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

Strip comment references from document.xml

Browse source
This commit is contained in:
Alex Marchant 2024-04-03 15:20:00 -04:00 • committed by jvoisin
parent f931a0ecee
commit f2c898c92d
2 changed files with 67 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

32
tests/test_libmat2.py
View file

@ -900,4 +900,34 @@ class TextDocx(unittest.TestCase):
self.assertIsNotNone(match)
os.remove('./tests/data/comment_clean.docx')
os.remove('./tests/data/comment_clean.cleaned.docx')
os.remove('./tests/data/comment_clean.cleaned.docx')
def test_comment_references_are_removed(self):
with zipfile.ZipFile('./tests/data/comment.docx') as zipin:
c = zipin.open('word/document.xml')
content = c.read()
r = b'w:commentRangeStart'
self.assertIn(r, content)
r = b'w:commentRangeEnd'
self.assertIn(r, content)
r = b'w:commentReference'
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/document.xml')
content = c.read()
r = b'w:commentRangeStart'
self.assertNotIn(r, content)
r = b'w:commentRangeEnd'
self.assertNotIn(r, content)
r = b'w:commentReference'
self.assertNotIn(r, content)
os.remove('./tests/data/comment_clean.docx')
os.remove('./tests/data/comment_clean.cleaned.docx')