email.generator.Generator ignores policy when using `multipart/signed` → ruins signing
还没有人认领这个 Issue。
评估
调研方向
从 Lib/email/generator.py 中的 _handle_multipart_signed 开始,使用包含 multipart/signed 和长文件名的 EmailMessage 示例重现该问题。检查 gh-100204 中链接的工作以及 email 包的相关测试;完成的标准是签名部分保留预期的 policy 和标头序列化,同时不破坏现有的 multipart 行为。
由索引模型根据 Issue 内容生成。
描述
Bug report
EmailMessage behaves differently when being set to multipart/signed mimetype.
from email.message import EmailMessage
inner = EmailMessage()
inner.add_attachment("some data", "text/plain", filename="*"*35)
outer1 = EmailMessage()
outer1.set_type("multipart/signed") # affected by generator.py/_handle_multipart_signed
outer1.attach(inner)
outer2 = EmailMessage()
outer2.set_type("multipart/signeX") # not affected by generator.py/_handle_multipart_signed
outer2.attach(inner)
# When accessing given submessage, nothing weird happens
outer1.get_payload()[0].as_string() == outer2.get_payload()[0].as_string() # True
# However, when accessing whole message at once, headers folding change for the outer1 `multipart/signed` message
inner.as_string() in outer1.as_string() # !False!
inner.as_string() in outer2.as_string() # True
This is due to a 13 years old generator.py code that for an unknown reason rewrites the policy so that no header was folded:
def _handle_multipart_signed(self, msg):
# The contents of signed parts has to stay unmodified in order to keep
# the signature intact per RFC1847 2.1, so we disable header wrapping.
# RDM: This isn't enough to completely preserve the part, but it helps.
p = self.policy
self.policy = p.clone(max_line_length=0)
try:
self._handle_multipart(msg)
finally:
self.policy = p
As a result, when I GPG-sign the inner message and attach it to a wrapping-outer message along with the signature (which is the right thing), the signature is void because policy being ignored, the headers folding got disabled on the output. I understand the method _handle_multipart_signed should help the message signing but it ruins it instead. One dirty solution would be to set the policy to max_line_length=0 which fails for whatever reason:
from email import policy
pol = policy.default.clone(max_line_length=0)
inner = EmailMessage(policy=pol)
inner.add_attachment("some data", "text/plain", filename="*"*35) # ValueError: maxlinelen must be at least 4
Therefore, I am not able to sign the inner message with the headers fold (as it is output unfold), not I am able to sign the inner message with the headers unfold (as ValueError prevents me to set the policy to not stop folding headers).
So my proposal is to remove _handle_multipart_signed altogether (which would be sufficient) or to find a use-case where it does make sense (I could not find any).
Your environment
- CPython versions tested on: Python 3.10.6
- Operating system and architecture: Ubuntu 22.04.1 LTS x86_64
Linked PRs
- gh-100204
- 主要语言
- Python
- 星标
- 77.2k
- 派生
- 36k
- 平均合并
- 1 天 9 小时
- 30 天内合并 PR
- 558
贡献指南
从这里开始
- 先读完整个 Issue,再读项目的贡献指南。
- 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
- Fork 仓库,在一个分支上完成修改。
- 提交 Pull Request,并在描述里引用这个 Issue 编号。
python/cpython 的其他 Issue
-
docs pending
难度 2/5 1-3 小时 新手友好度 78/100
-
stdlib type-feature
难度 2/5 1-3 小时 新手友好度 78/100
-
stdlib type-feature
难度 2/5 1-3 小时 新手友好度 72/100
-
build type-bug
难度 2/5 1-3 小时 新手友好度 76/100
-
stdlib topic-email type-feature
难度 2/5 1-3 小时 新手友好度 70/100
相似的 Issue
-
link-check link-check:sphinx-theme
难度 2/5 1-3 小时 新手友好度 72/100
-
难度 2/5 1-3 小时 新手友好度 65/100
qgis/QGIS-Documentation#11275 ·
-
bug priority:normal ready-for-dev
难度 2/5 1-3 小时 新手友好度 88/100
OpenHands/extensions#626 · 1 条评论 ·
-
难度 1/5 1 小时以内 新手友好度 90/100
CSCfi/sd-search-api#39 ·
-
难度 1/5 1 小时以内 新手友好度 90/100