utf8 range limiting.

compiler work for class variable access in the class defintion scope

lib/comp.c

@@ -760,7 +760,7 @@ static int emit_variable_access (hcl_t* hcl, int mode, const hcl_var_info_t* vi,
 	{
 		{ HCL_CODE_PUSH_CTXTEMPVAR_0, HCL_CODE_POP_INTO_CTXTEMPVAR_0, HCL_CODE_STORE_INTO_CTXTEMPVAR_0 },
 		{ HCL_CODE_PUSH_INSTVAR_0,    HCL_CODE_POP_INTO_INSTVAR_0,    HCL_CODE_STORE_INTO_INSTVAR_0    },
-		{ HCL_CODE_PUSH_OBJVAR_0,     HCL_CODE_POP_INTO_OBJVAR_0,     HCL_CODE_STORE_INTO_OBJVAR_0     },
+		{ HCL_CODE_PUSH_CLSVAR_X,     HCL_CODE_POP_INTO_CLSVAR_X,     HCL_CODE_STORE_INTO_CLSVAR_X     },
 	};
 
 	switch (vi->type)
@@ -774,7 +774,7 @@ static int emit_variable_access (hcl_t* hcl, int mode, const hcl_var_info_t* vi,
 
 		case VAR_CLASS:
 			HCL_ASSERT (hcl, vi->ctx_offset == 0);
-			//return emit_double_param_instruction(hcl, inst_map[2][mode], vi->index_in_ctx, <<index to the class name in literal table>>, srcloc);
+			return emit_single_param_instruction(hcl, inst_map[2][mode], vi->index_in_ctx, srcloc);
 	}
 
 	return -1;
@@ -2091,25 +2091,28 @@ static HCL_INLINE int compile_class_p1 (hcl_t* hcl)
 		{
 			/* instance variables */
 			hcl_oow_t checkpoint;
+
 			if (!HCL_CNODE_IS_CONS_CONCODED(tmp, HCL_CONCODE_VLIST)) break;
 			checkpoint = hcl->c->tv.s.len;
+
 			if (collect_vardcl(hcl, obj, &obj, tv_dup_check_start, &dclcount, "instance") <= -1) return -1;
 			nivars += dclcount;
 
-			if (ivar_len <= 0) ivar_start = checkpoint;
+			if (ivar_len <= 0) ivar_start = (cvar_len <= 0)? checkpoint: cvar_start;
 			ivar_len += hcl->c->tv.s.len - checkpoint;
 
 			if (cvar_len > 0)
 			{
+				/* place the instance variables before the class variables 
+				 * if class variables "a b" has been collected before instance variables "cc dd ee"
+				 * the rotation below manipulates the buffer to contain "cc dd ee a b".
+				 */ 
 				hcl_rotate_oochars (&hcl->c->tv.s.ptr[cvar_start], hcl->c->tv.s.len - cvar_start, -1, cvar_len); 
-				ivar_start = cvar_start;
 				cvar_start += hcl->c->tv.s.len - checkpoint;
 			}
 		}
 	}
 
-//HCL_DEBUG4 (hcl, "AAAAAAAAAAAAAAAAAAAAAAA>>>> [%.*js]  [%.*js]\n", ivar_len, &hcl->c->tv.s.ptr[ivar_start], cvar_len, &hcl->c->tv.s.ptr[cvar_start]);
-
 	if (nivars > 0)
 	{
 		hcl_oop_t tmp;
@@ -3710,7 +3713,7 @@ redo:
 			switch (HCL_CNODE_ELIST_CONCODE(oprnd))
 			{
 				case HCL_CONCODE_XLIST:
-					hcl_setsynerrbfmt (hcl, HCL_SYNERR_VARDCLBANNED, HCL_CNODE_GET_LOC(oprnd), HCL_NULL, "empty executable list");
+					hcl_setsynerrbfmt (hcl, HCL_SYNERR_BANNED, HCL_CNODE_GET_LOC(oprnd), HCL_NULL, "empty executable list");
 					return -1;
 
 				case HCL_CONCODE_ARRAY:
@@ -3730,7 +3733,7 @@ redo:
 					goto done;
 
 				case HCL_CONCODE_VLIST:
-					hcl_setsynerrbfmt (hcl, HCL_SYNERR_VARDCLBANNED, HCL_CNODE_GET_LOC(oprnd), HCL_NULL, "variable declaration disallowed");
+					hcl_setsynerrbfmt (hcl, HCL_SYNERR_BANNED, HCL_CNODE_GET_LOC(oprnd), HCL_NULL, "empty variable declaration");
 					return -1;
 
 				default:

lib/err.c

@@ -77,6 +77,9 @@ static hcl_ooch_t errstr_35[] = {'a','r','g','u','m','e','n','t',' ','n','u','m'
 static hcl_ooch_t errstr_36[] = {'r','e','t','u','r','n',' ','c','o','u','n','t',' ','e','r','r','o','r','\0'};
 static hcl_ooch_t errstr_37[] = {'t','o','o',' ','m','a','n','y',' ','s','e','m','a','p','h','o','r','e','s','\0'};
 static hcl_ooch_t errstr_38[] = {'e','x','c','e','p','a','i','o','n',' ','n','o','t',' ','h','a','n','d','l','e','d','\0'};
+static hcl_ooch_t errstr_39[] = {'s','t','a','c','k',' ','u','n','d','e','r','f','l','o','w','\0'};
+
+static hcl_ooch_t errstr_40[] = {'s','t','a','c','k',' ','o','v','e','r','f','l','o','w','\0'};
 
 static hcl_ooch_t* errstr[] =
 {
@@ -84,7 +87,8 @@ static hcl_ooch_t* errstr[] =
 	errstr_8, errstr_9, errstr_10, errstr_11, errstr_12, errstr_13, errstr_14, errstr_15,
 	errstr_16, errstr_17, errstr_18, errstr_19, errstr_20, errstr_21, errstr_22, errstr_23,
 	errstr_24, errstr_25, errstr_26, errstr_27, errstr_28, errstr_29, errstr_30, errstr_31,
-	errstr_32, errstr_33, errstr_34, errstr_35, errstr_36, errstr_37, errstr_38
+	errstr_32, errstr_33, errstr_34, errstr_35, errstr_36, errstr_37, errstr_38, errstr_39,
+	errstr_40
 };
 
 

lib/exec.c

@@ -3442,7 +3442,12 @@ if (do_throw(hcl, hcl->_nil, fetched_instruction_pointer) <= -1)
 			case HCL_CODE_CLASS_EXIT:
 			{
 				LOG_INST_0 (hcl, "class_exit");
-				/* TODO: stack underflow check? */
+				if (HCL_CLSTACK_IS_EMPTY(hcl))
+				{
+					hcl_seterrbfmt (hcl, HCL_ESTKUNDFLW, "class stack underflow");
+					supplement_errmsg (hcl, fetched_instruction_pointer);
+					goto oops;
+				}
 				HCL_CLSTACK_POP (hcl);
 				break;
 			}
@@ -3452,7 +3457,13 @@ if (do_throw(hcl, hcl->_nil, fetched_instruction_pointer) <= -1)
 				hcl_oop_t c;
 
 				LOG_INST_0 (hcl, "class_push_exit");
-				/* TODO: stack underflow check? */
+
+				if (HCL_CLSTACK_IS_EMPTY(hcl))
+				{
+					hcl_seterrbfmt (hcl, HCL_ESTKUNDFLW, "class stack underflow");
+					supplement_errmsg (hcl, fetched_instruction_pointer);
+					goto oops;
+				}
 				HCL_CLSTACK_POP_TO (hcl, c);
 				HCL_STACK_PUSH (hcl, c);
 
@@ -3633,6 +3644,12 @@ if (do_throw(hcl, hcl->_nil, fetched_instruction_pointer) <= -1)
 				hcl_oop_class_t t;
 				FETCH_PARAM_CODE_TO (hcl, b1);
 				LOG_INST_1 (hcl, "store_into_clsvar %zu", b1);
+				if (HCL_CLSTACK_IS_EMPTY(hcl))
+				{
+					hcl_seterrbfmt (hcl, HCL_ESTKUNDFLW, "empty class stack");
+					supplement_errmsg (hcl, fetched_instruction_pointer);
+					goto oops;
+				}
 				HCL_CLSTACK_FETCH_TOP_TO(hcl, t);
 				t->cvar[b1] = HCL_STACK_GETTOP(hcl);
 				break;
@@ -3643,6 +3660,12 @@ if (do_throw(hcl, hcl->_nil, fetched_instruction_pointer) <= -1)
 				hcl_oop_class_t t;
 				FETCH_PARAM_CODE_TO (hcl, b1);
 				LOG_INST_1 (hcl, "pop_into_clsvar %zu", b1);
+				if (HCL_CLSTACK_IS_EMPTY(hcl))
+				{
+					hcl_seterrbfmt (hcl, HCL_ESTKUNDFLW, "empty class stack");
+					supplement_errmsg (hcl, fetched_instruction_pointer);
+					goto oops;
+				}
 				HCL_CLSTACK_FETCH_TOP_TO(hcl, t);
 				t->cvar[b1] = HCL_STACK_GETTOP(hcl);
 				HCL_STACK_POP (hcl);

lib/hcl.h

@@ -90,7 +90,9 @@ enum hcl_errnum_t
 	HCL_ECALLARG,   /**< runtime error - wrong number of arguments to call */
 	HCL_ECALLRET,   /**< runtime error - wrong number of return variables to call */
 	HCL_ESEMFLOOD,  /**< runtime error - too many semaphores */
-	HCL_EEXCEPT    /**< runtime error - exception not handled */
+	HCL_EEXCEPT,    /**< runtime error - exception not handled */
+	HCL_ESTKOVRFLW, /**< runtime error - stack overflow */
+	HCL_ESTKUNDFLW  /**< runtime error - stack overflow */
 };
 typedef enum hcl_errnum_t hcl_errnum_t;
 

lib/obj.c

@@ -360,7 +360,7 @@ hcl_oop_t hcl_makeclass (hcl_t* hcl, hcl_oop_t superclass, hcl_ooi_t nivars, hcl
 	hcl_pushvolat (hcl, &superclass);
 	hcl_pushvolat (hcl, &ivars_str);
 	hcl_pushvolat (hcl, &cvars_str);
-	c = (hcl_oop_class_t)hcl_allocoopobj(hcl, HCL_BRAND_CLASS, HCL_CLASS_NAMED_INSTVARS);
+	c = (hcl_oop_class_t)hcl_allocoopobj(hcl, HCL_BRAND_CLASS, HCL_CLASS_NAMED_INSTVARS + ncvars);
 	hcl_popvolats (hcl, 3);
 	if (HCL_UNLIKELY(!c)) return HCL_NULL;
 

lib/utf8.c

@@ -26,6 +26,8 @@
 
 #include "hcl-prv.h"
 
+/*#define RETAIN_RFC2279 1*/
+
 /*
  * from RFC 2279 UTF-8, a transformation format of ISO 10646
  *
@@ -36,6 +38,12 @@
  * 4:4 00010000-001FFFFF  11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
  * inv 00200000-03FFFFFF  111110xx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
  * inv 04000000-7FFFFFFF  1111110x 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
+ *
+ * RFC3629 limits the ranges like this:
+ * 1:2 00000000-0000007F  0xxxxxxx
+ * 2:2 00000080-000007FF  110xxxxx 10xxxxxx
+ * 3:2 00000800-0000FFFF  1110xxxx 10xxxxxx 10xxxxxx
+ * 4:4 00010000-0010FFFF  11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
  */
 
 struct __utf8_t
@@ -55,9 +63,13 @@ static __utf8_t utf8_table[] =
 	{0x00000000ul, 0x0000007Ful, 0x00, 0x80, 0x7F, 1},
 	{0x00000080ul, 0x000007FFul, 0xC0, 0xE0, 0x1F, 2},
 	{0x00000800ul, 0x0000FFFFul, 0xE0, 0xF0, 0x0F, 3},
+#if defined(RETAIN_RFC2279)
 	{0x00010000ul, 0x001FFFFFul, 0xF0, 0xF8, 0x07, 4},
 	{0x00200000ul, 0x03FFFFFFul, 0xF8, 0xFC, 0x03, 5},
 	{0x04000000ul, 0x7FFFFFFFul, 0xFC, 0xFE, 0x01, 6}
+#else
+	{0x00010000ul, 0x0010FFFFul, 0xF0, 0xF8, 0x07, 4}
+#endif
 };
 
 static HCL_INLINE __utf8_t* get_utf8_slot (hcl_uch_t uc)