Improve assembly error reporting (#323)

simulator/src/cpu/alu.ts

@@ -23,8 +23,11 @@ export type COMMANDS_ASM = typeof commandASMValues extends Set<infer S>
   ? S
   : never;
 
-export function isCommandAsm(command: unknown): command is COMMANDS_ASM {
-  return commandASMValues.has(command as COMMANDS_ASM);
+export function isCommandAsm(command: string): command is COMMANDS_ASM {
+  return (
+    commandASMValues.has(command as COMMANDS_ASM) ||
+    commandASMValues.has(command.replace("M", "A") as COMMANDS_ASM)
+  );
 }
 
 export type COMMANDS_OP =

simulator/src/languages/asm.ts

@@ -10,6 +10,8 @@ import {
   JUMP,
   JUMP_ASM,
   JUMP_OP,
+  isAssignAsm,
+  isCommandAsm,
 } from "../cpu/alu.js";
 import { KEYBOARD_OFFSET, SCREEN_OFFSET } from "../cpu/memory.js";
 import { makeC } from "../util/asm.js";
@@ -104,19 +106,56 @@ asmSemantics.addAttribute<AsmInstruction>("intermediateInstruction", {
   },
 });
 
+function getAsmAssign(assignN: ohm.Node) {
+  let assign = assignN.child(0)?.child(0)?.sourceString ?? "";
+
+  // The book (figure 4.5) specifies DM and ADM as the correct forms for destination,
+  // but since the desktop simulators accept only MD and AMD we have decided to accept both */
+  if (assign == "DM") {
+    assign = "MD";
+  }
+  if (assign == "ADM") {
+    assign = "AMD";
+  }
+  if (assign != "" && !isAssignAsm(assign)) {
+    const reversed = assign.split("").reverse().join("");
+    const suggestion = isAssignAsm(reversed)
+      ? `. Did you mean ${reversed}?`
+      : "";
+    throw createError(
+      `Invalid ASM target: ${assign}${suggestion}`,
+      span(assignN.source)
+    );
+  }
+
+  return assign;
+}
+
+function getAsmOp(opN: ohm.Node) {
+  const op = opN.sourceString;
+
+  if (!isCommandAsm(op)) {
+    const reversed = op.split("").reverse().join("");
+    const suggestion = isCommandAsm(reversed)
+      ? `. Did you mean ${reversed}?`
+      : "";
+    throw createError(
+      `Invalid ASM value: ${opN.sourceString}${suggestion}`,
+      span(opN.source)
+    );
+  }
+
+  return op;
+}
+
 asmSemantics.addAttribute<AsmInstruction>("instruction", {
   aInstruction(_at, name): AsmAInstruction {
     return A(name.value, span(this.source));
   },
   cInstruction(assignN, opN, jmpN): AsmCInstruction {
-    let assign = assignN.child(0)?.child(0)?.sourceString ?? "";
-    if (assign == "DM") {
-      assign = "MD";
-    }
-    if (assign == "ADM") {
-      assign = "AMD";
-    }
-    const op = opN.sourceString as COMMANDS_ASM;
+    const assign = getAsmAssign(assignN);
+
+    const op = getAsmOp(opN) as COMMANDS_ASM;
     const jmp = (jmpN.child(0)?.child(1)?.sourceString ?? "") as JUMP_ASM;
     return C(assign as ASSIGN_ASM, op, jmp, span(this.source));
   },

simulator/src/languages/grammars/asm.ohm

@@ -8,32 +8,12 @@ ASM <: Base {
   label = openParen identifier closeParen
   aInstruction = at (identifier | decNumber)
   cInstruction = assign? op jmp?
-
-  /* The book (figure 4.5) specifies DM and ADM as the correct forms for destination,
-     but since the desktop simulators accept only MD and AMD we have decided to accept both */
-  assign = (
-      "AMD" | "ADM"
-      | "AM"
-      | "AD"
-      | "MD" | "DM"
-      | "M"
-      | "D"
-      | "A"
-      ) equal
-
-  op =
-      | "0" | "1" | "-1"
-      | "!D" | "!A" | "!M"
-      | "-D" | "-A" | "-M"
-      | "D+1" | "A+1" | "M+1"
-      | "D-1" | "A-1" | "M-1"
-      | "D+A" | "D+M"
-      | "D-A" | "D-M"
-      | "A-D" | "M-D"
-      | "D&A" | "D&M"
-      | "D|A" | "D|M"
-      | "D" | "A" | "M"
 
+  assignChar = "A" | "M" | "D"
+  opChar = assignChar | "0" | "1" | "!" | "-" | "+" | "|" | "&"
+
+  assign = assignChar+ equal
+  op = opChar+
 
   jmp = semi ("JGT" | "JEQ" | "JGE" | "JLT" | "JNE" | "JLE" | "JMP")
 }

simulator/src/languages/grammars/asm.ohm.js

@@ -1,5 +1,4 @@
-const asm = `
-ASM <: Base {
+const asm = `ASM <: Base {
   Root := ASM
   ASM = intermediateInstruction* instruction?
   
@@ -9,32 +8,12 @@ ASM <: Base {
   label = openParen identifier closeParen
   aInstruction = at (identifier | decNumber)
   cInstruction = assign? op jmp?
-  
-  /* The book (figure 4.5) specifies DM and ADM as the correct forms for destination,
-     but since the desktop simulators accept only MD and AMD we have decided to accept both */
-  assign = (
-      "AMD" | "ADM"
-      | "AM"
-      | "AD"
-      | "MD" | "DM"
-      | "M"
-      | "D"
-      | "A"
-      ) equal
-      
-  op =
-      | "0" | "1" | "-1"
-      | "!D" | "!A" | "!M"
-      | "-D" | "-A" | "-M"
-      | "D+1" | "A+1" | "M+1"
-      | "D-1" | "A-1" | "M-1"
-      | "D+A" | "D+M"
-      | "D-A" | "D-M"
-      | "A-D" | "M-D"
-      | "D&A" | "D&M"
-      | "D|A" | "D|M"
-      | "D" | "A" | "M"
 
+  assignChar = "A" | "M" | "D"
+  opChar = assignChar | "0" | "1" | "!" | "-" | "+" | "|" | "&"
+  
+  assign = assignChar+ equal
+  op = opChar+
 
   jmp = semi ("JGT" | "JEQ" | "JGE" | "JLT" | "JNE" | "JLE" | "JMP")
 }`;

simulator/src/util/asm.ts

@@ -87,7 +87,7 @@ function cop(asm: string): number {
     parts?.[0] != asm || // match is not exhaustive
     !isAssignAsm(assign) ||
     !isJumpAsm(jump) ||
-    (!isCommandAsm(operation) && !isCommandAsm(operation.replace("M", "A")))
+    !isCommandAsm(operation)
   ) {
     // TODO: This should return Result<> instead of throw
     throw new Error("Invalid c instruction");