Add error messages for out of range numeric literals (#188)

* Add error messages for out of range numeric literals * Add test
spicelang · Aug 23, 2022 · 529c03a · 529c03a
1 parent 52d2361
commit 529c03a
Show file tree

Hide file tree

Showing 14 changed files with 201 additions and 61 deletions.
diff --git a/media/test-project/os-test.spice b/media/test-project/os-test.spice
@@ -10,8 +10,72 @@ f<int> main() {
     printf("Hello %s!", p1.getSecond());
 }*/
 
-f<int> main() {
-    foreach dyn item : { 1, 2, 3, 4, 5 } {
-        printf("Item %d: %d\n", idx, item);
+import "std/type/int" as integer;
+
+const int vertexCount = 9;
+
+f<int> minDistance(int[] dist, bool[] sptSet) {
+    int min = integer.MAX_VALUE;
+    int minIndex;
+
+    for int v = 0; v < vertexCount; v++ {
+        if !sptSet[v] && dist[v] <= min {
+            min = dist[v];
+            minIndex = v;
+        }
     }
+
+    return minIndex;
+}
+
+p printSolution(int[] dist) {
+    printf("Vertex \t\t Distance from source\n");
+    for int i = 0; i < vertexCount; i++ {
+        printf("%d \t\t %d\n", i, dist[i]);
+    }
+}
+
+p dijkstra(int[vertexCount][vertexCount] graph, int src) {
+    int[vertexCount] dist;
+    bool[vertexCount] sptSet;
+
+    // Fill with default values
+    for int i = 0; i < vertexCount; i++ {
+        dist[i] = integer.MAX_VALUE;
+        sptSet[i] = false;
+    }
+
+    // Set distance to starting node to 0
+    dist[src] = 0;
+
+    for int count = 0; count < vertexCount - 1; count++ {
+        int u = minDistance(dist, sptSet);
+        sptSet[u] = true;
+        for int v = 0; v < vertexCount; v++ {
+            if (!sptSet[v] && graph[u][v] != 0 && dist[u] != integer.MAX_VALUE &&
+                dist[u] + graph[u][v] < dist[v]) {
+                dist[v] = dist[u] + graph[u][v];
+            }
+        }
+    }
+
+    // Print the solution
+    printSolution(dist);
+}
+
+f<int> main() {
+    int[vertexCount][vertexCount] graph = {};
+    graph[0] = { 0, 4, 0, 0, 0, 0, 0, 8, 0 };
+    graph[1] = { 4, 0, 8, 0, 0, 0, 0, 11, 0 };
+    graph[2] = { 0, 8, 0, 7, 0, 4, 0, 0, 2 };
+    graph[3] = { 0, 0, 7, 0, 9, 14, 0, 0, 0 };
+    graph[4] = { 0, 0, 0, 9, 0, 10, 0, 0, 0 };
+    graph[5] = { 0, 0, 4, 14, 10, 0, 2, 0, 0 };
+    graph[6] = { 0, 0, 0, 0, 0, 2, 0, 1, 6 };
+    graph[7] = { 8, 11, 0, 0, 0, 0, 1, 0, 7 };
+    graph[8] = { 0, 0, 2, 0, 0, 0, 6, 7, 0 };
+
+    printf("Computing shortest paths with Dijkstra's algorithm ...\n");
+    dijkstra(graph, 0);
+    printf("Done.\n");
 }
diff --git a/src/exception/AntlrThrowingErrorListener.cpp b/src/exception/AntlrThrowingErrorListener.cpp
@@ -2,13 +2,11 @@
 
 #include "AntlrThrowingErrorListener.h"
 
-#include "LexerParserError.h"
+#include <exception/LexerParserError.h>
+#include <util/CodeLoc.h>
 
 void AntlrThrowingErrorListener::syntaxError(antlr4::Recognizer *recognizer, antlr4::Token *offendingSymbol, size_t line,
                                              size_t charPositionInLine, const std::string &msg, std::exception_ptr e) {
-  // Create message fragments
-  std::string errorType = mode == LEXER ? "Lexer error" : "Parser error";
-  std::string codePos = std::to_string(line) + ":" + std::to_string(charPositionInLine);
-  // Throw error
-  throw LexerParserError(errorType + " at line " + codePos + " " + msg);
+  CodeLoc codeLoc = CodeLoc("", line, charPositionInLine);
+  throw LexerParserError(codeLoc, mode == LEXER ? TOKENIZING_FAILED : PARSING_FAILED, msg);
 }
diff --git a/src/exception/ErrorFactory.cpp b/src/exception/ErrorFactory.cpp
@@ -4,6 +4,10 @@
 
 #include <util/CodeLoc.h>
 
+LexerParserError ErrorFactory::get(const CodeLoc &codeLoc, LexerParserErrorType type, const std::string &message) {
+  return LexerParserError(codeLoc, type, message);
+}
+
 SemanticError ErrorFactory::get(const CodeLoc &codeLoc, SemanticErrorType type, const std::string &message) {
   return SemanticError(codeLoc, type, message);
 }

diff --git a/src/exception/ErrorFactory.h b/src/exception/ErrorFactory.h
@@ -7,6 +7,7 @@
 
 #include <exception/CliError.h>
 #include <exception/IRError.h>
+#include <exception/LexerParserError.h>
 #include <exception/LinkerError.h>
 #include <exception/SemanticError.h>
 
@@ -19,9 +20,10 @@ class ErrorFactory {
   ErrorFactory() = default;
 
   // Public methods
-  [[nodiscard]] static SemanticError get(const CodeLoc &codeLoc, SemanticErrorType type, const std::string &message) ;
-  [[nodiscard]] static IRError get(const CodeLoc &codeLoc, IRErrorType type, const std::string &message) ;
-  [[nodiscard]] static IRError get(IRErrorType type, const std::string &message) ;
+  [[nodiscard]] static LexerParserError get(const CodeLoc &codeLoc, LexerParserErrorType type, const std::string &message);
+  [[nodiscard]] static SemanticError get(const CodeLoc &codeLoc, SemanticErrorType type, const std::string &message);
+  [[nodiscard]] static IRError get(const CodeLoc &codeLoc, IRErrorType type, const std::string &message);
+  [[nodiscard]] static IRError get(IRErrorType type, const std::string &message);
   [[nodiscard]] static CliError get(CliErrorType type, const std::string &message);
   [[nodiscard]] static LinkerError get(LinkerErrorType type, const std::string &message);
 };
diff --git a/src/exception/LexerParserError.cpp b/src/exception/LexerParserError.cpp
@@ -2,9 +2,42 @@
 
 #include "LexerParserError.h"
 
+#include <util/CodeLoc.h>
+
+/**
+ * Constructor: Used in case that the exact code position where the error occurred is known
+ *
+ * @param codeLoc Code location where the error occurred
+ * @param type Type of the error
+ * @param message Error message suffix
+ */
+LexerParserError::LexerParserError(const CodeLoc &codeLoc, const LexerParserErrorType &type, const std::string &message) {
+  errorMessage = "Error in " + codeLoc.toPrettyString() + ": " + getMessagePrefix(type) + ": " + message;
+}
+
 /**
  * Get the message for this particular error instance
  *
  * @return Error message in form of a char array
  */
-const char *LexerParserError::what() const noexcept { return errorMessage.c_str(); }
+const char *LexerParserError::what() const noexcept { return errorMessage.c_str(); }
+
+/**
+ * Get the prefix of the error message for a particular error
+ *
+ * @param type Type of the error
+ * @return Prefix string for the error type
+ */
+std::string LexerParserError::getMessagePrefix(LexerParserErrorType type) {
+  switch (type) {
+  case TOKENIZING_FAILED:
+    return "Tokenizing failed";
+  case PARSING_FAILED:
+    return "Parsing failed";
+  case NUMBER_OUT_OF_RANGE:
+    return "Number is out of range";
+  case INVALID_CHAR_LITERAL:
+    return "Invalid char literal";
+  }
+  return "Unknown error"; // GCOV_EXCL_LINE
+}
diff --git a/src/exception/LexerParserError.h b/src/exception/LexerParserError.h
@@ -6,18 +6,26 @@
 #include <string>
 #include <utility>
 
+// Forward declarations
+struct CodeLoc;
+
+enum LexerParserErrorType { TOKENIZING_FAILED, PARSING_FAILED, NUMBER_OUT_OF_RANGE, INVALID_CHAR_LITERAL };
+
 /**
  * Custom exception for errors, occurring in the semantic analysis phase
  */
 class LexerParserError : public std::exception {
 public:
   // Constructors
-  explicit LexerParserError(std::string message) : errorMessage(std::move(message)) {} // GCOV_EXCL_LINE
+  explicit LexerParserError(const CodeLoc &codeLoc, const LexerParserErrorType &type, const std::string &message);
 
   // Public methods
   [[nodiscard]] const char *what() const noexcept override;
 
 private:
   // Members
   std::string errorMessage;
+
+  // Private methods
+  [[nodiscard]] static std::string getMessagePrefix(LexerParserErrorType errorType);
 };
diff --git a/src/exception/SemanticError.h b/src/exception/SemanticError.h
@@ -82,5 +82,6 @@ class SemanticError : public std::exception {
   // Members
   std::string errorMessage;
 
+  // Private methods
   [[nodiscard]] static std::string getMessagePrefix(SemanticErrorType errorType);
 };
diff --git a/src/parser/AstBuilderVisitor.cpp b/src/parser/AstBuilderVisitor.cpp
@@ -7,6 +7,12 @@
 #include <ast/AstNodes.h>
 #include <util/CommonUtil.h>
 
+AstBuilderVisitor::AstBuilderVisitor(AstNode *rootNode, std::string fileName)
+    : currentNode(rootNode), fileName(std::move(fileName)) {
+  // Create error factory for this specific file
+  this->err = std::make_unique<ErrorFactory>();
+}
+
 std::any AstBuilderVisitor::visitEntry(SpiceParser::EntryContext *ctx) {
   auto entryNode = dynamic_cast<EntryNode *>(currentNode);
   for (const auto &subTree : ctx->children) {
@@ -1197,17 +1203,17 @@ std::any AstBuilderVisitor::visitPrimitiveValue(SpiceParser::PrimitiveValueConte
       primitiveValueNode->data.doubleValue = std::stod(t->toString());
     } else if (auto t = dynamic_cast<antlr4::tree::TerminalNode *>(subTree); t->getSymbol()->getType() == SpiceParser::INT_LIT) {
       primitiveValueNode->type = PrimitiveValueNode::TYPE_INT;
-      primitiveValueNode->data.intValue = parseInt(t->toString());
+      primitiveValueNode->data.intValue = parseInt(t);
     } else if (auto t = dynamic_cast<antlr4::tree::TerminalNode *>(subTree);
                t->getSymbol()->getType() == SpiceParser::SHORT_LIT) {
       primitiveValueNode->type = PrimitiveValueNode::TYPE_SHORT;
-      primitiveValueNode->data.shortValue = parseShort(t->toString());
+      primitiveValueNode->data.shortValue = parseShort(t);
     } else if (auto t = dynamic_cast<antlr4::tree::TerminalNode *>(subTree); t->getSymbol()->getType() == SpiceParser::LONG_LIT) {
       primitiveValueNode->type = PrimitiveValueNode::TYPE_LONG;
-      primitiveValueNode->data.longValue = parseLong(t->toString());
+      primitiveValueNode->data.longValue = parseLong(t);
     } else if (auto t = dynamic_cast<antlr4::tree::TerminalNode *>(subTree); t->getSymbol()->getType() == SpiceParser::CHAR_LIT) {
       primitiveValueNode->type = PrimitiveValueNode::TYPE_CHAR;
-      primitiveValueNode->data.charValue = parseChar(ctx->CHAR_LIT()->toString());
+      primitiveValueNode->data.charValue = parseChar(ctx->CHAR_LIT());
     } else if (auto t = dynamic_cast<antlr4::tree::TerminalNode *>(subTree);
                t->getSymbol()->getType() == SpiceParser::STRING_LIT) {
       primitiveValueNode->type = PrimitiveValueNode::TYPE_STRING;
@@ -1478,59 +1484,76 @@ void AstBuilderVisitor::replaceEscapeChars(std::string &string) {
   CommonUtil::replaceAll(string, "\\?", "\?");
 }
 
-int32_t AstBuilderVisitor::parseInt(const std::string &input) {
-  std::function<int32_t(const std::string &, int)> cb = [](const std::string &substr, int base) {
+int32_t AstBuilderVisitor::parseInt(antlr4::tree::TerminalNode *terminal) {
+  std::function<int32_t(const std::string &, int)> cb = [this](const std::string &substr, int base) {
     return std::stoi(substr, nullptr, base);
   };
-  return parseNumeric(input, cb);
+  return parseNumeric(terminal, cb);
 }
-int16_t AstBuilderVisitor::parseShort(const std::string &input) {
-  std::function<int16_t(const std::string &, int)> cb = [](const std::string &substr, int base) {
+int16_t AstBuilderVisitor::parseShort(antlr4::tree::TerminalNode *terminal) {
+  std::function<int16_t(const std::string &, int)> cb = [this](const std::string &substr, int base) {
     return (int16_t)std::stoi(substr, nullptr, base);
   };
-  return parseNumeric(input, cb);
+  return parseNumeric(terminal, cb);
 }
 
-int64_t AstBuilderVisitor::parseLong(const std::string &input) {
-  std::function<int64_t(const std::string &, int)> cb = [](const std::string &substr, int base) {
+int64_t AstBuilderVisitor::parseLong(antlr4::tree::TerminalNode *terminal) {
+  std::function<int64_t(const std::string &, int)> cb = [this](const std::string &substr, int base) {
     return std::stoll(substr, nullptr, base);
   };
-  return parseNumeric(input, cb);
+  return parseNumeric(terminal, cb);
 }
 
-int8_t AstBuilderVisitor::parseChar(const std::string &input) { return input[1]; }
+int8_t AstBuilderVisitor::parseChar(antlr4::tree::TerminalNode *terminal) {
+  std::string input = terminal->toString();
+  if (input.length() != 3) {
+    CodeLoc codeLoc = CodeLoc(fileName, terminal->getSymbol());
+    throw err->get(codeLoc, PARSING_FAILED, "Invalid char literal " + input);
+  }
+  return input[1];
+}
 
 std::string AstBuilderVisitor::parseString(std::string input) {
   input = input.substr(1, input.size() - 2);
   replaceEscapeChars(input);
   return input;
 }
 
-template <typename T> T AstBuilderVisitor::parseNumeric(const std::string &input, std::function<T(const std::string &, int)> cb) {
-  if (input.length() >= 3) {
-    char c1 = input[0];
-    char c2 = input[1];
-    std::string substr = input.substr(2);
-    if (c1 == '0') {
-      switch (c2) {
-      case 'd':
-      case 'D':
-        return cb(substr, 10);
-      case 'b':
-      case 'B':
-        return cb(substr, 2);
-      case 'h':
-      case 'H':
-      case 'x':
-      case 'X':
-        return cb(substr, 16);
-      case 'o':
-      case 'O':
-        return cb(substr, 8);
-      default:
-        return cb(input, 10);
+template <typename T>
+T AstBuilderVisitor::parseNumeric(antlr4::tree::TerminalNode *terminal, std::function<T(const std::string &, int)> cb) {
+  std::string input = terminal->toString();
+  try {
+    if (input.length() >= 3) {
+      char c1 = input[0];
+      char c2 = input[1];
+      std::string substr = input.substr(2);
+      if (c1 == '0') {
+        switch (c2) {
+        case 'd':
+        case 'D':
+          return cb(substr, 10);
+        case 'b':
+        case 'B':
+          return cb(substr, 2);
+        case 'h':
+        case 'H':
+        case 'x':
+        case 'X':
+          return cb(substr, 16);
+        case 'o':
+        case 'O':
+          return cb(substr, 8);
+        default:
+          return cb(input, 10);
+        }
       }
     }
+    return cb(input, 10);
+  } catch (std::out_of_range &e) {
+    CodeLoc codeLoc = CodeLoc(fileName, terminal->getSymbol());
+    throw err->get(codeLoc, NUMBER_OUT_OF_RANGE, "The provided number is out of range");
+  } catch (std::invalid_argument &e) {
+    CodeLoc codeLoc = CodeLoc(fileName, terminal->getSymbol());
+    throw err->get(codeLoc, NUMBER_OUT_OF_RANGE, "You tried to parse '" + input + "' as an integer, but it was no integer");
   }
-  return cb(input, 10);
 }
diff --git a/src/parser/AstBuilderVisitor.h b/src/parser/AstBuilderVisitor.h
@@ -7,13 +7,15 @@
 #include <functional>
 #include <utility>
 
+#include <exception/ErrorFactory.h>
+
 // Forward declarations
 class AstNode;
 
 class AstBuilderVisitor : public SpiceVisitor {
 public:
   // Constructors
-  explicit AstBuilderVisitor(AstNode *rootNode, std::string fileName) : currentNode(rootNode), fileName(std::move(fileName)) {}
+  explicit AstBuilderVisitor(AstNode *rootNode, std::string fileName);
 
   // Public methods
   std::any visitEntry(SpiceParser::EntryContext *ctx) override;
@@ -85,13 +87,14 @@ class AstBuilderVisitor : public SpiceVisitor {
   // Members
   AstNode *currentNode;
   std::string fileName;
+  std::unique_ptr<ErrorFactory> err;
 
   // Private methods
-  static int32_t parseInt(const std::string &input);
-  static int16_t parseShort(const std::string &input);
-  static int64_t parseLong(const std::string &input);
-  static int8_t parseChar(const std::string &input);
+  int32_t parseInt(antlr4::tree::TerminalNode *terminal);
+  int16_t parseShort(antlr4::tree::TerminalNode *terminal);
+  int64_t parseLong(antlr4::tree::TerminalNode *terminal);
+  int8_t parseChar(antlr4::tree::TerminalNode *terminal);
   static std::string parseString(std::string input);
-  template <typename T> T static parseNumeric(const std::string &input, std::function<T(const std::string &, int)> cb);
+  template <typename T> T parseNumeric(antlr4::tree::TerminalNode *terminal, std::function<T(const std::string &, int)> cb);
   static void replaceEscapeChars(std::string &string);
 };