SpikeGLX: fix event_count returning channel 0's count for every digital channel - #1951

While checking how each reader times its segments for #1937 I noticed that SpikeGLXRawIO._event_count takes its arguments in a different order from the base class, so the block index is used as the channel. On spikeglx/DigitalChannelTest_g0 event_count gives 326 for all eight digital channels while get_event_timestamps returns 326 events for XD0 and none for the other seven.

This PR puts the arguments in the base class order. I added that file to entities_to_test, where the shared read_events check fails on master with Wrong shape 0, 326. Before this the file was only read by test_nidq_digital_channel and never went through the shared checks. This is another case for #1945.