Some optimization with "EX DE, HL \ JP next"

[redirectme]

The system has several combinations

	EX DE, HL 
        JP next

If we denote (e.g., ex_de_hl_next: ) one of them, we can put "JP ex_de_hl_next" instead of the others. This saves 1 byte per combination with no loss of speed.
For example,

over:              
	POP DE     
	PUSH DE    
	PUSH HL    
ex_de_hl_next:     
	EX DE, HL  
	JP next    

And then modify:

minus_rot:                
	POP DE            
	POP AF            
	PUSH HL           
	PUSH AF           
	JP ex_de_hl_next  

Modify other locations in the same way.
In addition, we can aslo put another label:

over:              
	POP DE     
	PUSH DE
dup_ex_de_hl_next:         
	PUSH HL    
ex_de_hl_next:     
	EX DE, HL  
	JP next    

Then replace the combination

        PUSH HL
	EX DE, HL 
        JP next

by "JP dup_ex_de_hl_next"

For example,

rot:                    
	POP AF          
	POP DE          
	PUSH AF         
	JP dup_ex_de_hl_next

There are other places for JP dup_ex_de_hl_next. This saves 2 bytes without losing speed.

Added: I guess I'm wrong about same speed. Certainly "JP ex_de_hl_next" is a bit slower than the original code.