Queries about records

[Jonathan Bluett-Duncan]

Hi Vikram,

On Thu, 24 Dec 2020, 17:08 Vikram Bakshi, <vab2048 at gmail.com> wrote:

> Hi Remi,
>
> > The idea is that, in a method, if you are able to see a record, you can
> create it.
>
> Ah, this really clarifies the reasoning!
> It should definitely be in the documentation somewhere (if it's not
> already) - because I think this will end up being a common question.
>
> > No, if we introduce a shorthand for avoiding the requireNonNull(), it
> will be something global to all/most of the types in Java, not just for
> record components.
>
> Out of curiosity is there any intention to revive the efforts of JSR 305
> (which seems to be abandoned at this point)?
> A standard set of @NonNull and @CheckForNull annotations supported by the
> JDK itself would be quite helpful.
> For now I will rely on lombok's @NonNull (
> https://projectlombok.org/features/NonNull) to remove the boiler plate -
> it
> works quite well.
>

Yep, there is! Well, there's a third-party effort anyway, called jspecify (
jspecify.dev), which is being lead by engineers at JetBrains, Google and
other organisations inc. the Checker Framework maintainers.

They hold regular email conversations over at
https://groups.google.com/g/jspecify-dev in case you're interested.

They also host all their work-in-progress docs publicly on
https://drive.google.com/drive/mobile/folders/1vZl1odNCBncVaN7EwlwfqI05T_CHIqN-
and their source code on https://github.com/jspecify/jspecify.

Cheers,
Jonathan


> > There are no difference, records are always static otherwise it means
> that a record would have access to the enclosing values so it would not be
> defined only by its components.
>
> Great - that's what I thought. Thanks for clarifying.
>
> Merry Christmas (or Christmas Eve depending on your time zone)!
>
> Regards,
> Vikram
>
>
> On Wed, Dec 23, 2020 at 11:30 PM Remi Forax <forax at univ-mlv.fr> wrote:
>
> > ----- Mail original -----
> > > De: "Vikram Bakshi" <vab2048 at gmail.com>
> > > À: "amber-dev" <amber-dev at openjdk.java.net>
> > > Envoyé: Mercredi 23 Décembre 2020 17:58:25
> > > Objet: Queries about records
> >
> > > Hello,
> >
> > Hi,
> >
> > >
> > > I have a number of queries about the new Java record feature.
> > >
> > > 1. What exactly is the reason that the canonical constructor must have
> > the
> > > same access as the record?
> > >
> > > Consider the following use case: the record is intended to be a
> dumb-data
> > > carrier of its validated fields but the validation cannot take place in
> > the
> > > record's constructor because we require access to some elaborate
> > validator
> > > objects to actually perform the validation.
> > >
> > > So I thought a way to solve this would be to create the record as
> public
> > > and the canonical constructor as package-private. Then we have a record
> > > 'factory' in the same package which performs the actual validation and
> > > creation of the record. However this is not allowed currently because
> it
> > > violates the constraint that the canonical constructor must have the
> same
> > > access as the record itself.
> > >
> > > I was just wondering why this constraint?
> >
> > Records are named tuples, they are defined only by their components, in a
> > transparent manner i.e. no encapsulation.
> > From a tuple, you can access to the value of each component and from all
> > component values, you can create a tuple.
> > The idea is that, in a method, if you are able to see a record, you can
> > create it. Thus the canonical constructor has the same visibility as the
> > record itself.
> >
> > Being able to de-construct/re-construct a record allows to transform any
> > data text/binary representation (think JSON by example) from/to a
> > record/list of record.
> >
> > >
> > > 2. Is there a shorthand to declaring a record field must be non-null?
> > >
> > > So for the case where we do not want to allow null in a record field we
> > > have to implement our own logic in the canonical constructor and call
> > > `Objects.requireNonNull(...)`. Is there a way to declare in the record
> > > signature itself that we want it to be non-null to avoid this
> > boilerplate?
> >
> > No, if we introduce a shorthand for avoiding the requireNonNull(), it
> will
> > be something global to all/most of the types in Java, not just for record
> > components.
> >
> > >
> > > 3. What exactly is the difference between a static record and
> non-static
> > > record ?
> > >
> > > I find myself creating a whole bunch of public records within a single
> > > class. Is there actually any need to mark them as static?
> >
> > There are no difference, records are always static otherwise it means
> that
> > a record would have access to the enclosing values so it would not be
> > defined only by its components.
> >
> > >
> > > Thanks very much for creating records - they make coding a lot more fun
> > and
> > > allow for writing much more beautiful Java.
> > >
> > > Regards,
> > > vab2048
> >
> > regards,
> > Rémi
> >
>