Revision on hspec-expectations-pretty-diff places a lower bound on aeson than the one we already have

[mihaimaruseac]

Looking at https://hackage.haskell.org/package/hspec-expectations-pretty-diff-0.7.2.6/revisions/ we see that r1 has the following description

Changed the test suite 'tests' component's library dependency on 'aeson' from

>=0

to

<2

However, we have aeson-2.2.2.0 and hspec-expectations-pretty-diff-0.7.2.6 in the snapshot and no bounds issue (though we expect a failure from hspec-expectations-pretty-diff -- valpackett/hspec-expectations-pretty-diff#7).

As a result, we now have

aeson-2.2.2.0 (changelog) (Adam Bergmark adam@bergmark.nl @bergmark) is out of bounds for:

• hspec-expectations-pretty-diff-0.7.2.6 (< 2). Greg V greg@unrelenting.technology @myfreeweb. Used by: test-suite

To fix this, I'll move hspec-expectations-pretty-diff from expected failure section to skipped test.

CC @Bodigrim as the author of the revision, in case a better solution exists.

[Bodigrim]

However, we have aeson-2.2.2.0 and hspec-expectations-pretty-diff-0.7.2.6 in the snapshot and no compilation failure

That's surprising.

test/Test/Hspec/Expectations/PrettySpec.hs:35:17: error: [GHC-83865]
    • Couldn't match expected type ‘Key’
                  with actual type ‘Data.Text.Internal.Text’
    • In the first argument of ‘(.=)’, namely ‘pack "foo"’
      In the expression: pack "foo" .= object [pack "bar" .= Number 123]
      In the first argument of ‘object’, namely
        ‘[pack "foo" .= object [pack "bar" .= Number 123]]’

[mihaimaruseac]

As you mentioned in the other threads, it is because we are not actually compiling. I was wrong to say "no compilation failure", the correct phrase would have been "no bounds issues".

Closing as there's not much we can do. The test was already broken.

Thank you for the updates and for doing the revisions