I'd just like to report something odd with the AnchoredSnapper.
When snapping around 350.000 Lines to correspondig endpoints (grouping via Feature IDs derived from a topology), I found that several lines that were very close in length to the snapping threshold are in neither the touched, nor the untouched output port.
Instead, they just disappear. Not in large numbers, only between 30 lines were dropped, but I visualized them from my database and found that those were valid lines with a length of slightly more than 10m. Snapping tolerance was just that, 10m.
I had categorized the features already according to their endpoint distances. Lines with less than 10m were entirely in another snapping category, where I had similar problems, e.g. with the snapping threshold of 10cm. Lines with less than 10cm weren't in any output ports there, too.
I minimized the problem by lowering the snapping tolerance to 9.95m. Then, different but less features were affected. I could re-integrate those affected into the line pool with a FeatureMerger, so in the end no line feature was lost.
However, I have no clue why these features are dropped in the first place. Shouldn't there be another output port for them?