News for hackers. Stuff that matters.

A comment on A New Development for Coverity and Heartbleed. Parent comment

Re:A New Development for Coverity and Heartbleed by skybrian on Saturday April 12, 2014 @05:08PM (#7579211)
I think it works even better if you can get help from the type system. For example, the SafeHtml interface in GWT [1] gives you some safety from Java's type checking and can also make additional static analysis easier. (Then it becomes an exercise in making sure the API is used as intended.)

Perhaps something similar could be done using typedefs in C?

[1] http://www.gwtproject.org/javadoc/latest/com/google/gwt/safe...

[ Parent | on HN ]
Re:A New Development for Coverity and Heartbleed by dbaupp on Saturday April 12, 2014 @06:54PM (#7579510)
Typedefs in C are just aliases, so given `typedef int foo;` one can freely use `int`s and `foo`s interchangeably, i.e. no checking by the compiler.

That said, one could use actual wrapper structs around the various types.

[ Parent | on HN ]
Re:A New Development for Coverity and Heartbleed by skybrian on Saturday April 12, 2014 @07:27PM (#7579583)
A static analysis framework could treat certain typedefs specially since it's parsing the C code anyway.
[ Parent | on HN ]
Re:A New Development for Coverity and Heartbleed by marshray on Saturday April 12, 2014 @09:30PM (#7579816)
Great idea! Too bad openssl uses #define instead
[ Parent | on HN ]
Re:A New Development for Coverity and Heartbleed by johnkeeping on Sunday April 13, 2014 @06:26AM (#7580896)
There's also sparse[0]'s address space annotation, which Linux uses for annotating data from userspace.

[0] https://git.kernel.org/cgit/devel/sparse/sparse.git/

[ Parent | on HN ]