OK, I think I get it now.
* For individual feeds, you have to give an ot value equal to that of
the firstitemmsec when querying /stream/items/ids
* For categories, you can do the request on /stream/items/ids with the
group streamId and no ot constraint.
I think the reason I got confused is that I'm actually creating a
psuedo-category in my application called "Other Feeds" that
essentially wraps up any feed not in a category. So in my case, I've
got a category that can contain many feeds and the only way to get an
unread count for those feeds is to query each feed individually,
getting their unread count with the appropriate ot parameter, and add
them up for the category. Although this is fairly inefficient to do,
I absolutely get that I'm doing something weird here that Google
Reader isn't built for.
I may end up abandoning the idea of an automated "Other Feeds"
category altogether.
Thanks again for your help!
-Matt
On Feb 22, 9:57 pm, Mihai Parparita <
mih...@google.com> wrote:
> Actually, I'm not seeing this behavior, if I do a request like:
>
>
https://www.google.com/reader/api/0/stream/items/ids?s=user/-/label/t...
>
> Where "test" is a folder I just made, with a newly-added subscription in
> it, I get the expected number of items back (10 if I leave off the "ot"
> parameter, fewer if I move it up to a more recent interval).
>
> Mihai
>
>
>
>
>
>
>
> On Wed, Feb 22, 2012 at 6:53 PM, Mihai Parparita <
mih...@google.com> wrote:
> > This may be a bug on the Reader side. In the meantime, I think you'll have
> > to do that filtering yourself. If you do a request of the form:
>
> >
https://www.google.com/reader/api/0/stream/items/ids?s=user/-/label/N...
>
> > Then you can look at individual item timestamps, see which feed they came
> > from, and exclude those items if they're before the first item timestamp.
>
> > Mihai
>