AstragonQC

It isn't abandoned, just Win32 have changed the main goals of the project. I already posted a comment here earlier explaining that he wants the EK to be more focused on making drivers/advanced things to work rather than just running programs, as it's a better deal to develop alternatives of popular programs like Chromium by making a version specifically dedicated to older platforms rather than backporting whole chunks of system functions to make the original program to work, that are sometimes requiring indecent amounts of work for something that can potentially never came out or be completely unstable. That's primarily why he started developing Supermium: backporting the memory sharing API from Windows 10 to make Chrome 112+ to run simply doesn't worth the hundreds of hours of research/development he could have to invest in, in order to make something kinda "operational" and which "is not conflicting with the rest of the system". It's cheaper and quicker to create Supermium, which is perfectly adapted to run on the OS you want it to run into rather than making the big popular software to "run" in a semi-unstable way because your implementation of the said backported API is flawed and needs several dozen of extra hours to completely fix and finally having something working. But here also I said that once he would have done a such job, the application may have been updated and now requires another set of APIs that must be also backported to finally have that program to work! This is an endless cycle. For the moment Win32's priorities are elsewhere, but I can ensure you that the day he would work on an update for this project. (Source: trust me bro but it'll happen for sure) -- <OT::GrainOfSalt> Also a parallel can be made at a certain degree with VxKex, because there wasn't any releases between August 2022 and February 2024. Many people said this project was dead, but one day the long rewrite we all waited for finally came out, after 1.5 year. Just remember that those individuals are doing that generally on their spare time, some also does it completely for free and with good faith, and thus, owe you nothing in any way, except eventually donations for the many headaches they got while debugging "why that a*s-programmed application isn't running while another one, based on the same framework, runs flawlessly?!". Anyway, for a such projects I can't complaint because they do a job a very few people accept to do, and generally without requesting anything back. Yes, the time gap between two releases can be long, but hey, they have a life and other projects, and like I said, they do all their work voluntarily, wanting nothing back from you except maybe testing and your support.</OT>

It's more the installer, which is a self-extracting archive that runs the installer once it's done that is flagged, not the VxKex files themselves. My AV (BitDefender) never complained about the files stored in Program Files after installation is done. It does not. It's written on the project README that VxKex does not modify any of your system files.

I already mentioned it works with VxKex, so on that side it's fine

GIMP 3.0 will ultimately not be available on Windows 7 (at least out of the box; the latest v2.99.18 released last month now requires Windows 10, because Python 3.11). For the Installer you'll need to spoof OS version as Windows 10. BUT I noticed that the main app itself works without any workaround, just you need to have VxKex enabled (version spoof is not needed) to prevent the crash of one dependency and to stop Python from throwing dozens of "cannot locate xxxxxxx.dll" errors. I tested how GIMP runs and everything works, even the non-destructive editing. A more deep test is needed tho, to ensure everything is working as expected. Images: https://imgur.com/a/sBsauRD

At the general surprise, a new version of VxKex just dropped today! (No joke, I was opening GitHub while drinking water, I read the release and I almost chocked!)

Seems that VS 2022 17.9 cannot be installed without workarounds on Windows 7 due to Edge WebView2 install that fails because MS updated it from 105 to 119 https://imgur.com/a/eEeYlfC

Search on the Catalog, the updates are there https://www.catalog.update.microsoft.com/Search.aspx?q=2024-02 Windows Server 2008

Not totally, you can still come across him on Discord from time to time. The atmosphere is healthier there, and after all, well... sorry guys, but someone have to opens it (and nobody wants to, so I'll take the job and wear the hat of "the bad one" that ruins the mood) and seriously say it: some conflicts we can read on that forum are more similar to quarrels between schoolchildren than real dialogs between responsible adults, so I think that win32 is somewhat right to avoid such places, focus on his work and dialoging with others on more peaceful/serious-looking* places. We're mature people (at least, I hope), so I think we all can discuss in a better atmosphere, and express their disagreements as adults with adult-like wordings. Nevertheless, on a in-topic and more positive note: A message he posted on Discord a few hours ago mentions the beginning of early experiments with Windows 2000 while he's waiting for people to try to replicate the remaining freezes in Windows 7 in order to fix them. * DISCLAIMER: I do not put in cause the serious of some people here, don't make me say what I didn't say. I respect the work some may have and will done here. Also I don't focus anyone, after all we're all responsible about what we say on the Internet.

I verified how long links from this "private-user-images" GitHub subdomain stays active (for absolutely no reason other than curiosity), and I found that a link lasts only 5 minutes before it expires... So, put links to images by this method is kinda pointless, it's better to directly put the link of the comment or the issue the image is in, or upload it elsewhere then post it here.

Image links on GitHub expires kinda quickly. Btw, the image is probably from this comment : https://github.com/win32ss/supermium/issues/150#issuecomment-1881566029

Some people have bet on the hypothetical "Premium Assurance" that MS sold to companies between 2017 and 2018 before being replaced with ESUs that stated "6 more years of updates after the original EoL date" and Windows Server 2008/R2 was mentioned in it. Also when MS replaced PA with ESUs, it was stated in a document (Page 5) that MS will honor the terms of PA for companies that have bough it, which "proves" that some companies may actually have bought those 6 additional years of support. Some people are confident about the fact at least one company took the Premium Assurance when it was available (as it was a "buy early, pay less" model as this document proves it in Page 4), but they're unsure on how MS will distribute the update packages. The most probable is that MS will just continue to publish them through the MSU Catalog with another layer of artificial locks to "prevent" installing it on systems that are not "targeted" to receive them. Those 6 years starts upon EoL date (so January 14, 2020), which means the actual deadline is January 13, 2026 (which is mentioned in this slide, at slide 11). (If not counting some institutions like the army that will (without any doubt) buy additional support to Microsoft in order to maintain some systems until their decommission because XP is not the only in this case, 7 will probably have a such treatment as well, and also 10 in the future). -- So to correct you on your supposition, the EoL schedule is more something like this (in case we can access Server 2008/R2 updates packages after January): October 2024: Windows 7 x86 (with the EoL of POSReady 7 as Server 2008 R2 is only available as x64). January 2026: Windows Vista, Windows 7 x64. October 2026: Windows 8 & 8.1. (Or maybe *Plot twist* some companies will request more updates to MS, but that's more going into a crack theory rather than a probable thing). -- (Oh and the funnier is on the slide I mentioned earlier is at slide 21 where they just copy/pasted the text because it's written that Svr2012/R2 to EOL in January 2023, which is not the case. But it also describes something more interesting: MS also considered to provide this Premium Assurance plan on Server 2012/R2 and get updates until (October) 2029). The NT6 OSes never-ending lifespans are just hilarious. And Vista will outpass XP as the longest-supported OS with slightly more than 19 years worth of updates (December 2006-January 2026).

Iirc on a Supermium issue report, he mentioned that the "Extended Kernel's scope has been greatly reduced", and next releases would more focused on Kernel functions to make drivers working rather than making programs running on it. That's also why Supermium exists : it's more easy to make a fork of that program that can run on the target OS rather than trying to backport syscalls that would require hundreds of work hours to achieve. In another issue report, he explained that making modern versions of Chrome working "natively" on Vista ExK would require to backport the shared memory API from Windows 10 1803/09 (if not making some weird things to replicate this behavior, with a significant risk of breaking everything during the process), which would imply a lot of work for not that much at the end. As soon he would have finished, Chrome probably received one or more updates and if we run out of "luck powder", it will need to backport even more syscalls. At this point, some would probably say "just run the other OS instead". A perpetual circle that would never ends. From my part, I consider this choice more appropriate, especially towards people that don't want to modify their system (from a side they have legitimate reasons to refuse to install untrusted system binaries) but still want to have some quite recent programs, especially for the browser. (it's more easy to make someone accept installing a standalone program specifically made to run on their OS rather than trying to convincing them to install a "Kind of Kernel Extender they don't understand anything about, with the risk of a problem that can happen during or after the installation and break their OS"). So to resume quickly, no, Vista Extended Kernel isn't dead, just the main scope of this project has changed to something more technical and kinda """advanced users friendly""". (This is probably badly explained but that's how I understand the current state of thing)

I'm doing a big edit on the message I posted in August, it seems that VS2022 17.7.x runs on W7. VS Installer is blocking the install but if you select "Download all, then install", VSI will show the popup when the download finishes, but if you click on Cancel, it will install VS and all component you selected as if nothing happened. Surprisingly, no modifications are required after that, VS just works as previous versions does. It also work for updating an existing install, by doing "More" > "Download, then update" So here again, another case of artificial limitations... Not tested for the current 17.8 pre-release, but it'll probably also work. EDIT : 17.8.0 works

Of course, here it is: Also as this thread is more dedicated to this specific section, so I have a little suggestion for the tray : being able (if possible) to change the gap between the clock and the date, as this is a detail that make me crazy me when I use the Windows 7 style, as I'm a bit of a perfectionist. I'm explaining this by the fact the gap is 8 pixels on W7, 8 and 8.1, but 9 pixels on W10/Legacy taskbar (yes, I see the difference, and it triggers me, I always want to... AAAARH! Remove this extra pixel to fix this thing and finally die peacefully). This image I made in January can explain what I have in mind if you want a more "imaged" context: (Click to open at full size) I doubt this is doable but if it is, it would be a nice touch, the minimum is just this extra gap pixel, if it can be completely adjusted by anyone with a setting, it would be even better.

It is funny to see that the first image showed in that video is a Windows XP with a Vista-looking theme.

Hi, am I the only one to have my FxNightly install got automatically updated to a newer build despite it shouldn't be possible? I mean, I kept my FxN on the v117 July 18 build (as I still use it for some things), and when I wanted to open the browser this morning I noticed it self-updated to a v119 build (2023-09-06) (which obviously doesn't run due to missing DLL entries). So I manually downgraded my install to the July 13 build, and as soon I restarted, the browser immediately started to download and install a newer build (2023-09-12), definitely proving the update ability is not blocked anymore... I put the policy file to completely block updates but now I'm wondering why (and especially how) it happened, as it never behaved like this during the two past months (It just showed the orange triangle like on my FxDE install) (Maybe it happened because it was a v117 build and not a v115 one. Or there's something else I didn't understand yet)

Visual Studio 2022 v17.7 seems to be incompatible with Windows 7 as VSI refuses to update my installation from 17.6.5 to 17.7.0

The latest Canary build completely broke the system tray. I suspect the fact the latest build (25921) cleared some things from explorer.exe so I think MS is probably going to remove the old taskbar completely in the next builds.

I already had this issue by the past, and I somewhat fixed it by installing the version 472.12 from September 2021, then manually updating to 474.11 (the last WHQL certified version). I can't really certify it will work for you, but for me It solved the issues I have with the driver installer.

UPDATE ! Mozilla pushed an update this afternoon, and it was for the move to ESR.

For the moment, the only thing to modify in order to make Firefox runnable again is the required subsystem version strings, which can be done using a tool like CFF Explorer. This is also a sign they will probably start to implement Win10+ functions into the browser. I also tested with the VxKex version spoofer and it doesn't work, so if you really need to run newer builds of Firefox, you will need to modify the binary by yourself.

The issue have been fixed in v113.0.1

Actually in this Qt forum thread, some people provided some Windows 7-compatible Qt6 DLLs. With these DLLs you can run some Qt6 application like OBS v28.x (v29 also requires a modification to OBS.dll to change IsWow64Process2 to IsWow64Process). So if your application is based on a version of Qt6 that have custom DLLs (6.2.2, 6.3.0, 6.3.1, 6.4.2 or 6.5.1), just try them. If not, as Qt is open source, maybe you can retrieve the source for the specific version your app uses and modify them create DLLs that works on W7.

I installed updates on my W7, seems to be okay after 1.5 hour. .NET updates also installed successfully.

That's wrong. No end of support date have been decided yet. Also please note that this is not because Chromium ends support for Win7 that Mozilla will immediately follow. As Firefox does not rely on a Chromium base, they can continue providing support as long they want. For example if they wants to continue supporting W7 until 2026, they can (I'm not saying they will, but they're capable to). As you can see, the latest Nightly build runs on W7. If Mz planned to end the support in February as you said, Nightly should have be blocked on v109, as v111 represents the release that will be released in March. This scenario happened on Chrome & consorts, they blocked W7 on v109 on Canary builds in December when 110.0.xxxx.xx builds arrived, because if you plan to stop support on a product, providing alpha builds of the next version on it is just senseless.