Turning on or off the candle (light) with the inline switch trips the GFI on the hair dryer - but never the wall mounted one. Push the reset button, and all's well.
I've measured some receptacle GFCI's tripping at 6mA and others at 4.2mA, which tend to nuisance trip more with motor loads and magnetic-ballasted lights. Inrush current is a common factor between such induction loads, and 1800W+ hair dryers.
Since, voltage drop increases with amperage, with high inrush current the GFCI may momentary detect more IČR losses across the (Hot with switch leg) than neutral without switch-leg, due to different impedance.
GFCI's may not always mix well (nuisance trip) with switch legs and high inrush currents. If the GFCI can't be replaced, try feeding the light (switch-leg wiring) from line side of GFCI.
Does anyone know if Residual Current Devices (RCD) and GFCI's trip at similar mA settings, or what the different standards are?
This is a very interesting problem! Does this happen if, while switched on, you unplug the candle and plug it back in? Or if you unscrew the light bulb? Does it happen for any other cord-and-plug loads, or just this candle?
With the hair dryer switched off, there should be no current possible at all through the CT in the GFCI. I suspect high-frequency arcing in the switch might be triggering some of the circuitry in the GFCI. If this is the case, it would only happen when you turn the switch, and not via any other way of turning the light on/off.
The hair dryer plug is usually not really a GFCI, it is an immersion detector. There is a ring in the barrel of the blower exhaust that is connected via the 3d wire to the plug and it looks for current on that wire. (at least in the ones I have taken apart). I have no clue as to why a 7w lamp would affect that. Do you have another dryer you can try in the same outlet?