Thu, 5 Feb 2009, 02:40

To my students:

Now that you all have access to the correctly operating assembler, one student
wants to be sure we really, really want you to deliver what the correct
assembler delivers.  That is,

A student writes:

	1. When testing our assembler against the one posted on the course
	website, is it important that our terminal error messages (i.e. Error:
	invalid opcode) match those of the example? Or is it only important that
	the error code match (i.e. the numbers 0 to 4)?

I encourage you to read the clarifications on the web site, and then, 
if the clarifications are not clear, please ask.  In this case, please 
read Clarification 12.  If it is not clear, then please ask me again.

	2. When my program executes the second pass, it writes to the output 
	file line by line. When an error is detected, the appropriate error 
	code is reported and the assembler exits. However, the output file 
	contains the instructions that were assembled up to that point. 
	The example assembler posted on the website does not write anything 
	to the file when an error occurs. Is this the desired behavior, or 
	is my implementation sufficient?

The student wants to know if it is enough to do what the correct assembler

If the program does not assemble, we are interested in why it didn't.  So, we
will check only the error code generated by your assembler.  We are not 
interested in looking at a partial binary produced for a buggy program.

	<<name withheld to protect the student who does not trust easily>>

Good luck finishing the program by Sunday night.

Yale Patt