This happens every freaking time something "new" is released (whether its fake by EP or full major new release), and over and over and over again....
Yes, but custom builds don't have to be 19045 or 26340, they can be 19044 or 26100 just like the real thing. Hackjobs can be identical to the real thing, the same way uupdumps is. Both are still frankenbuilds cause they're unofficial and not built or approved by Microsoft. The only non-frankenbuilds are the ones officially released by Microsoft, which neither hackjobs or builds by uupdump are. Everything that's not made by MIcrosoft is a frankenbuild, regardless of the compatibility with updates or how close is it to the official thing.
This was the same with the original stuff (https://forums.mydigitallife.net/posts/1890049), who knows why MSFT made it possible, the EP's didn't get offered on WU and again, this is not the LTSC thread so it is offtopic by default. I asked for some moderation.
Please stop talking about constructed/reconstructed builds here. Stay in you topic for that. The difference between UUPDump created builds and the reconstructed ones is that the former follow official build rules, they are SKUs that actually exist. (Re)construction creates build that don't follow these rules and these SKUs don't officially exist. So, all bets are off when it comes to update them.
Re(constructions) can use official SKUs, and can be updatable. Xinso and others incl me constantly makes 19044 LTSC/1607 LTSB and Server even Pro/Home 19045 versions that are updatable just like the real thing. Your info that re(constructions) must use SKUs that don't exist officially / cannot be update officially is simply not true and is misinformation. Have a nice day.
Windows 11 Insider Experimental build 26340.9233 reproducibly crashes on sign-out/restart with SYSTEM_SERVICE_EXCEPTION 0x3B. WinDbg bucket: AV_win32kfull!SetMagnificationInputTransform Faulting function: win32kfull!SetMagnificationInputTransform+0x39 Failure hash: {527223d9-4b8b-e502-e7e6-16fc4649fd38} The crash occurs during desktop teardown: DestroyDesktop -> MagpDecomposeDesktop -> MagpRevokeInputTransfrom -> CancelMagnificationInputTransform -> SetMagnificationInputTransform Workaround found and reproduced on two physical machines: vivetool /enable /id:27829265 vivetool /enable /id:61457898 Feature 61457898 is FindMyMouse_Prototype; 27829265 is the prerequisite mouse pointer/crosshair feature. Important: the first restart after enabling 61457898 still crashes, because the current session was apparently created with the old state. After that reboot, subsequent sign-outs and restarts no longer crash. A/B confirmation: vivetool /reset /id:61457898 -> crash returns on sign-out/restart. vivetool /enable /id:61457898 -> after one crashing reboot, subsequent sign-outs/restarts are stable. This has been reproduced on two separate physical machines. Some other physical headless machines do not reproduce the crash even with 61457898 undefined, so the feature appears to be a workaround rather than the sole trigger.
The last few test releases of windows crash with me on VirtualBox 7.2.4 at the same point. But restarting the VM continues successfully.
Isn't that just a screenshot of what you fully posted here? Or do you want to show us it is on another place too? What do you want to tell me with this?
The message I posted here was submitted to the Feedback Hub and acknowledged by Microsoft. I also posted it on ElevenForum, and there's confirmation of the solution there. (It's important to note that this only happens on physical machines, not virtual machines.)
Not going to waste time to install this level of test OS on a real machine, they are releasing way to much crap these days, i keep it to the publishing of the ISOs .
Even the developers at Microsoft can't keep track of all of those. No human could. After Skynet kills all of the humans, then life will be much simpler.
Not sure what this is all about but I'm finding this build feels way better than 25H2, Feels like 26H1. No problems at all, in fact it's less buggy than 25H2 "I'm having a WinME to Win2000 moment" lol
There is s few issues, one being dns over https errors at logon. I also found there was no certificate bound to 443, i fixed that but it doesn't stop the errors at logon, could be trying to obtain before its ready as everything works fine.