added code to handle symbols that have no section and need to be

allocated with zeros...

CMakeLists.txt

@@ -55,8 +55,17 @@ set(Theodosius_SOURCES "")
 
 list(APPEND Theodosius_SOURCES
 	"include/comp/comp.hpp"
+	"include/comp/obf.hpp"
 	"include/comp/reloc.hpp"
 	"include/comp/symbol_table.hpp"
+	"include/comp/transform/add_op.hpp"
+	"include/comp/transform/and_op.hpp"
+	"include/comp/transform/operation.hpp"
+	"include/comp/transform/or_op.hpp"
+	"include/comp/transform/rol_op.hpp"
+	"include/comp/transform/ror_op.hpp"
+	"include/comp/transform/sub_op.hpp"
+	"include/comp/transform/xor_op.hpp"
 	"include/decomp/decomp.hpp"
 	"include/decomp/routine.hpp"
 	"include/decomp/symbol.hpp"

src/tests/demo/main.cpp

@@ -47,5 +47,5 @@ int main(int argc, char* argv[]) {
   spdlog::info("decomposed {} symbols...", res.value());
   auto entry_pnt = t.compose("EntryPoint");
   spdlog::info("entry point address: {:X}", entry_pnt);
-  reinterpret_cast<void (*)()>(entry_pnt)();
+  // reinterpret_cast<void (*)()>(entry_pnt)();
 }

src/tests/demolib/main.cpp

@@ -1,3 +1,5 @@
+#define OBF __declspec(code_seg(".obf"))
+
 extern "C" int MessageBoxA(void* hWnd,
                           char* lpText,
                           char* lpCaption,
@@ -9,7 +11,7 @@ struct test_t {
 
 test_t t = {};
 
-extern "C" void EntryPoint() {
+OBF extern "C" void EntryPoint() {
   t.buff[0] = 1;
   t.buff[1] = 2;
   MessageBoxA(nullptr, "Hello World", "Hello World", nullptr);

src/theo/comp/comp.cpp

@@ -33,6 +33,10 @@ void comp_t::allocate() {
   //
   m_dcmp->syms()->for_each([&](theo::decomp::symbol_t& sym) {
     if (sym.sym_type() == decomp::sym_type_t::data) {
+      // if the symbol has a section then we will refer to the allocation made
+      // for that section...
+      //
+      if (sym.scn()) {
         auto scn_sym =
             m_dcmp->syms()->sym_from_hash(m_dcmp->scn_hash_tbl()[sym.scn()]);
 
@@ -44,6 +48,19 @@ void comp_t::allocate() {
         }
 
         sym.allocated_at(scn_sym.value().allocated_at() + sym.offset());
+      } else {  // else if there is no section then we allocate based upon the
+                // size of the symbol... this is only done for symbols that are
+                // bss...
+                //
+
+        // bss is read write...
+        //
+        coff::section_characteristics_t prot = {};
+        prot.mem_read = true;
+        prot.mem_write = true;
+
+        sym.allocated_at(m_allocator(sym.size(), prot));
+      }
     }
   });
 }

src/theo/decomp/decomp.cpp

@@ -48,15 +48,16 @@ std::optional<comp::symbol_table_t*> decomp_t::decompose() {
       // if the symbol is a function then we are going to decompose it...
       // data symbols are handled after this...
       //
-      if (sym->has_section() &&
-          sym->derived_type == coff::derived_type_id::function) {
+      if (sym->has_section()) {
+        if (sym->derived_type == coff::derived_type_id::function) {
           auto scn = img->get_section(sym->section_index - 1);
-        auto dcmp_type =
-            scn->name.to_string(img->get_strings()) == INSTR_SPLIT_SECTION_NAME
+          auto dcmp_type = scn->name.to_string(img->get_strings()) ==
+                                   INSTR_SPLIT_SECTION_NAME
                                ? decomp::sym_type_t::inst_split
                                : decomp::sym_type_t::function;
           auto fn_size = scn->size_raw_data;
-        auto fn_bgn = scn->ptr_raw_data + reinterpret_cast<std::uint8_t*>(img);
+          auto fn_bgn =
+              scn->ptr_raw_data + reinterpret_cast<std::uint8_t*>(img);
 
           spdlog::info("decomposing function: {} size: {}",
                        sym->name.to_string(img->get_strings()), fn_size);
@@ -67,13 +68,13 @@ std::optional<comp::symbol_table_t*> decomp_t::decompose() {
           auto syms = rtn.decompose();
           spdlog::info("decomposed routine into {} symbols...", syms.size());
           m_syms->add_symbols(syms);
-      } else if (sym->has_section() &&
-                 sym->storage_class == coff::storage_class_id::public_symbol) {
+        } else if (sym->storage_class ==
+                   coff::storage_class_id::public_symbol) {
           auto scn = img->get_section(sym->section_index - 1);
           auto scn_sym = m_syms->sym_from_hash(m_scn_hash_tbl[scn]);
 
-        // if the section doesnt have a symbol then make one and put it into the
-        // symbol table...
+          // if the section doesnt have a symbol then make one and put it into
+          // the symbol table...
           //
           if (!scn_sym.has_value()) {
             auto scn_sym_name =
@@ -102,9 +103,9 @@ std::optional<comp::symbol_table_t*> decomp_t::decompose() {
 
           // create a symbol for the data...
           //
-        decomp::symbol_t new_sym(sym->name.to_string(img->get_strings()).data(),
-                                 sym->value, {}, scn, sym, {},
-                                 sym_type_t::data);
+          decomp::symbol_t new_sym(
+              sym->name.to_string(img->get_strings()).data(), sym->value, {},
+              scn, sym, {}, sym_type_t::data);
 
           spdlog::info("adding data symbol: {} located inside of section: {}",
                        new_sym.name(),
@@ -114,6 +115,24 @@ std::optional<comp::symbol_table_t*> decomp_t::decompose() {
 
           m_syms->add_symbol(new_sym);
         }
+      } else if (sym->storage_class ==
+                 coff::storage_class_id::
+                     external_definition) {  // else if the symbol has no
+                                             // section... these symbols require
+                                             // the linker to allocate space for
+                                             // them...
+
+        // the data is init to all zeros...
+        //
+        std::vector<std::uint8_t> data(sym->value, 0);
+
+        // the symbol is data, it has no section...
+        //
+        decomp::symbol_t bss_sym(sym->name.to_string(img->get_strings()).data(),
+                                 {}, data, {}, sym, {}, sym_type_t::data);
+
+        m_syms->add_symbol(bss_sym);
+      }
     }
   });