David Bremner <[email protected]> writes: > > This looks plausible to me, and of course the test passes. I did wonder > if this means that people actually using notmuch-user-agent would run > the hook twice now?
After a certain amount of re-reading the docstring of define-mail-user-agent, and the source for compose-mail, I see that the HOOKVAR parameter is something that users of the mail-agent can tweak, rather than something that e.g. compose-mail promises to run. So I think your patch is fine. And I'm 99% sure the issue with the spurious output is just a little bug in test-lib.sh d _______________________________________________ notmuch mailing list [email protected] https://notmuchmail.org/mailman/listinfo/notmuch
