)]}'
{
  "commit": "dbb357dae2227250b6cbf110d8a02f83934eb5c9",
  "tree": "08e2943ffdfeb16a09a3b21b2b88e88c1c4e6228",
  "parents": [
    "59830ec1d93be038ffe883b69229753bb18ff59e"
  ],
  "author": {
    "name": "Martin Storsjö",
    "email": "martin@martin.st",
    "time": "Fri Nov 21 16:41:04 2025 +0200"
  },
  "committer": {
    "name": "Martin Storsjö",
    "email": "martin@martin.st",
    "time": "Fri Nov 28 23:03:29 2025 +0200"
  },
  "message": "github: Cache the GNU downloads\n\nThe downloads from gnu.org have occasionally failed (probably due\nto being under load from AI bots).\neea8f577e02a08d90e7a7c1c60251e7760f6a880 helped a bit by doing the\ndownload from a mirror, but despite that, these downloads still\nfail occasionally.\n\nTo avoid having to fetch the source tarballs for each build (both\nfor reducing unnecessary load on the upstream servers, and for\nimproved reliability), store them in github action caches.\n\nUnless we had a cache hit (a fully matching cache key), download\nall the tarballs we\u0027re interested in (which then gets stored,\nimplicitly, the end of the job) upfront. This uses the same cache\nkey, and downloads the same tarballs, for both the initial \"gcc\" jobs\n(that only need binutils and gcc) and the later \"gcc-cross\" (which\nalso need gmp, mpfr and mpc), to allow sharing a single cache\nfor both of them.\n\nSigned-off-by: Martin Storsjö \u003cmartin@martin.st\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "81846460b8bb02dc0f8ff086f3b8b9980bdec8c3",
      "old_mode": 33188,
      "old_path": ".github/workflows/build.yml",
      "new_id": "8a9110f0a79e682ea98ae0e20d8c4324410e0af2",
      "new_mode": 33188,
      "new_path": ".github/workflows/build.yml"
    }
  ]
}