Oh this is great to know! I actually used this before when writing an arena allocator, since it seemed to be relevant and was already built into a system that was relying on WIN32 and HRESULT errors to begin with. I always had fun trying to find existing error codes among Windows' header files to use for other things.
I wonder if performing a system scan or file check has ever actually fixed any errors. Rebooting on the other hand, fixes basically any transient problem I encounter, even on non-Windows machines (a friend who has a Mac doesn't always believe me when I tell them to reboot to fix random unusual slowness/hangs, but they have only 8GB of memory and it has always worked so far!)
I will say though, non-Windows machines rarely need a reboot while Windows often should practically be rebooted daily.
Oh this is great to know! I actually used this before when writing an arena allocator, since it seemed to be relevant and was already built into a system that was relying on WIN32 and HRESULT errors to begin with. I always had fun trying to find existing error codes among Windows' header files to use for other things.
I wonder if performing a system scan or file check has ever actually fixed any errors. Rebooting on the other hand, fixes basically any transient problem I encounter, even on non-Windows machines (a friend who has a Mac doesn't always believe me when I tell them to reboot to fix random unusual slowness/hangs, but they have only 8GB of memory and it has always worked so far!)
I will say though, non-Windows machines rarely need a reboot while Windows often should practically be rebooted daily.
Since nearly every consumer machine uses non-ECC RAM it's probably best to just do a full shutdown at night and boot up the next day.
It reminds me of "bitsquatting" where you can get a lot of hits for domains 1 bit off really popular domains (separate from likely typos).
I doubt random bitflips are the source of most NT invariant violations. A reboot does fix them all the same though.