Fixes #2162.

Before this, AllrightsReservered had two separate meanings; the author
explicitly chose this license or not license was specified and therefore
defaults to AllRightsReserved.

The default license when no license is specified is now UnspecifiedLicense.

Closes: #2141

Unlike GHC, Cabal will still try to find a globally consistent choice for
all package names, i.e. this does *not* implement private dependencies.

Signed-off-by: Edward Z. Yang <ezyang@cs.stanford.edu>

The initial support for module re-exports relied on ghc-pkg to resolve
user-specified re-exports to references to specific installed packages.
This resolution is something that can fail so it's better for Cabal to
do it during the package configure phase.

In addition, it is inconvenient in ghc-pkg to be doing this resolution,
and it just seems fishy as a design. Also, the same ModuleExport type
was being used both for user-specified source re-exports and also for
the specific re-exports in the package db.

This patch splits the type into two: one for source level, and one for
resolved ones for use in the package db. The configure phase resolves
one to the other.

One minor change: it is now possible to re-export a module defined in
the same package that is not itself exported (ie it's in other-modules,
rather than exposed-modules). Previously for modules definied in the
same package they had to be themselves exported. Of course for
re-exports from other packages they have to be exposed.

Re-exported modules allow packages to reexport modules from their
dependencies without having to create stub files.  Reexports of the same
original module don't count as ambiguous imports when module finding
occurs.  The syntax is:

    "orig-pkg" OrigName as NewName

You can omit 'as NewName', in which case it is reexported as the same
name.  Self referential aliases work too; however, they're only visible
to packages which depend on this package.

Left to future work: just provide a module name 'OrigName', where ghc-pkg
figures out what the source package is.

Signed-off-by: Edward Z. Yang <ezyang@cs.stanford.edu>

It's just noise that duplicates information in the 'LICENSE' file.

Based closely on the patches by Mathieu Boespflug <mboes@cs.mcgill.ca>

This field is intended to be used instead of (or in addition to) the
normal 'license-file' field by packages that have multiple files for
their license material. This is useful when eg the license is
supplemented by additional permissions and/or conditions. Notably,
the LGPL is structured in this way: it amends the GPL with additional
permissions, therefore one should distribute both the GPL in COPYING
and the LGPL in COPYING.LESSER.

So we keep both the license-file and license-files fields (rather than
deprecating one) and packages can use either or a mixture.

Fixes #1483.

It was verified that a generated license file is identical to the
original copy at http://www.gnu.org/licenses/agpl-3.0.html.

For example, you might have an images/ directory in your project, with
images that you refer to from Haddock with:

    <<images/foo.png>>

Then in your Cabal file you would include:

    extra-html-files: images/*.png

And these would both be packaged by sdist, and "cabal haddock" will
install them in:

    ~/.cabal/share/doc/PROJECT/html/images/*.png

Fixes #1167

For compatability, we need to make sure people using C files for the
main-is field, also have to specify the right cabal version.

This allows specifying the main-is field as a C file.

This is closely based on patches by Edward Z. Yang, who in turn credits
and earlier set of patches by Irene Knapp. The slight difference in
this version of the patch is that it is adjusted to work with the new
approach where we have separate hs compile; c compile; and link phases.

It should not be used in packages ready for being distributed.

I'm looking at you, ansi-terminal.

This patch adds the Apache license (http://www.apache.org/licenses/)
as a versioned license. Only version 2.0 is added, as the others are
historical and should not be used.

I've also updated slightly the license parse test, but I'm not sure
that the instructions for running the tests are correct (I don't see
the "license parsers" test being shown when running
./test/dist/build/suite/suite).

Refactor duplicate names check to avoid having to manually write all
O(n^2) possible collision cases between executables, test suites, and
benchmarks.

Makes things a little simpler in GHC's build system if libraries are in
the same directory as their name.

At the top level we now have cabal/ and cabal-install/

Only Cabal-1.10 and later can use test suites. Versions of Cabal prior
to 1.8 actually barf on test-suite sections, while Cabal-1.8 will
ignore these sections with a warning. Previously the QA check enforced
that packages with test-suite section specify 'cabal-version: >= 1.10'
but strictly speaking we only need to require 'cabal-version: >= 1.8'.
This relaxation allows people to write packages using test suites such
that people using Cabal-1.8 will be able to build and install the
package, just not run the test suite. Clear as mud?

It now works with Extensions rather than KnownExtensions

The knownExtensions list had already got out of sync with the datatype.
Now we use an Enum instance to get the list of constructors.
Also, having the actual extensions in a separate type than UnknownExtension
and EnableExtension makes it easier to add extension disabling support
later.

New fields default-language and other-languages for specifying the
base languages for the package, e.g. Haskell98, Haskell2010
New fields default-extensions and other-extensions for the language
extensions. Separate from the old extensions field.
The separation lets us express the difference between declaring to
the outside world that a package uses certain languages or extensions
and whether certain languages or extensions should be applied to
all modules in the package component.

Means we only need to have the version in one place now. Yay.

For historical reasons the cabal-version is specified with a version range,
to indicate the range of versions of tools that the package will work with.
We now think it makes more sense to specify the version of the Cabal spec
that the package follows. Old Cabal versions will not be able to parse simple
versions in this field. So we initially make the parser allow plain versions
but then we add a check to warn about using it prior to Cabal-1.12 at which
point it will be allowed.
Added a check about using version ranges that are not of the form '>= x.y'.
Also change behaviour to ignore upper bounds in the given version range.

Starting with Marc Weber's code and just moving
it about and making it fit in a bit nicer.