fix(handlers): handle dangling symlinks in MultiFile handlers.

MultiFile handlers would collect files within a directory corresponding
to a specific schema without checking if those files are actually
present.

For example, a directory could contain dangling symlinks with a name
corresponding to the glob search. This would lead to FileNotFoundError
being thrown by the multi-file handlers.

Author: qkaiser

tests/integration/archive/sevenzip/__input__/multi-sevenzip-dangling-symlinks.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:70f2e1d6ef9b9954fde7f76de1026dce3726ef16e3eb5b944d03f45394675e11
+size 10240

tests/integration/archive/sevenzip/__output__/multi-sevenzip-dangling-symlinks.tar_extract/dangling-symlink.7z.001

@@ -0,0 +1 @@
+hello

tests/integration/archive/sevenzip/__output__/multi-sevenzip-dangling-symlinks.tar_extract/dangling-symlink.7z.002

@@ -0,0 +1 @@
+hello

tests/integration/archive/sevenzip/__output__/multi-sevenzip-dangling-symlinks.tar_extract/dangling-symlink.7z.003

@@ -0,0 +1 @@
+hello

tests/integration/compression/gzip/__input__/multi-gzip-dangling-symlinks.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4e4c5188742aa0848dd13878d4f996465dddd3a1c5376cb3af872149d714daa0
+size 10240

tests/integration/compression/gzip/__output__/multi-gzip-dangling-symlinks.tar_extract/dangling-symlink.gz.a

@@ -0,0 +1 @@
+hello

tests/integration/compression/gzip/__output__/multi-gzip-dangling-symlinks.tar_extract/dangling-symlink.gz.b

@@ -0,0 +1 @@
+hello

tests/integration/compression/gzip/__output__/multi-gzip-dangling-symlinks.tar_extract/dangling-symlink.gz.c

@@ -0,0 +1 @@
+hello

unblob/handlers/archive/sevenzip.py

@@ -106,6 +106,12 @@ class MultiVolumeSevenZipHandler(DirectoryHandler):
     PATTERN = Glob("*.7z.001")
 
     def calculate_multifile(self, file: Path) -> Optional[MultiFile]:
+        paths = sorted(
+            [p for p in file.parent.glob(f"{file.stem}.*") if p.resolve().exists()]
+        )
+        if not paths:
+            return None
+
         with file.open("rb") as f:
             header_data = f.read(HEADER_SIZE)
 
@@ -117,8 +123,6 @@ def calculate_multifile(self, file: Path) -> Optional[MultiFile]:
         size = calculate_sevenzip_size(header)
         logger.debug("Sevenzip header", header=header, size=size, _verbosity=3)
 
-        paths = sorted(file.parent.glob(f"{file.stem}.*"))
-
         files_size = sum(path.stat().st_size for path in paths)
         logger.debug(
             "Multi-volume files", paths=paths, files_size=files_size, _verbosity=2

unblob/handlers/compression/gzip.py

@@ -166,12 +166,14 @@ def is_valid_gzip(self, path: Path) -> bool:
         return True
 
     def calculate_multifile(self, file: Path) -> Optional[MultiFile]:
-        paths = sorted(file.parent.glob(f"{file.stem}.*"))
+        paths = sorted(
+            [p for p in file.parent.glob(f"{file.stem}.*") if p.resolve().exists()]
+        )
 
         # we 'discard' paths that are not the first in the ordered list,
         # otherwise we will end up with colliding reports, one for every
         # path in the list.
-        if file != paths[0]:
+        if not paths or file != paths[0]:
             return None
 
         if self.is_valid_gzip(file):