security-and-privacy/mat2
security-and-privacy/mat2
1
0
Fork 0
mirror of synced 2025-07-04 04:17:29 +02:00
Code Releases Activity

Remove dangling references from document.xml.rels

Browse source 
The file `word/_rels/document.xml.rels` is similar to `[Content_Types].xml` and
has references to other files in the archive. If those references aren't
removed Word refuses to open the document. # Please enter the commit message
for your changes. Lines starting
This commit is contained in:
Alex Marchant 2024-04-05 18:45:58 +02:00 • committed by jvoisin
parent 09672a2dcc
commit 156855ab7e
2 changed files with 58 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

21
tests/test_libmat2.py
View file

@ -931,3 +931,24 @@ class TextDocx(unittest.TestCase):
os.remove('./tests/data/comment_clean.docx')
os.remove('./tests/data/comment_clean.cleaned.docx')
def test_clean_document_xml_rels(self):
with zipfile.ZipFile('./tests/data/comment.docx') as zipin:
c = zipin.open('word/_rels/document.xml.rels')
content = c.read()
r = b'Target="comments.xml"'
self.assertIn(r, content)
shutil.copy('./tests/data/comment.docx', './tests/data/comment_clean.docx')
p = office.MSOfficeParser('./tests/data/comment_clean.docx')
self.assertTrue(p.remove_all())
with zipfile.ZipFile('./tests/data/comment_clean.cleaned.docx') as zipin:
c = zipin.open('word/_rels/document.xml.rels')
content = c.read()
r = b'Target="comments.xml"'
self.assertNotIn(r, content)
os.remove('./tests/data/comment_clean.docx')
os.remove('./tests/data/comment_clean.cleaned.docx')