bug#66795: 29.1; debbugs hashcash behavior discrepancy

[Eric Abrahamsen]

Michael Albinus <michael.albinus@gmx.de> writes:

> Forward to Eric Abrahamsen; it seems to be rather a Gnus problem
>
> From: Bruno Victal <mirai@makinata.eu>
> Subject: bug#66795: 29.1; debbugs hashcash behavior discrepancy
> To: 66795@debbugs.gnu.org
> Date: Sat, 28 Oct 2023 15:27:11 +0100 (4 days, 1 hour, 8 minutes ago)
>
> 1. Have `(setq message-generate-hashcash t)' in init.el.
> 2. Using debbugs (elpa), open any issue with more than one recipient to
> Cc. (e.g. #66561) with `M-x debbugs-gnu' (optionally followed by `M-x
> debbugs-gnu-get-bug-by-id <RET> …')
> 3. Select message and start replying with `S w'.
> 4. X-Hashcash headers are opportunistically generated and displayed,
> in contrast to a plain 'C-x m' (Message mode) where it only starts
> doing so after `C-c C-c'.
>
> It's strange that the value of `message-generate-hashcash' is ignored
> when debbugs is used. This is undesired behavior since it results in
> redundant or unused X-Hashcash headers when I edit the addresses in the
> Cc or To lines.

I'm not sure where debbugs comes into it -- Gnus will generate hashcash
headers for any reply, during message setup. You don't see anything in
"C-x m" because there's nothing in the To or Cc headers yet. But "S w"
in a regular Gnus summary buffer will also generate the headers.