Without much forethought, I used my DMM to measure the resistance of the 24k resistor in the following circuit under development (red lead to the V+ pin end, and black lead to the + input pin end). There was nothing else attached to the LM258 or resistor at the time.
Later, I realized that the DMM applied a negative voltage to the + input relative to the the Vcc pin. The DMM apparently uses 100mV to check resistance (measured using my other multimeter - an analog one - by just attaching lead to lead).
Judging from the schematic, I would have forward biased the emitter-base junctions of the input transistors and the current source. I suspect that the voltage drop across those junctions is larger than 100 mV; if so, then there would not have been any current flow…and no damage. Right?
Also, this situation would be no different than having a voltage lower than V+ at one of the inputs (normal powered operation), except that in my case, the chip wasn’t being powered in the normal way.
PS: The first attachment is the schematic shown above. The last two attachments are excerpts from the datasheet with maximum values - I don’t think the negative voltage limit is applicable, since it’s with respect to ground.
PPS: I haven’t finished the circuit or tried it, yet.