[Response received] Explorer (as shell) does not start on boot - 22000.527

[ArthurWolfhound]

I've just turned on my laptop after receiving an update, I think it was a cumulative patch. And I was greeted with fully black screen.

It took me a moment to understand that it is just explorer failing to start. And then I see a notification from EP about symbols for build being downloaded and me needing to restart explorer. I mean, users of EP mostly would know about Ctrl+Shift+Esc (though, I haven't tried Ctrl+Alt+Del, to be fair), but some help would be nice.

Is there a way to implement some kind of warning system, that if explorer does not start, there's ExplorerPatcher popping up and providing help with options: Try to restart explorer, Check for updates, Temporarily disable EP and restart original explorer.

[valinet]

Notifications don’t work when Explorer is not running, so maybe something slightly different happened…?

[ArthurWolfhound]

No, I just started explorer manually via Task Manager and it just, went on running. Will restart my laptop now to see if it fails again.
Question - do you host "symbols" elsewhere? Because it is strange that it said something akin to loaded symbols for the build 22000.527, but that does not match even your "highest" pre-release version?

but I can’t work on them either without any means to reproduce, crash dumps, image of the problematic OS install etc.
I can understand that

[ArthurWolfhound]

Restarted the laptop. Nothing. Just works as intended.

Sorry for all of that. You're right - should've checked event viewer before writing here. Apparently dxgi.dll faulted and explorer did die.
Weirdly enough - There's only one error message in the event viewer and I seem to be unable to get any of the files referenced in the event viewer except one, others are already missing.

explorercrash.zip have a look if you'll have a spare minute for that, that's all that I managed to salvage from event viewer.

[valinet]

So why is the title labelled "Not EPs fault" if Faulting module path: C:\Windows\dxgi.dll which is exactly ExplorerPatcher...? It was indeed an EP problem.

Thanks for the Event Viewer output. It lead me to an area in the code which could produce a fault under certain circumstances, indeed. I have now fixed that. I will push an update soon. Again, thanks very much for that, hints like these help a lot in identifying and diagnosing problems. The nature of this application is prone to potentially having small mistakes like that in the code, that's why extensive testing is always welcome.

[ArthurWolfhound]

Haha, I seem to be all over the place when communicating with people. Glad to help!

[valinet]

I have written a tutorial on the matter, to help users more easily diagnose these types of faults: #881