unit test failure in Core PR 15796 #610

issue instagibbs opened this issue on April 12, 2019
  1. instagibbs commented at 9:34 PM on April 12, 2019: none
    make[6]: Entering directory '/home/travis/build/bitcoin/bitcoin/build/bitcoin-x86_64-unknown-linux-gnu/src/secp256k1'
    ./build-aux/test-driver: line 107: 26441 Aborted                 (core dumped) "$@" > $log_file 2>&1
    FAIL: tests
    PASS: exhaustive_tests
    ========================================
       libsecp256k1 0.1: ./test-suite.log
    ========================================
    
    # TOTAL: 2
    # PASS:  1
    # SKIP:  0
    # XFAIL: 0
    # FAIL:  1
    # XPASS: 0
    # ERROR: 0
    
    .. contents:: :depth: 2
    
    FAIL: tests
    ===========
    
    src/tests.c:405: test condition failed: ((~x[m][shift]) << (64 - (1 << usebits))) == 0
    test count = 64
    random seed = 54d904804b5e0db448021441be754c10
    FAIL tests (exit status: 134)
    
    ============================================================================
    Testsuite summary for libsecp256k1 0.1
    ============================================================================
    # TOTAL: 2
    # PASS:  1
    # SKIP:  0
    # XFAIL: 0
    # FAIL:  1
    # XPASS: 0
    # ERROR: 0
    ============================================================================
    See ./test-suite.log
    ============================================================================
    make[6]: *** [test-suite.log] Error 1
    Makefile:1213: recipe for target 'test-suite.log' failed
    make[6]: Leaving directory '/home/travis/build/bitcoin/bitcoin/build/bitcoin-x86_64-unknown-linux-gnu/src/secp256k1'
    make[5]: *** [check-TESTS] Error 2
    Makefile:1319: recipe for target 'check-TESTS' failed
    make[5]: Leaving directory '/home/travis/build/bitcoin/bitcoin/build/bitcoin-x86_64-unknown-linux-gnu/src/secp256k1'
    make[4]: *** [check-am] Error 2
    Makefile:1532: recipe for target 'check-am' failed
    make[4]: Leaving directory '/home/travis/build/bitcoin/bitcoin/build/bitcoin-x86_64-unknown-linux-gnu/src/secp256k1'
    Makefile:13330: recipe for target 'check-local' failed
    make[3]: *** [check-local] Error 2
    make[3]: Leaving directory '/home/travis/build/bitcoin/bitcoin/build/bitcoin-x86_64-unknown-linux-gnu/src'
    make[2]: *** [check-am] Error 2
    Makefile:13078: recipe for target 'check-am' failed
    make[2]: Leaving directory '/home/travis/build/bitcoin/bitcoin/build/bitcoin-x86_64-unknown-linux-gnu/src'
    make[1]: *** [check-recursive] Error 1
    Makefile:12766: recipe for target 'check-recursive' failed
    make[1]: Leaving directory '/home/travis/build/bitcoin/bitcoin/build/bitcoin-x86_64-unknown-linux-gnu/src'
    make: *** [check-recursive] Error 1
    Makefile:775: recipe for target 'check-recursive' failed
    
  2. instagibbs commented at 9:34 PM on April 12, 2019: none

    Probably random but felt best to report since I've never seen one in a long time.

  3. instagibbs cross-referenced this on Apr 12, 2019 from issue CReserveKey should not allow passive key re-use, debug assert in destructor by instagibbs
  4. MarcoFalke commented at 9:39 PM on April 12, 2019: none

    Last time I reproduced them they were non-deterministic even with the same seed used

  5. sipa commented at 9:41 PM on April 12, 2019: contributor

    The randomness tests are known to fail in a small number of cases (there is an old PR about this, #367 but iirc it didn't actually improve things). @MarcoFalke It should be deterministic if the source code, seed, and test count are identical.

  6. gmaxwell commented at 9:41 PM on April 12, 2019: contributor

    @MarcoFalke Can you clarify what you mean? Unless there is a serious bug there should be no source of nondeterminism in the tests (other than the seed)

  7. instagibbs commented at 9:44 PM on April 12, 2019: none

    @MarcoFalke is the seed dumped anywhere?

  8. sipa commented at 9:49 PM on April 12, 2019: contributor

    @instagibbs From your output:

    test count = 64
    random seed = 54d904804b5e0db448021441be754c10
    
  9. jonasnick commented at 9:20 AM on April 13, 2019: contributor

    It can be reproduced:

    $ ./tests 1 54d904804b5e0db448021441be754c10
    test count = 3
    random seed = 54d904804b5e0db448021441be754c10
    src/tests.c:450: test condition failed: ((~x[m][shift]) << (64 - (1 << usebits))) == 0
    
  10. instagibbs commented at 12:11 PM on April 13, 2019: none

    Good to know, closing unless Marco has a non-reproducible failure with seed

  11. instagibbs closed this on Apr 13, 2019

  12. MarcoFalke cross-referenced this on Jul 13, 2021 from issue Make m_mempool optional in CChainState by jamesob
  13. real-or-random cross-referenced this on Jul 17, 2023 from issue subtree: update libsecp256k1 to latest master by fanquake

github-metadata-mirror

This is a metadata mirror of the GitHub repository bitcoin-core/secp256k1. This site is not affiliated with GitHub. Content is generated from a GitHub metadata backup.
generated: 2026-04-23 01:15 UTC

This site is hosted by @0xB10C
More mirrored repositories can be found on mirror.b10c.me