`--use-unicode` is added to `cabal haddock` and `cabal haddock-project`.
One cannot simply use `--haddock-option="--use-unicode"` because it
makes haddock to fail when building indexes.

We no longer define the `__HADDOCK_VERSION__` macro when invoking GHC
through Haddock, since doing so essentially guarantees recompilation
during documentation generation.

We audited all uses of `__HADDOCK_VERSION__` in hackage, ensuring there
was a reasonable path forward to migrate away from using
`__HADDOCK_VERSION__` for each, while generating the same documentation
as it did before.

If you are a user of `__HADDOCK_VERSION__`, please take a look at the
discussion in https://github.com/haskell/cabal/pull/9177 and reach out
to us if your use case is not covered.

Reverts the version-cpp flag introduced in the previous commit to avoid
a workaround. Instead, we get rid of the problem (`__HADDOCK_VERSION__`) at its root.

See the discussion in #9177

Haddock no longer writes compilation files by default, so we do not need to pass
tmp dirs for `-hidir`, `-stubdir`, and `-odir` via `--optghc`. Indeed, we do not
*want* to do so, since it results in recompilation for every invocation of
Haddock via Cabal. This commit stops this from happening for haddock versions >=
2.28 (when Hi Haddock was introduced).

This commit also stops the default definition of the `__HADDOCK_VERSION__` macro
when invoking GHC through haddock. Since a very limited set of users may still
depend on this macro, we introduce the `--haddock-version-cpp` flag and
`haddock-version-cpp:` cabal.project field, which enable the definition of
the `__HADDOCK_VERSION__` macro when invoking GHC through Haddock. This will
almost guarantee recompilation during documentation generation due to the macro
definition.

This commit also renames the `--haddock-lib` flag to `--haddock-resources-dir`
(and `haddock-lib:` cabal.project field to `haddock-resources-dir:`), and adds
this flag to the users guide since it was missing an entry. This also allows us
to add this field to `cabal-install:test:integration-tests2`, since it is no
longer ambiguous with the `--lib` flag.

This commit also causes `documentation: true` or `--enable-documentation` to
imply `-haddock` for GHC.

Also, since Haddock >= 2.29 is renaming `--lib` to `--resources-dir`, this
commit switches the flag provided to Haddock using a backwards compatible
condition based on the Haddock version.

Adds a changelog entry.

Very simple, backward-compatible change.  Not sure how this never
happened before, it would be a shame to have worked on the
preprocessor interface and not exposed it to users at all.

Took this chance to move the test preprocessor out of the
PreProcessors module and into a test case.

User hooks look like this now:
     preConf  :: Args -> ConfigFlags -> IO HookedBuildInfo,
     postConf :: IO ExitCode,

type HookedBuildInfo = (Maybe BuildInfo, [(String, BuildInfo)])

this means that only stuff in BuildInfo can be overridden.  A
buildinfo file now looks like this:

name: libraryname
include-dirs: /foo/bar, /bang/baz
extensions: CPP

executable: executablename
extensions: TemplateHaskell

if ./setup haddock is run w/ verbosity > 4, it adds -v to haddock command

Introduced new data type, Library which is symmetric to Executable.
Removed "executable only" and "library only" fields from
PackageDescription and moved them to their respective internal types.

Moved Build-depends out of buildinfo and into the package description
proper, as per Ross's suggestion.

Cleaned up parser and test cases accordingly.

Now it searhes for *.cabal in the current directory, if it finds more
than one, there's an error.  For now, it allows Setup.description with
a warning.

** Previously, "hidden-modules" had nothing to do with what gets
   built, and we had to compute "exposed-modules" with "modules `setdiff`
   hidden modules".  At Ross Paterson's suggestion, I added
   Exposed-Modules.

** Since the executable sections also used the "modules" field, I
   changed it to use "executable-modules", and got rid of the
   ambiguous "modules" field.  This should hopefully be more clear to
   everyone.

End-to-end preprocessing solution, including some test cases.
One test case fails:
  right now, preprocessors just stick files into the current
  directory, then copies them into the destination.  this is no good
  since we now won't be able to tell them from real source files the
  next time around.  clean is broken here...

** removed hugs test case that never works
** added -DDEBUG flags to HUnit and A
** fixed A's description file up a bit
** tried to cut back on cascading failures in moduleTest