Updated the truncation to parse each feed

[TeckniX]

Updated the read method to apply filter, sorting and truncation on each feed, instead of the overall feed results. If provided with 2+ feeds before, the results would truncate to the size of the first feed (10+ items always requested). The updated code, will truncate each feed to the size of $query['limit'], providing you with each feed filtered and sorted to the query limit count.

2 feed with a limit of 5, will return 10 feed entries (5 per feed).

resulting array is: array('feed-1-date'=>array(),       'feed-1-date'=>array(),       'feed-1-date'=>array(),       'feed-1-date'=>array(),       'feed-1-date'=>array(),       'feed-2-date'=>array(),       'feed-2-date'=>array(),       'feed-2-date'=>array(),       'feed-2-date'=>array(),       'feed-2-date'=>array()); Instead of: array('feed-1-date'=>array(),       'feed-1-date'=>array(),       'feed-1-date'=>array(),       'feed-1-date'=>array(),       'feed-1-date'=>array());

[milesj]

The order of the feeds is based on the timestamp: https://github.com/milesj/cake-utility/blob/master/Model/Datasource/FeedSource.php#L313

If the last 10 items in the feed are from a single source because of a newer timestamp, then those 10 should be returned during truncate. It shouldn't split it between multiple sources disregarding the date.

[TeckniX]

I see what you're saying - It's 2 different ways of looking at it, not sure if it's possible to support both. Basically you return the latest 10 items from an aggregation of feeds. I return the latest 10 items from each feed. I guess I could simply loop through the feeds and concatenate to an overall array.

Just odd that the FeedSource does it on its own, instead of the logic being in aggregation.

[milesj]

Well it's a datasource and needs to support the same options as if it was a regular database. But I could remove the limit of 20 restriction, and just have it return all if limit is null.

[TeckniX]

I think it's fine as-is then. I see the distinction. Thanks!