mrb_str_new_static(): zero copy string creation

5 files changed, 49 insertions, 13 deletions

diff --git a/include/mruby.h b/include/mruby.h
index 8c811c58f..28eea6929 100644
--- a/include/mruby.h
+++ b/include/mruby.h
@@ -231,6 +231,7 @@ void *mrb_free(mrb_state*, void*);
 
 mrb_value mrb_str_new(mrb_state *mrb, const char *p, size_t len);
 mrb_value mrb_str_new_cstr(mrb_state*, const char*);
+mrb_value mrb_str_new_static(mrb_state *mrb, const char *p, size_t len);
 
 mrb_state* mrb_open(void);
 mrb_state* mrb_open_allocf(mrb_allocf, void *ud);
diff --git a/include/mruby/string.h b/include/mruby/string.h
index c62d79563..a1b38d039 100644
--- a/include/mruby/string.h
+++ b/include/mruby/string.h
@@ -37,7 +37,8 @@ struct RString {
 #define RSTRING_LEN(s)    (RSTRING(s)->len)
 #define RSTRING_CAPA(s)   (RSTRING(s)->aux.capa)
 #define RSTRING_END(s)    (RSTRING(s)->ptr + RSTRING(s)->len)
-#define MRB_STR_SHARED      256
+#define MRB_STR_SHARED    1
+#define MRB_STR_STATIC    (1<<1)
 
 void mrb_str_decref(mrb_state*, mrb_shared_string*);
 void mrb_str_modify(mrb_state*, struct RString*);
diff --git a/src/gc.c b/src/gc.c
index 0e9fd698c..fa6c6e424 100644
--- a/src/gc.c
+++ b/src/gc.c
@@ -531,7 +531,7 @@ obj_free(mrb_state *mrb, struct RBasic *obj)
   case MRB_TT_STRING:
     if (obj->flags & MRB_STR_SHARED)
       mrb_str_decref(mrb, ((struct RString*)obj)->aux.shared);
-    else
+    else if ((obj->flags & MRB_STR_STATIC) == 0)
       mrb_free(mrb, ((struct RString*)obj)->ptr);
     break;
 
diff --git a/src/string.c b/src/string.c
index 371d3faf6..34769222a 100644
--- a/src/string.c
+++ b/src/string.c
@@ -34,6 +34,7 @@ static mrb_value mrb_str_subseq(mrb_state *mrb, mrb_value str, mrb_int beg, mrb_
 void
 mrb_str_decref(mrb_state *mrb, mrb_shared_string *shared)
 {
+  if (shared->refcnt < 0) return;
   shared->refcnt--;
   if (shared->refcnt == 0) {
     mrb_free(mrb, shared->ptr);
@@ -44,6 +45,18 @@ mrb_str_decref(mrb_state *mrb, mrb_shared_string *shared)
 void
 mrb_str_modify(mrb_state *mrb, struct RString *s)
 {
+  if (s->flags & MRB_STR_STATIC) {
+    char *p = s->ptr;
+
+    s->ptr = (char *)mrb_malloc(mrb, (size_t)s->len+1);
+    if (p) {
+      memcpy(s->ptr, p, s->len);
+    }
+    s->ptr[s->len] = '\0';
+    s->flags &= ~MRB_STR_STATIC;
+    return;
+  }
+
   if (s->flags & MRB_STR_SHARED) {
     mrb_shared_string *shared = s->aux.shared;
 
@@ -68,6 +81,7 @@ mrb_str_modify(mrb_state *mrb, struct RString *s)
       mrb_str_decref(mrb, shared);
     }
     s->flags &= ~MRB_STR_SHARED;
+    return;
   }
 }
 
@@ -249,6 +263,19 @@ mrb_str_new_cstr(mrb_state *mrb, const char *p)
   return mrb_obj_value(s);
 }
 
+mrb_value
+mrb_str_new_static(mrb_state *mrb, const char *p, size_t len)
+{
+  struct RString *s;
+
+  s = mrb_obj_alloc_string(mrb);
+  s->len = len;
+  s->aux.capa = len;
+  s->ptr = (char *)p;
+  s->flags |= MRB_STR_STATIC;
+  return mrb_obj_value(s);
+}
+
 char *
 mrb_str_to_cstr(mrb_state *mrb, mrb_value str0)
 {
@@ -271,12 +298,19 @@ str_make_shared(mrb_state *mrb, struct RString *s)
   if (!(s->flags & MRB_STR_SHARED)) {
     mrb_shared_string *shared = (mrb_shared_string *)mrb_malloc(mrb, sizeof(mrb_shared_string));
 
-    shared->refcnt = 1;
-    if (s->aux.capa > s->len) {
-      s->ptr = shared->ptr = (char *)mrb_realloc(mrb, s->ptr, s->len+1);
+    if (s->flags & MRB_STR_STATIC) {
+      shared->refcnt = -1;      /* should never be freed */
+      shared->ptr = s->ptr;
+      s->flags &= ~MRB_STR_STATIC;
     }
     else {
-      shared->ptr = s->ptr;
+      shared->refcnt = 1;
+      if (s->aux.capa > s->len) {
+        s->ptr = shared->ptr = (char *)mrb_realloc(mrb, s->ptr, s->len+1);
+      }
+      else {
+        shared->ptr = s->ptr;
+      }
     }
     shared->len = s->len;
     s->aux.shared = shared;
diff --git a/src/symbol.c b/src/symbol.c
index 24db3a4b6..e913f95c8 100644
--- a/src/symbol.c
+++ b/src/symbol.c
@@ -213,7 +213,7 @@ mrb_sym_to_s(mrb_state *mrb, mrb_value sym)
   size_t len;
 
   p = mrb_sym2name_len(mrb, id, &len);
-  return mrb_str_new(mrb, p, len);
+  return mrb_str_new_static(mrb, p, len);
 }
 
 /* 15.2.11.3.4  */
@@ -381,14 +381,14 @@ mrb_sym2str(mrb_state *mrb, mrb_sym sym)
 {
   size_t len;
   const char *name = mrb_sym2name_len(mrb, sym, &len);
-
+  mrb_value str;
+  
   if (!name) return mrb_undef_value(); /* can't happen */
+  str = mrb_str_new_static(mrb, name, len);
   if (symname_p(name) && strlen(name) == len) {
-    return mrb_str_new(mrb, name, len);
-  }
-  else {
-    return mrb_str_dump(mrb, mrb_str_new(mrb, name, len));
+    return str;
   }
+  return mrb_str_dump(mrb, str);
 }
 
 const char*
@@ -402,7 +402,7 @@ mrb_sym2name(mrb_state *mrb, mrb_sym sym)
     return name;
   }
   else {
-    mrb_value str = mrb_str_dump(mrb, mrb_str_new(mrb, name, len));
+    mrb_value str = mrb_str_dump(mrb, mrb_str_new_static(mrb, name, len));
     return RSTRING(str)->ptr;
   }
 }