Add ragel machine generators to SCONS

src/trusted/validator_ragel/unreviewed/gen-dfa.cc

Index: src/trusted/validator_ragel/unreviewed/gen-dfa.cc
===================================================================
--- src/trusted/validator_ragel/unreviewed/gen-dfa.cc	(revision 8170)
+++ src/trusted/validator_ragel/unreviewed/gen-dfa.cc	(working copy)
@@ -12,12 +12,27 @@
 #include <string.h>
 
 #include <algorithm>
+#include <iterator>
 #include <map>
 #include <set>
 #include <string>
 #include <tuple>
 #include <vector>
 
+#ifdef __GNUC__
+/* We use operand number formats here.  They are defined in SUSv2 and supported
+   on all POSIX systems (including all versions of Linux and MacOS), yet
+   -pedantic warns about them.  */
+#pragma GCC diagnostic ignored "-Wformat"
+/* GCC complains even when we explicitly use empty initializer list.  Can be
+   easily fixed with data member initializers, but this requires GCC 4.7+.  */
+#pragma GCC diagnostic ignored "-Wmissing-field-initializers"
+/* Default is perfectly valid way to handle missing cases.  Especially if we
+   only care about very few non-default ones as often here.  */
+#pragma GCC diagnostic ignored "-Wswitch-enum"
+#pragma GCC diagnostic error "-Wswitch"
+#endif
+
 template <typename T, size_t N>
 char (&ArraySizeHelper(T (&array)[N]))[N];
 #define arraysize(array) (sizeof(ArraySizeHelper(array)))
@@ -189,7 +204,7 @@
 
   class Instruction {
    public:
-    static bool check_flag_valid(const std::string &flag) {
+    static void check_flag_valid(const std::string &flag) {
       if (all_instruction_flags.find(flag) == all_instruction_flags.end()) {
         fprintf(stderr, "%s: unknown flag: '%s'\n",
                 short_program_name, flag.c_str());
@@ -568,7 +583,7 @@
       fprintf(out_file, "  action rel8_operand {\n"
 "    operand0 = JMP_TO;\n"
 "    base = REG_RIP;\n"
-"    index = REG_NONE;\n"
+"    index = NO_REG;\n"
 "    scale = 0;\n"
 "    disp_type = DISP8;\n"
 "    disp = p;\n"
@@ -576,7 +591,7 @@
 "  action rel16_operand {\n"
 "    operand0 = JMP_TO;\n"
 "    base = REG_RIP;\n"
-"    index = REG_NONE;\n"
+"    index = NO_REG;\n"
 "    scale = 0;\n"
 "    disp_type = DISP16;\n"
 "    disp = p - 1;\n"
@@ -584,7 +599,7 @@
 "  action rel32_operand {\n"
 "    operand0 = JMP_TO;\n"
 "    base = REG_RIP;\n"
-"    index = REG_NONE;\n"
+"    index = NO_REG;\n"
 "    scale = 0;\n"
 "    disp_type = DISP32;\n"
 "    disp = p - 3;\n"
@@ -685,23 +700,23 @@
       if (ia32_mode) {
         fprintf(out_file, "  action modrm_only_base {\n"
 "    disp_type = DISPNONE;\n"
-"    index = REG_NONE;\n"
+"    index = NO_REG;\n"
 "    base = (*p) & 0x07;\n"
 "    scale = 0;\n"
 "  }\n"
 "  action modrm_base_disp {\n"
-"    index = REG_NONE;\n"
+"    index = NO_REG;\n"
 "    base = (*p) & 0x07;\n"
 "    scale = 0;\n"
 "  }\n"
 "  action modrm_pure_disp {\n"
-"    base = REG_NONE;\n"
-"    index = REG_NONE;\n"
+"    base = NO_REG;\n"
+"    index = NO_REG;\n"
 "    scale = 0;\n"
 "  }\n"
 "  action modrm_pure_index {\n"
 "    disp_type = DISPNONE;\n"
-"    base = REG_NONE;\n"
+"    base = NO_REG;\n"
 "    index = index_registers[((*p) & 0x38) >> 3];\n"
 "    scale = ((*p) & 0xc0) >> 6;\n"
 "  }\n"
@@ -714,27 +729,27 @@
       } else {
         fprintf(out_file, "  action modrm_only_base {\n"
 "    disp_type = DISPNONE;\n"
-"    index = REG_NONE;\n"
+"    index = NO_REG;\n"
 "    base = ((*p) & 0x07) |\n"
 "          ((rex_prefix & 0x01) << 3) |\n"
 "          (((~vex_prefix2) & 0x20) >> 2);\n"
 "    scale = 0;\n"
 "  }\n"
 "  action modrm_base_disp {\n"
-"    index = REG_NONE;\n"
+"    index = NO_REG;\n"
 "    base = ((*p) & 0x07) |\n"
 "          ((rex_prefix & 0x01) << 3) |\n"
 "          (((~vex_prefix2) & 0x20) >> 2);\n"
 "    scale = 0;\n"
 "  }\n"
 "  action modrm_rip {\n"
-"    index = REG_NONE;\n"
+"    index = NO_REG;\n"
 "    base = REG_RIP;\n"
 "    scale = 0;\n"
 "  }\n"
 "  action modrm_pure_index {\n"
 "    disp_type = DISPNONE;\n"
-"    base = REG_NONE;\n"
+"    base = NO_REG;\n"
 "    index = index_registers[(((*p) & 0x38) >> 3) |\n"
 "                           ((rex_prefix & 0x02) << 2) |\n"
 "                           (((~vex_prefix2) & 0x40) >> 3)];\n"
@@ -987,7 +1002,7 @@
       T { "XB",         0x80 },
       T { "RXB",        0x00 }
     };
-    for (int vex_it = 0; vex_it < arraysize(vex_fields); ++vex_it) {
+    for (size_t vex_it = 0; vex_it < arraysize(vex_fields); ++vex_it) {
       auto vex = vex_fields[vex_it];
       fprintf(out_file, "  VEX_%2$s = %3$s%1$s;\n"
 "", enabled(Actions::kVexPrefix) && !ia32_mode ? " @vex_prefix2" : "",
@@ -1007,7 +1022,7 @@
       T { "01001",      9       },
       T { "01010",      10      },
     };
-    for (int vex_it = 0; vex_it < arraysize(vex_map); ++vex_it) {
+    for (size_t vex_it = 0; vex_it < arraysize(vex_map); ++vex_it) {
       auto vex = vex_map[vex_it];
       fprintf(out_file, "  VEX_map%1$s = %2$s;\n"
 "", vex.first, chartest((c & 0x1f) == vex.second));
@@ -1062,7 +1077,7 @@
           T { "dreg",                   "DebugRegister"                 },
           T { "selector",               "Selector"                      }
         };
-        for (int size_it = 0; size_it < arraysize(sizes); ++size_it) {
+        for (size_t size_it = 0; size_it < arraysize(sizes); ++size_it) {
           auto size = sizes[size_it];
           fprintf(out_file, "  action operand%1$d_%2$s {\n"
 "    operand%1$d_type = Operand%3$s;\n"
@@ -1072,8 +1087,8 @@
         if (ia32_mode) {
           fprintf(out_file, "  action operand%1$d_absolute_disp {\n"
 "    operand%1$d = REG_RM;\n"
-"    base = REG_NONE;\n"
-"    index = REG_NONE;\n"
+"    base = NO_REG;\n"
+"    index = NO_REG;\n"
 "    scale = 0;\n"
 "  }\n"
 "  action operand%1$d_from_opcode {\n"
@@ -1098,7 +1113,7 @@
         } else {
           fprintf(out_file, "  action operand%1$d_absolute_disp {\n"
 "    operand%1$d = REG_RM;\n"
-"    base = REG_NONE;\n"
+"    base = NO_REG;\n"
 "    index = REG_RIZ;\n"
 "    scale = 0;\n"
 "  }\n"
@@ -1141,7 +1156,7 @@
           T { "rm",                     "REG_RM"        },
           T { "st",                     "REG_ST"        }
         };
-        for (int type_it = 0; type_it < arraysize(types); ++type_it) {
+        for (size_t type_it = 0; type_it < arraysize(types); ++type_it) {
           auto type = types[type_it];
           fprintf(out_file, "  action operand%1$d_%2$s {\n"
 "    operand%1$d = %3$s;\n"
@@ -1153,20 +1168,20 @@
     if (enabled(Actions::kParseOperandsStates)) {
       for (auto i = 0 ; i < 5; ++i) {
         fprintf(out_file, "  action operand%1$d_unused {\n"
-"    operand%1$d_read = false;\n"
-"    operand%1$d_write = false;\n"
+"    operand%1$d_read = FALSE;\n"
+"    operand%1$d_write = FALSE;\n"
 "  }\n"
 "  action operand%1$d_read {\n"
-"    operand%1$d_read = true;\n"
-"    operand%1$d_write = false;\n"
+"    operand%1$d_read = TRUE;\n"
+"    operand%1$d_write = FALSE;\n"
 "  }\n"
 "  action operand%1$d_write {\n"
-"    operand%1$d_read = false;\n"
-"    operand%1$d_write = true;\n"
+"    operand%1$d_read = FALSE;\n"
+"    operand%1$d_write = TRUE;\n"
 "  }\n"
 "  action operand%1$d_readwrite {\n"
-"    operand%1$d_read = true;\n"
-"    operand%1$d_write = true;\n"
+"    operand%1$d_read = TRUE;\n"
+"    operand%1$d_write = TRUE;\n"
 "  }\n"
 "", i);
       }
@@ -1189,7 +1204,7 @@
     explicit MarkedInstruction(Instruction instruction_) :
         Instruction(instruction_),
         instruction_class(get_instruction_class(instruction_)),
-        opcode_in_modrm(false), opcode_in_imm(false), rex { } {
+        rex { }, opcode_in_modrm(false), opcode_in_imm(false) {
       if (has_flag("branch_hint")) {
         optional_prefixes.insert("branch_hint");
       }
@@ -1237,7 +1252,7 @@
                   (*opcode) = "(";
                   opcode->append(saved_opcode);
                   static const char cc[] = {'9', 'a', 'b', 'c', 'd', 'e', 'f'};
-                  for (int c_it = 0; c_it < arraysize(cc); ++c_it) {
+                  for (size_t c_it = 0; c_it < arraysize(cc); ++c_it) {
                     auto c = cc[c_it];
                     opcode->push_back('|');
                     (*(saved_opcode.rbegin())) = c;
@@ -1509,7 +1524,7 @@
                    operand_it != operands.end(); ++operand_it) {
                 auto &operand = *operand_it;
                 static const char cc[] = {'H', 'L', 'U', 'V', 'W'};
-                for (int c_it = 0; c_it < arraysize(cc); ++c_it) {
+                for (size_t c_it = 0; c_it < arraysize(cc); ++c_it) {
                   auto c = cc[c_it];
                   if ((operand.source == c) &&
                       (*(operand.size.rbegin()) == 'x') &&
@@ -1548,7 +1563,7 @@
       }
       bool modrm_memory = false;
       bool modrm_register = false;
-      char operand_source;
+      char operand_source = ' ';
       for (auto operand_it = operands.begin();
            operand_it != operands.end(); ++operand_it) {
         auto &operand = *operand_it;
@@ -1672,7 +1687,7 @@
         T { " operand_sib_pure_index",  true,   false   },
         T { " operand_sib_base_index",  true,   true    }
       };
-      for (int mode_it = 0; mode_it < arraysize(modes); ++mode_it) {
+      for (size_t mode_it = 0; mode_it < arraysize(modes); ++mode_it) {
         auto mode = modes[mode_it];
         print_operator_delimeter();
         if (mod_reg_is_used()) {
@@ -1752,7 +1767,7 @@
           return true;
         }
         static const char cc[] = { 'G', 'P', 'V' };
-        for (int c_it = 0; c_it < arraysize(cc); ++c_it) {
+        for (size_t c_it = 0; c_it < arraysize(cc); ++c_it) {
           auto c = cc[c_it];
           if (operand.source == c) {
             return true;
@@ -1767,7 +1782,7 @@
            operand_it != operands.end(); ++operand_it) {
         auto &operand = *operand_it;
         static const char cc[] = { 'E', 'M', 'N', 'Q', 'R', 'U', 'W' };
-        for (int c_it = 0; c_it < arraysize(cc); ++c_it) {
+        for (size_t c_it = 0; c_it < arraysize(cc); ++c_it) {
           auto c = cc[c_it];
           if (operand.source == c) {
             return true;
@@ -1918,12 +1933,13 @@
         fprintf(out_file, " & VEX_map%s) ", opcodes[1].c_str() + 4);
         auto third_byte = opcodes[2];
         static const char* symbolic_names[] = { "cntl", "dest", "src1", "src" };
-        for (int symbolic_it = 0;
+        for (size_t symbolic_it = 0;
              symbolic_it < arraysize(symbolic_names); ++symbolic_it) {
           auto symbolic = symbolic_names[symbolic_it];
           for (auto it = third_byte.begin(); it != third_byte.end(); ++it) {
-            if ((third_byte.end() - it) >= strlen(symbolic) &&
-                !strncmp(&*it, symbolic, strlen(symbolic))) {
+            if (static_cast<size_t>(third_byte.end() - it) >=
+                                                          strlen(symbolic) &&
+                                   !strncmp(&*it, symbolic, strlen(symbolic))) {

[pasko-google - do not use, 2012/04/05 14:16:38]

I do not like how this is formatted. Should be shifted by 4, mot aligned to the
right border.

btw, this can look less ugly if you added in front of the loop:
size_t symbolic_len = strlen(symbolic);

[khim, 2012/04/05 16:38:43]

Whatever. I don't really care about this file (I have clean version anyway).

Done.

               third_byte.replace(it, it + strlen(symbolic), "XXXX");
               break;
             }
@@ -1945,7 +1961,8 @@
         auto third_byte_ok = (arraysize(third_byte_check) ==
                                                            third_byte.length());

[pasko-google - do not use, 2012/04/05 14:16:38]

spacing, better be:
bool third_byte_ok =
    (arraysize(third_byte_check) == third_byte.length());

[khim, 2012/04/05 16:38:43]

Done.

         if (third_byte_ok) {
-          for (int set_it = 0; set_it < arraysize(third_byte_check); ++set_it) {
+          for (size_t set_it = 0; set_it < arraysize(third_byte_check);
+                                                                     ++set_it) {

[pasko-google - do not use, 2012/04/05 14:16:38]

shiftwidth = 4

[khim, 2012/04/05 16:38:43]

Done.

             auto &set = third_byte_check[set_it];
             if (set.find(third_byte[&set - third_byte_check]) == set.end()) {
               third_byte_ok = false;