Use const object parameters in getter API functions #1047
Conversation
LaszloLango
added
enhancement
| @@ -392,7 +392,7 @@ ecma_op_function_object_get_own_property (ecma_object_t *obj_p, /**< the functio | |||
| } | |||
| else if (!ecma_get_object_is_builtin (obj_p)) | |||
| { | |||
| prop_p = ecma_op_function_try_lazy_instantiate_property (obj_p, property_name_p); | |||
| prop_p = ecma_op_function_try_lazy_instantiate_property ((ecma_object_t *) obj_p, property_name_p); | |||
There was a problem hiding this comment.
IMHO, this cast makes the patch ugly and non-secure. Any suggestion to avoid this?
There was a problem hiding this comment.
You cannot avoid this, since reading a property (length for example) might also trigger its construction. Hence the object is extended, and an out-of-memory error can stop the engine.
|
This is a WIP patch. There are few const to non-const cast in the code that shouldn't be there. |
LaszloLango
force-pushed
the
use-const-in-api-getters
branch
2 times, most recently
from
dfdda5f to
f63c3f2
Compare
|
What is the plan with this patch? Shall we move the property hash map creation into some other place? Or make it error tolerant? |
|
@zherczeg, good question. :) I think the API change is still needed and the referred issue is valid and open in a long time, so we should do something with it. Moving the property hash map creation sounds good to me, but currently no idea how to do it. Any specific idea? |
|
Perhaps we can move it when a property is assigned. But that does not easy either. |
LaszloLango
force-pushed
the
use-const-in-api-getters
branch
from
f63c3f2 to
56dd68c
Compare
Related issue: jerryscript-project#668 JerryScript-DCO-1.0-Signed-off-by: László Langó llango.u-szeged@partner.samsung.com
LaszloLango
force-pushed
the
use-const-in-api-getters
branch
from
56dd68c to
6471e03
Compare
|
Won't fix. |
Related issue: #668
JerryScript-DCO-1.0-Signed-off-by: László Langó llango.u-szeged@partner.samsung.com