add write block w/o length, renaming read_char

[elral]

read_char() is renamed to read_byte() as mainly a byte is required.
Within OnGetInfo() a char is required to send the config via the command messenger, at this point read_byte() is casted.
Additional functions to save/read a complete array or struct. This could be helpfull for future use (e.g. parameters for an analog calibration should be in a struct and this struct should be saved to the EEPROM). These additional functions and the existing functions for reading/saving arrays with fixed lenght are moved to EEPROM.h in a template.

storeName() and restoreName() are cleaned up slightly

Fixes #221

[github-actions]

Firmware for this pull request:
Firmware.zip

[github-actions]

Firmware for this pull request:
Firmware.zip

[DocMoebiuz]

TY!

[github-actions]

Firmware for this pull request:
Firmware.zip

[GioCC]

Not sure about the benefit of templating these functions.
I'm afraid that the compiler might generate redundant instances of the same function (unless we trust it to be smart enough - which I admit it might well be) if the template is instantiated for different types.
I see that even the basic EEPROM.get() uses templating - but it's still a framework library, so redefining it would get in the way of porting.
Passing by reference has no particular purpose other than mask the fact that we are getting its raw pointer anyway. What's bad in plainly invoking the basic functions with & and sizeof()?
Besides, defining the templates in the .h file rather than in the .cpp should inline them - is that desired?

(BTW, casting to uint8_t* would not work for non-POD objects - which we are not currently using and won't probably be using anytime soon, but still something to be aware of.)