You're right, that master ACK comment might not really make sense.If your STM32 master sends a read command, check if the slave acknowledges it (some NACK bit in ISR? Still don't have the manual here...).If it doesn't the STM will probably not throw ...
Everything looks alright on the scope? SCL & SDA timing?SDA sends what it should, slave really ACKs?And STM32 ACKs in case slave sends?Maybe it's the timeout?I can't remember right now, maybe the enable bit in CR1 must be set after setting the addres...