On Thu, 06 Oct 2011 09:21:48 -0400, Jesse Rosenthal <jrosenthal@jhu.edu> wrote:

> morning's project. In retrospect, I think the main issue was that I was
> trying to figure out how history would be kept. By using git, though,
> your idea would be that we get history for free, right?

Yeah. Of course, one would have to decide if the history of the current
dump file format is intelligible, or if some alternative format should
be used.

> of git? I.e., all we need something to just keep different dumps in
> different namespaces, so that when I pull your dump, the tags come in as
> `bremner/to-fix` (or whatever our separator is), and then when I dump
> all `bremner` tags, it removes the namespace and puts it in
> `bremner.tagfile`.

something like that sounds plausible. Currently the query parser doesn't
handle searches like "tag:bremner/to-fix" very well, because it
helpfully splits at '/' (aiui; maybe somebody else can explain it
better).  "notmuch search "tag:notmuch*" does not seem to work either,
so my first idea to kludge around the problem by using
e.g. "notmuchPushed" doesn't seem to work either.

> So using
> either the library or the bindings would probably speed things up quite
> a bit, by only dumping the tags you care about. Which means it couldn't
> just be a shell script: it would have to be c/ruby/python.

Without looking at the code, I'd guess it wouldn't be too hard to have
notmuch-dump take a standard search term; it's just as I mentioned I
don't see an elegant way of doing what we want with a standard search
term. Of course I'd be happy to be proved wrong.

David