我尝试在 Flask API 中设置 pades 签名流程。
当我们在客户端计算机上使用 PKCS11 设备时,我们需要使用中断的签名流程:
/pades/start
上发布其证书(PEM 文件)和要签名的 PDF。/pades/complete
上发布其任务 ID 和计算出的签名。 API 使用此签名来创建数字签名的 PDF目前,该流程有效。但生成的 PDF 被认为具有无效签名,并显示此消息“定义签名数据范围的意外字节范围值。 详细信息:签名字节范围无效”
# Relevant part in the /pades/start route
with open(task_dir / "certificate.pem", "w") as f:
f.write(body["certificate"])
cert = load_cert_from_pemder(task_dir / "certificate.pem")
with open(task_dir / "document.pdf", "rb+") as f:
writer = IncrementalPdfFileWriter(f)
fields.append_signature_field(
writer,
sig_field_spec=fields.SigFieldSpec("Signature", box=(200, 600, 400, 660)),
)
meta = signers.PdfSignatureMetadata(
field_name="Signature",
subfilter=fields.SigSeedSubFilter.PADES,
md_algorithm="sha256",
)
ext_signer = signers.ExternalSigner(
signing_cert=cert,
cert_registry=registry.CertificateRegistry(),
signature_value=bytes(8192), # I tried to adjust this with many different values without success
)
pdf_signer = signers.PdfSigner(meta, signer=ext_signer)
prep_digest, tbs_document, _ = pdf_signer.digest_doc_for_signing(writer)
post_sign_instructions = tbs_document.post_sign_instructions
def async_to_sync(awaitable):
loop = asyncio.new_event_loop()
asyncio.set_event_loop(loop)
return loop.run_until_complete(awaitable)
signed_attrs: asn1crypto.cms.CMSAttributes = async_to_sync(
ext_signer.signed_attrs(
prep_digest.document_digest, "sha256", use_pades=True
)
)
task = {
**(body or {}),
"id": task_id,
"prep_digest": prep_digest,
"signed_attrs": signed_attrs,
"psi": post_sign_instructions,
}
redis.set(
f"task:{task_id}",
pickle.dumps(task),
)
writer.write_in_place()
return {"task": task_id, "digest": prep_digest.document_digest.hex()}
# Relevant part in the /pades/complete route
task_id = body["task"]
task_str = redis.get(f"task:{task_id}")
task = pickle.loads(task_str) if task_str else None
task_dir = Path(get_task_dir(settings.WORKDIR, task_id))
if not task:
return {"error": "Task not found"}, 404
ext_signer = signers.ExternalSigner(
signing_cert=load_cert_from_pemder(task_dir / "certificate.pem"),
signature_value=bytes.fromhex(body["signature"]),
cert_registry=registry.CertificateRegistry(),
)
sig_cms = ext_signer.sign_prescribed_attributes(
"sha256", signed_attrs=task["signed_attrs"]
)
with open(task_dir / "document.pdf", "rb+") as f:
PdfTBSDocument.finish_signing(
f,
prepared_digest=task["prep_digest"],
signature_cms=sig_cms,
post_sign_instr=task["psi"],
)
redis.delete(f"task:{task_id}")
return "ok"
我可以尝试什么来修复此错误消息?
经过几个小时的搜索,我找到了解决方案。我把它放在这里给下一个需要它的人
出现问题的原因是:
我们需要返回
signed_attrs.dump()
而不是 prep_digest.document_digest
(并对它进行哈希处理!)
我们应该在摘要计算期间使用
pdf_signer.digest_doc_for_signing(pdf_out=writer,in_place=True)
,以允许正确更新pdf