Re: [PATCH 0/3] composing patches

[Sebastian Spaeth <Sebastian@sspaeth.de>]

[-- Attachment #1: Type: text/plain, Size: 1532 bytes --]

On Fri, 10 Dec 2010 17:35:17 +0200, Felipe Contreras wrote:
> On Fri, Dec 10, 2010 at 1:20 AM, Sebastian Spaeth <Sebastian@sspaeth.de> wrote:
> > But notmuch is not the User-Agent. That's why the emacs UI fills in the
> > User-Agent value itself. I am not sure on the policy of creating
> > Message-IDs, perhaps notmuch-reply could actually fill in some defaults there.
> 
> Why it's not? The emacs UI is not really doing anything on top of
> 'notmuch reply'.

Well, it is taking care of the editing that I do, the formatting, the
attaching of files, and it also decides how the file is being sent
out. Isn't that what an UA does? That having said, I don't mind to
prefill a value that the emacs UI can override. I was just pointing out
that what I send out in a few seconds is not a product of notmuch
(mostly).

> And if it is, it can very well override that value.
> Besides, what about other users (vim)? Why not make the output of
> 'notmuch reply' ready to be dumped to sendmail?

When would it make sense to have notmuch reply be dumped to sendmail? Do
you often send quoted replies back to the original sender? :-)
 
> It seems right now there's a lot of reliance on emacs UI, and gnus.

Besides that I happen to use message mode for editing my mails, I don't
use gnus at all. If adding a message-uid and a User-Agent header counts
as *a lot*, then yes, it's a lot :).

But all that is not worth spending more electrons on, prefill the values
and lets override them where needed.

Sebastian

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]