Is it safe

Bombaywalla is correct in stating that "the guys here are full of $$$", especially Elizabeth who leaves her gear on all the time, and myself being fabulously wealthy. True, if you're only using one source at a time you're OK...sadly you won't be able to play 2 sources of disparate music at the same time thus robbing you of bewilderingly cacophonous headache inducing events. Otherwise, get a pro audio mixing board and forget about Assumed Diligence.

I suspect that what Bombaywalla said will be correct in most and perhaps nearly all cases. But I would not feel 100.0% certain in any given case without analyzing a schematic for both sources, which is unlikely to be available, and/or obtaining clear confirmation that it is ok from the manufacturers of both sources.

I say that in part because bipolar transistors, which might be used in the output stages of the source components, commonly have specified "absolute maximum" reverse breakdown voltages for their base-to-emitter junctions in the area of 5 to 7 volts. A well designed output stage will incorporate diodes or other protection mechanisms to assure that number isn't exceeded, but I would not assume that all designs provide that protection. If the analog output of the sources is specified as 2 volts maximum, presumably on an "rms" basis, that corresponds to 2.8 volts peak, which is still comfortable in relation to 5 volts, but conceivably a significantly higher voltage could be present for a brief amount of time while the component is powering down. I doubt that would cause any immediate damage, but I wouldn't rule out the possibility of effects that are cumulative over time.

Also, some integrated circuit devices that conceivably could be used in the output stages have absolute maximum ratings on the voltage that can be applied to their inputs and outputs that are specified in terms of, and have values close to, the value of the power supply voltage that is applied to the device. When the device is off the power supply voltage is zero, so any significant voltage that is applied will exceed the absolute maximum spec in those cases. While the spec is probably not intended to be applicable to the powered down situation, and from a physical standpoint I doubt that there would be a problem, again I wouldn't be 100% certain.

And of course there is always the possibility of making a mistake and having both components turned on at the same time, which would result in one component having to drive the low output impedance of the other component. Most designs would be unharmed in that situation, but I would be hesitant to assert that all designs would be unharmed.

The bottom line: It would most likely be ok, but it is not good practice, and I personally would not do it.

Regards,
-- Al

One further thought, adding to my previous post: Noise that may enter the amp via the RCA cable while the XLR input is being listened to will not be rejected by the amp's balanced input circuit, since it will be present on only one of the two signals in the balanced signal pair. That might impact sonics. The amount of noise that is introduced into the RCA input in that situation may be much greater than when the RCA input is being listened to, since the powered down RCA source will presumably have a much higher output impedance than when it is powered up.

Regards,
-- Al

Don't do it unless the vendor's tech expert assures it is OK.