Refactoring: unify upb_msg.

The cost is that a upb_msg will now always have an overhead
of 2*sizeof(void*).  This is comparable to proto2 overhead.
The benefit is that upb_msg is now self-describing, and
read-only algorithms can now operate on a upb_msg regardless
of the memory-management scheme.

Also, upb_array and upb_string now know inherently if they
own their associated memory, and upb_array has a generic
pointer for memory management purposes like upb_msg does.

1 files changed, 37 insertions, 38 deletions

diff --git a/tools/upbc.c b/tools/upbc.c
index f593ae9..f9890c6 100644
--- a/tools/upbc.c
+++ b/tools/upbc.c
@@ -5,8 +5,6 @@
  * easier to understand, but by its nature it is doing some very "meta"
  * kinds of things.
  *
- * TODO: compiler currently has memory leaks (trivial to fix with valgrind).
- *
  * Copyright (c) 2009 Joshua Haberman.  See LICENSE for details.
  */
 
@@ -22,19 +20,19 @@
  * the string (and thus never need to re-allocate). */
 
 /* Convert to C identifier: foo.bar.Baz -> foo_bar_Baz. */
-static void to_cident(struct upb_string str)
+static void to_cident(struct upb_string *str)
 {
-  for(uint32_t i = 0; i < str.byte_len; i++)
-    if(str.ptr[i] == '.' || str.ptr[i] == '/')
-      str.ptr[i] = '_';
+  for(uint32_t i = 0; i < str->byte_len; i++)
+    if(str->ptr[i] == '.' || str->ptr[i] == '/')
+      str->ptr[i] = '_';
 }
 
 /* Convert to C proprocessor identifier: foo.bar.Baz -> FOO_BAR_BAZ. */
-static void to_preproc(struct upb_string str)
+static void to_preproc(struct upb_string *str)
 {
   to_cident(str);
-  for(uint32_t i = 0; i < str.byte_len; i++)
-    str.ptr[i] = toupper(str.ptr[i]);
+  for(uint32_t i = 0; i < str->byte_len; i++)
+    str->ptr[i] = toupper(str->ptr[i]);
 }
 
 static int memrchr(char *data, char c, size_t len)
@@ -64,7 +62,7 @@ static void write_h(struct upb_symtab_entry *entries[], int num_entries,
                     char *outfile_name, char *descriptor_cident, FILE *stream)
 {
   /* Header file prologue. */
-  struct upb_string include_guard_name = upb_strdupc(outfile_name);
+  struct upb_string *include_guard_name = upb_strdupc(outfile_name);
   to_preproc(include_guard_name);
   fputs("/* This file was generated by upbc (the upb compiler).  "
         "Do not edit. */\n\n", stream),
@@ -90,21 +88,21 @@ static void write_h(struct upb_symtab_entry *entries[], int num_entries,
     struct upb_enum *e = entry->ref._enum;
     google_protobuf_EnumDescriptorProto *ed = e->descriptor;
     /* We use entry->e.key (the fully qualified name) instead of ed->name. */
-    struct upb_string enum_name = upb_strdup(entry->e.key);
+    struct upb_string *enum_name = upb_strdup(&entry->e.key);
     to_cident(enum_name);
 
-    struct upb_string enum_val_prefix = upb_strdup(entry->e.key);
-    enum_val_prefix.byte_len = memrchr(enum_val_prefix.ptr,
+    struct upb_string *enum_val_prefix = upb_strdup(&entry->e.key);
+    enum_val_prefix->byte_len = memrchr(enum_val_prefix->ptr,
                                        UPB_SYMBOL_SEPARATOR,
-                                       enum_val_prefix.byte_len);
-    enum_val_prefix.byte_len++;
+                                       enum_val_prefix->byte_len);
+    enum_val_prefix->byte_len++;
     to_preproc(enum_val_prefix);
 
     fprintf(stream, "typedef enum " UPB_STRFMT " {\n", UPB_STRARG(enum_name));
     if(ed->set_flags.has.value) {
       for(uint32_t j = 0; j < ed->value->len; j++) {  /* Foreach enum value. */
         google_protobuf_EnumValueDescriptorProto *v = ed->value->elements[j];
-        struct upb_string value_name = upb_strdup(*v->name);
+        struct upb_string *value_name = upb_strdup(v->name);
         to_preproc(value_name);
         /* "  GOOGLE_PROTOBUF_FIELDDESCRIPTORPROTO_TYPE_UINT32 = 13," */
         fprintf(stream, "  " UPB_STRFMT UPB_STRFMT " = %" PRIu32,
@@ -128,7 +126,7 @@ static void write_h(struct upb_symtab_entry *entries[], int num_entries,
     if(entries[i]->type != UPB_SYM_MESSAGE) continue;
     struct upb_symtab_entry *entry = entries[i];
     /* We use entry->e.key (the fully qualified name). */
-    struct upb_string msg_name = upb_strdup(entry->e.key);
+    struct upb_string *msg_name = upb_strdup(&entry->e.key);
     to_cident(msg_name);
     fprintf(stream, "struct " UPB_STRFMT ";\n", UPB_STRARG(msg_name));
     fprintf(stream, "typedef struct " UPB_STRFMT "\n    " UPB_STRFMT ";\n\n",
@@ -143,9 +141,11 @@ static void write_h(struct upb_symtab_entry *entries[], int num_entries,
     struct upb_symtab_entry *entry = entries[i];
     struct upb_msgdef *m = entry->ref.msg;
     /* We use entry->e.key (the fully qualified name). */
-    struct upb_string msg_name = upb_strdup(entry->e.key);
+    struct upb_string *msg_name = upb_strdup(&entry->e.key);
     to_cident(msg_name);
     fprintf(stream, "struct " UPB_STRFMT " {\n", UPB_STRARG(msg_name));
+    fputs("  struct upb_msgdef *def;\n", stream);
+    fputs("  void *gptr;\n", stream);
     fputs("  union {\n", stream);
     fprintf(stream, "    uint8_t bytes[%" PRIu32 "];\n", m->set_flags_bytes);
     fputs("    struct {\n", stream);
@@ -153,7 +153,7 @@ static void write_h(struct upb_symtab_entry *entries[], int num_entries,
       static char* labels[] = {"", "optional", "required", "repeated"};
       struct google_protobuf_FieldDescriptorProto *fd = m->field_descriptors[j];
       fprintf(stream, "      bool " UPB_STRFMT ":1;  /* = %" PRIu32 ", %s. */\n",
-              UPB_STRARG(*fd->name), fd->number, labels[fd->label]);
+              UPB_STRARG(fd->name), fd->number, labels[fd->label]);
     }
     fputs("    } has;\n", stream);
     fputs("  } set_flags;\n", stream);
@@ -170,14 +170,14 @@ static void write_h(struct upb_symtab_entry *entries[], int num_entries,
           type_name_ref.ptr++;
           type_name_ref.byte_len--;
         }
-        struct upb_string type_name = upb_strdup(type_name_ref);
+        struct upb_string *type_name = upb_strdup(&type_name_ref);
         to_cident(type_name);
         if(f->label == GOOGLE_PROTOBUF_FIELDDESCRIPTORPROTO_LABEL_REPEATED) {
           fprintf(stream, "  UPB_MSG_ARRAY(" UPB_STRFMT ")* " UPB_STRFMT ";\n",
-                  UPB_STRARG(type_name), UPB_STRARG(*fd->name));
+                  UPB_STRARG(type_name), UPB_STRARG(fd->name));
         } else {
           fprintf(stream, "  " UPB_STRFMT "* " UPB_STRFMT ";\n",
-                  UPB_STRARG(type_name), UPB_STRARG(*fd->name));
+                  UPB_STRARG(type_name), UPB_STRARG(fd->name));
         }
         upb_strfree(type_name);
       } else if(f->label == GOOGLE_PROTOBUF_FIELDDESCRIPTORPROTO_LABEL_REPEATED) {
@@ -193,7 +193,7 @@ static void write_h(struct upb_symtab_entry *entries[], int num_entries,
           "struct upb_int64_array*"
         };
         fprintf(stream, "  %s " UPB_STRFMT ";\n",
-                c_types[fd->type], UPB_STRARG(*fd->name));
+                c_types[fd->type], UPB_STRARG(fd->name));
       } else {
         static char* c_types[] = {
           "", "double", "float", "int64_t", "uint64_t", "int32_t", "uint64_t",
@@ -202,7 +202,7 @@ static void write_h(struct upb_symtab_entry *entries[], int num_entries,
           "int32_t", "int64_t"
         };
         fprintf(stream, "  %s " UPB_STRFMT ";\n",
-                c_types[fd->type], UPB_STRARG(*fd->name));
+                c_types[fd->type], UPB_STRARG(fd->name));
       }
     }
     fputs("};\n", stream);
@@ -230,7 +230,7 @@ struct strtable_entry {
 struct typetable_entry {
   struct upb_strtable_entry e;
   struct upb_msg_fielddef *field;
-  struct upb_string cident;  /* Type name converted with to_cident(). */
+  struct upb_string *cident;  /* Type name converted with to_cident(). */
   /* A list of all values of this type, in an established order. */
   union upb_value *values;
   int values_size, values_len;
@@ -251,7 +251,7 @@ struct msgtable_entry {
 int compare_entries(const void *_e1, const void *_e2)
 {
   struct strtable_entry *const*e1 = _e1, *const*e2 = _e2;
-  return upb_strcmp((*e1)->e.key, (*e2)->e.key);
+  return upb_strcmp(&(*e1)->e.key, &(*e2)->e.key);
 }
 
 /* Mutually recursive functions to recurse over a set of possibly nested
@@ -297,7 +297,6 @@ static void add_strings_from_msg(void *data, struct upb_msgdef *m,
  *
  * TODO: make these use a generic msg visitor. */
 
-
 struct typetable_entry *get_or_insert_typeentry(struct upb_strtable *t,
                                                 struct upb_msg_fielddef *f)
 {
@@ -306,7 +305,7 @@ struct typetable_entry *get_or_insert_typeentry(struct upb_strtable *t,
   struct typetable_entry *type_e = upb_strtable_lookup(t, &type_name);
   if(type_e == NULL) {
     struct typetable_entry new_type_e = {
-      .e = {.key = type_name}, .field = f, .cident = upb_strdup(type_name),
+      .e = {.key = type_name}, .field = f, .cident = upb_strdup(&type_name),
       .values = NULL, .values_size = 0, .values_len = 0,
       .arrays = NULL, .arrays_size = 0, .arrays_len = 0
     };
@@ -323,7 +322,7 @@ static void add_value(union upb_value value, struct upb_msg_fielddef *f,
 {
   struct typetable_entry *type_e = get_or_insert_typeentry(t, f);
   if(type_e->values_len == type_e->values_size) {
-    type_e->values_size = max(type_e->values_size * 2, 4);
+    type_e->values_size = UPB_MAX(type_e->values_size * 2, 4);
     type_e->values = realloc(type_e->values, sizeof(*type_e->values) * type_e->values_size);
   }
   type_e->values[type_e->values_len++] = value;
@@ -343,7 +342,7 @@ static void add_submsgs(void *data, struct upb_msgdef *m, struct upb_strtable *t
       struct typetable_entry *arr_type_e =
           get_or_insert_typeentry(t, f);
       if(arr_type_e->arrays_len == arr_type_e->arrays_size) {
-        arr_type_e->arrays_size = max(arr_type_e->arrays_size * 2, 4);
+        arr_type_e->arrays_size = UPB_MAX(arr_type_e->arrays_size * 2, 4);
         arr_type_e->arrays = realloc(arr_type_e->arrays,
                                      sizeof(*arr_type_e->arrays)*arr_type_e->arrays_size);
       }
@@ -492,7 +491,7 @@ static void write_message_c(void *data, struct upb_msgdef *m,
         for(unsigned int j = 0; j < m->num_fields; j++) {
           struct upb_msg_fielddef *f = &m->fields[j];
           google_protobuf_FieldDescriptorProto *fd = m->field_descriptors[j];
-          fprintf(stream, "    ." UPB_STRFMT " = ", UPB_STRARG(*fd->name));
+          fprintf(stream, "    ." UPB_STRFMT " = ", UPB_STRARG(fd->name));
           if(upb_msg_isset(msgdata, f))
             fprintf(stream, "true");
           else
@@ -505,7 +504,7 @@ static void write_message_c(void *data, struct upb_msgdef *m,
           struct upb_msg_fielddef *f = &m->fields[j];
           google_protobuf_FieldDescriptorProto *fd = m->field_descriptors[j];
           union upb_value val = upb_msg_get(msgdata, f);
-          fprintf(stream, "    ." UPB_STRFMT " = ", UPB_STRARG(*fd->name));
+          fprintf(stream, "    ." UPB_STRFMT " = ", UPB_STRARG(fd->name));
           if(!upb_msg_isset(msgdata, f)) {
             fputs("0,   /* Not set. */", stream);
           } else if(upb_isstring(f)) {
@@ -516,7 +515,7 @@ static void write_message_c(void *data, struct upb_msgdef *m,
               struct strtable_entry *str_e = upb_strtable_lookup(&strings, val.str);
               assert(str_e);
               fprintf(stream, "&strings[%d],   /* \"" UPB_STRFMT "\" */",
-                      str_e->num, UPB_STRARG(*val.str));
+                      str_e->num, UPB_STRARG(val.str));
             }
           } else if(upb_isarray(f)) {
             /* Find this submessage in the list of msgs for that type. */
@@ -642,15 +641,15 @@ int main(int argc, char *argv[])
   if(!outfile_base) outfile_base = input_file;
 
   /* Read input file. */
-  struct upb_string descriptor;
-  if(!upb_strreadfile(input_file, &descriptor))
+  struct upb_string *descriptor = upb_strnew();
+  if(!upb_strreadfile(input_file, descriptor))
     error("Couldn't read input file.");
 
   /* Parse input file. */
   struct upb_context c;
   upb_context_init(&c);
-  google_protobuf_FileDescriptorSet *fds =
-      upb_msg_parsenew(c.fds_msg, &descriptor);
+  struct upb_msg *fds_msg = upb_msg_parsenew(c.fds_msg, descriptor);
+  google_protobuf_FileDescriptorSet *fds = (void*)fds_msg;
   if(!fds)
     error("Failed to parse input file descriptor.");
   if(!upb_context_addfds(&c, fds))
@@ -686,7 +685,7 @@ int main(int argc, char *argv[])
     write_message_c(fds, c.fds_msg, cident, h_filename, argc, argv, input_file, c_file);
     fclose(c_file);
   }
-  upb_msg_free(fds, c.fds_msg);
+  upb_msg_free(fds_msg);
   upb_context_free(&c);
   upb_strfree(descriptor);
   fclose(h_file);