https://github.com/v1t3ls0n/assembler-openu-project/files/8794964/university_project_specs_doc.pdf
Task Summary: Create an assembler for an assembly programming language defined in The Open University (of Israel) System Programming Lab course final project.
-
memory efficiency: It was important for us to allocate the memory of the binary image output dynamically so that the memory use of the program would be efficient and that the memory allocation of the image the assembler generates would fit the minimum required for compiling source code.
-
Separation of concern: we tried to split the program into independent components that will be responsible for well-defined tasks and that the sharing of data between files will be through some kind of encapsulation method such as state getters/setters and scoping global variables to be static so that they will be scoped to the single file that they belong to.
-
Abstraction of data structures and variables used to fit our needs by defining different typedef structs types enums and constants.
The assembler parses the original source (.as) file and generates the expanded (.am) source file that contains the same source code content, the only difference is that the macros defined in the original source file (.as) are replaced with their content. To do so the assembler uses a hash table (Macro Table) to store the macro name with its starting and ending indexes (FILE position) of its content. If it finds any error in the syntax of the macros it yields a relevant error message and jumps straight to Next, otherwise, it continues to the First Run stage.
In the first run, the assembler counts the size of the data image and the instruction image (without allocating any memory), adds all the symbols to the symbol table, and last but not least the assembler verifies that there are no errors in code before it moves on to step 3. If the assembler encounters any error in the code it outputs an error/warning message to stderr and error.log (file which will be created in the root folder if any errors/warnings occur). To report any error in the source code in the first run, the assembler parses all the content of the source code regardless of any occurrence of an error. If any error (warnings do not count as errors) occurs, the assembler will not continue to stage 3 but it will go straight on to the Next stage.
If the first run ended successfully (else: Next) the assembler will update the address of each data symbol in the symbol table, and then it will allocate the exact amount of memory it needs to write for the target image (binary of compiled code) then it will reset the counters (IC/DC) and continue to the second run.
In the second run, our assembler writes the words of each line of code in binary format, inserts the words (binary bits) to the memory image in the correct position depending on the type of word (data/instruction) and adds each external operand that appeared in the correct spot in the external operands table.
If the assembler encounters a label operand that is not within the symbol table and is not external it yields an error message and then continues to check the rest of the code to discover all errors of this kind and report them (it is the only kind of error that can appear in the second run stage since any other type of error already been reported in the previous stages by now and has already prevented the assembler from entering the second run stage) and in this case, the assembler will finish the second run but will not export any files.
If the second run is finished without any error (else: Next) we will generate all of the required outputs (.ob, .ext .ent files).
The assembler then moves on to handle the next source file (which means that it will repeat over these 5 steps for each source file passed to it) until the last one then it ends the program.