Compressing Files (Zip and Tar archives)

diagram extracting an archive you did not create mermaid

  flowchart TD
  A["an archive from somewhere else"] --> B["namelist() -- read the names FIRST"]
  B --> C{"any entry with .. or a leading /?"}
  C -->|yes| D["it is trying to write outside the destination"]
  C -->|no| E["extract normally"]
  F["zipfile.extractall"] --> G["strips .. and leading slashes -- lands inside"]
  H["tarfile.extractall"] --> I["applies the 'data' filter"]
  I --> J["raises OutsideDestinationError rather than escaping"]
  D --> K["reject the archive; do not 'fix up' the path and continue"]

An archive is a list of names chosen by whoever built it, and those names can point outside the directory you are extracting into. Python's zipfile normalises them away; tarfile historically did not, which is why modern versions apply a filter and refuse. Either way the rule for anything downloaded is to look at the names before extracting.

What actually happens with a hostile archive — measured

I built an archive containing three members: ok.txt, ../escaped.txt and /abs.txt.

zipfile: extractall() wrote all three inside the destination, as ok.txt, escaped.txt and abs.txt. Nothing escaped — extractall strips leading slashes and .. segments.

tarfile: extractall() refused:

text

OutsideDestinationError: '../escaped.txt' would be extracted to ...

Python 3.14 applies the data filter by default. Older versions extracted such members happily, which is the original “tar slip” vulnerability.

So the common advice — “never use extractall, it is vulnerable to path traversal” — is not accurate for zipfile.extractall on any modern Python, and is now handled for tarfile too. What remains genuinely unsafe is a hand-rolled loop:

python

for name in z.namelist():                 # do not do this
    with open(os.path.join(dest, name), "wb") as f:
        f.write(z.read(name))

That has no sanitisation at all, and os.path.join(dest, "/abs.txt") discards dest entirely.

Zip with shutil (quick)

make_zip.py

import shutil
 
# creates my_backup.zip
shutil.make_archive("my_backup", "zip", root_dir="./demo")

Tar archives

make_tar.py

import tarfile
 
with tarfile.open("demo.tar.gz", "w:gz") as tar:
    tar.add("./demo", arcname="demo")

Extracting

extract_zip.py

import zipfile
 
with zipfile.ZipFile("my_backup.zip") as z:
    z.extractall("./restored")

Tips

• store archives outside the source folder
• validate extracted size/location

🧪 Try It Yourself

Exercise 1 – List Files with os.listdir

pch.exerciseTag pch.exerciseTitle

Exercise 2 – Join Paths with os.path.join

pch.exerciseTag pch.exerciseTitle

Exercise 3 – Write and Read a File

pch.exerciseTag pch.exerciseTitle