)]}'
{
  "commit": "e4298a38e617f2269d16270883b647aeb836b99e",
  "tree": "1ec91d5e09fbcff4f4fbd09288a473f4874c0cbf",
  "parents": [
    "e52d85596c0862a4f0bc1a40de84756017afc8b8"
  ],
  "author": {
    "name": "Jacob Rutherford",
    "email": "jruthe@fuchsia.infra.roller.google.com",
    "time": "Wed Nov 12 15:17:47 2025 -0800"
  },
  "committer": {
    "name": "Copybara-Service",
    "email": "copybara-worker@google.com",
    "time": "Wed Nov 12 15:19:55 2025 -0800"
  },
  "message": "[roll] Roll fuchsia [debugger] Capture full stacks on some exceptions\n\nThis changes exception handling in the backend to have different\nbehavior depending on the type of exception we\u0027re handling. General\nexceptions (e.g. crashes, page faults, etc) now always send full stacks,\nwhich helps clients make decisions automatically.\n\nAlso changed are software breakpoints exceptions that are not associated\nwith any of our breakpoints. These can come from things like the gTest\nframework\u0027s assertion failures when running with\n--gtest_break_on_failure, for example, in addition to being manually\ntyped by a developer.\n\nWe want cases like these to already have a complete stack trace, when\npossible. In cases where unwind tables have been stripped from the\ndebugged binary, we may not be able to make full progress with the\nsynchronous target unwinder, and the client will have to do unwinding\nwith the debug info on the host anyway, so this is best effort.\n\nOriginal-Bug: 444070494\nOriginal-Reviewed-on: https://fuchsia-review.googlesource.com/c/fuchsia/+/1421434\nOriginal-Revision: 27b56e72dded5b11266f1b4463f7c490a4876e4c\nGitOrigin-RevId: 796bfc608a02457db150db2eea194517a847588c\nChange-Id: I46880fe663e3d7ffdb77c099f0dbcc3a42097ac2\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "9c6e87d5bd1abf9d6581bbf3e5b5a82bfcfed428",
      "old_mode": 33188,
      "old_path": "stem",
      "new_id": "3dc9bb744bb4a62a2641d07187462a8f65544f13",
      "new_mode": 33188,
      "new_path": "stem"
    }
  ]
}