fix: make hash_paths work when base_path is set

Since #664 we started returning paths relative to the repository root from `_find_matching_files`. This meant that when we join `base_path` with `path` in `hash_paths`, we end up with `base_path` in there twice.

Fixing this without breaking `mozpath.match` means we need to join together the repository root and matched files after `mozpatch.match`. This, in turn, requires that some tests are able to call `get_repository`, which requires faking a repository being present.

Author: bhearsum

src/taskgraph/util/hash.py

@@ -38,20 +38,20 @@ def hash_paths(base_path, patterns):
             found.update(matches)
         else:
             raise Exception(f"{pattern} did not match anything")
-    for path in sorted(found):
+    for (abs_path, path) in sorted(found):
         h.update(
-            f"{hash_path(mozpath.abspath(mozpath.join(base_path, path)))} {mozpath.normsep(path)}\n".encode()
+            f"{hash_path(abs_path)} {mozpath.normsep(path)}\n".encode()
         )
     return h.hexdigest()
 
 
 @functools.lru_cache(maxsize=None)
 def _find_matching_files(base_path, pattern):
-    files = _get_all_files(base_path)
-    return [path for path in files if mozpath.match(path, pattern)]
+    repo = get_repository(os.getcwd())
+    files = _get_all_files(base_path, repo)
+    return [(mozpath.join(repo.path, path), path) for path in files if mozpath.match(path, pattern)]
 
 
 @functools.lru_cache(maxsize=None)
-def _get_all_files(base_path):
-    repo = get_repository(os.getcwd())
+def _get_all_files(base_path, repo):
     return repo.get_tracked_files(base_path)

test/test_transforms_run_toolchain.py

@@ -2,12 +2,14 @@
 Tests for the 'toolchain' transforms.
 """
 
+import os.path
 from pprint import pprint
 
 import pytest
 
 from taskgraph.transforms.run import make_task_description
 from taskgraph.util.templates import merge
+from taskgraph.util.vcs import GitRepository
 
 TASK_DEFAULTS = {
     "description": "fake description",
@@ -33,14 +35,23 @@ def run_task_using(mocker, run_transform):
         "taskcluster/scripts/toolchain/run.sh",
         "taskcluster/scripts/toolchain/run.ps1",
     ]
+    # fake a repository during the test, which is needed for `_get_all_files` to
+    # work correctly
+    mocker.patch("taskgraph.util.hash.get_repository", new=lambda path: GitRepository(path))
 
     def inner(task):
         task = merge(TASK_DEFAULTS, task)
         m = mocker.patch(
             "taskgraph.transforms.run.toolchain.configure_taskdesc_for_run"
         )
-        run_transform(make_task_description, task)
-        return m.call_args[0]
+        repo_dir = os.path.abspath(os.path.join(os.path.dirname(__file__), "data"))
+        cwd = os.getcwd()
+        try:
+            os.chdir(repo_dir)
+            run_transform(make_task_description, task)
+            return m.call_args[0]
+        finally:
+            os.chdir(cwd)
 
     return inner