As stated in this feature request, we work agile and the tester gets new applications versions (almost) on a daily basis.
The tester does not know which objects have been changed/deleted. And does ...
I'm pretty ok with what you say, but still not on the "on run" analysis... If i understand well your process, you have a first run which will highlights changes (with all corresponding tests failed as they didn't run), you then update your tests, and run again your tests. For me, the first run should not happen (the results are not trustworthy), but an analysis step instead! From a very logical point of view, the 2 ways are quite the same, except that you don't have a useless run