I've put up a web page that explains how I setup my Visual Studio settings. This includes downloads for .vsix extensions (the "Download" under "Erweiterungen installieren", "Indent Guides") or .config files (further down as "CodeMaid.config").
Internet Explorer and Edge are the only two browsers that deliberately alter the extensions of these download files into something that fails to work after download. Firefox simply does what the web server tells it and saves the downloaded file under the specified name.
I could handle the .vsix case by providing the MIME type "application/vsix" for it. Otherwise IE would have interpreted such files as .zip (which they technically are). They must not be renamed anyway because a .zip is opened totally differently than a .vsix file.
The .config file doesn't have a special MIME type though. It really is just an XML file and IE renames it to .xml. Unfortunately, it is meant to be used in a program that wants .config files.
Why is Internet Explorer breaking all kinds of things? How can it be stopped from doing such nonsense? I've found something about an HTTP header "X-Content-Type-Options: nosniff" for IE 8 but that doesn't change anything today so I suppose it's wrong or also broken. Or maybe I just need three more headers by now?