Tim,
The answers to 1) and 2) are both the same. One of the many problems I was having in attempting to narrow down the cause of the 'one definition rule' warning is that the compiler/linker was pulling in what appeared to be extraneous i2cdev-related libraries, in particular some RTIMULib files. I wanted to prevent this from happening to see if that was the cause of the problem (it wasn't).
In reading the documentation associated with Jeff Rowberg's i2cdev library, I saw that it recommended placing the appropriate library files either in the Arduino Libraries folder or directly into the sketch folder, so I elected the latter approach.
The reason I disabled the 'Deep Search' option was to simplify the problem. With it disabled I was reasonably sure the compiler won't try and pull in extraneous files. For the same reason, I moved all the i2cdev library files to a completely different non-Arduino folder on my PC, so even if the compiler searched the Arduino/Libraries folder, it wouldn't be able to find any other i2cdev library files.
Another reason for this setup is I can, if necessary, modify the local i2cdev files with no fear that my modifications will result in corrupted/unusable 'regular' library files - I can always delete my screwed up local version and start over again with 'pristine' copies from my hidden i2cdev folder.
The bottom line on all of this is I am trying to achieve a stable, constrained-scope configuration to figure out what causes compiling in the VS/VM environment to fail to link, while the same source code assembly (.INO and .CPP/.H files) compiles and links properly, and try and pinpoint the cause of the 'one definition rule' warning.
As it stands now, both compile setups issue the same two warnings. One (the overflow warning) is understandable and easily fixed, but the other one (the 'one definition rule' warning) is still a mystery. However, this mystery (which is where I started this odyssey) is clearly a minor problem compared to not being able to compile at all in the VS/VM environment.
I hope this answers your questions. I've done a lot of troubleshooting in my 50+ year engineering career, and I'm doing my very best to approach this problem in a methodical fashion. If you think I'm screwing up and should be doing something different, please let me know and I'll be happy to change.

.
TIA,
Frank