These are expected to be bytes, not str.

this does not work in the general case since there is no (recursive)
convertion of objects used as model object initialization.

We can only check when using the from_dict() factory.

for better clarity on the code author's intention.

it was still here for bw-compat but should not be necessary any more.

this function does not need to be a local function of the to_dict namespace.

This will allow deduplicating code across loaders.

This generates more realistic contents and avoids spurious HashCollisions when
generating a set of objects using these hypothesis strategies, at the cost of
slightly worse "boundary checking" (i.e. we won't check contents with a length >
4096 bytes).

This lets us generate Content objects directly from a bytestring, with the
proper set of hashes auto-generated from the contents.

Required by loaders, when they can't parse the fullname.

Will be used by loaders.

It just makes more sense that way.

eg. before this change, all leafs would be equal to each other.

They will be used by loaders, so they can deal only with
model objects, instead of having to do the same conversion themselves.

This removes the `data` and `save_path` arguments of `from_file` and
`from_disk`, as data loading is always deferred from now on.
To access it, users are now expected to either open the data files
themselves, or us `.to_model().with_data()`.

Regular files were created, as the 'mode' argument of os.mknod was missing.
However, creating devices requires root; so we can't reasonably do that in
tests.

Instead, we're using /dev/null instead of creating one.
And while we're at it, let's also use /dev/zero (which, if not handled
properly, will result in an infinite read).

It should be cheap enough to do it, and it makes tests easier.

It will be useful to loaders, as they currently load the entire
content in memory before deciding to skip it.

This allows the `swh identify` command to work again.

Close T2288

Can be useful to deduplicate code in swh-storage.

It already is in practice.

The current postgresql model refuses NULL values.

Related to P589

This allows for example to type:

  curl -s https://archive.softwareheritage.org/browse/content/sha1_git:64582b78792cd6c2d67d35da5a11bb80886a6409/raw/ | swh identify
  swh:1:cnt:64582b78792cd6c2d67d35da5a11bb80886a6409	-

Related to T2188

Add support to automatically compute identifier in the following object models:
Directory, Release, Revision, Snapshot.

If the identifier is not provided as parameter, it will be computed when the model
is initialized.

Allows us to call tests on things other than python3, with the same settings.

they will remain for internal use only, so should not be mentioned in this
public spec

this reverts the documentation part of 67fade5f

currently provides mainly 2 generators:
- gen_origins()
- gen_contents()

we do not really need them to be mutable, plus we gain their instances now
being hashable, so we can add them in set() for example.