Add a --github-repository argument to the local repo cache

Author: vrigal

bot/code_review_bot/__init__.py

@@ -191,16 +191,30 @@ def hash(self):
         We make the assumption that the message does not contain the line number
         If an error occurs reading the file content (locally or remotely), None is returned
         """
+        from code_review_bot.revisions import GithubRevision, PhabricatorRevision
+
         assert self.revision is not None, "Missing revision"
 
+        local_repository = None
+        if (
+            isinstance(self.revision, PhabricatorRevision)
+            and settings.mercurial_cache_checkout
+        ):
+            local_repository = settings.mercurial_cache_checkout
+        elif isinstance(self.revision, GithubRevision):
+            assert (
+                settings.gtihub_cache
+            ), "Github cache repository is mandatory to analyse a github revision"
+            local_repository = settings.gtihub_cache
+        else:
+            raise NotImplementedError
+
         # Build the hash only if the file is not autogenerated.
         # An autogenerated file resides in the build directory that it has the
         #  format `obj-x86_64-pc-linux-gnu`
         file_content = None
         if "/obj-" not in self.path:
-            file_content = self.revision.get_file_content(
-                self.path, settings.mercurial_cache_checkout
-            )
+            file_content = self.revision.get_file_content(self.path, local_repository)
 
         if file_content is None:
             self._hash = None

bot/code_review_bot/cli.py

@@ -26,7 +26,7 @@
 )
 from code_review_bot.config import settings
 from code_review_bot.report import get_reporters
-from code_review_bot.revisions import PhabricatorRevision, Revision
+from code_review_bot.revisions import GithubRevision, PhabricatorRevision, Revision
 from code_review_bot.tools.libmozdata import setup as setup_libmozdata
 from code_review_bot.tools.log import init_logger
 from code_review_bot.workflow import Workflow
@@ -64,6 +64,13 @@ def parse_cli():
         type=Path,
         default=None,
     )
+    parser.add_argument(
+        "--github-repository",
+        help="Optional path to a up-to-date github repository matching the analyzed revision.\n"
+        "This argument is required for Github reviusions in order to compute issues' hashes based on file content.",
+        type=Path,
+        default=None,
+    )
     parser.add_argument("--taskcluster-client-id", help="Taskcluster Client ID")
     parser.add_argument("--taskcluster-access-token", help="Taskcluster Access token")
     return parser.parse_args()
@@ -116,6 +123,7 @@ def main():
         taskcluster.secrets["repositories"],
         taskcluster.secrets["ssh_key"],
         args.mercurial_repository,
+        args.github_repository,
     )
 
     # Setup statistics
@@ -205,6 +213,11 @@ def main():
             )
             return 1
 
+    if isinstance(revision, GithubRevision):
+        assert (
+            args.github_repository is not None
+        ), "Girhub revision analysis requires the --github-repository argument to be set"
+
     # Run workflow according to source
     w = Workflow(
         reporters,

bot/code_review_bot/config.py

@@ -58,6 +58,7 @@ def __init__(self):
 
         # Cache to store whole repositories
         self.mercurial_cache = None
+        self.github_cache = None
 
         # SSH Key used to push on try
         self.ssh_key = None
@@ -78,6 +79,7 @@ def setup(
         repositories,
         ssh_key=None,
         mercurial_cache=None,
+        github_cache=None,
     ):
         # Detect source from env
         if "TRY_TASK_ID" in os.environ and "TRY_TASK_GROUP_ID" in os.environ: