The concern, for me, is whether the onus is on the developer or on the
tools. For example, if I write a method:
public String getTheValue() { return map.get("value");}
with no annotation, the assumption is the return value of getTheValue is
always non-null even though java.util.Map does not make any such guarantee.
So now, the onus is on me, as the developer to write
public @Nullable String getTheValue() { return map.get("value");}
I can do this, but the reality is that I'm lazy and I'm probably going to
forget and there's a high likelihood that nobody is going to call me out on
it during code review because it's one small function in a 5,000 line PR :)
For me, I'd much rather have the static analysis tool assume that anything
not annotated might possibly be null so that when I write the following, I
get a bug during the build:
String foo = getTheValue();
return foo.length();
Now the tool is actually doing its job of reminding me that I'm using a
value without having checked for nullability and if I want to avoid that
warning (maybe because I'm using a special java.util.Map), I need to update
my annotations. Given that the Iceberg codebase doesn't use these
annotations, I think we'll be dealing with nullable values far more often
than non-nullable ones.
Mike
On Tue, May 27, 2025 at 2:25 PM Eric Maynard <eric.w.mayn...@gmail.com>
wrote:
> Right, but I think Michael is correct that when something comes from
> outside of Polaris, we have to assume it’s nullable even if it’s not
> annotated as such.
>
> Also as we are using CDI, it may not be clear whether an implementation of
> an interface is coming from “Polaris code” and therefore what assumptions
> can be made about the nullability of its members.
>
> Given these considerations, I think developers are forced to do their own
> due diligence for nullability when working with non-annotated members. If
> we would go with option 1, this means developers are encouraged to do the
> opposite and forego a nullability check on non-annotated members. That
> seems dangerous.
>
> Another argument against option 1 is that for better or for worse most
> variables in Polaris are in fact nullable. If we have the choice between
> annotating 90% of variables and annotating 10%, I prefer the latter.
>
> On Tue, May 27, 2025 at 2:12 PM Dmitri Bourlatchkov
> <dmitri.bourlatch...@dremio.com.invalid> wrote:
>
> > Yes, the idea is to simplify things by using only one of @Nullable
> > and @Nonnull in Polaris code.
> >
> > However, I lean toward using @Nullable when necessary and assume @Nonnull
> > by
> > default. Again this is only for Polaris code.
> >
> > Cheers,
> > Dmitri.
> >
> > On Tue, May 27, 2025 at 5:07 PM Eric Maynard <eric.w.mayn...@gmail.com>
> > wrote:
> >
> > > If we've checked something is non-null or received a non-null value
> from
> > a
> > > library, we can continue annotating it as @Nonnull within Polaris. We
> can
> > > also use Optional to get similar functionality to typescript's ? and
> > handle
> > > missing values that way.
> > >
> > > In all other cases, developers must assume any given reference can be
> > null
> > > and do their own checks. It seems to me that there is no need for
> > @Nullable
> > > in that case. That is, if we label everything that cannot be null as
> > > @Nonnull and everything that can be null as @Nullable, then every
> > reference
> > > should be annotated as one or the other. We can simplify things by just
> > > removing the @Nullable annotations.
> > >
> > > --EM
> > >
> > > On Tue, May 27, 2025 at 1:56 PM Dmitri Bourlatchkov <di...@apache.org>
> > > wrote:
> > >
> > > > Your example is quite valid, Mike. In my proposal (several emails
> ago)
> > > I'd
> > > > classify this as "inputs into Polaris code".
> > > >
> > > > However, I guess the topic of this thread is slightly different. The
> > > > question we're trying to reach consensus on, is about what we should
> > > > annotate as @Nullable in Polaris code.
> > > >
> > > > I'd say if a value is received from external code, which is not
> > > annotated,
> > > > and returned from a Polaris method without validation, and the
> returned
> > > > value can reasonably be null, then the Polaris method should be
> > annotated
> > > > as @Nullable. However, if Polaris code checks the value to be
> non-null,
> > > the
> > > > return value need not be annotated in Polaris.
> > > >
> > > > Does this sound reasonable?
> > > >
> > > > Thanks,
> > > > Dmitri.
> > > >
> > > > On Tue, May 27, 2025 at 4:34 PM Michael Collado <
> > collado.m...@gmail.com>
> > > > wrote:
> > > >
> > > > > I generally assume the other way around. Many (most) libraries
> don't
> > > > > annotate the return values of their methods, so I assume everything
> > is
> > > > > nullable unless specifically told otherwise. I would prefer
> > everything
> > > be
> > > > > non-nullable unless specifically stated (the ? is the one thing I
> > would
> > > > > steal from typescript if I could), but given that library code
> can't
> > be
> > > > > modified to be explicit when null return values are possible, I
> don't
> > > > think
> > > > > we can make that assumption.
> > > > >
> > > > > Mike
> > > > >
> > > > > On Thu, May 22, 2025 at 10:57 AM Dmitri Bourlatchkov <
> > di...@apache.org
> > > >
> > > > > wrote:
> > > > >
> > > > > > My preference is for option 1 below.
> > > > > >
> > > > > > On Thu, May 22, 2025 at 1:53 PM Alex Dutra
> > > > <alex.du...@dremio.com.invalid
> > > > > >
> > > > > > wrote:
> > > > > >
> > > > > > > Hi,
> > > > > > >
> > > > > > > My proposal was centered around compile-time checks and targets
> > > > mostly
> > > > > > > developers and contributors. I am not questioning the
> usefulness
> > of
> > > > > > > runtime checks when these make sense.
> > > > > > >
> > > > > > > Maybe an example is better than a thousand words. Let's
> imagine a
> > > > > > > simple getOrDefault() method. Which version do you prefer?
> > > > > > >
> > > > > > > 1. Annotate only nullable items:
> > > > > > > public String getOrDefault(@Nullable String s, String def) {
> > > return s
> > > > > > > == null ? def : Objects.requireNonNull(def); }
> > > > > > >
> > > > > > > 2. Annotate only non-null items:
> > > > > > > @Nonnull public String getOrDefault(String s, @Nonnull String
> > def)
> > > {
> > > > > > > return s == null ? def : Objects.requireNonNull(def); }
> > > > > > >
> > > > > > > 3. Annotate everything:
> > > > > > > @Nonnull public String getOrDefault(@Nullable String s,
> @Nonnull
> > > > > > > String def) { return s == null ? def :
> > > Objects.requireNonNull(def); }
> > > > > > >
> > > > > > > Many places in Polaris are using option 3, which is too verbose
> > and
> > > > > > > leads to visual fatigue. What I was suggesting to the community
> > is
> > > to
> > > > > > > adopt option 1, that reduces the visual clutter and also
> assumes
> > > > > > > non-null by default.
> > > > > > >
> > > > > > > (You will notice that I added a runtime check to all three
> > > versions.)
> > > > > > >
> > > > > > > Hope that helps to clarify the discussion.
> > > > > > >
> > > > > > > Alex
> > > > > > >
> > > > > > > On Wed, May 21, 2025 at 8:54 PM Yufei Gu <flyrain...@gmail.com
> >
> > > > wrote:
> > > > > > > >
> > > > > > > > >
> > > > > > > > > In my opinion, assuming everything is nullable by default
> > isn't
> > > > the
> > > > > > > > > best approach for writing robust code. I believe a bias
> > towards
> > > > > > > > > non-nullness leads to more reliable systems.
> > > > > > > >
> > > > > > > > I agreed with the intention, but am concerned that assuming
> > > > > everything
> > > > > > is
> > > > > > > > non-nullness may discourage null-checking, which is
> problematic
> > > as
> > > > > > > runtime
> > > > > > > > null-checking isn't a thing.
> > > > > > > >
> > > > > > > > Yufei
> > > > > > > >
> > > > > > > >
> > > > > > > > On Wed, May 21, 2025 at 9:52 AM Alex Dutra
> > > > > > <alex.du...@dremio.com.invalid
> > > > > > > >
> > > > > > > > wrote:
> > > > > > > >
> > > > > > > > > Hi Eric,
> > > > > > > > >
> > > > > > > > > You're right that annotations don't change the bytecode at
> > > > runtime.
> > > > > > > > > Their real value comes during compilation, as many static
> > > > analysis
> > > > > > > > > tools use them to flag potential issues. They can even
> cause
> > > > build
> > > > > > > > > failures depending on how you configure them. My IDE
> > (IntelliJ)
> > > > > > > > > frequently warns me when I forget to handle a potential
> NPE;
> > if
> > > > > > you're
> > > > > > > > > not seeing similar feedback, it might be worth checking
> your
> > > IDE
> > > > > > > > > settings.
> > > > > > > > >
> > > > > > > > > While the annotations are primarily for compile-time
> checks,
> > > that
> > > > > > > > > doesn't mean we can't also incorporate runtime checks. We
> > > should
> > > > > aim
> > > > > > > > > to include these whenever this makes sense, for example by
> > > using
> > > > > > > > > Guava's Preconditions. This is especially useful if we
> can't
> > > > > > guarantee
> > > > > > > > > that a method parameter, for instance, will never be null,
> > > > because
> > > > > > > > > it's being provided by some external system.
> > > > > > > > >
> > > > > > > > > In my opinion, assuming everything is nullable by default
> > isn't
> > > > the
> > > > > > > > > best approach for writing robust code. I believe a bias
> > towards
> > > > > > > > > non-nullness leads to more reliable systems.
> > > > > > > > >
> > > > > > > > > I am also a big fan of Optional and think we should strive
> to
> > > use
> > > > > it
> > > > > > > > > as much as possible. That said, it's not always possible,
> > > > > especially
> > > > > > > > > if you are implementing a third-party interface that
> doesn't
> > > use
> > > > > it.
> > > > > > > > > Using Optional in class fields and method parameters is
> also
> > > > > > > > > controversial: Optional was designed primarily as a signal
> > from
> > > > the
> > > > > > > > > callee to the caller, to signify: "no result". In other
> > words,
> > > > its
> > > > > > > > > main purpose is to clarify method return types. This post
> on
> > > > Stack
> > > > > > > > > Overflow by Brian Goetz is worth reading: [1].
> > > > > > > > >
> > > > > > > > > Thanks,
> > > > > > > > >
> > > > > > > > > Alex
> > > > > > > > >
> > > > > > > > > [1]:
> > > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
> https://stackoverflow.com/questions/26327957/should-java-8-getters-return-optional-type/26328555#26328555
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > On Wed, May 21, 2025 at 4:37 PM Eric Maynard <
> > > > > > eric.w.mayn...@gmail.com
> > > > > > > >
> > > > > > > > > wrote:
> > > > > > > > > >
> > > > > > > > > > Thanks for bringing this up as I’ve been confused by
> this a
> > > few
> > > > > > > times.
> > > > > > > > > >
> > > > > > > > > > Before Polaris I hadn’t really encountered these
> > annotations
> > > > and
> > > > > I
> > > > > > > was
> > > > > > > > > > surprised to learn they don’t “do anything” — that is,
> > there
> > > is
> > > > > no
> > > > > > > > > > additional safety you get at runtime when a null value is
> > > > passed
> > > > > > > into a
> > > > > > > > > > parameter marked non-null. Similarly nothing enforces
> that
> > > you
> > > > > > handle
> > > > > > > > > null
> > > > > > > > > > values when something is annotated as nullable.
> > > > > > > > > >
> > > > > > > > > > For that reason, I tend to assume everything is nullable
> > > > > regardless
> > > > > > > of
> > > > > > > > > > annotation and I would be in favor of standardizing
> around
> > > that
> > > > > > > > > assumption.
> > > > > > > > > > Iff something is annotated as Non-null a developer should
> > > feel
> > > > > safe
> > > > > > > > > > skipping a check for null, but otherwise they should
> handle
> > > > null.
> > > > > > > > > >
> > > > > > > > > > I am a big fan of Optional and of trying to reduce the
> > usage
> > > of
> > > > > > null
> > > > > > > as
> > > > > > > > > > much as possible though.
> > > > > > > > > >
> > > > > > > > > > On Wed, May 21, 2025 at 3:02 PM Alex Dutra
> > > > > > > <alex.du...@dremio.com.invalid
> > > > > > > > > >
> > > > > > > > > > wrote:
> > > > > > > > > >
> > > > > > > > > > > Hi all,
> > > > > > > > > > >
> > > > > > > > > > > A while ago, we had a discussion regarding which
> nullness
> > > > > > > annotations
> > > > > > > > > > > to use and whether we should consider favoring non-null
> > by
> > > > > > > default. I
> > > > > > > > > > > would like to revive that discussion.
> > > > > > > > > > >
> > > > > > > > > > > We are currently using the `jakarta.annotation` package
> > > > > > > consistently,
> > > > > > > > > > > but the defaults are not clear: should we consider
> > > everything
> > > > > as
> > > > > > > > > > > non-null by default and only annotate the nullable
> > things,
> > > or
> > > > > the
> > > > > > > > > > > other way around? Some classes are cluttered with both
> > > > > > annotations,
> > > > > > > > > > > which seems unnecessary and confusing.
> > > > > > > > > > >
> > > > > > > > > > > I would personally be in favor of considering
> everything
> > as
> > > > > > > non-null by
> > > > > > > > > > > default.
> > > > > > > > > > >
> > > > > > > > > > > Please let me know your thoughts.
> > > > > > > > > > >
> > > > > > > > > > > Alex
> > > > > > > > > > >
> > > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>
