See the green line.
If the drawing (from the first post of this thread) really resembles the C8F, then you can take away some of the threading. Just make sure you have
- enough threading on the retaining ring left to hold the ring under pressure and
.
.
Pretty sure he means this (Pic courtesy of ImA4Wheelr). This would be the “underside” of the ring, notice, no holes for pliers cause it’s upside down.
No, pretty sure he doesn’t 
What I mean is filing the threading of the retaining ring.
Underside, yes, but from the outside, to remove some threading so you can turn the ring a bit more.
Your picture, contactcr, shows filing down the inside of the retaining ring. You do that when the ring collides with parts on the spring side of the driver. This wouldn’t be the case here with the C8F.
Wouldn’t it be easier to just put a few solder blobs on the outside of the driver? I built mine with a 1.2mm thick 20mm FET+1 and had no issue but I guess YMMV. The ring is ridiculously thick so i’ll leave up the pic even though it isn’t what you meant.
I did file down a little of the lower threads, but it had no effect on where the ring stopped rotating. Very strange. I may have to look at it closer.
Unfortunately I don’t have the 21700-C8F yet, so I can’t check myself.
I did this kind of filing several times with brass-pills (before the pill-less design took over) and it worked well to get the pill in deeper.
Of course you can do a comfy-style solder addon to the driver. Preferably the component side, as solder does not hold pressure well, especially not with a shearing force like from a retaining ring.
Btw, comfychair used a drill press to ‘mill’ these solder blobs flat. That was a classic.
Hey Sofirn next let give us C8F 26650 and the single emitter throwy version either 21700/26650 .
Why do you want 26650 when 21700 has almost the same performance but more compact?
You get longer run times, plus it feels better in the hand. The smaller 21700 is a bit awkward for me.
Then there is the SP33.
I talked Sofirn again. They told me that talk with the engineer and will think about the songle led 21700 version. I said that make it like C8F so same double switch just change the front end to single led.
Like ZozzV6 said, the old SP33 was a single emitter 26650 light. I’m not sure why they changed it to xhp50.2 with boost driver. Maybe low sales? It might have been better to have 2 seperate models.
That good I like the 2 switch with forward click tail of 21700 C8F , hope for bigger head (55-60 mm.) with XHP35 for better throw and please make the light switch to stay on as locator .
Now your asking for a completely different light.
You want Sofirn to make a light like the thrunite catapult v6 or lumintop odl20c, but with 2 switches?
Yes that I mean Jason but with Sofirn price. 
I recently received 2 C8F 21700 hosts, along with a driver kindly sent by Sofirn, and today I found the time to assemble 1.
I used 2 pieces of 16AWG wire (about 4-5cm each), 3x XP-L2 V4 4500k 80CRI 3-step, and the Sofirn driver.
And I was really amazed by the result, I measured 17.8A max at the tailcap, with a 30T @ 4.2V
THANKS Sofirn.
That will be just under 5000 on-LED lumens, right? 
Anyone know if the 21700 host will show back up in aliexpress store?
From a test to ceiling bounce, I estimate about 4700-4800 lumens.
Not too shabby!
I modded up a C8F 21700 host with 3 351D 4000K 90 CRI's and also measured 17.8A at the tail! Mine does 4150 at start, 3750 at 30 secs, but that's not ANSI lumens, more like manufacturer lumens. Measured 33 kcd throw. I used 20 AWG LED wires (long to re-program), and swapped the tail bypass to 20 AWG.
I got one more 21700 host on the way (hope so), and got in 3 C8F 18650 hosts, 2 modded up.
One C8F with XP-L2 V6 5000K actually read 19.1 amps on a VTC5D cell, 5640-5050 lumens, 42.5 kcd.
I hope they get in more - doesn't sound or look good. They don't seem to sell the 18650 versions any longer as well.
