On 10/26/07, Matthew Montgomery <[EMAIL PROTECTED]> wrote: > > On Oct 26, 2007, at Oct 26, 1:56 PM, Dave wrote: > > > On 10/26/07, Matthew Montgomery <[EMAIL PROTECTED]> wrote: > >> I have noticed an issue with the weblogger site-wide feeds. For > >> example, in the recent entries Atom feed pulling from the default > >> planet, the "updated" element is older than the most recent entry. > >> This came up as an issue with our separate use of the planet > >> aggregator. Any configured planet that consumes one of the site-wide > >> feeds from the weblogger will only update when the "updated" date is > >> incremented. NetNewsWire on the other hand could care less, and only > >> looks at the actual entry dates.
There may be a quick fix for this. I think I can improve the situation by replacing the main blog last-update time with last-update time associated with the site-wide cache in the feeds. It won't be perfect, but it will be better what what we have now. - Dave > >> > >> Options: > >> > >> 1. Alter the weblogger code such that the updated or lastBuildDate > >> value matches the most recent entry in the feed. > >> > >> 2. Alter the planet aggregator code such that it ignores the updated > >> or lastBuildDate values of the target feed. > >> > >> I would vote for number one as based on the Atom spec, am entirely > >> new entry would satisfy the significant verbiage of the updated > >> element. > > > > Yes, there is a bug and I agree with #1. > > > > The problem is that we use the front-page weblog's last-updated date > > as the feed's update date, which is wrong. On first glance, this does > > not look to be a trivial fix. > > Yep I just got in there enough to see what was going on with the > lastModified value. Thanks for the comments Dave, I filed the following. > > http://opensource.atlassian.com/projects/roller/browse/ROL-1587 > > -- > Matthew Montgomery > .Sun Engineering > Sun Microsystems, Inc. > > >
