Hacker News Reader

Comment by driver733 | Hacker News Reader

You are right, but you still have to catch the checked IOException.

> You are right, but you still have to catch the checked

> IOException.

Rather than which alternative? There are tonnes of IO errors that can occur and are captured in Java that you may not ever even consider. IO is exceptionally tough and programmers should be aware that there could be 1 of a possible million reasons for failure, even if they don't care exactly why.

It's not a massive ask either:

    try{
      /* Perform some IO that 99.999999% of the time */
    }catch(Exception e){
      /* Something went wrong and we don't know the state of the disk */
    }

Personally I believe applications should have their own wrapper around IO and handle it accordingly. I.e. not caring, re-trying, show-stopper, etc, etc.

A lot of code I've written will have the following if I really don't care if it works or not:

    try{
      /* IO code */
    }catch(Exception e){
      /* Do nothing */  // <-- Let others know that this was on purpose
    }

99% of the time I want to do the same thing for IOException as I would for NoSuchElementException or a DOMException.

That said, even if typed checked exception handling is important...it quickly becomes untenable.

   Thing thing = create();
   try {
     ...
   } finally {
     cleanup(thing);
   }

You might try to abstract this

   withThing(thing -> ...)

But what if "..." potentially throws IOException? What if it potentially throws IOException or AWTException?

You have to give up all your exception type safety to make an abstraction like withThing. It's just not composable.

    interface ThrowingConsumer<T, E extends Exception> {
        void accept(T value) throws E;
    }

    // ...

    <E extends Exception> void withThing(ThrowingConsumer<Thing,E> callback) throws E { ... }

The problems are that (a) you need to add an extra type parameter for exceptions all over the place, (b) exception unification (sum type) doesn't work generically - you can say E1 | E2 in a catch block but it's not a real type, and (c) it composes poorly with existing libraries that expect Consumer<T> throughout.

For one level deep callbacks (for resource handling and the like) it works reasonably well though.

(b) is what I had in mind as the problem.

You can always opt out of the typed checked exceptions.

    try {
        ...
    } except (IOException|NoSuchElementException e) {
        throw new RuntimeException(e);
    }

If you decide you don't want checked exceptions in your code, wrapping the exceptions at the boundaries is not a huge deal.

Or switch to a language that does that for you, like Scala or Kotlin.

For a lot of scripts you want the program to crash with a stack trace if you get an IOException, which is the default behavior if it's uncaught. The user is the developer, and if there's an error the developer fixes it by munging something on their filesystem.

Java was designed for "production" software, which in the late 90s and early 00s meant software that ran for long periods of time as a server, servicing thousands of mission-critical customers. Crashing was not acceptable behavior for that. But now a lot of production software often requires a lot of ancillary one-off tasks that are just done by the developers - testing, data-munging, exploratory code, migrations, demos, etc. That's a very different environment from where you code to spec, the spec never changes, and once the software is done it's supposed to run for years without crashing.

Early Java was designed for set-top boxes and then web browsers. Large-scale servers came later; in fact a lot of libraries still don't support the async futures that were officially added five years ago.

I actually like it. I also create my own exceptions for rare (exceptional) cases, just not to forget to handle them.

But some Java SDK exceptions drive me crazy. For example `URL.parse("http://example.com")` over hard-coded strings that I know 100% won't throw exceptions, but I still to catch 3 of them.

`URI.create("https://example.com")`

That's why URI has a factory method that you're only supposed to use for known strings (although of course this can't be checked):

https://docs.oracle.com/javase/7/docs/api/java/net/URI.html#...

After URI appeared in 1.4, the only reason to use URL was to create a URLConnection from a URI. Since openConnection() throws IOException, it's not a big deal that toURL() throws a MalformedURLException - just catch it along with all the other IOExceptions.

Since 11, there's no reason to use URL at all, because you can use HttpClient to actually do HTTP.

Ok I'm a veteran Java developer that understands exactly what you're saying but try explaining that to people unfamiliar with Java or junior engineers and watch their brains glaze over.

The JDK is filled with a ton of dumb "once upon a time we thought this was okay..." things that don't properly encode the correct modern idiom

For a long time I've wanted a sort of "local deprecation" tool where we could have a list of things in the JDK that shouldn't be used, and any direct reference to them would cause our build to fail.

https://errorprone.info/ can be configured to look at the AST and warn or reject a lot of specific problems.

We use https://github.com/policeman-tools/forbidden-apis for that.

Yes, it absolutely is! But then so are all other languages of its vintage. Part of learning a programming language is learning that.

Newer languages definitely have an advantage here, in that they haven't been around long enough for people to have figured out which bits of them are dumb.

It's worse. In 1.4 they added URI because URL is very flawed. You should never use URL. The equals method actually does a DNS resolve and compares the ip addresses. This means that comparing two URL's on the same server will always return true when compared. A lot of URL's you compare will give you true as there are a lot of sites run on the same machine with the same ip address.

Had this happen in production in a third party library we used many moons ago. Sysop came and asked us why we did six figure DNS lookups over a short period of time (24h? Less? Don't remember).

Would probably have gone unoticed at most AWS/GCP/Azure shops today.

Worse was when I had somebody adding the ip address as an Inet4Address on every message passed between machines in a production environment that explicitly didn't have DNS (banks have occasionally very odd ideas about securing subnets). Every single message was doing a reverse DNS lookup and then timing out. And there were a _lot_ of messages.

Wait, seriously? Did whoever designed that have any idea what URLs actually were? Even without DNS, comparing the hosts is no way to compare URLs. I can't think of a single example of two 'equal' URLs whose strings don't match (up until the parameters, at least)...

Yes - bear in mind the URL class has been around since 1995, when the landscape was somewhat different!

At that point, URLs resolving to the same IP were considered equal, even if the host names were different. Even now, there’s no real difference between say “http://example.com” and “http://example.com:80”; it would have been reasonable to consider these equal.

http://internalserver/resource

http://internalserver.initech.com/resource

Those are certainly not equivalent, as the 1st would not even resolve in the majority of environments. I'm assuming the 2nd host is a reverse proxy for the 1st? Unless that fact is part of your design assumptions (which it certainly can't be for a general URL class in a standard library), those should be considered different.

If the two hostnames resolve to the same IP address, they are equivalent. This could happen if initech.com was a default domain. This is a very standard situation in internal networks which i am surprised you are not familiar with.

I am quite familiar with that kind of setup and it's even further from equivalent than what I assumed you were referring to. Shared web hosting is the common example, but some other protocols are capable of distinguishing between hostnames.

google.com and maps.google.com resolve to the same address but are not equivalent. Even if your example and many others are equivalent, there are many that aren't. A general library should not make assumptions like that, unless they hold 100% of the time.

You asked for a single example of two 'equal' URLs whose strings don't match up, and now you've got one.

Looking back, I did phrase it ambiguously. What I meant was two URLs that are equal by definition, (no matter the server config, assuming standards compliance). A (general) library should make no more assumptions than the standard it implements. What I was asking about were cases, similar to http://h/p?x1=a&x2=b being equal to http://h/p?x2=b&x1=a since query parameter order is defined to not matter.

The spec I am mostly familiar with is, however, rather new, and whatever standard controlled URLs at the time (if any) might have made more assumptions.

How about http://localhost http://::1 and http://0:0:0:0:0:0:0:1 ?

Wow. And of course .equals is a synchronous method, so the whole thread blocks on a network operation whenever you do that.

Helluva way to thread starve your application. :(

https://docs.oracle.com/javase/7/docs/api/java/net/URL.html#...

Absolutely wild how people are all in favour of Either/Result, but still look down on checked exceptions.

You might find this article interesting. https://www.pragmaticobjects.com/chapters/001_checked_except...

I've been reading articles like this for twenty years now, and i've never found one anything other than idiotic.

I love both, but Java's implementation of checked exceptions cause harder and harder problems, especially when combined with tools such as lambdas, since there's no way to generically compose or handle checked exceptions inside them.

Java's implementation of checked exceptions looks pretty minimal and sound to me.

How would you implement them?

It's certainly fair to say I have no alternative recommendation. I enjoyed using them prior to Java 8, and they gave the guarantees I was looking for. They have just had a much harder time integrating with newer features than could be hoped for. I'm not sure how much of that is intrinsic to checked exceptions, and how much is intrinsic to Java's implementation.

One example of this: there is no way to encode the type of a checked exception in a generic. I would like to be able to express something like the following:

  interface ExceptionHandler<E extends Exception, S, T> {
    T wrap(
      Function<S, T, throwing E> fn,
      Function<E, T> exceptionMapper
    );
  }

But there is no way to express that 'throwing E' part. The type of a checked exception is firmly embedded in the interface. So I can't supply, say, an IO method as a Function<S, T> parameter, since the IOException causes a mismatch, and I'd have to write a handler specifically for methods that throw IOException, another for those that throw TimeoutException, a third for those that throw IOException AND TimeoutException, etc; a fairly fruitless goal without automatic code generation.

You can actually use a generic parameter in a throws clause, so you can write something like:

    interface FunctionThrows1<S, T, E1 extends Throwable> {
      T apply(S in) throws E1
    }
    interface FunctionThrows2<S, T, E1 extends Throwable, E2 extends Throwable> {
      T apply(S in) throws E1, E2
    }
    ...

But you still have to write one version of your method for each arity of throwing that you want to be able to wrap.

Interesting. I tried this years ago, and it never worked. Maybe that's changed in recent versions? Arity versions are still more feasible at least, and many languages are willing to pay the cost for them.

Yeah, I'm pretty sure there were some compiler bugs involving generic exceptions around 1.5 - 1.6. (I don't know if they ever got fixed -- I switched to Kotlin.)

AFAICR this has always worked. But you rapidly run into limitations - as i say in another comment, catch and throw are not themselves generic. Plus, none of the interfaces used in the streams API have exception parameters.

> Plus, none of the interfaces used in the streams API have exception parameters.

That's partly because there's another problem. Potentially, you could have streams of the form, say, `...map(A::foo).map(A::bar).map(A::baz).collect(...)`, and each of foo, bar, baz adds another exception type to the set that can be thrown by collect.

Result/Either-based streams have exactly the same problem. An exception-supporting stream API would solve it the same way: a function can either throw no exception, or an exception of the same type as thrown by a previous operation. You could have a .mapException operation which could change the exception type, like Result::map_err in Rust.

In practice, you would end up with the stream throwing an exception of some general type (in 95% of cases, IOException!), but you could still write specific catch blocks for each possible subtype.

Let me put it this way: the message I got from the language team is that the problem is certainly acknowledged, and various alternatives to dealing with it are known, but none of them has so far been shown the clearly preferable, "good" choice, so until one presents itself, the decision is to do nothing rather than add/remove features that could later have other negative consequences.

Maybe that was it. That means there's no way to generically remove or translate the exceptions, e.g. by wrapping it into a runtime exception, or into a specific Either value based on the exception type.

The problem boils down to the fact that you can have a disjunction of any number of checked exception types (including zero). No other party of the type system allows disjunctions, so it causes a lot of problems. The checked exception is conceptually part of the return type, but is split out.

I wish Kotlin had, instead of ignoring the existence of checked exceptions, instead translated them into part of the return type. I use Kotlin a lot these days, and one annoyance for me is dealing with code that throws exceptions. They fixed the annoying "(almost) anything can be null" problem of java and replaced it with an equivalent problem. Why can't nullability and failure results both be part of the static type?

(The workaround it to manually use an Either type yourself, but it doesn't help you with calling anyone else's code, since virtually everything throws exceptions on failure.)

Checked exceptions are just sum types and you unpack it with a try catch block.

Yes, but they are the only part of Java's type system that allows sum types.

How do you declare a method that generically takes a function that in turn takes a K, returns a V, and can throw whatever checked exceptions it wants, and you'll rethrow them? Last I checked, this wasn't possible in Java.

If checked exceptions were instead replaced with sum type return values, then it becomes trivial.

You can do that and propagate the exception:

    @FunctionalInterface
    interface ThrowingFunction<T, R, E extends Exception> {
        R apply(T argument) throws E;
    }

    static <T, R, E extends Exception> R apply(T argument, ThrowingFunction<T, R, E> function) throws E {
        return function.apply(argument);
    }

But not catch and rethrow it, because throw and catch aren't generic, and the type parameter isn't reified. You can emulate reified generics with the usual trick of passing a Class object:

    static <T, R, E extends Exception> R apply(T argument, ThrowingFunction<T, R, E> function, Class<E> witness) throws E {
        try {
            return function.apply(argument);
        } catch (Exception e) {
            throw witness.cast(e);
        }
    }

But this is pretty horrid.

Right, it's horrid, and it only works for exactly one checked exception type. So there's no way to define a single `Stream.map` (for example) that takes a function that throws whatever number of checked exceptions and just propagates them.

Either/Result also only works for exactly one error type.

Well, yes. In a way that's another example of the kind of problem I'm talking about.

This thread stemmed from https://news.ycombinator.com/item?id=21808522 :

> Java's implementation of checked exceptions looks pretty minimal and sound to me. > > How would you implement them?

Java's checked exceptions aren't "minimal" and are problematic because they behave unlike the rest of the types system. I'm not saying that they should have ditched checked exceptions and left it at that, though. What they should have done was make sum types a first-class part of the type system, and then checked exceptions become redundant.

In languages that actually have general sum types rather than special-casing the way Java's exceptions do, you can get multiple types of errors with an Either by making a union of the different types of errors. You can alternatively use an N-way sum type as your return value in place of the 2-way Either.

Kotlin has sealed classes, which can be used to create a sum type, and you can even do the same in Java with enough boilerplate, but in either of those languages you run into another problem if you try this: it doesn't interoperate well with the vast majority of existing library code that throws exceptions.

The problem is that try catch is so verbose

Incorrect for two reasons:

1) They follow a different code path (which is what makes them a superior way to handle errors over return values).

2) They cannot be emulated by sum types when all you want to do is not handle the exception and let it bubble up the stack frames.