Quoth David Edmondson on Jan 10 at  3:47 pm:
> On Tue, 10 Jan 2012 10:36:50 -0500, Austin Clements <amdragon@MIT.EDU> wrote:
> > LGTM, though would it be easier to put this in the else clause of the
> > if after the setq count?
> 
> Agreed. I got confused thinking about it due to the empty elements in
> the matrix, but perhaps that doesn't matter. I'll continue to think, but
> would rather not delay fixing the bug.

Fair enough.

> > Is it possible for a tag in the last column to be just long enough to
> > make the line still wrap?  Somehow my current tag set doesn't trigger
> > this bug, so I can't test this case (and I admit I can't follow
> > notmuch-hello-insert-tags well enough to reason this out).
> 
> With a sufficiently narrow window it's always possible to generate wrap,
> of course. I couldn't make it happen for any window width that seemed
> reasonable.

I should have specified more than one column.  Clearly if you're down
to one column it's always possible to wrap, but if you have more than
one, does the code always reduce the number of columns in preference
to allowing a particularly long tag name to wrap a line?  Based on
your testing, it sounds like it handles this correctly.  (Even if it
didn't, this shouldn't block this patch; it's related, but not the
same issue.)