Monthly Downloads: 20
Programming language: Haskell
License: BSD 3-clause "New" or "Revised" License
Tags: Distribution     DevOps    
Latest version: v0.3.5

dynamic-cabal alternatives and similar packages

Based on the "Distribution" category.
Alternatively, view dynamic-cabal alternatives based on common mentions on social networks and blogs.

Do you think we are missing an alternative of dynamic-cabal or a related project?

Add another 'Distribution' Package



Build Status

Note: The problem this library solves no longer exists starting with GHC 7.10, since versions later than that do not depend on the Cabal library. You only need this library if you like to support older GHC versions.

If you've ever used Cabal together with the GHC-API, you know the problem. Because GHC depends on a version of Cabal, which is often outdated, there is no way to parse the setup-config file generated by newer cabal versions. This library attemps to solve the problem by dynamically generating code that performs the action you want, and then compiling and loading that with GHC. With this method, you don't need to depend on Cabal at compile time and so you can use any version of Cabal.


Currently, the library only allows two queries: Getting the targets (along with their dependencies, ghc options, etc) and the package databases. The first is easily achieved using the targets query provided by the library. To run the query, you can use the runQuery function, which takes the path to the setup-config file as an argument. For example, the following little program prints out the names of all test suites, when run in a configured cabal project root directory:

import Distribution.Client.Dynamic

main :: IO ()
main = do
  tgs <- runQuery (on localPkgDesc targets) "dist/setup-config"
  mapM_ putStrLn [ n | TestSuite n <- map name tgs ]

Because targets works on a PackageDescription, on localPkgDesc is used to get the current PackageDescription.


At the moment, I only implemented the functions I need myself. If you have more functions you want to implement, just send a pull request or open an issue.