Don't let find() mess up dircache (#531)

Author: martindurant

.github/workflows/ci.yml

@@ -49,8 +49,6 @@ jobs:
     name: lint
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3.1.0
+      - uses: actions/checkout@v3
       - uses: actions/setup-python@v4
-        with:
-          python-version: "3.9"
-      - uses: pre-commit/[email protected]
+      - uses: pre-commit/[email protected]

.isort.cfg

@@ -0,0 +1,2 @@
+[settings]
+known_third_party = aiohttp,click,decorator,fsspec,fuse,google,google_auth_oauthlib,pytest,requests,setuptools

.pre-commit-config.yaml

@@ -18,15 +18,11 @@ repos:
     rev: 6.0.0
     hooks:
       - id: flake8
-        files: gcsfs/
-  - repo: https://github.com/pycqa/isort
-    rev: 5.10.1
+  - repo: https://github.com/asottile/seed-isort-config
+    rev: v2.2.0
     hooks:
-      - id: isort
-        args: ["--profile", "black", "--filter-files"]
-  - repo: https://github.com/asottile/pyupgrade
-    rev: v3.2.3
+      - id: seed-isort-config
+  - repo: https://github.com/pre-commit/mirrors-isort
+    rev: v5.7.0
     hooks:
-      - id: pyupgrade
-        args:
-          - --py37-plus
+      - id: isort

gcsfs/core.py

@@ -1220,7 +1220,10 @@ async def _find(
                     "size": 0,
                 }
 
-                cache_entries.setdefault(parent, []).append(previous)
+                print(parent, previous)
+                listing = cache_entries.setdefault(parent, [])
+                if previous not in listing:
+                    listing.append(previous)
 
                 previous = dirs[parent]
                 parent = self._parent(parent)

gcsfs/tests/test_core.py

@@ -14,13 +14,17 @@
 from fsspec.utils import seek_delimiter
 
 import gcsfs.checkers
+import gcsfs.tests.settings
 from gcsfs import __version__ as version
 from gcsfs.core import GCSFileSystem, quote
 from gcsfs.credentials import GoogleCredentials
 from gcsfs.tests.conftest import a, allfiles, b, csv_files, files, text_files
-from gcsfs.tests.settings import TEST_BUCKET, TEST_PROJECT, TEST_REQUESTER_PAYS_BUCKET
 from gcsfs.tests.utils import tempdir, tmpfile
 
+TEST_BUCKET = gcsfs.tests.settings.TEST_BUCKET
+TEST_PROJECT = gcsfs.tests.settings.TEST_PROJECT
+TEST_REQUESTER_PAYS_BUCKET = gcsfs.tests.settings.TEST_REQUESTER_PAYS_BUCKET
+
 
 def test_simple(gcs):
     assert not GoogleCredentials.tokens
@@ -1339,3 +1343,24 @@ def test_cp_two_files(gcs):
         target + "/file0",
         target + "/file1",
     ]
+
+
+def test_multiglob(gcs):
+    # #530
+    root = TEST_BUCKET
+
+    ggparent = root + "/t1"
+    gparent = ggparent + "/t2"
+    parent = gparent + "/t3"
+    leaf1 = parent + "/foo.txt"
+    leaf2 = parent + "/bar.txt"
+    leaf3 = parent + "/baz.txt"
+
+    gcs.touch(leaf1)
+    gcs.touch(leaf2)
+    gcs.touch(leaf3)
+    gcs.invalidate_cache()
+
+    assert gcs.ls(gparent, detail=False) == [f"{root}/t1/t2/t3"]
+    gcs.glob(ggparent + "/")
+    assert gcs.ls(gparent, detail=False) == [f"{root}/t1/t2/t3"]