ProjectTypes

A project type is how eproject determines what project a file belongs to. When a file is first visited, eproject tests each project type until it finds one that matches, and then treats that file as belonging to that type of project. The eproject-root and eproject-name are then determined. (Eproject doesn’t actually maintain “projects” anywhere, it is usually assumed that buffers with the same eproject-root are in the same project.)

To define a project type, you need to run some Lisp code. Generally you will do this in your .emacs, although at some point, project definitions will be included with eproject, and you can just load those.

Example definition:

(define-project-type haskell (generic) (look-for "Setup.lhs"))

Here, we define a haskell project type. We determine the project root by looking for a file called Setup.lhs. If we find one, we know we have a haskell project, and we know that the project root is the directory that contains Setup.lhs.

The (look-for ...) above is a magic function that starts with the visited file’s directory, and visits each parent until it finds the named file. The place where (look-for ...) is can be any Lisp form, though, so you can write a complex function to determine project membership, if you like.

The (generic) part is a list of “supertypes” to inherit attributes from. You can supply attributes as a property list (:key value pairs) after the supertypes. Eproject understands a few attributes internally (see InternalEprojectAttributes, but extensions that use eproject can use any key they want. The attributes can be values that are shared between all projects of the same type, or they can be functions that are evaluated in a project file buffer.

For more complex example see Eproject-for-Perl