Mark Walters <> writes:

> Hi
> Would it be worth putting the run-hook after the (insert chosen) rather
> than before? That would mean that the hook had access to the full new
> header. It would also mean that it wouldn't matter if the hook changed
> the buffer -- as it is I think the replace might go wrong as we replace
> beg to end and those seem to be integer-points not markers


> One final query -- this function will be called when completing any of
> To: Cc: Bcc: From: and some other less common headers. We could pass an
> argument which says which header we are on but that is probably more
> complexity than necessary. However, it is probably worth documenting
> that it may be called from these headers in the defcustom for the hook.

Just changing the doc sounds sensible to me. Adding extra arguments
sounds complicated since we are also calling the hook from within
the nothmuch-company backend.


notmuch mailing list

Reply via email to