A minor point, but given the buff to AS shouldn't Glyph of Arcane Shot be mandatory, rather than recomended?

Looking solely at pet autoattack damage Mastery v Haste. On gear all rating points are equal value. It takes 128.05701 rating points for 1% of haste, and it takes 179.28 rating points to gain 1 masterypoint, which is 1.7% damage increase. Or, it takes 128 rating points to gaing 1% pet autoattack damage via haste, and it takes 105 rating points to gaing 1% pet autoatack damage via mastery, or to make them equal, 1% pet autoattack increase via haste is equal to 1.2% damage increase via mastery.
Looking at the other effects, 1% haste means 1% more damage from hunter autoshots, 1% more base focus regeneration for the hunter and 1% more base focus regeneration for the pet. Apart from this haste also gives a stepped benefit to Cobra Shot - stepped as in the benefit only arrives at thos amounts of haste where it allows another CoS cast to be an overall increase in output (even when pushing Kill Command), so the benefit does arrive a bit earlier than a clean 1 additional CoS.
Mastery increase benefits pet basic attacks and Kill Command in addition to autoattack, so for haste you are trading 1% pet autoattack damage + 1% autoshotdamage + 1% base focus gain for both hunter and pet for 1.2% pet autoattack damage + 1.2% pet basic attack damage + 1.2% Kill Command damage.

To try to draw a conclusion out of that, I am not saying haste is bad, I am saying that; at a given, nonspecific point, mastery is the better choice of the 2. Referring back to the OP, currently mastery outperforms haste gains in the various replacements I've tinkered with 3:2 - mastery is by no means the strongest secondary (crit is) - and 3:2 reanslates into 1.2 dps benefit to 0.8 dps benefit (or similar amounts depending on ap and crit). NB!, I am not talking about the calculated or projected gains, I am talking of actually replacing X Mastery with X haste or vice versa. The gained benefit takes into account the haste you have on your current gear and all the benefits you have gained from it so far, most importantly any plateaus you have already reached (and passed) for amounts of shots. The projected gains should be less misleading, but in reality I have found that to not be the case when you switch stats (they are very accurate if you simply add X rating, but not if you switch X rating).

The focus on fitting in another cobra shot is, for my part, because that is the point where haste outperforms mastery - when gaining enough haste to consistently fit in another CoS in a rotation the net gain from teh additional cobra shot will consistently outperform the similar mastery. This agian leads to haste being largely ignored for BM, because we start out at 2 second cast which is perfect for 2 cobras in a rotation.
I am not dismissing haste, haste helps, and we get enough born in on gear and through FF that we don't have to worry about reforging to haste or gemming for haste except possibly with the 4t11 bonus where we might need ~100 haste rating to fit in another CoS if we have reforged haste to crit or mastery wherever possilble, at which point it is definately the priority over mastery, and even crit (though I would chose to forego mastery over crit).
The main thing with haste in my opinion is that it is subjectively deceptive - more haste lets you pull of your rotation with greater ease - which will significantly increase a players output (moreso the less perfect the players reactions are) significantly more than mastery or crit, but its is a very subjective thing, and cannot reasonably be suggested IMO.
Thinking about it, I am inclined to add a caveat to haste rating noting this, so players that will see this benefit don't feel like bad players for choosing to reforge contrary to the indications - in the same way that the previous BM thread noted Rowan's Rifle as a very good ilvl 251 ranged for BM, regardless of itemization being bad, I'm just not sure how to word it right now.

This is the part that is confusing. For a lot of people, myself included, haste always projects to be more damage than mastery. I am not talking about calculating the benefit of reforging, but simply looking at the projected stat ratings and the actual simulated dps changes that come from adding X haste vs. X mastery. As near as I can tell, our theorycrafting tools show that at a given, nonspecific point, haste is better than mastery.

According to femaledwarf.com (which is an online version of our tried and true spreadsheet, unless I'm mistaken), with zero buffs/debuffs using a haste elixir results in 13,746.84 dps. Using a mastery elixir shows 13,730.35 dps. That's not much of a difference, but it favors haste. Using all raid debuffs/buffs except for flasks, haste elixir gives 19,291.12 dps and mastery gives 19,134.61. Other people have been getting similar results, or else we wouldn't be talking about it in the first place.

The problem is that at various points, on FD.com haste projects ahead of mastery, which always is dead last in dps per budget, if nothing else. That also would suggest that if you're adding a new piece of gear with mastery on it, reforging the mastery into haste (assuming crit already is on it) should make that item more powerful than if you didn't reforge it. If what you're saying is that in real world testing, adding mastery always results in more damage, then it means our theorycrafting tools are flat out wrong (including the projections, which you said should be more accurate) and we need to figure out how to fix them.

AoTH = 2000AP
OwN = AoTH * 1.1 or 1.2. or 1.3 = 2000 * 1.1 or 1.2 or 1.3 = 2200 or 2400 or 2600
In BM spec u have Animal Handler = previou result * 1.25 = 2600 * 1.25 = 3250