On Tue, 02 Nov 2010 17:33:08 -0400, Jameson Rollins <jrollins@finestructure.net> wrote:
> On Tue, 02 Nov 2010 21:03:55 +0000, Darren McGuicken <mailing-notmuch@fernseed.info> wrote:
> > A call to 'notmuch search-tags' from the command line does indeed
> > return an empty string as the first entry for me[1].
> 
> fwiw, I don't personally get any empty strings in the output of
> search-tags (with 0.4).  So I don't think it's anything inherent to
> that function.

Thanks for the confirmation Jamie - I had a potter with my data and sure
enough, a search for 'tag:""' returned one hit.  I was able to use the
emacs interface to remove the empty string tag from the matching message
('-' then return with no parameters).

So it's probably fair to say that we'll want to handle at least the
empty string scenario as part of our standard tag addition logic since
it can lead to unexpected side effects in searching.

Do we have any other 'known bad' scenarios in tag addition that we want
to catch before I try my hand at a patch?