The Lines->Linestrips conditioner produces correct lines, but incorrect normals, UVs, other non-vertex semantics.
Looking at my test files, it looks like during the Lines->Linestrips conversion, the <p> stream is shifted by one.
Steps to Reproduce:
0) Open COLLADA Refinery 2.03 (also tested internal SVN copy)
1) Use the following settings:
-- input: duck.dae (from Basic_Samples)
-- MACRO -- toLines
-- conditioner -- Lines->Linestrips
-- output: duck_linestrips.dae
2) Open duck_linestrips.dae in the CTS reference viewer (Feeling Viewer 3.02) or examine differences between duck_lines.dae and duck_linestrip.dae. I've included the equivalent with cube.dae for easier reading.