stensal

4 ms·

1. 1.▲
   by stensal 8y ago

   I changed it to "dynamic pointer typing". Thanks for your questions.
2. 2.▲
   by stensal 8y ago

   I want to avoid to do comparison with the said X tool without fully understanding their constraints because different approaches have made different design trade off and each has its own strength. That having been said, the academia often p
3. 3.▲
   by stensal 8y ago

   Heap is a centralized resource. Each allocation/free will have to acquire/release lock. It's much more costly than stack allocation/deallocation. You might get more efficient machine code, but using heap will likely cost
4. 4.▲
   by stensal 8y ago

   Thanks a lot for reposting it here. Our original post https://news.ycombinator.com/item?id=18894032 has answered some of the questions you might have.
5. 5.▲
   by stensal 8y ago

   If ASAN/MSAN works for you well that's great. You want to give Stensal SDK a try, you will find it's easier to use provided Stensal SDK can build your code. If you have any problems, feel free to send bugs report to suppor
6. 6.▲
   by stensal 8y ago

   Ok, The following is the highly level goal. It dictates the design choice and hence the user experience of Stensal SDK. I want to make C/C++ more user friendly and less scary to beginners. If you love C, you should keep enjoying codin
7. 7.▲
   by stensal 8y ago

   Nice try bro! That will make C like Java, it will not happen. C should be like just C.
8. 8.▲
   by stensal 8y ago

   As you have rightly pointed out the intended use case. I want to add more constraints to it. It's intended to be used in production deployments that are security critical, constant applying security patches is not an optimal solution,
9. 9.▲
   by stensal 8y ago

   (1) It is not designed to find more memory errors than said X tool. It is designed to find complete runtime memory errors that can cause security concerns, more specifically, like remote code execution and information leakage (like heartble
10. 10.▲
    by stensal 8y ago

    No, the pointer always has the native pointer size. I think the type in my use of "dynamic type system" confused you. The type has a very narrow meaning, it has only two values. It is either POINTER or NON_POINTER. A POINTER type
11. 11.▲
    by stensal 8y ago

    Yes, you have stated it correctly. The size and liveness information are from allocations (either stack or heap). ASAN and Valgrind do not associate these information with pointers, but add redzones (extra paddings with special bit pattern
12. 12.▲
    by stensal 8y ago

    I didn't know SillyMUD. But I did read the paper about an interpretation based memory checking tool many years ago. Unfortunately, I forgot the name. I believe the interpretation based memory checking tool assigns each pointer a type.
13. 13.▲
    by stensal 8y ago

    I guess you ask how this dynamic type system works. It basically gives each valid pointer a type, which includes the size information and liveness information of memory pointed by the pointer. When a pointer is dereferenced, the runtime che
14. 14.▲
    by stensal 8y ago

    I'm the founder and built the SDK. Agreed, ASAN and Valgrind cover the most memory errors and are readily available for almost all platforms. The SDK is designed for a different purpose -- to catch all runtime memory issues with a ver