows Vista (to fix Chrome 150 crashes due to Windows 11 compatibility check), fixes for the Second System for Services PE patcher, incorporate wtsapi32 and setupapi wrappers on Windows Vista (for Parsec Portable), add new DLL definition for bcryptprimitives to fix issues with QBitTorrent on Windows Vista and 7, and add a new DLL definition for mmdevapi to fix Discord voice calling on Windows 7.

Updated on 2026-07-07 at 1246 (UTC-4) to fix issues on Windows Vista with the wtsapi32 wrapper.Updated on 2026-07-08 at 2230 (UTC-4) to add Windows Vista support for the MMDEVAPI wrapper.

Comments

Templayer

Windows 7 64-bit Ultimate, AMD FX-83XX, Nvidia Quadro FontForge is failing to start (the newest Chocolatey version of it), fails silently without producing any error, even in event viewer (where are the logs for Second System?) Which is a problem, since it fails due to a missing Win8+ function when launched without Second System. I might have to downgrade it back to 2023 VxKex NEXT, meanwhile, has a different error for FontForge (a nullpointer happens in its innards instead, but at least I get an error window while using it!) Faulting application name: fontforge.exe, version: 2025.10.9.0, time stamp: 0x68e807e7 Faulting module name: pwrs_k32.dll, version: 10.0.17763.5000, time stamp: 0x6a2c2cd8 Exception code: 0xc0000005 Fault offset: 0x0000000000041fe3 Faulting process id: 0x2bc0 Faulting application start time: 0x01dcfea42c4e1bb6 Faulting application path: C:\Program Files\FontForgeBuilds\bin\fontforge.exe Faulting module path: C:\SecondSystem\x64\pwrs_k32.dll Report Id: 6dc19068-6a97-11f1-aa78-b42e99468bb7 nheko fails to start. Which is a problem, since I finally forced myself to install Second System primarily for nheko. It's a qt6 app. I need a Matrix client that is not electron-based for low-resource usage use-case scenario. The old Chocolatey 2023 version is qt5 and fails silently with Second System (works without VxKex Next and without Second System, but images are not shown, only their salt hash). 0.12.1 is qt6 and it looks like it only launches with the gaming directx11 profile, but crashes with an error window upon launch. Faulting application name: nheko.exe, version: 0.12.1.0, time stamp: 0x68989db2 Faulting module name: unknown, version: 0.0.0.0, time stamp: 0x00000000 Exception code: 0xc0000005 Fault offset: 0x0000000000000000 Faulting process id: 0x2f28 Faulting application start time: 0x01dcfea2bf14060e Faulting application path: C:\Program Files (x86)\nheko\nheko.exe Faulting module path: unknown Report Id: fda20334-6a95-11f1-aa78-b42e99468bb7 FluffyChat fails: (haven't tried all profiles, though) Faulting application name: fluffychat.exe, version: 1.0.0.0, time stamp: 0x6454e003 Faulting module name: ucrtbase.DLL, version: 10.0.14393.2990, time stamp: 0x5caeb96f Exception code: 0x40000015 Fault offset: 0x000000000006e01f Faulting process id: 0x19d0 Faulting application start time: 0x01dcfea54d84c450 Faulting application path: C:\ProgramData\chocolatey\lib\fluffychat\tools\fluffychat.exe Faulting module path: C:\windows\system32\ucrtbase.DLL Report Id: 8bcfc40b-6a98-11f1-aa78-b42e99468bb7 Faulting application name: fluffychat.exe, version: 1.0.0.0, time stamp: 0x6454e003 Faulting module name: permission_handler_windows_plugin.dll, version: 0.0.0.0, time stamp: 0x6454dfb4 Exception code: 0xc0000005 Fault offset: 0x000000000000404c Faulting process id: 0x33b4 Faulting application start time: 0x01dcfea540edc60f Faulting application path: C:\ProgramData\chocolatey\lib\fluffychat\tools\fluffychat.exe Faulting module path: C:\ProgramData\chocolatey\lib\fluffychat\tools\permission_handler_windows_plugin.dll Report Id: 7f0a13fb-6a98-11f1-aa78-b42e99468bb7 Datovka is a qt6 application that launches with the Win11 profile, but fails to get autheticated due to an SSL error: https://imgur.com/ensyzY1 NeoChat fails to launch. Faulting application name: neochat.exe, version: 0.0.0.0, time stamp: 0x671ceaa5 Faulting module name: neochat.exe, version: 0.0.0.0, time stamp: 0x671ceaa5 Exception code: 0xc00000fd Fault offset: 0x0000000000a6b6a7 Faulting process id: 0x2954 Faulting application start time: 0x01dcfea72ba087ea Faulting application path: C:\ProgramData\chocolatey\lib\neochat.portable\tools\bin\neochat.exe Faulting module path: C:\ProgramData\chocolatey\lib\neochat.portable\tools\bin\neochat.exe Report Id: 697ff301-6a9a-11f1-aa78-b42e99468bb7 This doesn't mention what it was for - maybe Datovka? PFX operation failed as AuthSafes count doesn't lie in expected range. Maximum permissible value: 200. Erroneous value: 456. It looks like many applications just fail to launch without any sort of error with the default config. I kinda wish I could set it up for Second System not to be used by default, as I will now have to go through hundreds of applications and games that work fine without it to disable it (I've installed this Win7 in 2017). I'm still attempting to find a single instance where it actually works properly. :( There was no error during the installation, everything reported success. I did restart the PC, of course. And my explorer.exe just crashed and restarted, and 7+ taskbar tweaker was unable to hook it. :S That said, thumbnail loading is pretty much instant now even for a folder with many 2K and 4K music videos on an nonSSD (i.e. HDD) drive (probably loaded from the system thumbnail store, which is on an SSD, but still! This bothered me for years... and suddenly, poof!)

Templayer

Also checked the antivirus, no hits.

win32

You can set a REG_SZ value in HKCU\Software\LSC\SecondSystem named DefaultProfile to Passive. I have just found the issue with FontForge, and I will submit the crashing fix soon, but it turns out that it also has a problem with my newer DirectWrite, so it will have to be run in Win10NoDWrite mode or Second System disabled entirely. Neochat (unstable master-16064-windows-cl-msvc2022-x86_64) is starting up fine for me, but some of the other QT6 apps you've listed don't work on my D3D feature level 10_0-only testbed and I will get to them later. But the issue with Neochat appears to be a recursive call in a DLL, and a crash dump (which would be created in %LOCALAPPDATA%\CrashDumps if reg key HKLM\SOFTWARE\Microsoft\Windows\Windows Error Reporting\LocalDumps exists and WER is enabled) would probably reveal the source of the issue there.

Templayer

About half-way of configuring every application not to use Second System, I found out that now every single .exe file (i.e. anything for which it wasn't set up yet) is set to the profile that leaves Second System disabled! YAAY! And if I set an application to actually use it, it persists WHILE everything is still set not to use it! I might have just encountered a bug that's actually quite the feature! :p Either that or I am very dumb, which is also possible. EDIT: AND THEN I read your comment. I have no clue how I did it, though, without editing the registry.

Jaguarek62

your second system is almost, almost ready to launch rockstar games launcher. I started digging, it is really obfuscated, but I got here: https://snipboard.io/fWYQ24.jpg Would you help debug it further?

Templayer

Also, just to clear the confusion up - my second reply to my own post wasn't a reply to you, as I didn't know you had replied to the post yet (thanks, Patreon!) So there's also the bug/feature of disabling Second System globally by setting it to the disabled profile by default WITHOUT using the registry. I don't know how exactly it happened, maybe I have managed to launch the config dialog WITHOUT a proper exe target and this was the result of that? It just started to be disabled by default in the middle of me going through all the applications in my PC, setting it up (or rather - unsetting it)

win32

The "Second System Disabled" is the default state of any application without an entry. Perhaps it should simply be blank. And I also found that Chocolatey distributes some very outdated versions of Neochat. However I couldn't get the version 24 you were using, just version 23, as well as the latest versions from other sources like here: https://cdn.kde.org/ci-builds/network/neochat/release-26.04/windows/. So I haven't been able to reproduce the bug.

2LENSFLARE4U

Is there anyway to add the emojis onto the official discord app? I just see squares on where the emojis should be. Thanks.

bsp

I wish I had known about it sooner. Great job ;) For now, everything seems to be working fine except for VSCode 1.124, which for some reason crashes if the "release notes" tab is opened.

win32

Currently on my system, launcher.exe returns 0xFFFF7001. Now trying to find the best way to freeze execution without tripping the anti-debug stuff in there.

Templayer

Have you checked the dumps I sent you yet? I made a few too many replies, I don't want those to get lost in the ether. It's a bit back in the thread now. My neochat has been installed through chocolatey, it's its portable variant, and it is 24.8.2, see here: https://imgur.com/vBLFLc8 If you need anything else, just ask. Afaik, primary for testing purposes, it would be great if the dialog window with choosing profiles had a checkbox for staying open. Trying out all profiles & VxKex NEXT configuration combinations takes an insane amount of time per each application. Damn those permutations! ✊

win32

I have now looked at all dumps. FontForge: problem found, will be resolved in next release (and is also not compatible with dwritecore, so fontforge.exe should be either disabled or set to Win10NoDWrite) Explorer: no Second System modules are loaded, so it may be caused by one of the several shell extensions that are loaded Nheko: looks like an attempt to load the 8.1 dxgi which was not present, in either dump (it should be named dxg8.dll, along with 81d11.dll for the 8.1 d3d11). However, my own enhancements to dxgi and d3d11 should be at a similar or even higher level so you may want to revert to the other profiles (except for Win10NoDWrite, since Qt 6 needs dwritecore or equivalent). Neochat: will now attempt with the portable release, since the crash has not been reproducible so far, and the stack in the dump is illegible.

2LENSFLARE4U

When attempting to install google chrome. the setup gets stuck at no disk space. for microsoft edge, it crashes at 42% of installing edge. What could be the problem?

Templayer

I tried nheko with every single profile, and every VxKex NEXT configuration combination (including disabling VxKex and trying every Second System profile, of course). Only the directx11 profile produces an error window - other profiles just do something in the background (busy mouse cursor) and then they die, and the process disappears without a trace (maybe literally). If I remember correctly, I didn't get much rest since I tried it. But nheko is the one I tried to get working the most, as it is also what I am using on my Linux systems, so I am well used to it. I even set a handwritten font for nheko on my personal laptop at work so that my boss can no longer read my personal messages. <_<

win32

Google Chrome installs fine for me, but it needs to be given a wide berth (at least 1.5-2 GB free on the volume that holds %PROGRAMFILES%). Edge seems to have just added some unimplemented win8 functions in its installer, which I am currently adding. Chrome's installer also writes a detailed log to Program Files (x86)\Google\GoogleUpdater.

win32

Which Windows version is this? I can't replicate this issue on Windows 7 with 2022 updates.

win32

Edge installer should now be fixed.

2LENSFLARE4U

It says on event viewer for MicrosoftEdgeUpdate.exe that "pwrb_k32.dll" is the faulting module. The installer closes itself after briefly seeing the setup with no errors.

2LENSFLARE4U

After removing leftover files the setup had left over, I get another error into event viewer saying that kernelbase.dll is the faulting module. The path is "C:\Program Files (x86)\Microsoft\EdgeUpdate\Install\{61949868-2021-4C9B-9DDB-E1229F774CD8}\EDGEMITMP_FB167.tmp\setup.exe"

win32

Did you install the latest version that I uploaded yesterday? I made changes to both modules.

win32

Was pwrs_k32.dll in %LOCALAPPDATA%\SecondSystem\x64 modified on June 18?

bsp

Windows 8.1 Pro

win32

That 0xFFFF7001 was indeed a crash in BCryptCreateHash. I did get the post-login crash but the only useful information was that the last win32 error was ERROR_NO_TOKEN.

Templayer

Also, if somebody asks if I am editing copyrighted fonts to add support for Czech diacritics, etc., it's total slander! Especially if the police is the one asking. https://imgur.com/stNq0qP Nope, perfectly normal, as all fonts should be! :V ...but yeah, I need FontForge. :>

bsp

On the other hand, should DRM in Firefox work? Or am I doing something wrong? Edit: Yes, my bad. Using the correct profile 8.1 DRM works again

2LENSFLARE4U

No it says 5/21/2026. I ran the .bat file and also the regedit win 7 defs.

win32

Did you uninstall the previous version, log off/reboot and then install the new version? That is how it should be done to ensure that the files are no longer locked by applications using them and can be replaced.

2LENSFLARE4U

It finally works now. The problem was that I never uninstalled second system and only ran the bat file and the defs the whole time. What if there are minor revisions such as this one (not major version updates) Would running the definitions only work?

win32

Some revisions only add definitions, but most do have changes to DLLs, although some of them are mostly for specific programs and/or specific applications.

Jaguarek62

Hi, Thank you for checking this. Yes, we also confirmed that the earlier 0xFFFF7001 crash was in BCryptCreateHash. That part seems to be past now in my current Second System 3.1 / Windows 8.1 setup. The launcher now gets much farther than before: - Rockstar Launcher starts visually. - The Rockstar logo animation appears. - The UI transitions work. - The account signs in. - The launcher goes online. - It reaches the post-login flow. - Then it closes/crashes later. So at this point it no longer looks like the original BCryptCreateHash issue. About ERROR_NO_TOKEN: You mentioned that in your post-login crash, the only useful information was that the last Win32 error was ERROR_NO_TOKEN. In our current traces, token probing did not confirm that as the active root cause. The token probe mostly showed normal OpenProcessToken / GetTokenInformation behavior. It may still be related or a side effect, but our current crash path looks different. Current technical findings: 1. Protected / virtualized dispatcher Launcher.exe appears to use a protected exception-driven dispatcher / virtualized control flow. There are many stack-thunk patterns and high-entropy executable sections. Per-crash AV skipping was abandoned because it was just walking through dispatcher symptoms. 2. Tagged mirror / lazy sync layer We implemented a guard DLL: rgl_rip_guard.dll The current important source file is: rgl_tag_mirror.c The guard implements lazy page-granular mirror coherency: - mirror pages start PAGE_NOACCESS - on execute/read/write fault inside a tagged mirror page, exactly one 4K page is copied from the live Launcher.exe image - protection is restored - FlushInstructionCache is called - execution resumes at the original translated RIP This fixed earlier stale mirror problems. 3. Dynamic tag windows and remapping We hit a collision where one requested tagged mirror window overlapped shell32.dll. We implemented requested_tag_base -> actual_mirror_base remapping. That part now works. We also added generic irregular tag-window discovery. Table pollution from false stack candidates was fixed. 4. Ret-chain dispatcher frames We found that some protected exceptions are not normal crashes but dispatcher transitions. One important frame format was: normal frame: [rsp+0x28] = target [rsp+0x30] = ret gadget [rsp+0x38] = context flagged frame: [rsp+0x28] = 1 [rsp+0x30] = target [rsp+0x38] = ret gadget [rsp+0x40] = context The flagged ret-chain variant successfully consumed a previous C0000096 privileged-instruction dispatcher event and the launcher continued farther. 5. Current blocker The current blocker is no longer BCryptCreateHash, C0000096, stale mirror bytes, shell32 collision, or unknown tag base. The current proven failfast path is: Launcher+0x1876D -> call qword ptr [rax+0x10] -> Launcher+0x18730 -> jmp Launcher+0x1E86DE -> jmp Launcher+0x18747 -> Launcher+0x410580 -> Launcher+0x410598 / int29 -> c0000409, FAST_FAIL_INVALID_ARG Launcher+0x410580 / +0x410598 is not the root cause. Runtime bytes show it is only a failfast helper: - it probes IsProcessorFeaturePresent(PF_FASTFAIL_AVAILABLE), ECX=0x17 - then calls int 29h with subcode 5 So the real issue is earlier. 6. Error object path At Launcher+0x1876D, RAX/RDX were initially actual mirror/remap pointers. We tested narrow normalization to live Launcher pointers: - RAX -> Launcher+0x4599B0 - RDX -> Launcher+0x45FA90 The normalization worked and the normalized pointers reached Launcher+0x18730. However, the launcher still reached the same FAST_FAIL_INVALID_ARG path. That means the root cause is probably not just pointer leakage at +0x1876D. The suspicious part is that the object being dispatched through [RAX+0x10] already appears to be an invalid/error object. The relevant data: - Launcher+0x4599B0 looks like a C++ error/exception dispatch table. It contains pointers to +0x18630, +0x18810, +0x18730 and text such as “bad array new length”. - Launcher+0x45FA90 looks like a static string/error block, including “Invalid presentation type for pointer”. So the current working hypothesis is: The wrong error/invalid object is selected or constructed earlier, likely between protected events 2112..2116, before Launcher+0x1876D. The next debugging goal is to find the first point where: - RDX becomes Launcher+0x45FA90 or its mirror/remap equivalent - RAX becomes Launcher+0x4599B0 or its mirror/remap equivalent Questions for you: 1. Does ERROR_NO_TOKEN still look relevant to this path, or could it have been a stale/side-effect Win32 error from an earlier run? 2. Are you willing to look at my Launcher folder and current guard/mirror source/logs? 3. If yes, what exactly would you prefer? - clean Launcher folder only - Launcher.exe + SocialClubHelper versions/hashes - rgl_rip_guard.dll - rgl_tag_mirror.c - sanitized rgl_tag_mirror.log - sanitized launcher.log - WER event details - Ghidra decompile as address map only Important note: The Ghidra decompile is not reliable by itself because the launcher is protected/obfuscated and runtime-transformed. Runtime memory logs and live disassembly are more reliable than the static pseudocode. Also, I am not trying to use a cracked version. I own the game and I am trying to get my legitimate Steam/Rockstar copy running on Windows 8.1 with Second System 3.1 Extended Kernel. If you are willing to look at it, I can prepare a sanitized archive with the exact files and logs.

win32

So you compiled another DLL that hooks the launcher to deal with the bad pointers that it eventually calls (like one that ends in 1DD3)? And yes, I would like to see that DLL, the launcher.log and possibly the last win32 error on that fail fast exception.

Jaguarek62

Yes, I compiled another diagnostic DLL: rgl_rip_guard.dll. It is a user-mode helper loaded early into Launcher.exe. It does not patch Rockstar/Social Club logic, skip auth/entitlement, skip AVs, or bypass failfast. It mainly handles/logs tagged mirror execution: lazy page sync from live Launcher.exe, dynamic tag-window remapping, guarded dispatcher-frame continuation tests, and failfast/caller diagnostics. The package contains the DLL, PDB, source, build command/log, current launcher.log/crash logs, the full current rgl_tag_mirror.log, and a separate failfast-reference run. For the failfast path itself I do not have a meaningful GetLastError value. The actionable failfast detail is RCX=5 / FAST_FAIL_INVALID_ARG at Launcher+0x410598. The earlier ERROR_NO_TOKEN note belongs to the older BCryptCreateHash path, not this int29 failfast.

2LENSFLARE4U

I'm having this issue where steam doesn't fully close when trying to shut down or restart my windows 7 machine. I get the dialog saying it's trying to close steam but it just hangs there. Even when hitting force restart or shutdown. I hit cancel on the dialog to close it myself and then steam doesn't respond, once steam is fully closed then the pc restarts normally. On the last compatible version of steam on vanilla Windows 7, this doesn't happen. Only on the latest version. Is this secondsystem related?

win32

It does appear to be related, however it will take some time to find a resolution to this issue.

win32

I have now determined the likely cause of the RGL issue. A global variable in launcher.exe (at RVA offset 0x70AF40 in the version with a digital signature dated June 17 2026) determines whether valid code paths related to DXGI are executed. It is set to 0x80000005 on Windows 10 and it creates a DXGI factory. On Windows 8.1 it remains set to 0 and it results in very different code paths being taken, leading to the relevant code not being copied to certain areas, ultimately resulting in the hashes of those code sections not being equivalent to the expected hash value and then crashing. I still have not finished figuring out the triggering condition for setting the variable.

ben0258

I've still been trying to run Parsec on Windows Vista, it seems to be happy putting 8.1 DLLs that it requires next to the executable, but I'm stuck on "The procedure entry point pSetupDiEnumSelectedDrivers could not be located in SETUPAPI.dll." even though the correct version should be there. Windows Vista Ultimate x64, ThinkPad X220, Intel Core i5-2520M.

win32

Most of my attention has been diverted to the Rockstar Games Launcher obfuscation recently, but I did also resolve every dependency in Parsec and its service. I briefly got the logon window to attempt to open (but could not on that system as it needs D3D feature level 11_0) but after further changes to the wtsapi32 wrapper, it sits in purgatory. Windows 7 DLLs will be sufficient, as it does run on 7.

ben0258

I wonder, is there any other way to use Parsec on Vista? In the meantime, I have gotten the Remote Desktop Connection client from Windows 8.1 and 10 to work on Vista, and are currently evaluating it as a Parsec alternative.

Navi Tentan

Tried to install discord but gave me the following: "System.ComponentModel.Win32Exception (0x80004005): The specified executable is not a valid application for this OS platform."

win32

Which Windows version are you running?

win32

You may have to right-click DiscordSetup.exe and open "Second System Config", and run the application using the "DefaultConfig" profile.

Navi Tentan

That doesn't appear even though I ran InstallSecondSystem as an Administrator.

win32

That Second System Config right-click menu entry should be written to the registry under HKCU\Software\Classes\exefile\open\Second System Config. I've never heard of it failing to write the key before.

Navi Tentan

exefile doesn't appear for me

win32

Is HKCU\Software\LSC\SecondSystem even there? If you run the installer again, do you see any errors jump out at you?

Navi Tentan

Yep, it's not there and the installer doesn't give me an error.

Jaguarek62

Hi, uhm.. I bypassed it, but I have no idea what to do now. I broke their DRM you could technically remove anticheat etc.. how to handle this properly? I can share the files in DM. Your dxgi clue was amazingly helpful

Jaguarek62

Also this bypass is INSANE. AI stopped helping in the middle, cause it detected arxan. It works only on AMD ryzen cpus at this point, but you could theoretically debug second system to behave the expected way. You are fighting arxan at cpu instruction level, that’s why it likely works only on my system atm.

win32

Yes, send anything that way.

Rocky

Happy to report that discord installs and works perfectly on this version on windows 7

Navi Tentan

Sorry for starting a new thread but I managed to install it properly and get discord + firefox working. However I keep running into "The install operation failed" for Visual Studio 2026 Community.

win32

There appears to be an issue with the XML parser in .NET making the VS installer fail. I am working to resolve it.

win32

I don't have access to view the image link. But if it is error 107, then it is non-fatal.

win32

I am working on some changes to winhttp and schannel, which may help.

2LENSFLARE4U

What exactly does second system bring to windows 10? Will I be able to install modern apps on say Windows 10 1507?

win32

The Windows 10 version was originally created for testing purposes only, but there are some applications such as Discord that have dropped 1507, in addition to the Qt 6.10+ applications that now need 1903, so it should allow them to run.

win32

I have added setupapi and wtsapi32 wrappers to the Vista Second System definitions, but you will still need newdev.dll and devrtl.dll from Windows 7 in the application folder.

win32

I have now updated the DLL definitions to fix the bcryptprimitives issue and added a new service-level wrapper for dnsapi.dll. However, to run the AmneziaVPN service, you will need to use Second System's sespatch.exe to patch AmneziaVPN-service.exe, Qt6Core.dll, Qt6Network.dll and amnezia_xray.dll. These binaries can be dragged and dropped on sespatch.exe.

Артем Пономарев

qBittorrent is working propetly now, thank you. But Amnezia, despite the fact that the service has started and seems to be working, refuses to connect.

Артем Пономарев

I solved the connection problem by patching tunnel.dll using sespatch.exe and running AmneziaVPN.exe with Win7 config. I hope it'll work stably

ashbee

Hey there! does second system work for windows 8.1 embedded or does it need to be a pro/core version?

win32

Any SKU should work.

ashbee

Alright, thanks for your response!

Артем Пономарев

Can Cubase 14 be fixed on Windows 7? It falls on graphics2d.dll with error 0x80004001 From windbg: ExceptionAddress: 0000000000446194 (graphics2d!createGraphicsFactory+0x000000000002dc64) ExceptionCode: c0000005 (Access violation) ExceptionFlags: 00000000 NumberParameters: 2 Parameter[0]: 0000000000000001 Parameter[1]: 0000000000000000 Attempt to write to address 0000000000000000

FL

Okay not sure if I am doing something wrong or what. But this release seemingly just does not work for me (I am on Vista) Going to 3.0 works fine, but this minor release has nothing launch that 3.0 does (For example steam never starts, Vesktop doesn't work, those are the only 2 things I did try as of now.)

win32

I have now updated 3.01 for Vista. Looks like there were some problems with the wtsapi32 wrapper.

win32

It looks like it is a DXGI issue. I am currently making various major adjustments to DXGI and I will try when I am able to return to a win7 system.

FL

You sure you updated the package correctly? I redownloaded and while the archive file says 20260707, the folder inside is still the original 20260703, and it acts just like the version of 3.01 I described where nothing really works.

FL

Never mind, just needed to restart again.

win32

It appears that I had not updated some directory names, but I did add p_wtsapi.dll to the x64 DLL directory. Until users started to request Parsec on Vista, this wrapper was only used on Windows 7.

FL

I have found a issue with the update version still. Everything else works, but steamwebui just doesn't start. I tried with -cef, -cef-disable-gpu, tried without launch options, on Vista it does not come up. At most the logging on screen can come up, but once it gets to the main screen, nothing. I know you are probably working hard on the rest of second system so I don't think it needs to be the most major of priority (Plus I don't really need 3.0.1 since it doesn't make some of the stuff that doesn't work function so 3.0 is fine for me for now.)

Einar

Hi there, Firefox 152.0.5 won't install properly on Windows 7 Ultimate 64-Bit, it stopps installing at ca. 80-90 %.

Templayer

I had to uninstall Second System yesterday. (SecondSystem301_Win7_20260617) At first, about a week ago, my JDownloader 2 stopped being able to use FFMPEG to combine audio and video DASH streams from Youtube. Since it occurred after an update of JD, I was sure it was its fault (the youtube plugin breaking is rather frequent, as Google really wants you to watch the ads!), but nobody else reported it even a week after, which had me suspicious. Well, yesterday, I have cleaned my PC after turning it off, turned it on, and... about 80+% of applications were working correctly, but the others had strange behavior. At that point, I thought I have moved a cable or something, and one of my many drives was hiccuping, but no. JDownloader no longer launched with an error of "Cannot initialize main class" (paraphrasing). Further testing revealed that all applications using VxKex NEXT (which I haven't updated in a long time - most certainly way before I even installed the Second System) would crash on launch, without any GUI, just a process appears, and then it disappears. Not only were their log files timestamped with the last time I successfully launched them, the logs weren't metadata touched upon these failed launches, keeping their last changed date to the datetime before the restart. Namely, qbittorrent and HandBrake. Some others were broken as well. The system log had NOTHING WRONG with anything, there was NOTHING regarding VxKex NEXT or other non-working applications, which has raised my brow. Some very complex and fragile software, however, did work correctly, like Magix Vegas 16. I returned back to JDownloader, and I found out that while the wrapped exe version wrote the main class initialization error, double-clicking the jar file worked perfectly. (it is not under VxKex, as most of my apps are not) C# / .NET applications were broken. Like TwitchDownloader. They gave out a classic "the program stopped working" error screen, but again, nothing in the logs. Godot has also failed with this generic message. (again, not using VxKex for that, as I am using an old version that is still compatible with win7) I was surprised that Stacher, a rather fragile application (at least under Win7), DID work. What finally got me a clue what was going on were my recording applications. My main recording application, Mirillis Action, told me that it cannot run under Safe Mode (!!), which was really, REALLY worrying since I wasn't in safe mode. xD My fallback recording application, Bandicam, however, told me that it cannot run with a kernel debugger active (!!!). I rechecked, and all of these applications had Second System disabled as their profile (or whatever that is called, obviously, I cannot see it anymore after uninstalling it). I also went to take a look at system variables, specifically the PATH variable, and the folders with Second System were before everything else, I assume that is intentional? Restarting didn't help. I have uninstalled Second System and lo & behold, everything, EVERYTHING started working again. I do not know why these problems didn't happen sooner, it makes no sense. Back when I have installed Second System (which feels like months ago), I made sure to restart afterwards, and then I painstakingly went through each, and every application to make sure nothing broke. To make sure this exact situation wouldn't happen later, when I had no clue what was going on. And yet, here we are. Unless something modified the PATH system variable wrong, I have no clue as to what triggered this, or why it didn't occur sooner. My current theory is that even after restarting after installing it, the Second System didn't start working right away - that would explain why Second System worked on nothing, as I have tried many Matrix clients with it (via the context menu, and setting a profile, testing all profiles), and even you were unsure why there are some early nullpointers in those (and unreadable stuff in the stack). Why it suddenly "properly activated", however, I am unsure of. If it even was a proper "activation" (for a lack of a better term) at all. While browsers DID work (r3dfox), I had to order my lunch through my Linux nucbox server, because I was scared of what else could be happening at that point. I do not own a smartphone. :ú Here's a list of affected applications (I haven't tried all of my applications, though), might be worth a check: HandBrake (doesn't launch, no error anywhere, VxKex NEXT) qbittorrent (doesn't launch, no error anywhere, VxKex NEXT) JDownloader 2 (wrapper only, cannot instantiate main class error) Godot (generic "The application has stopped..." error, version 4.3 32-bit) TwitchDownloader (generic "The application has stopped..." error, VxKex NEXT I think) AltDrag (it appeared in the systray, but its mode could not be changed. Actually, it couldn't be clicked on at all. This was the first thing I have tested upon restarting after an uninstall, and it started working correctly) Mirillis Action (yes, I have a proper lifetime license, cannot run in Safe Mode error) Bandicam (System/Kernel debugger active error, refuses to launch further) Chatty (I am not sure if it worked or crashed with a generic error, it is a Java based app, and I only tried the exe wrapper. Or maybe it gave the same exact error as JDownloader? I don't remember. I was panicking at that point...) I had to postpone going for a walk with my senior dog for several hours, so I was very, VERY angry, yelling at my screen, etc. I am not willing to test this further than I have already. I guess I'll wait a few years until I try this software again. What really bothers me, however, is that it started giving me issues LONG after I proactively went through and tested everything for broken stuff (I do realize that this might not be your fault, however), as I DO NOT use sleep mode nor hibernation! It was a long time ago since I went and thoroughly tested everything! Arrgh! Originally, I had those hours scheduled so that I could make an offline backup of my crappy website (including the database and everything) at www.Templayer.cz, since the webhosting prices have gone up by more than 3x (and they want me to pay even more than the current 3k CZK webhosting price for a certificate, YUCK), and make my own hosting instead (I already have a server that I am running things like a Matrix server, and a TURN server for end-to-end encrypted calls / videosharing, Chat Control my ***). Not much time left to do that, so having to solve yet another seemingly randomly occurring problem (I mean... why now? Why not before? Why not after? WHY RIGHT NOW?!) was frustrating. :( :(