Björn Gustavsson

> There is some strange failure in the doc build, that seems to be unrelated to my changes. Yes, it is unrelated to your changes.

Thanks for your pull request.

Thanks for your pull request. Please squash the commits.

Thanks! Added to our daily builds.

I have re-classified this issue as a feature request. Dialyzer uses `none()` as the bottom type in the sense described in "Computer science applications" mentioned in the [Wikipedia article](https://en.wikipedia.org/wiki/Bottom_type), not...

@jhogberg The code very explicitly returns `?nil` when given impossible contents: ``` t_list(?none) -> ?nil; t_list(?unit) -> ?nil; t_list(Contents) -> ?list(Contents, ?nil, ?unknown_qual). ``` Is that an oversight by the...

@kostis Yes, I quoted the wrong code. I quoted the code that I had updated. It would be useful for us if you could find time later to explain why...

The arguments given in this discussion thread have convinced that this behaviour is indeed a bug. The fix is in the linked pull request.

The return type of `foo()` is `none()`. The pull request does not change that. Dialyzer is able to distinguish the [the two differenct senses for the bottom type](https://en.wikipedia.org/wiki/Bottom_type). I initially...

I have updated the PR.