gh-86650: get_obs_local_part() fails to handle empty local part

[ZackerySpytz]

https://bugs.python.org/issue42484

• Issue: get_obs_local_part fails to handle empty local part #86650

[github-actions]

This PR is stale because it has been open for 30 days with no activity.

[spacefreak86]

What exactly is missing for this PR to get applied? It would be great because this bug is affecting my productive system and I have to monkey-patch the email library by myself.

[MaxwellDupre]

Ran 1654 tests in 6.561s
OK (skipped=1)
Looks ok.

[lesleyxyz]

I am getting this bug in a production environment.
It seems emails from the sender [email protected] have a message id starting with <[

How is this PR not yet reviewed after- all this time? How can we notify the right people to review this?

[arhadthedev]

@python/email-team

[mdavranche]

Same here, got this on my production environment, preventing me to upgrade from Python 3.7.3 (bug not present) to a more recent version (bug confirmed in 3.8, 3.9, 3.10, 3.11).

Easy to trigger:

import email.parser
import email.policy
faulty_header = "Message-ID: <[9f83415a58f84675bac8ab3e6047a157-JFBVALKQOJXWILKCJQZFA7CJIFGVGU2QKJ6FGU2QKJCW2YLJNR6EK6DPKNWXI4A=@example.com]>"
message = email.parser.BytesParser(policy=email.policy.default).parsebytes(faulty_header.encode("utf-8"))
message.as_bytes()

[lesleyxyz]

Is there an ETA for this please?

[fsc-eriker]

I have a competing PR which does not simply silently ignore the problem, but instead raises a defect. My proposed implementation returns a partial message-id which is perhaps less than ideal.

#108133

[serhiy-storchaka]

Thank you for your contribution, @ZackerySpytz, but I think that it is not a correct solution. Invalid Message-ID should not be silently truncated. The parsing code should raise HeaderParseError, this will allow to preserve its literal value. See #117934.