Haskell Tooling Infrastructure

Now that I am finally turning back to HaRe, the question of infrastructure again comes up. Basically I woudl like to focus in the tool itself, but it can only be useful if it can deal with real world environments. This means properly managing all the different project and compiler environments, and being integrated into IDEs. The current version uses ghc-mod for the lower level interface, and a simple command line interface for the IDE integration. But I have a definite sense that as a community every tool maker is solving the same problems over and over, from ghc-mod to ide-backend to ghci-ng, and onwards. To start a discussion around some kind of architeccture that tool writers can develop agains, ide's can integrat against and ghc/cabal can eventually internalise, I have put some basic points down at https://github.com/RefactoringTools/HaRe/wiki/Requirements-for-IDE-GHC-inter... Regards Alan

"Alan & Kim Zimmerman"
Now that I am finally turning back to HaRe, the question of infrastructure again comes up.
Basically I woudl like to focus in the tool itself, but it can only be useful if it can deal with real world environments.
This means properly managing all the different project and compiler environments, and being integrated into IDEs.
The current version uses ghc-mod for the lower level interface, and a simple command line interface for the IDE integration.
But I have a definite sense that as a community every tool maker is solving the same problems over and over, from ghc-mod to ide-backend to ghci-ng, and onwards.
Filed issues for the three projects: https://github.com/kazu-yamamoto/ghc-mod/issues/494 https://github.com/fpco/ide-backend/issues/287 https://github.com/chrisdone/ghci-ng/issues/20 -- respectfully, Косырев Серёга

As a follow-up to this[1], I have written a post[2] detailing how ghc-mod
provides the downwards (BIOS) functions for the next version of HaRe[3].
[1]
https://github.com/RefactoringTools/HaRe/wiki/Requirements-for-IDE-GHC-inter...
[2]
http://alanz.github.io/haskell%20refactorer/2015/10/02/ghc-mod-for-tooling/
[3] https://github.com/alanz/HaRe/tree/wip
Alan
On Sun, Jun 7, 2015 at 2:02 PM, Alan & Kim Zimmerman
Now that I am finally turning back to HaRe, the question of infrastructure again comes up.
Basically I woudl like to focus in the tool itself, but it can only be useful if it can deal with real world environments.
This means properly managing all the different project and compiler environments, and being integrated into IDEs.
The current version uses ghc-mod for the lower level interface, and a simple command line interface for the IDE integration.
But I have a definite sense that as a community every tool maker is solving the same problems over and over, from ghc-mod to ide-backend to ghci-ng, and onwards.
To start a discussion around some kind of architeccture that tool writers can develop agains, ide's can integrat against and ghc/cabal can eventually internalise, I have put some basic points down at
https://github.com/RefactoringTools/HaRe/wiki/Requirements-for-IDE-GHC-inter...
Regards Alan
participants (2)
-
Alan & Kim Zimmerman
-
Kosyrev Serge