Compiler: Improve codegen on PowerPC and fix ppc/x64/64k assemblers.

Signed-off-by: Amlal El Mahrouss <amlal.elmahrouss@icloud.com>

12 files changed, 131 insertions, 57 deletions

diff --git a/.gitignore b/.gitignore
index d56db1e..48a4a74 100644
--- a/.gitignore
+++ b/.gitignore
@@ -8,6 +8,9 @@ local.properties
 .loadpath
 .recommenders
 
+# NewOS/MP-UX executable
+*.exec
+
 docs/
 
 # Compiled class file
@@ -220,4 +223,4 @@ _deps
 # The C/C++ compiler.
 mcc
 mpcc
-mpc++
\ No newline at end of file
+mpc++
diff --git a/Documentation/TODO.txt b/Documentation/TODO.txt
index 9c5214e..0e4f807 100644
--- a/Documentation/TODO.txt
+++ b/Documentation/TODO.txt
@@ -5,6 +5,7 @@
 	- Add support 'mov' which encodes a,b,c,d,e and 8,15 registers.
 
 - C compilers
-	- Add structs/enums, and a preprocssing program.
+    - Fix code generation (ppc done)
+	- Add structs/enums, and a preprocessing program (ppc todo)
 
 - Extended Syntax Checker (esc.exe -c-asm foo.s)
diff --git a/Examples/ExampleCDialect.S b/Examples/ExampleCDialect.S
index d9613ef..2a5203a 100644
--- a/Examples/ExampleCDialect.S
+++ b/Examples/ExampleCDialect.S
@@ -1,23 +1,28 @@
 # Path: Examples/ExampleCDialect.c
-# Language: PowerPC Assembly (Generated from ANSI C)
-# Build Date: 2024-4-4
+# Language: PowerPC Assembly (Generated from C)
+# Build Date: 2024-4-16
 
-dword export .code64 main
 
 
-	lda r2,0x1000
 
-	lda r12, import __MPCC_IF_PROC_417129812448
-	#r12 = Code to jump on, r11 right cond, r10 left cond.
-	beq r10, r11, r12
-dword export .code64 __MPCC_IF_PROC_417129812448
 
 
 
-	ldw r19,  import foo
-	jlr
+dword export .code64 __ImageStart
 
 
-	ldw r19,  57
-	jlr
+	li r3,0x1000
+
+	cmpw r10, r11
+	beq import__MPCC_IF_PROC_6168864856 
+dword export .code64 __MPCC_IF_PROC_6168864856
+
+
+
+	mr r31, r3
+	blr
+
+
+	mr r31, r0
+	blr
 
diff --git a/Examples/ExampleCDialect.c b/Examples/ExampleCDialect.c
index 899062c..8c65bf0 100644
--- a/Examples/ExampleCDialect.c
+++ b/Examples/ExampleCDialect.c
@@ -1,4 +1,10 @@
-int main(int argc, char const* argv[]) {
+struct {
+	int a;
+	int b;
+	int c;
+};
+
+int __ImageStart(int argc, char const* argv[]) {
   int* foo = 0x1000;
 
   if (foo == 57) {
diff --git a/Headers/AsmKit/CPU/ppc.hpp b/Headers/AsmKit/CPU/ppc.hpp
index 873552d..dea5b00 100644
--- a/Headers/AsmKit/CPU/ppc.hpp
+++ b/Headers/AsmKit/CPU/ppc.hpp
@@ -15,6 +15,7 @@
 #define OPTIONAL 0x4
 #define VMX 0x8
 #define CPU970 0x10 /* added to OPTIONAL insts that the 970 has */
+#define CPUMAHROUSS 0x12 /* optional mahrouss insts. */
 
 enum optype {
   NONE,    /* no operand */
diff --git a/Headers/StdKit/PEF.hpp b/Headers/StdKit/PEF.hpp
index 8d8cc4f..d5379c4 100644
--- a/Headers/StdKit/PEF.hpp
+++ b/Headers/StdKit/PEF.hpp
@@ -11,23 +11,23 @@
 // @file PEF.hpp
 // @brief Preferred Executable Format
 
-#define kPefMagic "PEF"
-#define kPefMagicFat "FEP"
+#define kPefMagic "Joy!"
+#define kPefMagicFat "yoJ!"
 
-#define kPefExt ".exe"
+#define kPefExt ".exec"
 #define kPefDylibExt ".lib"
 #define kPefLibExt ".slib"
 #define kPefObjectExt ".obj"
 #define kPefDebugExt ".dbg"
 
-#define kPefMagicLen 3
+#define kPefMagicLen 5
 
 #define kPefVersion 2
 #define kPefNameLen 64
 
 #define kPefBaseOrigin 0x1000000
 
-#define kPefStart "__start"
+#define kPefStart "__ImageStart"
 
 namespace CompilerKit {
 enum {
diff --git a/Sources/64asm.cc b/Sources/64asm.cc
index b955ad6..1d6404e 100644
--- a/Sources/64asm.cc
+++ b/Sources/64asm.cc
@@ -310,6 +310,14 @@ static bool asm_read_attributes(std::string &line) {
 
     auto name = line.substr(line.find("import") + strlen("import"));
 
+    /// sanity check to avoid stupid linker errors.
+    if (name.size() == 0)
+    {
+        detail::print_error("Invalid import",
+                              "ppcasm");
+          throw std::runtime_error("invalid_import");
+    }
+
     std::string result = std::to_string(name.size());
     result += kUndefinedSymbol;
 
diff --git a/Sources/ccplus.cc b/Sources/ccplus.cc
index ee73093..7e15afc 100644
--- a/Sources/ccplus.cc
+++ b/Sources/ccplus.cc
@@ -151,22 +151,23 @@ static bool kOnForLoop = false;
 static bool kInBraces = false;
 static size_t kBracesCount = 0UL;
 
-/* @brief C compiler backend for Mahrouss Logic C */
-class CompilerBackendClang final : public ParserKit::CompilerBackend {
+/* @brief C++ compiler backend for Mahrouss Logic C++ */
+class CompilerBackendCPlusPlus final : public ParserKit::CompilerBackend {
  public:
-  explicit CompilerBackendClang() = default;
-  ~CompilerBackendClang() override = default;
+  explicit CompilerBackendCPlusPlus() = default;
+  ~CompilerBackendCPlusPlus() override = default;
 
-  MPCC_COPY_DEFAULT(CompilerBackendClang);
+  MPCC_COPY_DEFAULT(CompilerBackendCPlusPlus);
 
   bool Compile(const std::string& text, const char* file) override;
 
-  const char* Language() override { return "C++20 based dialect for NewOS."; }
+  const char* Language() override;
+
 };
 
 /// compiler variables
 
-static CompilerBackendClang* kCompilerBackend = nullptr;
+static CompilerBackendCPlusPlus* kCompilerBackend = nullptr;
 static std::vector<detail::CompilerType> kCompilerVariables;
 static std::vector<std::string> kCompilerFunctions;
 
@@ -181,6 +182,8 @@ union number_cast final {
 };
 }  // namespace detail
 
+const char* CompilerBackendCPlusPlus::Language() override { return "C++"; }
+
 /////////////////////////////////////////////////////////////////////////////////////////
 
 // @name Compile
@@ -188,7 +191,7 @@ union number_cast final {
 
 /////////////////////////////////////////////////////////////////////////////////////////
 
-bool CompilerBackendClang::Compile(const std::string& text, const char* file) {
+bool CompilerBackendCPlusPlus::Compile(const std::string& text, const char* file) {
   if (text.empty()) return false;
 
   // if (expr)
@@ -370,7 +373,7 @@ MPCC_MODULE(CompilerCPlusPlus) {
   bool skip = false;
 
   kFactory.Mount(new AssemblyMountpointClang());
-  kCompilerBackend = new CompilerBackendClang();
+  kCompilerBackend = new CompilerBackendCPlusPlus();
 
   for (auto index = 1UL; index < argc; ++index) {
     if (argv[index][0] == '-') {
diff --git a/Sources/i64asm.cc b/Sources/i64asm.cc
index f7fa5e5..d19c515 100644
--- a/Sources/i64asm.cc
+++ b/Sources/i64asm.cc
@@ -367,6 +367,13 @@ static bool asm_read_attributes(std::string &line) {
 
     auto name = line.substr(line.find("import") + strlen("import") + 1);
 
+    if (name.size() == 0)
+    {
+        detail::print_error("Invalid import",
+                              "ppcasm");
+          throw std::runtime_error("invalid_import");
+    }
+
     std::string result = std::to_string(name.size());
     result += kUndefinedSymbol;
 
diff --git a/Sources/link.cc b/Sources/link.cc
index 8c88411..7a58083 100644
--- a/Sources/link.cc
+++ b/Sources/link.cc
@@ -193,6 +193,7 @@ MPCC_MODULE(NewOSLinker) {
   pef_container.Magic[0] = kPefMagic[kFatBinaryEnable ? 2 : 0];
   pef_container.Magic[1] = kPefMagic[1];
   pef_container.Magic[2] = kPefMagic[kFatBinaryEnable ? 0 : 2];
+  pef_container.Magic[3] = kPefMagic[3];
   pef_container.Version = kPefVersion;
 
   // specify the start address, can be 0x10000
diff --git a/Sources/ppc-cc.cc b/Sources/ppc-cc.cc
index 4e2f162..bb38439 100644
--- a/Sources/ppc-cc.cc
+++ b/Sources/ppc-cc.cc
@@ -320,16 +320,35 @@ bool CompilerBackendCLang::Compile(const std::string &text, const char *file) {
             value += tmp;
           }
 
-          syntaxLeaf.fUserValue = "\tldw r19, ";
+          syntaxLeaf.fUserValue = "\tmr r31, ";
 
           // make it pretty.
-          if (value.find('\t') != std::string::npos)
+          while (value.find('\t') != std::string::npos)
             value.erase(value.find('\t'), 1);
 
-          syntaxLeaf.fUserValue += value + "\n";
+          while (value.find(' ') != std::string::npos)
+            value.erase(value.find(' '), 1);
+
+          while (value.find("import") != std::string::npos)
+            value.erase(value.find("import"), strlen("import"));
+
+          bool found = false;
+
+          for (auto& reg : kState.kStackFrame)
+          {
+              if (value.find(reg.fName) != std::string::npos)
+              {
+                  found = true;
+                  syntaxLeaf.fUserValue += reg.fReg;
+                  break;
+              }
+          }
+
+          if (!found)
+                      syntaxLeaf.fUserValue += "r0";
         }
 
-        syntaxLeaf.fUserValue += "\tjlr";
+        syntaxLeaf.fUserValue += "\n\tblr";
 
         kState.fSyntaxTree->fLeafList.push_back(syntaxLeaf);
 
@@ -352,12 +371,13 @@ bool CompilerBackendCLang::Compile(const std::string &text, const char *file) {
       kIfFunction = "__MPCC_IF_PROC_";
       kIfFunction += std::to_string(time_off._Raw);
 
-      syntaxLeaf.fUserValue = "\tlda r12, import ";
-      syntaxLeaf.fUserValue +=
-          kIfFunction +
-          "\n\t#r12 = Code to jump on, r11 right cond, r10 left cond.\n\tbeq "
-          "r10, r11, r12\ndword export .code64 " +
-          kIfFunction + "\n";
+      syntaxLeaf.fUserValue =
+          "\tcmpw "
+          "r10, r11";
+
+      syntaxLeaf.fUserValue += "\n\tbeq import" + kIfFunction + " \ndword export .code64 " +
+      kIfFunction + "\n";
+
       kState.fSyntaxTree->fLeafList.push_back(syntaxLeaf);
 
       kIfFound = true;
@@ -421,11 +441,11 @@ bool CompilerBackendCLang::Compile(const std::string &text, const char *file) {
       if (textBuffer.find('=') != std::string::npos && kInBraces && !kIfFound) {
         if (textBuffer.find("*") != std::string::npos) {
           if (textBuffer.find("=") > textBuffer.find("*"))
-            substr += "\tlda ";
+            substr += "\tli ";
           else
-            substr += "\tldw ";
+            substr += "\tli ";
         } else {
-          substr += "\tldw ";
+          substr += "\tli ";
         }
       } else if (textBuffer.find('=') != std::string::npos && !kInBraces) {
         substr += "stw export .data64 ";
@@ -512,22 +532,33 @@ bool CompilerBackendCLang::Compile(const std::string &text, const char *file) {
                          return type.fName.find(substr) != std::string::npos;
                        });
 
-      if (kRegisterCounter == 5 || kRegisterCounter == 6) ++kRegisterCounter;
+    kCompilerVariables.push_back({.fName = substr});
+
+    if (textBuffer[text_index] == ';') break;
 
       std::string reg = kAsmRegisterPrefix;
+
+      ++kRegisterCounter;
       reg += std::to_string(kRegisterCounter);
 
-      if (var_to_find == kCompilerVariables.cend()) {
-        ++kRegisterCounter;
+      auto newSubstr = substr.substr(substr.find(" "));
+
+      std::string symbol;
+
+      for (size_t start = 0; start < newSubstr.size(); ++start) {
+        if (newSubstr[start] == ',')
+            break;
+
+        if (newSubstr[start] == ' ')
+            continue;
 
-        kState.kStackFrame.push_back({.fName = substr, .fReg = reg});
-        kCompilerVariables.push_back({.fName = substr});
+        symbol += (newSubstr[start]);
       }
 
-      syntaxLeaf.fUserValue += substr;
-      kState.fSyntaxTree->fLeafList.push_back(syntaxLeaf);
+      kState.kStackFrame.push_back({.fName = symbol, .fReg = reg});
 
-      if (textBuffer[text_index] == '=') break;
+      syntaxLeaf.fUserValue += "\n\tli " + reg + substr.substr(substr.find(','));
+      kState.fSyntaxTree->fLeafList.push_back(syntaxLeaf);
     }
 
     // function handler.
@@ -576,7 +607,7 @@ bool CompilerBackendCLang::Compile(const std::string &text, const char *file) {
           }
 
           args += args_buffer;
-          args += "\n\tlda r19, ";
+          args += "\n\tli r19, ";
         }
       }
 
@@ -598,7 +629,8 @@ bool CompilerBackendCLang::Compile(const std::string &text, const char *file) {
       if (kInBraces) {
         syntaxLeaf.fUserValue = args;
         syntaxLeaf.fUserValue += substr;
-        syntaxLeaf.fUserValue += "\n\tjrl\n";
+
+        syntaxLeaf.fUserValue += "\n\tblr\n";
 
         kState.fSyntaxTree->fLeafList.push_back(syntaxLeaf);
 
@@ -1139,7 +1171,7 @@ class AssemblyMountpointCLang final : public CompilerKit::AssemblyInterface {
 
     (*kState.fOutputAssembly) << "# Path: " << src_file << "\n";
     (*kState.fOutputAssembly)
-        << "# Language: PowerPC Assembly (Generated from ANSI C)\n";
+        << "# Language: PowerPC Assembly (Generated from C)\n";
     (*kState.fOutputAssembly) << "# Build Date: " << fmt << "\n\n";
 
     ParserKit::SyntaxLeafList syntax;
@@ -1204,12 +1236,12 @@ class AssemblyMountpointCLang final : public CompilerKit::AssemblyInterface {
             }
 
             if (ParserKit::find_word(leaf.fUserValue, needle)) {
-              if (leaf.fUserValue.find("import " + needle) !=
+              if (leaf.fUserValue.find("import ") !=
                   std::string::npos) {
-                std::string range = "import " + needle;
+                std::string range = "import ";
                 leaf.fUserValue.replace(
-                    leaf.fUserValue.find("import " + needle), range.size(),
-                    needle);
+                    leaf.fUserValue.find(range), range.size(),
+                    "");
               }
 
               if (leaf.fUserValue.find("ldw r6") != std::string::npos) {
diff --git a/Sources/ppcasm.cc b/Sources/ppcasm.cc
index eadedec..4bab2c4 100644
--- a/Sources/ppcasm.cc
+++ b/Sources/ppcasm.cc
@@ -315,6 +315,13 @@ static bool asm_read_attributes(std::string &line) {
 
     auto name = line.substr(line.find("import") + strlen("import") + 1);
 
+    if (name.size() == 0)
+    {
+        detail::print_error("Invalid import",
+                              "ppcasm");
+          throw std::runtime_error("invalid_import");
+    }
+
     std::string result = std::to_string(name.size());
     result += kUndefinedSymbol;