Formats 1 to 4
The opcode table records the base format, while a leading plus selects extended format 4.
Systems programming
A two pass SIC/XE assembler in one C++ program
This course project translates SIC/XE assembly into object code. The first pass assigns addresses and builds a symbol table. The second resolves operands, chooses addressing modes, and emits header, text, modification, and end records.
Assembly pipeline
Source becomes a relocatable object program
Parse source lines
Pass 1 and SYMTAB
Pass 2 encoding
Text records
H, T, M, E output
Pass 1
Addresses and symbols come first
The location counter starts at the operand of START. Each
instruction advances it by format length; directives such as
WORD, BYTE, RESW, and
RESB apply their own sizes. Labels are inserted into a
chained hash table, and duplicate symbols are reported as the listing
is built.
Directives
START, END, BASE, NOBASE, WORD, BYTE, RESW, and RESB affect state without all producing object bytes.
Symbols
A compact hash table maps labels to six digit addresses for operand resolution in pass 2.
Pass 2
Addressing flags determine the final instruction
The encoder handles immediate, indirect, indexed, program counter relative, base relative, and extended addressing. Format 4 symbol references also produce modification records. Text records stop at 30 bytes and split across reserved storage gaps.
| Syntax | Meaning | Encoding effect |
|---|---|---|
#VALUE |
Immediate | n = 0, i = 1 |
@VALUE |
Indirect | n = 1, i = 0 |
VALUE,X |
Indexed | x = 1 |
+JSUB TARGET |
Extended | e = 1 with a 20 bit address and modification record |
Object program
The output follows the standard record layout
H^COPY ^000000^00107A
T^000000^1E^...
M^000004^05
E^000000The sample COPY program exercises device I/O, subroutines, base addressing, format 4 calls, reserved memory, and byte constants.
Build and run
A Make target is enough
make
./assembler input.asm
The assembler writes the generated records to
output.obj and prints both pass listings to the terminal.