)]}'
{
  "commit": "678c841cbe0a8980148d97913c25c730744261d8",
  "tree": "83b7d11a21d31e874287854b17337fe500a309c1",
  "parents": [
    "4e446f27d05996309eec2ac1c2a71b66020cdfb2"
  ],
  "author": {
    "name": "David Benjamin",
    "email": "davidben@google.com",
    "time": "Fri Aug 10 08:53:34 2018 -0500"
  },
  "committer": {
    "name": "CQ bot account: commit-bot@chromium.org",
    "email": "commit-bot@chromium.org",
    "time": "Fri Aug 10 14:18:28 2018 +0000"
  },
  "message": "Use -flto\u003dthin in the CFI bot.\n\nThe CFI bot is currently failing on a mysterious error message, coming from the\nrecent clang roll.\n\n  Called function is not the same type as the call!\n    call void @EVP_MD_CTX_init(%struct.env_md_ctx_st* %8), !dbg !72123\n  LLVM ERROR: Broken function found, compilation aborted!\n\nChromium actually passes -flto\u003dthin, which seems to avoid the error, testing\nlocally. Why it does, I haven\u0027t the slightest clue. The offending calls to\nEVP_MD_CTX_init (and EVP_MD_CTX_cleanup) are those buried in\nbssl::ScopedEVP_MD_CTX. However, not all calls are problematic, only the one in\ntest_config.cc. What\u0027s more, if I add a call in async_bio.cc, linked into all\nthe same targets, the copy in test_config.cc is suddenly fine!?\n\nMaybe there\u0027s just a bug in the LTO logic that ThinLTO avoids...\n\nChange-Id: I5266eec75edea2a38dee8ad5591db8d65d3bdede\nReviewed-on: https://boringssl-review.googlesource.com/30805\nReviewed-by: Steven Valdez \u003csvaldez@google.com\u003e\nCommit-Queue: David Benjamin \u003cdavidben@google.com\u003e\nCQ-Verified: CQ bot account: commit-bot@chromium.org \u003ccommit-bot@chromium.org\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "7ba4a951ab2eed703803c48ff9ccdf5755101193",
      "old_mode": 33188,
      "old_path": "CMakeLists.txt",
      "new_id": "5d067e159d3815e302c3e1bd4260149a2621e562",
      "new_mode": 33188,
      "new_path": "CMakeLists.txt"
    }
  ]
}