The signature of isA is completely wrong about javahamcrest

[durron597]

I think the problem is that the version in CoreMatchers never got fixed

[durron597]

I'm surprised this hasn't gotten more attention! This is a problem in the most obvious use case, i.e.

import static org.hamcrest.core.IsInstanceOf.*;
import static org.hamcrest.core.Is.*;

import org.junit.Test;

import static org.junit.Assert.*;

public class HamcrestTest {
    interface FooInterface { }
    class Foo implements FooInterface { }
    class FooBar extends Foo { }

    @Test
    public void isTest() {
        Foo someFoo = new FooBar();
        FooInterface interfaceRef = someFoo;
        assertThat(someFoo, isA(FooBar.class)); // Doesn't compile
        assertThat(interfaceRef, isA(Foo.class)); // Doesn't compile

        assertThat(someFoo, instanceOf(FooBar.class)); // Compiles correctly
        assertThat(interfaceRef, instanceOf(Foo.class)); // Compiles correctly
    }
}

[udalrich]

Any hope of this eventually being released?

[nhojpatrick]

junit:4.13 depend on hamcrest-core:1.3 also has this problem

feel free to raise PR with junit so v4.14 could be upgrade to the latest version, that being currently org:hamcrest:hamcrest:2.2

v1.x is dead and out of support
v2.x i'm still trying to learn the guts and help with patches... so semi in support

[time4tea]

"completely wrong" were Nat's exact words btw - nothing to do with me.... :-)

[scarytom]

This is interesting in conjunction with some of Nat's other discussions on the matter:
http://code.google.com/p/hamcrest/issues/detail?id=83#c13

[npryce]

It should be <T> Matcher<Object> isA(Class<T> expectedType, Matcher<? super T> downcastMatcher)

[time4tea]

Which is a matcher of object not T. That would do it I think.
On 29 Jan 2013 22:11, "Nat Pryce" [email protected] wrote:

It should be Matcher isA(Class expectedType, Matcher<? super T>
downcastMatcher)

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/27#issuecomment-12861355.

[scarytom]

I'll have a go at this soon... watch this space :)

[time4tea]

oddly all the angle brackets in the type signature didn't appear in the email i got from github... hence my comment about Object, not T - which you can see here but not in the email...

[EarthCitizen]

Has there ever been any progress on this issue? I made my own version of isA() to get around the issues discussed here.

[scheffield]

@EarthCitizen: would be nice to see it as a Gist.

[EarthCitizen]

isA() is an alias for instanceOf() is it not? What about something like isClass() for exact Class matching?

[bitzl]

I've submitted an Pull Request (#127) to fix this. The behavior of isA is documented as shortcut to instanceOf. I've enhanced tests to check if the beavior is similar and changed the methods return a Matcher<?> instead of Matcher<TheClass.class>.

[sf105]

Fixed

[massenz]

Could you please give us an indication of which version this was/will be released for?
I'm currently using java-hamcrest:2.0.0.0 and it still has the same incorrect signature.

Thanks.

[durron597]

This change hasn't been released yet. I'm not a Hamcrest member so I can't tell when they plan to release the next one, but for now it looks like you'll have to build from source.

Unfortunately master still has the version number listed as 2.0.0.0, so you'd probably have to change that in your build if you're going to add it to a local repo.

[micheal-hill]

hamcrest:2.1 has this problem...

[ProYI]

junit:4.13 depend on hamcrest-core:1.3 also has this problem

[nhojpatrick]

hamcrest:2.1 has this problem...

does org:hamcrest:hamcrest:2.2 have the problem?