jake/SplitBit-Emulator
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Various bug fixes to assembler, added more data pointers.

Browse Source
This commit is contained in:
Anachronaut
2026-08-13 23:41:22 -04:00
parent b50210d127
commit c2440ae5fa
38 changed files with 1028 additions and 236 deletions
Download Patch File Download Diff File Expand all files Collapse all files
Source/Assembler/Assm-util.c
+95 -31
View File
@@ -9,6 +9,7 @@
#include <ctype.h>
#include "Assm-util.h"
#include "assembly.h"
#include "../Emulator/cpu.h" // For DATA_POINTERS, so the CPU stays the one source of truth.
int debug = 0;
@@ -43,42 +44,105 @@ int checkIfKeyword(intermediateElement *currentElement) {
int checkIfInstruction(intermediateElement *currentElement) {
char token[32];
strncpy(token, currentElement->token, sizeof(token)-1); //copying over one less than the total size of the buffer ensures we wind up with a null terminated string.
toUppercase(token);
if (getOpcode(token) != 0xFE) {
// It's a valid instruction, save its value and set its type.
currentElement->type = INSTRUCTION;
currentElement->byteValue = getOpcode(token);
currentElement->byteLength = 1;
if (debug) printf("Token: %s is an instruction.\n", currentElement->token);
return 1;
if (strlen(currentElement->token) >= sizeof(token)) {
// Longer than any mnemonic could be, so it is not one.
return 0;
}
return 0;
strcpy(token, currentElement->token);
toUppercase(token);
// An instruction that works through a Data Pointer may name which one by
// hanging a selector off the mnemonic, as in LDA.2. Split that off before
// looking the mnemonic up.
char *selector = strchr(token, '.');
if (selector) {
*selector = '\0';
selector++;
}
uint8_t opcode = getOpcode(token);
if (opcode == 0xFE) {
// It's not an instruction.
return 0;
}
currentElement->type = INSTRUCTION;
currentElement->byteValue = opcode;
currentElement->byteLength = 1;
currentElement->dataPointer = 0;
if (instructionTakesDataPointer(opcode)) {
// The selector is emitted whether or not it was written, so these are
// always two bytes. Leaving it off is the same as writing 0.
currentElement->byteLength = 2;
if (selector) {
char *end;
long value = strtol(selector, &end, 10);
if (*selector == '\0' || *end != '\0' || value < 0 || value >= DATA_POINTERS) {
fprintf(stderr, RED "Error: \"%s\" does not name a Data Pointer.\n Selectors run from 0 to %d.\n" RESET, currentElement->token, DATA_POINTERS - 1);
printf(" File: %s at line %d.\n", currentElement->fileName, currentElement->lineNumber);
exit(1);
}
currentElement->dataPointer = (uint8_t)value;
}
} else if (selector) {
fprintf(stderr, RED "Error: %s does not work through a Data Pointer, so it cannot take a selector.\n" RESET, token);
printf(" File: %s at line %d.\n", currentElement->fileName, currentElement->lineNumber);
exit(1);
}
if (debug) printf("Token: %s is an instruction using Data Pointer %d.\n", currentElement->token, currentElement->dataPointer);
return 1;
}
int checkIfLiteralValue(intermediateElement *currentElement) {
char token[32];
strncpy(token, currentElement->token, sizeof(token)-1);
if (token[0] == '0') {
// It's a literal value. Check if it's hex or dec.
if(token[1] == 'x') {
// It's a hex literal. Set its value and type.
currentElement->type = VALUE;
currentElement->byteLength = 1;
memmove(token, token + 2, strlen(token)); // Shift the string over to get rid of the 0x.
currentElement->byteValue = (uint8_t)strtol(token, NULL, 16);
if (debug) printf("Token: %s is a hexadecimal literal. \n", currentElement->token);
} else if (token[1] == 'd') {
// It's a decimal literal. Set its value and type.
currentElement->type = VALUE;
currentElement->byteLength = 1;
memmove(token, token + 2, strlen(token)); // Shift the string over to get rid of the 0d.
currentElement->byteValue = (uint8_t)strtol(token, NULL, 10);
if (debug) printf("Token: %s is a decimal literal. \n", currentElement->token);
}
return 1;
const char *token = currentElement->token;
if (token[0] != '0') {
// It's not a literal value.
return 0;
}
return 0;
// A leading zero means the programmer was trying to write a literal, so anything
// malformed from here on is an error. Falling through to the label check instead
// would quietly emit the wrong number of bytes and shift the rest of the program.
int base;
const char *baseName;
if (token[1] == 'x') {
base = 16;
baseName = "hexadecimal";
} else if (token[1] == 'd') {
base = 10;
baseName = "decimal";
} else {
fprintf(stderr, RED "Error: Malformed literal value \"%s\".\n Literals must be prefaced with 0x for hexadecimal or 0d for decimal.\n" RESET, token);
printf(" File: %s at line %d.\n", currentElement->fileName, currentElement->lineNumber);
exit(1);
}
// Everything after the prefix has to be a digit in that base.
const char *digits = token + 2;
if (*digits == '\0') {
fprintf(stderr, RED "Error: Literal value \"%s\" has no digits after its prefix.\n" RESET, token);
printf(" File: %s at line %d.\n", currentElement->fileName, currentElement->lineNumber);
exit(1);
}
for (const char *c = digits; *c; c++) {
if (!(base == 16 ? isxdigit((unsigned char)*c) : isdigit((unsigned char)*c))) {
fprintf(stderr, RED "Error: \"%c\" is not a %s digit, in literal value \"%s\".\n" RESET, *c, baseName, token);
printf(" File: %s at line %d.\n", currentElement->fileName, currentElement->lineNumber);
exit(1);
}
}
// The digits are all valid, so the only thing left to get wrong is the range.
long value = strtol(digits, NULL, base);
if (value > 255) {
fprintf(stderr, RED "Error: Literal value \"%s\" is too large to fit in one byte.\n Values must be in the range 0x00 to 0xFF, or 0d0 to 0d255.\n" RESET, token);
printf(" File: %s at line %d.\n", currentElement->fileName, currentElement->lineNumber);
exit(1);
}
currentElement->type = VALUE;
currentElement->byteLength = 1;
currentElement->byteValue = (uint8_t)value;
if (debug) printf("Token: %s is a %s literal. \n", token, baseName);
return 1;
}
int checkIfLabel(intermediateElement *currentElement) {