def test_update_hybrid_twice(fname):
with open(os.path.join(PDF_DATA_DIR, fname), 'rb') as inf:
w = IncrementalPdfFileWriter(inf)
t_obj = w.trailer['/Info'].raw_get('/Title')
assert isinstance(t_obj, generic.IndirectObject)
w.objects[(t_obj.generation, t_obj.idnum)] \
= generic.pdf_string('Updated')
out = BytesIO()
w.write(out)
r = PdfFileReader(out)
assert r.trailer['/Info']['/Title'] == 'Updated'
container_info = r.xrefs.get_xref_container_info(1)
assert container_info.xref_section_type == XRefSectionType.HYBRID_MAIN
container_info = r.xrefs.get_xref_container_info(2)
assert container_info.xref_section_type == XRefSectionType.STANDARD
w = IncrementalPdfFileWriter(out)
w.add_object(generic.pdf_string('This is an object'))
w.write_in_place()
r = PdfFileReader(out)
assert '/XRefStm' not in r.trailer
assert '/XRefStm' not in r.trailer_view
assert r.trailer['/Info']['/Title'] == 'Updated'
container_info = r.xrefs.get_xref_container_info(1)
assert container_info.xref_section_type == XRefSectionType.HYBRID_MAIN
container_info = r.xrefs.get_xref_container_info(2)
assert container_info.xref_section_type == XRefSectionType.STANDARD
container_info = r.xrefs.get_xref_container_info(3)
assert container_info.xref_section_type == XRefSectionType.STANDARD
def test_xref_null_update():
buf = BytesIO(MINIMAL)
w = IncrementalPdfFileWriter(buf)
w.write_in_place()
r = PdfFileReader(buf)
assert r.xrefs.total_revisions == 2
assert r.xrefs.explicit_refs_in_revision(1) == set()
def test_rogue_backreferences():
w = IncrementalPdfFileWriter(BytesIO(MINIMAL))
# intentionally refer back to the contents of the first page
w.root['/DSS'] = w.root['/Pages']['/Kids'][0].get_object().raw_get(
'/Contents')
w.update_root()
meta = signers.PdfSignatureMetadata(field_name='Sig1', )
out = signers.sign_pdf(w, meta, signer=FROM_CA)
# pretend to add a new form field, but actually secretly do a page
# tree modification.
sp = fields.SigFieldSpec('SigNew',
box=(10, 74, 140, 134),
doc_mdp_update_value=fields.MDPPerm.FILL_FORMS)
w = IncrementalPdfFileWriter(out)
fields.append_signature_field(w, sp)
w.write_in_place()
w = IncrementalPdfFileWriter(out)
contents_ref = w.root['/Pages']['/Kids'][0].get_object().raw_get(
'/Contents')
content_stream: generic.StreamObject = contents_ref.get_object()
content_stream._data = content_stream.data + b"q Q"
content_stream._encoded_data = None
w.mark_update(contents_ref)
w.write_in_place()
r = PdfFileReader(out)
emb = r.embedded_signatures[0]
emb.compute_integrity_info()
assert isinstance(emb.diff_result, SuspiciousModification)
def test_no_changes_policy():
w = IncrementalPdfFileWriter(BytesIO(MINIMAL_ONE_FIELD))
out = signers.sign_pdf(
w,
signers.PdfSignatureMetadata(
field_name='Sig1',
certify=True,
docmdp_permissions=fields.MDPPerm.FILL_FORMS),
signer=FROM_CA,
)
w = IncrementalPdfFileWriter(out)
# do an /Info update
dt = generic.pdf_date(datetime(2020, 10, 10, tzinfo=pytz.utc))
info = generic.DictionaryObject({pdf_name('/CreationDate'): dt})
w.set_info(info)
w.write_in_place()
# check with normal diff policy
r = PdfFileReader(out)
s = r.embedded_signatures[0]
assert s.field_name == 'Sig1'
status = val_trusted(s, extd=True)
assert status.modification_level == ModificationLevel.LTA_UPDATES
assert status.docmdp_ok
# now check with the ultra-strict no-op policy
r = PdfFileReader(out)
s = r.embedded_signatures[0]
status = validate_pdf_signature(s, diff_policy=NO_CHANGES_DIFF_POLICY)
assert isinstance(s.diff_result, SuspiciousModification)
assert not status.docmdp_ok
def test_xref_stream_null_update():
buf = BytesIO(MINIMAL_XREF)
w = IncrementalPdfFileWriter(buf)
w.write_in_place()
r = PdfFileReader(buf)
assert r.xrefs.total_revisions == 2
# The xref stream itself got added
assert len(r.xrefs.explicit_refs_in_revision(1)) == 1
def test_image_content_embed(infile):
path = os.path.join(IMG_DIR, infile)
w = IncrementalPdfFileWriter(BytesIO(MINIMAL))
img_content = images.PdfImage(
image=path,
writer=w,
opacity=0.6,
)
w.add_content_to_page(0, img_content, prepend=True)
w.write_in_place()
def add_dss(cls, output_stream, sig_contents, *, certs=None,
ocsps=None, crls=None, paths=None, validation_context=None,
force_write: bool = False, embed_roots: bool = True):
"""
Wrapper around :meth:`supply_dss_in_writer`.
The result is applied to the output stream as an incremental update.
:param output_stream:
Output stream to write to.
:param sig_contents:
Contents of the new signature (used to compute the VRI hash), as
as a hexadecimal string, including any padding.
If ``None``, the information will not be added to any VRI
dictionary.
:param certs:
Certificates to include in the VRI entry.
:param ocsps:
OCSP responses to include in the VRI entry.
:param crls:
CRLs to include in the VRI entry.
:param paths:
Validation paths that have been established, and need to be added
to the DSS.
:param force_write:
Force a write even if the DSS doesn't have any new content.
:param validation_context:
Validation context from which to draw OCSP responses and CRLs.
:param embed_roots:
.. versionadded:: 0.9.0
Option that controls whether the root certificate of each validation
path should be embedded into the DSS. The default is ``True``.
.. note::
Trust roots are configured by the validator, so embedding them
typically does nothing in a typical validation process.
Therefore they can be safely omitted in most cases.
Nonetheless, embedding the roots can be useful for documentation
purposes.
.. warning::
This only applies to paths, not the ``certs`` parameter.
"""
pdf_out = IncrementalPdfFileWriter(output_stream)
dss = cls.supply_dss_in_writer(
pdf_out, sig_contents, certs=certs, ocsps=ocsps,
crls=crls, paths=paths, validation_context=validation_context,
embed_roots=embed_roots
)
if force_write or dss.modified:
pdf_out.write_in_place()
def test_append_sigfield_trivial_ap():
buf = BytesIO(MINIMAL)
w = IncrementalPdfFileWriter(buf)
spec = fields.SigFieldSpec(sig_field_name='Sig1', box=(20, 20, 80, 40))
fields.append_signature_field(w, sig_field_spec=spec)
w.write_in_place()
r = PdfFileReader(buf)
pg1 = r.root['/Pages']['/Kids'][0]
assert '/Annots' in pg1
assert len(pg1['/Annots']) == 1
annot = pg1['/Annots'][0]
assert '/AP' in annot
assert annot['/AP']['/N'].data == b''
def test_append_sigfield_second_page():
buf = BytesIO(MINIMAL_TWO_PAGES)
w = IncrementalPdfFileWriter(buf)
fields.append_signature_field(w, fields.SigFieldSpec('Sig1', on_page=1))
w.write_in_place()
r = PdfFileReader(buf)
pg1 = r.root['/Pages']['/Kids'][0]
assert '/Annots' not in pg1
pg2 = r.root['/Pages']['/Kids'][1]
assert '/Annots' in pg2
assert len(pg2['/Annots']) == 1
assert pg2['/Annots'][0]['/T'] == 'Sig1'
def test_no_clobbering_xref_streams():
# Test witnessing the limitation on our reader implementation
# that disallows references to the xref stream of a previous revision
# from being overridden.
# (this behaviour may change in the future, but for now, the test is in
# place to deal with it)
buf = BytesIO(MINIMAL_XREF)
w = IncrementalPdfFileWriter(buf)
# update the xref stream in the previous revision
stream_ref = w.prev.xrefs.get_xref_container_info(0).stream_ref
w.mark_update(stream_ref)
w.write_in_place()
with pytest.raises(misc.PdfReadError, match="XRef.*must not be clobbered"):
PdfFileReader(buf)
def test_tamper_sig_obj(policy, skip_diff):
w = IncrementalPdfFileWriter(BytesIO(MINIMAL))
meta = signers.PdfSignatureMetadata(field_name='Sig1')
out = signers.sign_pdf(w, meta, signer=FROM_CA)
w = IncrementalPdfFileWriter(out)
sig_obj = w.prev.embedded_signatures[0].sig_object
sig_obj['/Bleh'] = generic.BooleanObject(False)
w.update_container(sig_obj)
w.write_in_place()
r = PdfFileReader(out)
emb = r.embedded_signatures[0]
status = validate_pdf_signature(emb,
diff_policy=policy,
skip_diff=skip_diff)
if skip_diff:
assert emb.diff_result is None
else:
assert isinstance(emb.diff_result, SuspiciousModification)
assert status.coverage == SignatureCoverageLevel.CONTIGUOUS_BLOCK_FROM_START
assert status.modification_level == ModificationLevel.OTHER
def test_embed_subset():
w = IncrementalPdfFileWriter(BytesIO(MINIMAL))
with open(NOTO_SERIF_JP, 'rb') as ffile:
ga = GlyphAccumulator(w, ffile, font_size=10)
res = ga.shape('版')
assert b'[<66eb>] TJ' in res.graphics_ops
res = ga.shape('テスト')
assert b'[<0637> 40 <062a0639>] TJ' in res.graphics_ops
# check the 'ffi' ligature
res = ga.shape('difficult')
assert b'[<0045004ae9e200440056004d0055>] TJ' in res.graphics_ops
w.write_in_place()
font_ref = ga.as_resource()
font = font_ref.get_object()
df = font['/DescendantFonts'][0].get_object()
font_file = df['/FontDescriptor']['/FontFile3']
# assert no ToUnicode assignment for 'f'
assert b'\n<0066>' not in font['/ToUnicode'].data
# assert a ToUnicode assignment for the 'ffi' ligature
assert b'\n<e9e2> <006600660069>' in font['/ToUnicode'].data
assert len(font_file.data) < 4000
def test_image_direct_embed(infile):
img = Image.open(os.path.join(IMG_DIR, infile))
w = IncrementalPdfFileWriter(BytesIO(MINIMAL))
image_ref = images.pil_image(img, w)
page_ref, resources = w.find_page_for_modification(0)
resources[pdf_name('/XObject')] = generic.DictionaryObject(
{pdf_name('/Img'): image_ref})
w.update_container(resources)
content_stream: generic.StreamObject = page_ref.get_object()['/Contents']
content_stream._data = \
content_stream.data + b' q 50 0 0 50 5 5 cm /Img Do Q'
content_stream._encoded_data = None
w.update_container(content_stream)
w.write_in_place()
# TODO flatten and do a visual comparison
image_obj = image_ref.get_object()
if 'indexed' in infile:
assert '/SMask' not in image_obj
else:
assert '/SMask' in image_obj
