Gentlemen,
nice work, nice driver, nice builds! I love ’em triple channel. 
.
On the LDO blowup:
Vout on pin5 was to my knowledge the sole version here on BLF for the SOT-23-5. Mattaus used the LT1761 in his Knucklehead V3, wight proposed 4 different LDOs with the same pinout for his AxxDD-SO8+LDO and yes, RMM used it in the BLF SRK V2. I kept it that way in V3.
ImA4Wheelr made the famous need-no-zeners thread, but he used completely different packages. I wasn’t aware of the LDOs RMM is selling and would not have thought that he changed the pinout.
TomE, glad to hear only the 4R7 blew up. When I read the datasheet of your LDO, it effectively connected GND to + on your board…
.
7135 in 2S driver:
I was sceptical when I read that RMM discourages from using the 7135 with 2S, so I asked him in his thread, because in prior times he wrote differently on the subject. Unfortunately he did not come back to the topic. 
So I took a 2S zener modded light and left it on for 8 days solely on the single 7135 (PWM255, 350mA). Twice a day I changed the cells for fresh ones. After the time, everything was still fine. The 7135 did not die, output in mA on PWM-255 was exactly as it was 8 days prior and Moon with PWM-1 was still working (MT-G2!).
After that, I didn’t really bother anymore and simply use it.
I’m sure heatsinking helps, so I use a certain eagle-package to solder the complete back of the 7135, and the driver is well connected to the body of the light. And it was a 350mA version of the 7135, the 380mA version is surely more prone to failure.
So in the end, with 350mA and proper heatsinking, odds are pretty good the 7135 will at least last quite a while.
.
NXP YLD vs YLC:
The PSMN3R2-30YL*C* is the older version (date in datasheet), technology was called NextPower instead of NextPowerS3 with the YL*D* MOSFETs. Is the S3 a major technology evolution or just some marketing hype? I don’t know.
YLD tend to have better values for RDSon in the 4.5V segment, compared to the YLC, while the latter are cheaper. QDG and QD values should be better with the S3 versions, that’s what their marketing says, but that’s not always the case. As QDG and QD only get truly relevant at higher switching frequencies than we use, I mostly care for RDSon.
But, truth be told, 3.2mOhm or 3.75mOhm, that is a difference of 0.00055 Ohm. The equivalent is 1.5mm of a copper cable of 0.5mm².