Compare commits

..

2 Commits

Author SHA1 Message Date
bog 7f5b5486c0 ADD: array index. 2023-08-24 17:22:49 +02:00
bog 73cd8e0c4a ADD: array types better representation. 2023-08-24 14:29:13 +02:00
17 changed files with 415 additions and 85 deletions

View File

@ -30,6 +30,12 @@ void array_free(array* self)
free(self->type); self->type = NULL;
}
struct value* array_deref_copy(array* self, size_t index)
{
assert(index < self->children.size);
return (struct value*) value_new_clone((value*)self->children.data[index]);
}
void array_push(array* self, struct value* element)
{
assert(self);

View File

@ -18,6 +18,7 @@ typedef struct {
void array_init(array* self, type* elements_type);
void array_free(array* self);
struct value* array_deref_copy(array* self, size_t index);
void array_push(array* self, struct value* element);
size_t array_str(array* self, char* buffer, size_t size);

View File

@ -17,13 +17,18 @@ void compile_node(compiler* self, node* root, program* prog)
assert(root);
assert(prog);
if (root->type == NODE_MUL)
if (root->type == NODE_INDEX)
{
compile_children(self, root, prog);
program_add_instr(prog, OP_ADEREF, NO_PARAM);
}
else if (root->type == NODE_MUL)
{
cstatic cs;
cstatic_init(&cs);
int lhs = cstatic_resolve(&cs, root->children.data[0]);
int rhs = cstatic_resolve(&cs, root->children.data[1]);
int lhs = cstatic_resolve_base_type(&cs, root->children.data[0]);
int rhs = cstatic_resolve_base_type(&cs, root->children.data[1]);
cstatic_free(&cs);
@ -49,8 +54,8 @@ void compile_node(compiler* self, node* root, program* prog)
cstatic cs;
cstatic_init(&cs);
int lhs = cstatic_resolve(&cs, root->children.data[0]);
int rhs = cstatic_resolve(&cs, root->children.data[1]);
int lhs = cstatic_resolve_base_type(&cs, root->children.data[0]);
int rhs = cstatic_resolve_base_type(&cs, root->children.data[1]);
cstatic_free(&cs);
@ -195,7 +200,7 @@ void compile_number(compiler* self, node* root, program* prog,
cstatic cs;
cstatic_init(&cs);
int ty = cstatic_resolve(&cs, root);
int ty = cstatic_resolve_base_type(&cs, root);
switch (ty)
{

View File

@ -11,74 +11,115 @@ void cstatic_free(cstatic* self)
assert(self);
}
int cstatic_resolve(cstatic* self, node* ast)
type* cstatic_new_type(cstatic* self, int base_type)
{
assert(self);
type* ty = malloc(sizeof(type));
type_init(ty, base_type);
return ty;
}
int cstatic_resolve_base_type(cstatic* self, node* ast)
{
assert(self);
type* ty = cstatic_resolve_new(self, ast);
int res = ty->base_type;
type_free(ty);
free(ty);
return res;
}
type* cstatic_resolve_new(cstatic* self, node* ast)
{
assert(self);
assert(ast);
if (ast->type == NODE_MUL)
{
int lhs = cstatic_resolve(self, ast->children.data[0]);
int rhs = cstatic_resolve(self, ast->children.data[1]);
int lhs = cstatic_resolve_base_type(self, ast->children.data[0]);
int rhs = cstatic_resolve_base_type(self, ast->children.data[1]);
if (lhs == TY_STRING || rhs == TY_STRING)
{
return TY_STRING;
return cstatic_new_type(self, TY_STRING);
}
}
if (ast->type == NODE_INTEGER)
if (ast->type == NODE_INDEX)
{
return TY_INTEGER;
// find inner type using first array element
size_t dim = ast->children.data[1]->children.size;
node* iter = ast->children.data[0];
for (size_t i=0; i<dim; i++)
{
iter = iter->children.data[0];
}
return cstatic_resolve_new(self, iter);
}
else if (ast->type == NODE_INTEGER)
{
return cstatic_new_type(self, TY_INTEGER);
}
else if (ast->type == NODE_FLOAT)
{
return TY_FLOAT;
return cstatic_new_type(self, TY_FLOAT);
}
else if (ast->type == NODE_ADD)
{
int lhs = cstatic_resolve(self, ast->children.data[0]);
int rhs = cstatic_resolve(self, ast->children.data[1]);
int lhs = cstatic_resolve_base_type(self, ast->children.data[0]);
int rhs = cstatic_resolve_base_type(self, ast->children.data[1]);
if (lhs == TY_STRING && rhs == TY_STRING)
{
return TY_STRING;
return cstatic_new_type(self, TY_STRING);
}
return lhs;
return cstatic_new_type(self, lhs);
}
else if (ast->type == NODE_STRING)
{
return TY_STRING;
return cstatic_new_type(self, TY_STRING);
}
else if (ast->type == NODE_ARRAY)
{
return TY_ARRAY;
assert(ast->children.size > 0);
type* child = cstatic_resolve_new(self, ast->children.data[0]);
type* ty = malloc(sizeof(type));
type_init_array(ty, child);
type_free(child);
free(child);
return ty;
}
else if (ast->type == NODE_BOOLEAN
|| ast->type == NODE_EQ
|| ast->type == NODE_NE)
{
return TY_BOOLEAN;
return cstatic_new_type(self, TY_BOOLEAN);
}
else
{
for (size_t i=0; i<ast->children.size; i++)
{
int ty = cstatic_resolve(self, ast->children.data[i]);
type* ty = cstatic_resolve_new(self, ast->children.data[i]);
if (ty != TY_NIL)
if (ty != NULL)
{
return ty;
}
}
}
return TY_NIL;
return NULL;
}
int cstatic_check(cstatic* self, node* ast, char* msg, size_t size)
@ -98,11 +139,55 @@ int cstatic_check(cstatic* self, node* ast, char* msg, size_t size)
}
}
// Arrays
if (ast->type == NODE_ARRAY)
if (ast->type == NODE_INDEX)
{
// check dimension
size_t dim = ast->children.data[1]->children.size;
size_t real_dim = 0;
node* iter = ast->children.data[0];
while (iter->type == NODE_ARRAY)
{
iter = iter->children.data[0];
real_dim++;
}
if (real_dim != dim)
{
snprintf(msg, size,
"E: array dimension mismatch:"
" expected '%ld', got '%ld'.",
real_dim, dim);
return 0;
}
// check that all indexes are integers
for (size_t i=0; i<dim; i++)
{
int status = cstatic_check_type_base(
self,
ast->children.data[1]->children.data[i],
msg,
size,
TY_INTEGER,
TYPE_END
);
if (!status)
{
return 0;
}
}
return 1;
}
else if (ast->type == NODE_ARRAY)
{
assert(ast->children.size > 0);
int ty = cstatic_resolve(self, ast->children.data[0]);
type* ty = cstatic_resolve_new(self, ast->children.data[0]);
for (size_t i=1; i<ast->children.size; i++)
{
@ -112,23 +197,28 @@ int cstatic_check(cstatic* self, node* ast, char* msg, size_t size)
msg,
size,
ty,
TYPE_END
NULL
);
if (!status)
{
type_free(ty);
free(ty);
return status;
}
}
type_free(ty);
free(ty);
return 1;
}
// String Operations
if (ast->type == NODE_ADD)
{
int lhs = cstatic_resolve(self, ast->children.data[0]);
int rhs = cstatic_resolve(self, ast->children.data[1]);
int lhs = cstatic_resolve_base_type(self, ast->children.data[0]);
int rhs = cstatic_resolve_base_type(self, ast->children.data[1]);
if (lhs == rhs && lhs == TY_STRING)
@ -139,8 +229,8 @@ int cstatic_check(cstatic* self, node* ast, char* msg, size_t size)
if (ast->type == NODE_MUL)
{
int lhs = cstatic_resolve(self, ast->children.data[0]);
int rhs = cstatic_resolve(self, ast->children.data[1]);
int lhs = cstatic_resolve_base_type(self, ast->children.data[0]);
int rhs = cstatic_resolve_base_type(self, ast->children.data[1]);
if ((lhs == TY_STRING && rhs == TY_INTEGER)
|| (lhs == TY_INTEGER && rhs == TY_STRING))
@ -169,7 +259,7 @@ int cstatic_check(cstatic* self, node* ast, char* msg, size_t size)
else if (ast->type == NODE_ASSERT
|| ast->type == NODE_NOT)
{
int status = cstatic_check_type(self,
int status = cstatic_check_type_base(self,
ast->children.data[0],
msg,
size,
@ -197,7 +287,7 @@ int cstatic_check(cstatic* self, node* ast, char* msg, size_t size)
return status;
}
status = cstatic_check_type(self,
status = cstatic_check_type_base(self,
ast->children.data[0],
msg,
size,
@ -213,7 +303,7 @@ int cstatic_check(cstatic* self, node* ast, char* msg, size_t size)
else if (ast->type == NODE_UADD
|| ast->type == NODE_USUB)
{
int status = cstatic_check_type(self,
int status = cstatic_check_type_base(self,
ast->children.data[0],
msg,
size,
@ -245,7 +335,7 @@ int cstatic_check(cstatic* self, node* ast, char* msg, size_t size)
return status;
}
status = cstatic_check_type(self,
status = cstatic_check_type_base(self,
ast->children.data[0],
msg,
size,
@ -262,6 +352,67 @@ int cstatic_check(cstatic* self, node* ast, char* msg, size_t size)
}
int cstatic_check_type(cstatic* self, node* lhs,
char* msg, size_t size, type* types, ...)
{
assert(self);
assert(lhs);
char aze[512];
node_str(lhs, aze, 512);
va_list args;
va_start(args, types);
type* rhs = types;
type* left = cstatic_resolve_new(self, lhs);
type* all_types[TY_TYPE_COUNT];
for (int i=0; i<TY_TYPE_COUNT; i++)
{
all_types[i] = NULL;
}
size_t i = 0;
while (rhs != NULL)
{
all_types[i] = rhs;
if (type_equals(left, rhs))
{
va_end(args);
type_free(left);
free(left);
return 1;
}
rhs = va_arg(args, type*);
i++;
}
va_end(args);
size_t sz = snprintf(msg, size, "E(%d): type mismatch, got '", lhs->lineno);
sz += type_str(left, msg + sz, size - sz);
sz += snprintf(msg + sz, size - sz, "', expected: \n");
size_t j = 0;
while (all_types[j] != NULL)
{
sz += snprintf(msg + sz, size - sz, "\t - ");
sz += type_str(all_types[j], msg + sz, size - sz);
sz += snprintf(msg + sz, size - sz, "\n");
j++;
}
type_free(left);
free(left);
return 0;
}
int cstatic_check_type_base(cstatic* self, node* lhs,
char* msg, size_t size, int types, ...)
{
assert(self);
@ -273,7 +424,7 @@ int cstatic_check_type(cstatic* self, node* lhs,
va_start(args, types);
int rhs = types;
int left = cstatic_resolve(self, lhs);
int left = cstatic_resolve_base_type(self, lhs);
int all_types[TY_TYPE_COUNT];
memset(all_types, TYPE_END, TY_TYPE_COUNT * sizeof(int));
@ -297,7 +448,7 @@ int cstatic_check_type(cstatic* self, node* lhs,
size_t sz = snprintf(msg, size, "E(%d): type mismatch, got '%s', ",
lhs->lineno,
TypesStr[left]);
TypesStr[left] + strlen("TY_"));
sz += snprintf(msg + sz, size - sz, "expected: \n");
@ -306,7 +457,7 @@ int cstatic_check_type(cstatic* self, node* lhs,
while (all_types[j] != TYPE_END)
{
sz += snprintf(msg + sz, size - sz, "\t '%s'\n",
TypesStr[all_types[j]]);
TypesStr[all_types[j]] + strlen("TY_"));
j++;
}
@ -320,17 +471,27 @@ int cstatic_check_same_type(cstatic* self, node* lhs, node* rhs,
assert(lhs);
assert(rhs);
int left = cstatic_resolve(self, lhs);
int right = cstatic_resolve(self, rhs);
type* left = cstatic_resolve_new(self, lhs);
type* right = cstatic_resolve_new(self, rhs);
if (left != right)
if (!type_equals(left, right))
{
snprintf(msg, size, "E(%d): expected '%s', got '%s'.",
lhs->lineno,
TypesStr[left],
TypesStr[right]);
size_t sz = 0;
sz += snprintf(msg + sz, size - sz, "E(%d): expected '", lhs->lineno);
sz += type_str(left, msg + sz, size - sz);
sz += snprintf(msg + sz, size - sz, "', got '");
sz += type_str(right, msg + sz, size - sz);
sz += snprintf(msg + sz, size - sz, "'");
type_free(left); free(left);
type_free(right); free(right);
return 0;
}
type_free(left); free(left);
type_free(right); free(right);
return 1;
}

View File

@ -3,6 +3,7 @@
#include "commons.h"
#include "node.h"
#include "type.h"
#define TYPE_END (-1)
@ -13,10 +14,16 @@ typedef struct {
void cstatic_init(cstatic* self);
void cstatic_free(cstatic* self);
int cstatic_resolve(cstatic* self, node* ast);
type* cstatic_new_type(cstatic* self, int base_type);
int cstatic_resolve_base_type(cstatic* self, node* ast);
type* cstatic_resolve_new(cstatic* self, node* ast);
int cstatic_check(cstatic* self, node* ast, char* msg, size_t size);
int cstatic_check_type(cstatic* self, node* lhs,
char* msg, size_t size, type* types, ...);
int cstatic_check_type_base(cstatic* self, node* lhs,
char* msg, size_t size, int types, ...);
int cstatic_check_same_type(cstatic* self, node* lhs, node* rhs,

View File

@ -20,6 +20,7 @@ STRING \"[^"]*\"
{WHITESPACES} {}
"assert" { return ASSERT; }
"," { return COMMA; }
"+" { return ADD; }
"-" { return SUB; }
"*" { return MUL; }

View File

@ -60,13 +60,15 @@ int main(int argc, char** argv)
{
size_t const BUF = 1024;
char buffer[BUF];
memset(buffer, 0, BUF);
node_str(ast, buffer, BUF);
printf("-- ast ---\n%s\n\n", buffer);
memset(buffer, 0, BUF);
program_str(&prog, buffer, BUF);
printf("--- program ---\n%s\n", buffer);
memset(buffer, 0, BUF);
vm_str(&v, buffer, BUF);
printf("--- stack ---\n%s\n", buffer);
}

View File

@ -10,7 +10,7 @@
G(NODE_ASSERT), \
G(NODE_UADD), G(NODE_USUB), G(NODE_ADD), G(NODE_SUB), \
G(NODE_MUL), G(NODE_DIV), G(NODE_MOD), G(NODE_POW), \
G(NODE_ARRAY)
G(NODE_ARRAY), G(NODE_INDEX)
#include "mutils.h"

View File

@ -28,7 +28,8 @@
G(OP_FUSUB), \
G(OP_CAT), \
G(OP_SMUL), \
G(OP_MKARRAY)
G(OP_MKARRAY), \
G(OP_ADEREF)
enum Opcodes {
OPCODES(GEN_ENUM)

View File

@ -23,7 +23,8 @@
%left ASSERT
%token <str> BOOLEAN INTEGER FLOAT STRING
%type <n_children> expr exprs prog;
%type <n_children> expr exprs prog array arrays builtins
%type <n_children> expr_list
%left EQ NE
%left AND
%left OR
@ -31,7 +32,7 @@
%left MUL DIV MOD
%left POW
%left NOT
%token OPAR CPAR OSQUARE CSQUARE
%token OPAR CPAR OSQUARE CSQUARE COMMA
%%
@ -66,7 +67,36 @@ exprs:
expr:
ASSERT expr {
// INDEX
array array {
node *n = malloc(sizeof(node));
node_init(n, NODE_INDEX, "", line);
size_t const SZ = $1 + $2;
node* all[SZ];
for (size_t i=0; i<SZ; i++)
{
all[i] = stack_pop();
}
for (size_t i=0; i<SZ; i++)
{
node_add_child(n, all[SZ - 1 - i]);
}
stack_push(n);
$$ = 1;
}
// EXPRESSIONS
| array {
$$ = 1;
}
| builtins {
$$ = $1;
}
| ASSERT expr {
node *n = malloc(sizeof(node));
node_init(n, NODE_ASSERT, "", line);
node_add_child(n, stack_pop());
@ -213,27 +243,12 @@ expr:
$$ = $2;
}
| OSQUARE exprs CSQUARE {
node *n = malloc(sizeof(node));
node_init(n, NODE_ARRAY, "", line);
node* elements[$2];
;
for (size_t i=0; i<$2; i++)
{
elements[i] = stack_pop();
}
builtins:
for (size_t i=0; i<$2; i++)
{
node_add_child(n, elements[$2 - 1 - i]);
}
stack_push(n);
$$ = 1;
}
| STRING {
STRING {
node* n = malloc(sizeof(node));
size_t const SZ = strlen($1);
char str[SZ - 2];
@ -267,6 +282,34 @@ expr:
}
;
expr_list:
expr { $$ = $1; }
| expr_list COMMA expr { $$ = $1 + $3; }
;
array:
OSQUARE expr_list CSQUARE {
node *n = malloc(sizeof(node));
node_init(n, NODE_ARRAY, "", line);
node* elements[$2];
for (size_t i=0; i<$2; i++)
{
elements[i] = stack_pop();
}
for (size_t i=0; i<$2; i++)
{
node_add_child(n, elements[$2 - 1 - i]);
}
stack_push(n);
$$ = 1;
}
;
%%
void yyerror(char const* msg)

View File

@ -6,11 +6,26 @@ void type_init(type* self, int base_type)
{
assert(self);
self->base_type = base_type;
self->array_type = NULL;
}
void type_init_array(type* self, type* array_type)
{
assert(self);
assert(array_type);
type_init(self, TY_ARRAY);
self->array_type = type_new_clone(array_type);
}
void type_free(type* self)
{
assert(self);
if (self->array_type != NULL)
{
type_free(self->array_type);
free(self->array_type);
}
}
type* type_new_clone(type* self)
@ -20,6 +35,11 @@ type* type_new_clone(type* self)
type* clone = malloc(sizeof(type));
type_init(clone, self->base_type);
if (self->array_type)
{
clone->array_type = type_new_clone(self->array_type);
}
return clone;
}
@ -28,11 +48,34 @@ int type_equals(type* self, type* rhs)
assert(self);
assert(rhs);
return self->base_type == rhs->base_type;
if (self->base_type != rhs->base_type)
{
return 0;
}
if (self->base_type == TY_ARRAY)
{
assert(self->array_type);
assert(rhs->array_type);
return type_equals(self->array_type, rhs->array_type);
}
return 1;
}
size_t type_str(type* self, char* buffer, size_t size)
{
assert(self);
return snprintf(buffer, size, "%s", TypesStr[self->base_type]);
size_t sz = snprintf(buffer, size, "%s",
TypesStr[self->base_type] + strlen("TY_"));
if (self->base_type == TY_ARRAY)
{
sz += snprintf(buffer + sz, size - sz, "<");
sz += type_str(self->array_type, buffer + sz, size - sz);
sz += snprintf(buffer + sz, size - sz, ">");
}
return sz;
}

View File

@ -20,9 +20,11 @@ extern char const* TypesStr[];
typedef struct type {
int base_type;
struct type* array_type;
} type;
void type_init(type* self, int base_type);
void type_init_array(type* self, type* array_type);
void type_free(type* self);
type* type_new_clone(type* self);

View File

@ -51,8 +51,9 @@ void value_init_array(value* self, array* arr, int lineno)
self->val.array_val = array_new_clone(arr);
self->lineno = lineno;
self->type = malloc(sizeof(type));
type_init(self->type, TY_ARRAY);
type_init_array(self->type, arr->type);
}
void value_free(value* self)

View File

@ -70,6 +70,7 @@ void vm_exec(vm* self, program* prog)
case OP_CAT: vm_cat(self); break;
case OP_MKARRAY: vm_mkarray(self, param); break;
case OP_ADEREF: vm_aderef(self); break;
default: {
fprintf(stderr, "unknown opcode %s\n",
@ -568,8 +569,8 @@ void vm_mkarray(vm* self, int param)
array_push(arr, (struct value*) elements[i]);
}
value_init_array(arr_val, arr, elements[0]->lineno);
vm_push_value(self, arr_val);
for (int i=0; i<param; i++)
@ -583,3 +584,39 @@ void vm_mkarray(vm* self, int param)
self->pc++;
}
void vm_aderef(vm* self)
{
assert(self);
value* idx = vm_pop_value(self);
value* arr = vm_pop_value(self);
array* idx_array = idx->val.array_val;
value* val = NULL;
for (size_t i=0; i<idx_array->children.size; i++)
{
size_t index = ((value*) idx_array->children.data[i])->val.integer;
val = (value*) array_deref_copy(
arr->val.array_val,
index);
if (val->type->base_type == TY_ARRAY)
{
value_free(arr);
free(arr);
arr = val;
}
}
assert(val);
vm_push_value(self, val);
value_free(idx); free(idx);
value_free(arr); free(arr);
self->pc++;
}

View File

@ -54,4 +54,6 @@ void vm_smul(vm* self);
void vm_cat(vm* self);
void vm_mkarray(vm* self, int param);
void vm_aderef(vm* self);
#endif

View File

@ -1,2 +1,8 @@
[]
[1 2 3 "4"]
[1, 2, 3, "4"]
[ [0, 1], ["0", "1"] ]
[1, 2, 3]["a"]
[1, 2, 3][false]
[1, 2, 3][3.14]
[1, 2, 3][ [0, 1] ]
[1, 2, 3][0, 0]

View File

@ -1,11 +1,23 @@
assert [1 - 2 3] == [-1 3]
assert [1 -2 3] == [1 -2 3]
assert [1, 2, 3] == [1, 2, 3]
assert ["a", "b", "c"] == ["a", "b", "c"]
assert [1, 2, 3] != [7, 2, 3]
assert [1, 2, 3] != [1, 2, 3, 4]
assert ["a", "b", "c"] != ["ac", "b", "c"]
assert ["a", "b", "c"] != ["a", "b", "c", "d"]
assert 3 == [3, 8, 9][0]
assert 8 == [3, 8, 9][1]
assert 9 == [3, 8, 9][2]
assert "8" == ["3", "8", "9"][2 - 1]
assert 7 == [[4, 6], [9, 2, 7], [3, 3, 2]][1, 2]
assert 9 == [[4, 6], [9, 2, 7], [3, 3, 2]][[2, 9, 1][2], 0]
assert [1 2 3] == [1 2 3]
assert ["a" "b" "c"] == ["a" "b" "c"]
assert [1 2 3] != [7 2 3]
assert [1 2 3] != [1 2 3 4]
assert ["a" "b" "c"] != ["ac" "b" "c"]
assert ["a" "b" "c"] != ["a" "b" "c" "d"]