Bump the index-state for Hackage, fix ensuing compilation failures

[jorisdral]

No description provided.

[jorisdral]

A CI run is showing a test failure:

  Test.Database.LSMTree.UnitTests
    unit_blobs:                                                                   FAIL
      [FoundWithBlob (ResolveAsFirst "value1") (WeakBlobRef {weakBlobRefFile = WeakRef (BlobFile {blobFileHandle = <Handle <root>/active/0.wbblobs>, blobFileRefCounter = <RefCounter>}), weakBlobRefSpan = BlobSpan {blobSpanOffset = 0, blobSpanSize = 5}})]
      ["\177\187t\a\SOH"]
        test/Test/Database/LSMTree/UnitTests.hs:57:
        expected: ["blob1"]
         but got: ["\177\187t\a\SOH"]
      
      Use -p '/unit_blobs/' to rerun this test only.

[jorisdral]

A CI run is showing a test failure:

  Test.Database.LSMTree.UnitTests
    unit_blobs:                                                                   FAIL
      [FoundWithBlob (ResolveAsFirst "value1") (WeakBlobRef {weakBlobRefFile = WeakRef (BlobFile {blobFileHandle = <Handle <root>/active/0.wbblobs>, blobFileRefCounter = <RefCounter>}), weakBlobRefSpan = BlobSpan {blobSpanOffset = 0, blobSpanSize = 5}})]
      ["\177\187t\a\SOH"]
        test/Test/Database/LSMTree/UnitTests.hs:57:
        expected: ["blob1"]
         but got: ["\177\187t\a\SOH"]
      
      Use -p '/unit_blobs/' to rerun this test only.

I changed the unit test to a property test in 1cd7e28 that I run for 100_000 tests on each CI job (of which there are ~18), and this did not seem to exhibit the same bug. For now I'm considering to be a flaky test, because I don't have other ideas for finding a reproducer

[recursion-ninja]

📤 LGTM