Why can't the updater use a "standard" temp directory (c:\windows\temp) that everything else in windows uses?
Because the files aren't guaranteed to be there after the service stops (thus requiring redownload, extracting, and patching everytime the files are deleted).
Or create its own?
Well, that's what we try to do. We're using a "system" folder (the "userprofile" folder) and we create our own directory inside this folder. There are only so many places we can put the files that we can guarantee that the system (or 3rd party tools) won't delete the files between multiple runs.
Seems like a waste of time to me when it is such a simple thing that is broken in your software.
I'm sorry if this seems like a rant, it is not intended as such but simply an accurate description of the reality of what you guys are building and the lack of robustness that seems to have crept into the code.
I fully agree. Bugs like this embarrass us to no end. We're trying to make something that doesn't have any "mission critical" bugs like this.
Is there any chance we can remotely debug this problem on one of the computers that experienced (or is experiencing) this bug? This way we can speed up the bug fixing process?