The `File is not a zip file` error for the output of `git show` by GitPython-CodePudding

The script to reproduce the issue

Save this code as a shell script and run it. The code should report the File is not a zip file error.

#!/bin/bash

set -eu

mkdir foo
cd foo

pip install --user GitPython

echo foo > a
zip a.zip a

# -t option validates the zip file.
# See https://unix.stackexchange.com/questions/197127/test-integrity-of-zip-file
unzip -t a.zip

git init
git add a.zip
git commit -m 'init commit'

cat << EOF > test.py
from git import Repo
import zipfile
from io import StringIO

repo = Repo('.', search_parent_directories=True)

raw = repo.git.show("HEAD:a.zip")

z = zipfile.ZipFile(StringIO(raw), "r")
EOF

python3 test.py

Original Question

I'm writing a Krita plugin for viewing files in previous commits of a Git repository, and I want to get the thumbnail file of a Krita file. To do so, I'm trying to get a file with git show, unzip it because a Krita file is a Zip file, and get preview.png or mergedimage.png.

%unzip image.kra
Archive:  image.kra
 extracting: mimetype                
  inflating: maindoc.xml             
  inflating: documentinfo.xml        
  inflating: preview.png             
  inflating: image/layers/layer2     
  inflating: image/layers/layer2.defaultpixel  
  inflating: image/layers/layer2.icc  
  inflating: image/annotations/icc   
  inflating: mergedimage.png         
  inflating: image/animation/index.xml

We can get the .kra file from the Git repository as str with GitPython. However, I can't parse the file with zipfile.ZipFile as it says File is not a zip file. (This code is based on this SO answer)

from git import Repo
import zipfile
from io import StringIO

repo = Repo('.', search_parent_directories=True)

raw = repo.git.show("HEAD~:image.kra")

z = zipfile.ZipFile(StringIO(raw), "r")

will emit

Traceback (most recent call last):
  File "/home/hiroki/krita_question/test.py", line 11, in <module>
    z = zipfile.ZipFile(StringIO(raw), "r")
  File "/usr/lib/python3.9/zipfile.py", line 1257, in __init__
    self._RealGetContents()
  File "/usr/lib/python3.9/zipfile.py", line 1324, in _RealGetContents
    raise BadZipFile("File is not a zip file")
zipfile.BadZipFile: File is not a zip file

I believe this is a valid Krita file because I can restore the file with git show on the command line. So,

%git show HEAD~:image.kra > prev.kra
%krita prev.kra

works correctly. Unzipping the file works too.

Why can't I parse the git show output as a Zip file?

git log --stat|grep -v 'Author':

commit b96d915862b39a204a9f4350e7e56634b6fcfe0b
Date:   Wed Mar 30 14:44:02 2022  0900

    chore: add

 ls | 231                                                                      
 1 file changed, 231 insertions( )

commit 619984a842c6c2daf31559c1979f91227a323648
Date:   Wed Mar 30 14:43:58 2022  0900

    chore: add

 image.kra | Bin 0 -> 777685 bytes
 1 file changed, 0 insertions( ), 0 deletions(-)

Versions

Python: 3.9.9

GitPython: 3.1.27

Krita: 5.0.2

Linux 5.15.16-gentoo

All files are created in Linux.

CodePudding user response：

It seems that this is caused by the GitPython's bug. It truncated the last \n of the output of git show and made the file invalid.

I changed the code to use subprocess.Popen and ZipFile succeeded.

import zipfile
from io import BytesIO
import subprocess

p = subprocess.Popen(["git", "show", "HEAD:a.zip"], stdout = subprocess.PIPE)

out, _ = p.communicate()

z = zipfile.ZipFile(BytesIO(out), "r")