bpo-45006 Add a data_offset field to ZipInfo

Lib/test/test_zipfile.py

@@ -3196,6 +3196,19 @@ def test_inheritance(self, alpharep):
             file = cls(alpharep).joinpath('some dir').parent
             assert isinstance(file, cls)
 
+    def test_dataoffset(self):
+        data = io.BytesIO()
+        with zipfile.ZipFile(data, 'w', compression=zipfile.ZIP_STORED) as zfp:
+            zfp.writestr("a/b/c.txt", "random data for c")
+            zfp.writestr("a/b/b.txt", "random data for b")
+            zfp.writestr("a/b/c/d.txt", "random data for d")
+            zfp.writestr("a.txt", "random data for a")
+        zip_content = data.getvalue()
+        with zipfile.ZipFile(data, 'r') as zfp:
+            for entry in zfp.infolist():
+                expected_offset = zip_content.index(zfp.read(entry))
+                self.assertEqual(entry.data_offset, expected_offset)
+
 
 if __name__ == "__main__":
     unittest.main()

Lib/zipfile.py

@@ -364,6 +364,7 @@ class ZipInfo (object):
         'CRC',
         'compress_size',
         'file_size',
+        'data_offset',
         '_raw_time',
     )
 
@@ -406,6 +407,7 @@ def __init__(self, filename="NoName", date_time=(1980,1,1,0,0,0)):
         self.external_attr = 0          # External file attributes
         self.compress_size = 0          # Size of the compressed file
         self.file_size = 0              # Size of the uncompressed file
+        self.data_offset = None         # Offset to beginning of compressed data
         # Other attributes are set by class ZipFile:
         # header_offset         Byte offset to the file header
         # CRC                   CRC-32 of the uncompressed file
@@ -1340,6 +1342,16 @@ def __repr__(self):
         result.append('>')
         return ''.join(result)
 
+    def _ComputeDataOffset(self, zinfo: ZipInfo):
+        if self.fp.seekable():
+            self.fp.seek(zinfo.header_offset)
+            fheader = struct.unpack(structFileHeader, self.fp.read(sizeFileHeader))
+            if fheader[_FH_SIGNATURE] != stringFileHeader:
+                return
+            return zinfo.header_offset + \
+                fheader[_FH_FILENAME_LENGTH] + \
+                fheader[_FH_EXTRA_FIELD_LENGTH] + sizeFileHeader
+
     def _RealGetContents(self):
         """Read in the table of contents for the ZIP file."""
         fp = self.fp
@@ -1406,6 +1418,7 @@ def _RealGetContents(self):
 
             x._decodeExtra()
             x.header_offset = x.header_offset + concat
+            x.data_offset = self._ComputeDataOffset(x)
             self.filelist.append(x)
             self.NameToInfo[x.filename] = x
 

Misc/NEWS.d/next/Library/2021-08-26-02-39-16.bpo-45006.ld-Rwq.rst

@@ -0,0 +1 @@
+class ``zipfile.ZipInfo`` now exposes a data_offset field to indicate offset to beginning of compressed data for this entry.