If the cell has 3.65V and the charger is 3.65V then there is no exchange of current. No exchange of current does not boil electrolyte. Like I said, his failure makes no sense.
There is exchange of current if when you switch off the charger the cell voltage drops to 3.45 volts. The current will be small but never the less that current is what is holding the cell voltage high. Doing this means you are slowly overcharging the cell. At some point it will become a problem. Now it may be that you need a voltage higher than 3.65 volts to cause a problem, but there will be a voltage that does cause a problem and you need to know what this voltage is.
Maybe the voltage limiting is what you call the float charge?
Wikipedia has a good definition of float charge. http://en.wikipedia.org/wiki/Trickle_charging
If you are going to use lithium cells as a starting battery then you have to consider the possibility that you may be slowly over charging the cells if your alternator is always holding them at 3.65 volts. Just checking cell voltages to make sure they don't exceed some value probably isn't enough. What really needs to be done is some testing of cells at various voltages and see what happens. Hold 1 cell at 3.65 volts, another at 3.75 volts, another at 3.85 volts, etc and see what the long term effect is to the cell. Until someone does this and publishes the results you can't make blanket statements that 3.65 volts or any other voltage is safe to hold the cells at when running them in a vehicle with an alternator charging them. Without doing these tests, the best you can do is look for a float charge voltage. When I have found them on Lifepo4 cell spec sheets it has been around 3.40-3.45 volts. But mostly the information isn't there suggesting that the manufactures either don't know what the float voltage is or don't want you to float charge the cells.
A lot of electric cars have parasitic loads when switched "off" so that also means the main pack needs to keep the 12V battery topped up. Making bad, worse. The bricked Tesla Roadster's are a fine example.
The solution for this problem
is simple and is what I have implemented in my Matiz. I have a 12 volt lead acid battery that gets charged from the traction pack by a DC-DC converter. When the car is switched off a contactor in the middle of the pack opens meaning that there is no parasitic load on the traction pack from the DC-DC or anything else. The 12 volt battery handles the parasitic loads from the car and if it is parked long enough to flatten the 12 volt battery I can just recharge it. This has never happened. The only other parasitic load would be the bms cell monitors which is tiny and can easily be unplugged if necessary, again I have never had the need to do this.