It probably just happens to work in this case - but more by luck than judgement!

It is not safe to assume that an assignment made in 'C' (ie, outside the #pragma ASM) won't have been "interfered with" by the time you get into the inline assembler.


Yes - it's what I tried to explain before:
http://www.8052.com/forum/read.phtml?id=143382

Again, you were lucky in this case - but it is not a good practice to adopt in general.


No: a skeleton is just a framework with no "meat" on it - you have included the meat!

The "skeletons" would look something like this:

void cout(unsigned char dato)
{
   // No meat here!
   // (it will be written later in assembler)
}

unsigned char cin(void)
{
   // No meat here!
   // (it will be written later in assembler)
}

You then get the compiler to translate that to assembler source;
You then edit the assembler source to insert the "meat"

See this Keil support article: http://www.keil.com/support/docs/50.htm