Hi David,

Thanks for looking into this. Much appreciated. 

Yes, following your steps, emacs -q followed by "M-x load-library <return>
notmuch", I was able to send an email without the error. It prompted me for my smtp server, username, and password, then returned this:

Sending email 
Sending email done
Invalid image size (see `max-image-size')
Mark set
Saving file /Users/peter/Dropbox/mail/gmail/sent/tmp/1451367804.6837_72197_1.Infinity.local...
Wrote /Users/peter/Dropbox/mail/gmail/sent/tmp/1451367804.6837_72197_1.Infinity.local
Sending...done 

To answer your other question, yes, there is an Fcc: header when I try to send email from my regular Emacs configuration. It's the same header I saw when I sent email from the minimal configuration: 

Fcc: /Users/peter/Dropbox/mail/gmail/sent

That directory does indeed exist: 

drwxr-xr-x@ 9 peter  staff  306 Dec 27 21:54 sent

And it contains:

.DS_Store    .mbsyncstate .nnmaildir   .uidvalidity cur          new          tmp

Does this give us any leads? 

Thanks again.

On Mon, Dec 28, 2015 at 9:46 AM, David Bremner <david@tethera.net> wrote:
fauno <fauno@kiwwwi.com.ar> writes:

> Peter Salazar <cycleofsong@gmail.com> writes:
>
>> Hello,
>>
>> I'm using notmuch-mode from within Emacs to send email using async/mbsync
>> through Gmail. However, every time I send email from within Emacs, I get
>> this error:
>
> i got this message too for fcc with default values, i'm on emacs 24.5
> with notmuch 0.21
>

Can either of you replicate the problem with a minimal configuration,
ideally with "emacs -q", followed by "M-x load-library <return>
notmuch"?

If so, does the path in the Fcc header exist? If so, what is it? a file,
a directory, a symlink?

For me, in emacs 24.5 / notmuch 0.21, with emacs -q, if that path is
missing I am prompted to create it. If I refuse, then I later get a
message about not being a maildir.

I guess the tl;dr is that I can't duplicate this problem. Looking at the
traceback Peter provided, it looks like he is using
"send-message-without-bullets" to send the message. Since this isn't a
notmuch function, it's likely bypassing the notmuch fcc setup that
notmuch-mua-send and notmuch-mua-send-and-exit do.


d