= Access to a Haskell package defined in a flake  (without using a `.cabal` file)

////
$ mkdir tempwork
$ cd tempwork
$ cp ../flake.nix .
$ git add flake.nix
$ nix develop
$ git add flake.lock
$ git commit -a -m temp
////

Occasionally you might want to run a short Haskell program that depends on a Haskell library,
but you don't want to bother writing a cabal file.
In this example we will access a Haskell package called `pandoc-columns`)
that is defined in a flake in a remote git repo.

[NOTE]
====
For non-trivial Haskell development projects,
it's usually more convenient to use `haskell-flake` as described in <<#haskell-flake>>,
together with the _high-level workflow_ described in <<_development_workflows>>.
====

[source,nix,linenums,highlight='5,8,12,17']
.flake.nix
....
include::flake.nix[]
....

Line 5 adds `pandoc-columns` as an input to this flake.
Line 8 allows the output function to reference `pandoc-columns`.
Line 12 makes a custom GHC that knows about `pandoc-columns`,
and line 17 uses the custom GHC as a build input for this flake.

Here's a short Haskell program that uses the new flake.

[source,haskell,linenums]
.Main.hs
....
include::Main.hs[]
....

Here's a demonstration using the program.

....
$ runghc Main.hs      # Fails; dependency not available
$# start-shell nix develop <<EOL
$ runghc Main.hs      # Works in development environment
$# EOL
....

////
Good adoc0 scripts clean up after themselves.
$ cd .. ; rm -rf tempwork # clean up
////