This replaces the copy&paste with a helper function in hash.hh.

fetchTarball, fetchTree, and fetchGit all have *optional* hash attrs.
This means that we need to be careful with what we allow to avoid
accidentally making these defaults. When ‘hash = ""’ we assume the
empty hash is wanted.

follow up of https://github.com/NixOS/nix/pull/3544

This allows hash="" so that it can be used for debugging purposes. For
instance, this gives you an error message like:

  warning: found empty hash, assuming you wanted 'sha256:0000000000000000000000000000000000000000000000000000'
  hash mismatch in fixed-output derivation '/nix/store/asx6qw1r1xk6iak6y6jph4n58h4hdmbm-nix':
    wanted: sha256:0000000000000000000000000000000000000000000000000000
    got:    sha256:0fpfhipl9v1mfzw2ffmxiyyzqwlkvww22bh9wcy4qrfslb4jm429

Add error message when FileIngestionMethod is out of bounds

Alternative fix to #3661. The cause was that 'name' is a
std::string_view into a temporary which could get overwritten.

Add `src/libutil/tests/libutil-tests` to `.gitignore`

I gather this comes from the new unit tests.

bool coerces anything >0 to true, but in the future we may have other
file ingestion methods. This shows a better error message when the
“recursive” byte isn’t 1.

Fixes #3648.

This was a latent bug that just appeared because of the tests that were
added. Remember to wait for CI! :)

Remove `addToStore` variant as requested by `FIXME`

Co-authored-by: James Lee <jbit@jbit.net>

Improve ref validity checking in fetchgit

As `git fetch` may chose to interpret refspec to it's liking, ensure that we
only pass refs that begin with `refs/` as is, otherwise, prepend them with
`refs/heads`. Otherwise, branches named `heads/foo` (I know it's bad, but it's
allowed), would be fetched as `foo`, instead of `heads/foo`.

The previous regex was too strict and did not match what git was allowing. It
could lead to `fetchGit` not accepting valid branch names, even though they
exist in a repository (for example, branch names containing `/`, which are
pretty standard, like `release/1.0` branches).

The new regex defines what a branch name should **NOT** contain. It takes the
definitions from `refs.c` in https://github.com/git/git and `git help
check-ref-format` pages.

This change also introduces a test for ref name validity checking, which
compares the result from Nix with the result of `git check-ref-format --branch`.

The idea is it's always more flexible to consumer a `Source` than a
plain string, and it might even reduce memory consumption.

I also looked at `addToStoreFromDump` with its `// FIXME: remove?`, but
the worked needed for that is far more up for interpretation, so I
punted for now.

Add unit tests for config.cc

Replace some `bool recursive` with a new `FileIngestionMethod` enum

installer: don't require xz on darwin

There was an enum there that matched in perfectly.

On macOS the system tar has builtin support for lzma while xz isn't
available as a separate binary.  There's no builtin package manager
there available either so having to install lzma (without nix) would be
rather painful.

This is much less confusing since recursive is no longer a boolean.

This is a different recursive than used in makeFixedOutputPath.

Provides some general overview on the mechanics of Config/Setting and
comments for the public methods of Config.

This moves the actual parsing of configuration contents into applyConfig
which applyConfigFile is then going to call. By changing this we can now
test the configuration file parsing without actually create a file on
disk.

Use /etc/zshenv instead of /etc/zshrc for profile

Add unit tests for "json.hh"

Add unit tests for xml-writer

Add unit tests for hashing functions

- --no-channel-add didn't have effect on multi-user installation
- some new flags didn't work at all
- document all installer flags

install-multi-user: allow overriding user count