Test: Add script vectors for interpreter edges #36458

pull benthecarman wants to merge 1 commits into bitcoin:master from benthecarman:script-vectors-interpreter-edges changing 2 files +38 −1
  1. benthecarman commented at 7:35 PM on October 7, 2026: contributor

    Add script_tests.json and tx_invalid.json vectors for edge cases that other implementations have gotten wrong, including CSV with -2, IFDUP on negative zero, ROLL with duplicate values, a taproot output key that is not a valid x coordinate, and single-flag STRICTENC, LOW_S and NULLFAIL cases.

    Some policy cases were only caught by the tx_valid harness adding flags back, which implementations copying the JSON don't replicate.

    Related: #36455

    These were discovered by finding gaps in an alternate implementation (bitcoin-s)

  2. Test: Add script vectors for interpreter edges
    Add script_tests.json and tx_invalid.json vectors for edge cases that
    other implementations have gotten wrong, including CSV with -2, IFDUP
    on negative zero, ROLL with duplicate values, a taproot output key
    that is not a valid x coordinate, and single-flag STRICTENC, LOW_S and
    NULLFAIL cases.
    
    Some policy cases were only caught by the tx_valid harness adding
    flags back, which implementations copying the JSON don't replicate.
    5d2fb8d309
  3. DrahtBot added the label Tests on Oct 7, 2026
  4. DrahtBot commented at 7:35 PM on October 7, 2026: contributor

    <!--e57a25ab6845829454e8d69fc972939a-->

    The following sections might be updated with supplementary metadata relevant to reviewers and maintainers.

    <!--006a51241073e994b41acfe9ec718e94-->

    External sites

    <!--021abf342d371248e50ceaed478a90ca-->

    Reviews

    See the guideline and AI policy for information on the review process.

    Type Reviewers
    ACK darosior, sedited

    If your review is incorrectly listed, please copy-paste <code>&lt;!--meta-tag:bot-skip--&gt;</code> into the comment that the bot should ignore.

    <!--5faf32d7da4f0f540f40219e4f7537a3-->

  5. fametrano commented at 9:36 AM on October 9, 2026: contributor

    I ran the new vectors at 5d2fb8d through btclib, a Python implementation I maintain. Every new script_tests.json row gives the expected result and error code. Each of the 4 new tx_invalid.json rows fails under its listed flag with the error its comment describes, and is valid with no flags.

  6. fanquake commented at 6:22 AM on October 16, 2026: member
  7. darosior commented at 2:49 PM on October 22, 2026: member

    Concept ACK. It's useful to know what re-implementations got wrong, as it's often non-obvious stuff we can better cover in our automated testing.

  8. in src/test/data/tx_invalid.json:239 in 5d2fb8d309
     235 | @@ -224,6 +236,10 @@
     236 |  [[["0000000000000000000000000000000000000000000000000000000000000100", 0, "-1 CHECKSEQUENCEVERIFY"]],
     237 |  "020000000100010000000000000000000000000000000000000000000000000000000000000000000000000040000100000000000000000000000000", "CHECKSEQUENCEVERIFY"],
     238 |  
     239 | +["Argument -2 with by-blockheight txin.nSequence=0, even though -2 has the disable flag bit set"],
    


    darosior commented at 8:49 PM on October 22, 2026:

    Isn't this one redundant with the script_tests.json one you added above?


    benthecarman commented at 9:31 PM on October 22, 2026:

    yeah i guess, being redundant doesn't really hurt here though

  9. in src/test/data/script_tests.json:2616 in 5d2fb8d309
    2612 | @@ -2616,6 +2613,8 @@
    2613 |  ["CHECKSEQUENCEVERIFY tests"],
    2614 |  ["", "CHECKSEQUENCEVERIFY", "CHECKSEQUENCEVERIFY", "INVALID_STACK_OPERATION", "CSV automatically fails on an empty stack"],
    2615 |  ["-1", "CHECKSEQUENCEVERIFY", "CHECKSEQUENCEVERIFY", "NEGATIVE_LOCKTIME", "CSV automatically fails if stack top is negative"],
    2616 | +["-2", "CHECKSEQUENCEVERIFY", "CHECKSEQUENCEVERIFY", "NEGATIVE_LOCKTIME", "CSV fails for any negative operand, even though -2 has the disable flag bit set"],
    


    darosior commented at 8:51 PM on October 22, 2026:

    Heh, nice. (Though the -1 vector would also catch this, but good to have it spelled out.)

  10. in src/test/data/script_tests.json:1041 in 5d2fb8d309
    1037 | @@ -1036,6 +1038,8 @@
    1038 |  ["-2147483648", "1ADD 1", "P2SH,STRICTENC", "SCRIPTNUM", "Because we use a sign bit, -2147483648 is also 5 bytes"],
    1039 |  ["2147483647", "1ADD 1SUB 1", "P2SH,STRICTENC", "SCRIPTNUM", "We cannot do math on 5-byte integers, even if the result is 4-bytes"],
    1040 |  ["2147483648", "1SUB 1", "P2SH,STRICTENC", "SCRIPTNUM", "We cannot do math on 5-byte integers, even if the result is 4-bytes"],
    1041 | +["0x05 0xff00000000", "1ADD DROP 1", "P2SH,STRICTENC", "SCRIPTNUM", "We cannot do math on 5-byte integers, even if the value fits in 1 byte when minimally encoded"],
    


    darosior commented at 8:52 PM on October 22, 2026:

    Nice.

  11. darosior approved
  12. darosior commented at 8:53 PM on October 22, 2026: member

    ACK 5d2fb8d3098703bc4d23d582eb4987cb2dbeda7a

  13. sedited approved
  14. sedited commented at 9:07 PM on November 4, 2026: contributor

    LGTM ACK 5d2fb8d3098703bc4d23d582eb4987cb2dbeda7a, but waiting on resolution here: #36455#pullrequestreview-5616742117

  15. sedited commented at 11:16 AM on November 5, 2026: contributor

    Merging this. If there is more, it can just go into another pull request.

  16. sedited merged this on Nov 5, 2026
  17. sedited closed this on Nov 5, 2026

  18. benthecarman deleted the branch on Nov 5, 2026

github-metadata-mirror

This is a metadata mirror of the GitHub repository bitcoin/bitcoin. This site is not affiliated with GitHub. Content is generated from a GitHub metadata backup.
generated: 2026-11-18 16:51 UTC

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