bech32 "prefix" (exit -11): failed (after 141 tests)

[juhp]

Testsuite failure with ghc-8.10.3 in the Stackage Nightly build:

Test suite failure for package bech32-1.1.0
    bech32-test:  exited with: ExitFailure 1

    Failures:

      test/AppSpec.hs:49:21: 
      1) App, bech32 command-line, can re-encode encoded strings bech3219chzuvz5nks
           uncaught exception: IOException of type OtherError
           readCreateProcess: bech32 "prefix" (exit -11): failed
           (after 141 tests)
             MinString "\" )\1093809r\51859\bzgtJ\31250\r\SO )x+x\NUL\FS\1024098[<A\1005548\63534\SO\1933\1055812\49522\1029039-jx\1
079070u\34769\17176OB\RSs\1079611\68220\DC2\t3C\SI\1082032\66739u\1049222+pm\SOH\STX@w\992698]s\\_\2824\47465\12590\1103213C\STX\106
7834c+\1010826e\ENQ?"

      To rerun use: --match "/App/bech32 command-line/can re-encode encoded strings bech3219chzuvz5nks/"

    Randomized with seed 876929532

    Finished in 56.1419 seconds
    63 examples, 1 failure

[rvl]

Hi @juhp, thanks for reporting the issue.

In your error message, the way the bech32 subprocess exits (with a segfault) seems suspicious. This doesn't seem normal for a Haskell pure program. We never tested with ghc-8.10.3, but have not yet seen this test failure on ghc-8.10.4 or ghc-8.6.5.

We fixed the other ghc-8.10 test failure in #35, and intend to re-enable bech32 in Stackage after making a Hackage release.