Haskell Process Type should "do the right thing"

[alexanderkjeldaas]

I don't think it makes sense that one has to fiddle with elisp per project in order to set the Haskell Process Type for a given project.

Selecting between cabal-dev, cabal-ghci and ghci should be done using some smarts. For example cabal-ghci should override ghci if a cabal sandbox is detected.

The default setting should thus be 'auto' and the auto-function should "do the right thing".

[alexanderkjeldaas]

Sorry, cabal-repl should override ghci. Goes to show my confusion and the need for an 'auto' setting ;-)

[hvr]

#198 (comment) :-)

[purcell]

This is now done, with the following logic: when haskell-process-type is set to 'auto (the new default), then the presence of a *.cabal file or a .cabal-sandbox directory in a parent directory will cause cabal-repl to be used. Otherwise, ghci will be used. This is in accordance with the suggestions in #198.

[alexanderkjeldaas]

locate-dominating-file (lambda (f) .. doesn't seem to do the right thing for me.

[purcell]

@alexanderkjeldaas Thanks -- a stupid mistake on my part. The lambda is passed the directory, not the filename, so of course the code I committed was not going to work properly. I've pushed a fix.