E.g.

  $ nix-build -I nixpkgs=git://github.com/NixOS/nixpkgs '<nixpkgs>' -A hello

This is not extremely useful yet because you can't specify a
branch/revision.

The function builtins.fetchgit fetches Git repositories at evaluation
time, similar to builtins.fetchTarball. (Perhaps the name should be
changed, being confusing with respect to Nixpkgs's fetchgit function,
with works at build time.)

Example:

  (import (builtins.fetchgit git://github.com/NixOS/nixpkgs) {}).hello

or

  (import (builtins.fetchgit {
    url = git://github.com/NixOS/nixpkgs-channels;
    rev = "nixos-16.03";
  }) {}).hello

Note that the result does not contain a .git directory.

Other stores don't do this either. It's up to the caller to check
signatures.

This is to allow store-specific configuration,
e.g. s3://my-cache?compression=bzip2&secret-key=/path/to/key.

Substitution is now simply a Store -> Store copy operation, most
typically from BinaryCacheStore to LocalStore.

If --no-build-output is given (which will become the default for the
"nix" command at least), show the last 10 lines of the build output if
the build fails.

This was added to support Hydra, but Hydra no longer uses it.

This also gets rid of --log-type, since the nested log type isn't
useful in a multi-threaded situation, and nobody cares about the
"pretty" log type.

This replaces nix-push. For example,

  $ nix copy --to file:///tmp/cache -r $(type -p firefox)

copies the closure of firefox to the specified binary cache. And

  $ nix copy --from file:///tmp/cache --to s3://my-cache /nix/store/abcd...

copies between two binary caches.

It will also replace nix-copy-closure, once we have an SSHStore class,
e.g.

  $ nix copy --from ssh://alice@machine /nix/store/abcd...

This prevents copying a partial closure to a binary cache.

This allows commands like "nix verify --all" or "nix path-info --all"
to work on S3 caches.

Unfortunately, this requires some ugly hackery: when querying the
contents of the bucket, we don't want to have to read every .narinfo
file. But the S3 bucket keys only include the hash part of each store
path, not the name part. So as a special exception
queryAllValidPaths() can now return store paths *without* the name
part, and queryPathInfo() accepts such store paths (returning a
ValidPathInfo object containing the full name).

This allows running arbitrary Nix commands against an S3 binary cache.

To do: make this a compile time option to prevent a dependency on
aws-sdk-cpp.

Forgot to commit this earlier...

This re-implements the binary cache database in C++, allowing it to be
used by other Store backends, in particular the S3 backend.

Caching path info is generally useful. For instance, it speeds up "nix
path-info -rS /run/current-system" (i.e. showing the closure sizes of
all paths in the closure of the current system) from 5.6s to 0.15s.

This also eliminates some APIs like Store::queryDeriver() and
Store::queryReferences().

Also updates tests to check for new information. Fixes #799