fs: directory: check first for symlink before content/directory

Fix T2842.

swh/fuse/fs/artifact.py

@@ -85,8 +85,20 @@ class Directory(FuseDirEntry):
                 else entry["perms"]
             )
 
-            # 1. Regular file
-            if swhid.object_type == CONTENT:
+            # 1. Symlink (check symlink first because condition is less restrictive)
+            if mode == DentryPerms.symlink:
+                target = ""
+                try:
+                    # Symlink target is stored in the blob content
+                    target = await self.fuse.get_blob(swhid)
+                except Exception:
+                    pass  # Ignore error and create a (broken) symlink anyway
+
+                yield self.create_child(
+                    FuseSymlinkEntry, name=name, target=target,
+                )
+            # 2. Regular file
+            elif swhid.object_type == CONTENT:
                 yield self.create_child(
                     Content,
                     name=name,
@@ -96,19 +108,11 @@ class Directory(FuseDirEntry):
                     # attributes without additional Software Heritage API call
                     prefetch=entry,
                 )
-            # 2. Regular directory
+            # 3. Regular directory
             elif swhid.object_type == DIRECTORY:
                 yield self.create_child(
                     Directory, name=name, mode=mode, swhid=swhid,
                 )
-            # 3. Symlink
-            elif mode == DentryPerms.symlink:
-                yield self.create_child(
-                    FuseSymlinkEntry,
-                    name=name,
-                    # Symlink target is stored in the blob content
-                    target=await self.fuse.get_blob(swhid),
-                )
             # 4. Submodule
             elif swhid.object_type == REVISION:
                 try:

swh/fuse/tests/data/config.py

@@ -15,10 +15,12 @@ def remove_swhid_prefix(swhid_str: str) -> str:
 REGULAR_FILE = "swh:1:cnt:61d3c9e1157203f0c4ed5165608d92294eaca808"
 # Directory
 ROOT_DIR = "swh:1:dir:c6dcbe9711ea6d5a31429a833a3d0c59cbbb2578"
-DIR_WITH_SUBMODULES = "swh:1:dir:80ae84abc6122c47aae597fde99645f8663d1aba"
-# Revision
-ROOT_REV = "swh:1:rev:b8cedc00407a4c56a3bda1ed605c6fc166655447"
-SUBMODULES = [
+# Directory (symlinks entries)
+CNT_SYMLINK = "swh:1:cnt:76219eb72e8524f15c21ec93b9b2592da49b5460"  # from rust-clippy
+DIR_WITH_CNT_SYMLINK = "swh:1:dir:8f04a97403e13995c8790aef9158c8981b026223"
+DIR_SYMLINK = "swh:1:cnt:e8310385c56dc4bbe379f43400f3181f6a59f260"  # from Limnoria
+DIR_WITH_DIR_SYMLINK = "swh:1:dir:0653a2af52ebb771ef8ce0388c942d77f9d7aee8"
+REV_SYMLINKS = [  # from rust compiler
     "swh:1:rev:87dd6843678575f8dda962f239d14ef4be14b352",
     "swh:1:rev:1a2390247ad6d08160e0dd74f40a01a9578659c2",
     "swh:1:rev:4d78994915af1bde9a95c04a8c27d8dca066232a",
@@ -27,6 +29,9 @@ SUBMODULES = [
     "swh:1:rev:1c2bd024d13f8011307e13386cf1fea2180352b5",
     "swh:1:rev:92baf7293dd2d418d2ac4b141b0faa822075d9f7",
 ]
+DIR_WITH_REV_SYMLINK = "swh:1:dir:80ae84abc6122c47aae597fde99645f8663d1aba"
+# Revision
+ROOT_REV = "swh:1:rev:b8cedc00407a4c56a3bda1ed605c6fc166655447"
 REV_SMALL_HISTORY = "swh:1:rev:37426e42cf78a43779312d780eecb21a64006d99"
 # Release
 ROOT_REL = "swh:1:rel:874f7cbe352033cac5a8bc889847da2fe1d13e9f"
@@ -79,9 +84,13 @@ TARGET_DIR = "swh:1:dir:b24d39c928b9c3f440f8e2ec06c78f43d28d87d6"
 ALL_ENTRIES = [
     REGULAR_FILE,
     ROOT_DIR,
-    DIR_WITH_SUBMODULES,
+    CNT_SYMLINK,
+    DIR_SYMLINK,
+    *REV_SYMLINKS,
+    DIR_WITH_CNT_SYMLINK,
+    DIR_WITH_DIR_SYMLINK,
+    DIR_WITH_REV_SYMLINK,
     ROOT_REV,
-    *SUBMODULES,
     REV_SMALL_HISTORY,
     ROOT_REL,
     REL_TARGET_CNT,

swh/fuse/tests/data/gen-api-data.py

@@ -36,11 +36,21 @@ from swh.model.identifiers import (
 API_URL_real = "https://archive.softwareheritage.org/api/1"
 API_URL_test = "https://invalid-test-only.archive.softwareheritage.org/api/1"
 
+# Use your own API token to lift rate limiting. Note: this is not necessary to generate
+# the API data only once but can be useful when re-generating it multiple times.
+API_TOKEN = ""
+
+
 MOCK_ARCHIVE: Dict[str, Any] = {}
 # Temporary map (swhid -> metadata) to ease data generation
 METADATA: Dict[SWHID, Any] = {}
 
 
+def get_from_api(endpoint: str) -> str:
+    headers = {"Authorization": f"Bearer {API_TOKEN}"} if API_TOKEN else {}
+    return requests.get(f"{API_URL_real}/{endpoint}", headers=headers).text
+
+
 def get_short_type(object_type: str) -> str:
     short_type = {
         CONTENT: "cnt",
@@ -61,10 +71,8 @@ def generate_archive_web_api(
 
     url = swhid_to_web_url(swhid, raw)
 
-    if raw:
-        data = requests.get(f"{API_URL_real}/{url}").text
-    else:
-        data = requests.get(f"{API_URL_real}/{url}").text
+    data = get_from_api(url)
+    if not raw:
         data = json.loads(data)
 
     MOCK_ARCHIVE[url] = data
@@ -146,7 +154,7 @@ swh:1:rev:d6b7c96c3eb29b9244ece0c046d3f372ff432d04 swh:1:rev:c01efc669f09508b55e
 
 def generate_origin_archive_web_api(url: str):
     url_visits = f"origin/{url}/visits/"
-    data = requests.get(f"{API_URL_real}/{url_visits}").text
+    data = get_from_api(url_visits)
     data = json.loads(data)
     MOCK_ARCHIVE[url_visits] = data
     # Necessary since swh-fuse will check the origin URL using the get/ endpoint

swh/fuse/tests/test_directory.py

 import os
 
 from swh.fuse.tests.common import check_dir_name_entries
-from swh.fuse.tests.data.config import DIR_WITH_SUBMODULES, ROOT_DIR
+from swh.fuse.tests.data.config import (
+    DIR_WITH_CNT_SYMLINK,
+    DIR_WITH_DIR_SYMLINK,
+    DIR_WITH_REV_SYMLINK,
+    ROOT_DIR,
+)
 
 
 def test_list_dir(fuse_mntdir):
@@ -19,8 +24,14 @@ def test_access_subdir(fuse_mntdir):
     assert dir_path.is_dir()
 
 
-def test_access_submodule_entries(fuse_mntdir):
-    dir_path = fuse_mntdir / "archive" / DIR_WITH_SUBMODULES
+def test_access_symlinks(fuse_mntdir):
+    cnt_sym_path = fuse_mntdir / "archive" / DIR_WITH_CNT_SYMLINK / "LICENSE-MIT"
+    assert os.readlink(cnt_sym_path) == "../LICENSE-MIT"
+
+    dir_sym_path = fuse_mntdir / "archive" / DIR_WITH_DIR_SYMLINK / "supybot"
+    assert os.readlink(dir_sym_path) == "src"
+
+    rev_sym_dir_path = fuse_mntdir / "archive" / DIR_WITH_REV_SYMLINK
     submodules = {
         "book": "swh:1:rev:87dd6843678575f8dda962f239d14ef4be14b352",
         "edition-guide": "swh:1:rev:1a2390247ad6d08160e0dd74f40a01a9578659c2",
@@ -32,4 +43,4 @@ def test_access_submodule_entries(fuse_mntdir):
     }
     for filename, swhid in submodules.items():
         target = f"../../archive/{swhid}"
-        assert os.readlink(dir_path / filename) == target
+        assert os.readlink(rev_sym_dir_path / filename) == target

[Phabricator Migration user]

Some references in the commit message have been migrated:

• T2842 is now #2842 (closed)