🚀 go-pugleaf

>Could any experts out there educate me WHY and HOW does 
self-modifying code use? 

The first program I ever wrote was self-modifying. In 
1976 with the 8080 for class assignment I wrote a ping 
pong program with LEDs. The rest of the class wrote code 
for each direction, but I wrote one "shell" routine and 
then inserted instructions into it after calculating the 
"pong" direction. My program was about 1/5 as long as 
all the others, which was important when you had only 1 
or 2K of memory. (The teacher had never heard of doing 
such a thing, and I had to convince him I deserved an 
A+)

>What the advantage of using self-modifying code that 
non-self-modifying code >cannot achieve?

On DECsystem-10s and 20s I wrote code to check the 
serial number so that people who stole the code could 
not run it (easily) on their processor. But if you put 
the instructions in the code itself, a smart hacker 
could remove them from the EXE file or insert a jump 
statement to skip over them. So I generated the check at 
run time by moving "numbers" to memory and then 
executing them. This make cracking the security much 
more difficult. I moved numbers to registers and then 
bit shifted them to make the instruction, then moved the 
PC  to the register. (God forbid trying this on a VAX!)

>Is there any compiler which will generate code that 
self-modified?  

If you mean C or BASIC, I'm sure there are some "hacker" 
compilers that allow you to play, but in general you 
have no way of knowing address locations that your C 
compiler will create, or even what instructions your C 
compiler will generate, so how can you write C 
statements that will modify something that does not yet 
exist? You have to use machine code, since then you know 
what is where. You can do it with any assembler as far 
as I  know, but you have to fuss with the OS and/or 
compile switches to actually compile and run the code 
and suppress error messages.

>A small and useful example of self-modifying will be 
very helpful.

Read above.