What I would like to see in the bullet industry is some standard for generating BC values so that shooters can get a fair comparison of potential ballistics.
I'm finishing up a book that will include experimentally measured BC's for over 100 long range bullets that I've collected over the last 2 years, all measured using the same method. My plan is to have this book in print by this spring/summer.
As far as practical limits to bullet length, one challenge is the fabrication, specifically the jackets. Copper jackets are drawn from flat circles of copper into cups. The longer and deeper the cup is, the harder it is to make it a uniform thickness.
There are other challenges as well, such as max practical length of bearing surface you can scrape down a bore before copper fouling becomes prohibitive. The longer bullets need very fast twist rates, which exacerbate every component of dispersion that's related to spin rate (which is most of them). Then there's the flight quality of the bullet. Very long bullets are more 'tipsy'. Consider the extreme case of an arrow with no fletchings. You can spin that sucker as fast as the dickens, but it simply won't fly well because it's just too long. Bullets that are longer than the current 'heavy for caliber' bullets are not as bad as the arrow, but begin to approach that domain.
Your belief that bullets have different BC's from all rifles is a common one. It's true that if a bullet is fired with marginal stability, it will have some pitching/yawing motion before it goes to sleep, and that will affect the BC. However, if the bullet is adequately stabilized which is the case for most accurate rifles having a twist equal to or faster than recommended, the bullets will emerge with adequate stability, and fly with a tight spiral and minimum drag.
In other words, it's possible to have a depressed BC from pitching/yawing, but only if there's a problem. If everything is right
, a bullet will have the same BC from any rifle that imparts adequate stability.
Atmospheric variations will make a bullet fly differently of course. It's simply a matter of accounting for the variables that affect air density.
In cases of borderline stability, you can have a bullet that's well stabilized in thin air (high alt, warm air), and have it not be stable in denser (low alt cold) air. Again, if the bullet is properly stabilized, it can overcome the denser air, fly perfectly point forward, and only be affected by the very predictable effects of different air density.
As far as real world variations in BC, your more likely to see a problem between lots of a given bullet. For example, I've tested different lots of a given bullet and found up to ~4% difference in BC. The difference is usually always from wider or narrower meplat (tip) diameters which is the #1 dimension that vary's among lots and affects drag. The difference in meplat diameter should be obvious, in the case of 4% of BC variation, there was about 0.015" difference in tip diameter.