That is interesting, the ratio switch is something I would have never suspected.
I am currently working on one for my bassist and I have found some more things everyone facing this issue can (and probably should) check.
First of all, the version of the pedal I have is a revision D, and the issue with the pedal is that there is no signal from the output and the LEDs all light up like it's latching to the power rail.
After what the OP said about the output op-amp U8, I have replaced it just in case to be safe, however, that didn't fix the issue. After many many hours of trying to figure out the circuitboard, I have found that it is much more likely that the IC responsible for the compression has died, upon which it took out a couple transistors with its last gasp. U5 is a TL064 (14-pin) 4 channel op-amp and where most of the compression magic happens from what I can tell. I still have to determine if that is actually gone (or maybe it's a problem with the selector switch, as above, tbd) but I have found 3 Transistors that had been damaged due to whatever happened.
Q12, Q13 and Q16 are all the same PNP type transistor and I found they had a Base-Emitter short. I have replaced those from what I guessed they were based on the SMD code, see GS code for reference in your database of choice.
(Also, almost all the SMD code markings on that board are very very weird, I haven't found some of those codes to this day...and had to do a bit of guesswork with some help of the database on yooneedone, but I haven't found a single exact marking code of any of the components on that board other than the op-amps.)
Anyway, replacing those had the effect that the latch-up was gone and I was able to see a signal at the output again, but it seems like none of the actual compression works, which is why I now suspect U5 to be faulty (and some of the signals around it were missing or really weak and weird.) So I have to further investigate this and will post when I know what else is going on in the circuit. Hope this helps so far.
What annoys me the most is the way Dunlop/MXR treats their customers with this issue. I have emailed them and asked if it is possible to obtain the schematic and even offered to sign an NDA, as from the previous posts I could already tell that this wouldn't be a welcome request at their end. I had even made it clear that I had already replaced and checked U8, so that this was working again and they would see that I knew what I was doing (more or less), but the response was just ridiculous. They slapped the most generic reply at me, treating this almost like an isolated incident, asking if I had isolated the pedal from everything in the signal chain (when it should have been clear that it was already opened and on my workbench, hooked up to frequency generator and oscilloscope) when they probably have a very clear idea what is wrong with their pedals.
Anyway, just venting my frustration, good luck to you all troubleshooting this overpriced failure!
Importan Update (Edit):
I have now found that Q12, Q13, Q16 are NOT PNP transistors. They are NJFETs and they very likely aren't even broken. I am very sorry for the mix up, but theis was just my initial guesswork without a proper schematic
So please don't act on this info, I am still trying to figure out what's really causing the issues and which bits really are or aren't broken. For now I have a little better understanding about what the circuit does, also in relation to the test points, and I will eventually try and post about it when (or if) I succeed.
One more thing: the ICs that can be found on the tiny LED board, U2, U3, U4 are all comparators and are only dealing with a DC-part coming off of the compressor sidechain to indicate the gain reduction. So it is in theory possible to get the pedal working for audio without utilising the indicator LEDs. But at this point that's not really helpful to anybody yet.