b4 send --reflect does not work (but regular send and --preview-to does)

[javiercarrascocruz]

I just noticed that b4 send --reflect throws the following error:

Ready to:
  - send the above messages to just javier.carrasco.cruz@gmail.com (REFLECT MODE)
  - with envelope-from: javier.carrasco.cruz@gmail.com
  - via SMTP server smtp.gmail.com

REFLECT MODE:
    The To: and Cc: headers will be fully populated, but the only
    address given to the mail server for actual delivery will be
    javier.carrasco.cruz@gmail.com

    Addresses in To: and Cc: headers will NOT receive this series.

Press Enter to proceed or Ctrl-C to abort
Connecting to smtp.gmail.com:0
---
  [PATCH v2] docs: scheduler: completion: Document complete_on_current_cpu()
Traceback (most recent call last):
  File "/home/hb/repos/b4/src/b4/command.py", line 450, in <module>
    cmd()
  File "/home/hb/repos/b4/src/b4/command.py", line 433, in cmd
    cmdargs.func(cmdargs)
  File "/home/hb/repos/b4/src/b4/command.py", line 95, in cmd_send
    b4.ez.cmd_send(cmdargs)
  File "/home/hb/repos/b4/src/b4/ez.py", line 2245, in cmd_send
    sent = b4.send_mail(smtp, send_msgs, fromaddr=fromaddr, patatt_sign=sign,
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/hb/repos/b4/src/b4/__init__.py", line 4213, in send_mail
    smtp.sendmail(fromaddr, [envpair[1]], bdata)
  File "/usr/lib/python3.12/smtplib.py", line 890, in sendmail
    raise SMTPRecipientsRefused(senderrs)
smtplib.SMTPRecipientsRefused: {'None': (553, b'5.1.3 The recipient address <None> is not a valid RFC 5321 address. For more\n5.1.3 information, go to\n5.1.3  https://support.google.com/a/answer/3221692 and review RFC 5321\n5.1.3 specifications. 38308e7fff4ca-32cd2ee7be0sm17662751fa.78 - gsmtp')}

I immediately tried to send the same patch from the same account with --preview-to (to my own address i.e. simulating --reflect) and it worked as expected, with no error output. Then I sent the patch upstream (b4 send), and it worked as well.

I am using the current master branch (HEAD at a9b2aba), but it also failed before fast-forwarding. I guess the key point is that "The recipient address is not a valid...". First I thought it was some misconfiguration, but given that the other tests worked fine, I thought that it might be something else.

If more information is required, please let me know.