I'd say that my results so far are similar to yours Randy. The XP-L HIs have really nice tint--cleaner than almost any other dedome I've seen. The total lumens output seems to be in between a high binned XP-G2 and a domed XP-L (with the throw being similar to a domed XP-G2).
I built a few triples at 5.8A-6A over the past few weeks and tested an XP-G2 S4 2B against the XP-L U5 5000K. I got about 200-300 lumens more out of the XP-L HIs, but that's through Carclo optics, which always eat some lumens regardless of the setup.
I don't know how the V5/V6 dedomed XP-Ls would have measured in that setup, since the ones I've built before and tested were with an FET driver pulling 10A+, so it's not really comparable.
At the end of the day, I will have to say that the tint is fantastic on all of the HIs I've tested so far. A very nice option for sure, especially for the triples. Now all we need is a factory dedomed XM-L2 and XP-G2. That would be heaven! (I would want the XM-L2 so that I didn't have to deal with trying to center an XP die in lights that are set up already for XM dies).
Also, it could have just been me because I didn't measure it, so take this anecdote for what it's worth, but it seems like at the same drive currents the triple XP-L HIs are actually running slightly cooler than the XP-G2s. I have built a ton of the triple S3s so I'm pretty used to how they feel, and it feels like the XP-L is producing a little less heat at the same drive current (more efficient).
I little birdy told me that some real XP-L HI V3s are coming soon, so we'll see how those compare.
--------
Regarding the light pipes, integrating spheres, milk cartons, etc.... I have one of the pipes and it has been fairly consistent for me, and the readings I've got from it have lined up very closely with those from others (even with some of the same lights that I've sent to others with different kinds of setups). It isn't a perfect solution, but it is close enough to give a pretty good idea. You do have to be consistent with how you measure the lights. I have different foam cutouts so that the lights stay centered and flat regardless of what light is used, because the readings do change if you tilt the light. It is an obvious flaw, but one that isn't fatal as long as you are consistent. Consistency counts.
I wouldn't use the numbers from this for ANSI style publication, which is why I rarely publish exact lumens numbers (and when I do, I usually underrate them a bit), but it does give me a good indication of how the lights perform and what the modifications have done.
As imperfect as it is, it is infinitely better than my two built-in light meters--which are easily fooled!