1. 4

    Even stranger for me is that 30% of bugs were filed by MinGW users. It’s hard to draw anything more from that figure alone (we don’t know how many unique users there are, for example), but it still strikes me as surprising.

    1. 2

      My normal Windows build with GUI and everything identifies as “i686-pc-mingw32.” It’s just build the way regardless if you’re using MinGW or not.

      1. 4

        Sorry, was a bit unclear there - it was more the fact that there are so many Windows users (well, bugs filed by Windows users). My naive assumption was that a minority of Emacs users are Windows users.

        1. 5

          Perhaps there simply are more bugs that occur on Windows? ;)

          1. 1

            That would probably make sense! My assumption is that there are more contributors contributing from a {u,Lin}nix machine than there are from Windows. This wouldn’t have anything to do with Windows itself, though.

          2. 1

            There were a lot of Unix developers forced to work on Windows for market reasons in the 90’s and such. So I’m guessing they took their text editor with them.

            I worked at all-Windows shops in game dev where at least one person used Emacs.

            The same dynamic applied to Python. I started out with Python on Windows when I worked in games. Proprietary game SDKs only work on Windows, at least back then. Even though like 95% of the Python devs use Unix, probably over 50% of users use it on Windows (as judged by their download stats). It was super useful on Windows, and probably the same goes for Emacs.

        2. 1

          Looks like it was actually closer to 10% (from a reply in the same thread):

          I’ve now done this with a bit more rigour. And graphs! Now each bug reporter email is only counted once per year. (The biggest effect of this is a drop in the mingw %.)

          See: http://debbugs.gnu.org/stats/emacs.html