tarfile: Make it possible to extract nested tarfiles in memory.

woju · woju · commit f2217df5b275 · 2025-07-14T11:01:56.000+02:00
FileSection.skip() (see below the diff) uses 2-argument readinto, so
attempting to recursively extract archives throws an error. This commit
adds optional second argument to fix this problem. After this commit, it
is possible to extract nested archives in roughly this fashion:

    with open(path, 'rb') as file:
        tar_outer = tarfile.TarFile(fileobj=file)
        for ti_outer in tar_outer:
            tar_inner = tarfile.TarFile(
                fileobj=tar_outer.extractfile(ti_outer))
            for ti_inner in tar_inner:
                ...

Nested archives are used in some embedded contexts, for example Mender
artifacts.

Signed-off-by: Wojciech Porczyk &lt;wojciech.porczyk@connectpoint.pl&gt;
diff --git a/python-stdlib/tarfile/tarfile/__init__.py b/python-stdlib/tarfile/tarfile/__init__.py
@@ -41,11 +41,13 @@ def read(self, sz=65536):
         self.content_len -= sz
         return data
 
-    def readinto(self, buf):
+    def readinto(self, buf, size=None):
         if self.content_len == 0:
             return 0
         if len(buf) > self.content_len:
             buf = memoryview(buf)[: self.content_len]
+        if size is not None and len(buf) > size:
+            buf = memoryview(buf)[:size]
         sz = self.f.readinto(buf)
         self.content_len -= sz
         return sz
