Yeah, that makes sense. I will see how to construct such a unit test(or call it benchmark)
Thanks, Charles. 2016-06-17 10:44 GMT-07:00 Christos Zoulas <chris...@zoulas.com>: > On Jun 16, 9:28pm, charles.cui1...@gmail.com (Charles Cui) wrote: > -- Subject: Re: _SC_SIGQUEUE_MAX > > | Hi Christos, > | > | I have completed some unit tests for ad's work and run these unit > tests > | on two versions (my version and your version). > | Your initial thinking is correct. There are bugs when getting and setting > | different fields and these bugs can be > | verified via my unit tests (it does not pass in my version, but pass in > | your version). > | Here are these patches. > | > https://github.com/ycui1984/posixtestsuite/commit/1a5bafe2691daf1755f7156d246f297568f7768b > | > https://github.com/ycui1984/posixtestsuite/commit/927e1ea5a3c9aa072ef7a18316903f6cf1a20056 > > Thanks; these look great but only test the basic functionality of > getting and setting fields. They don't test any of the actual > functionality of the feature being added. What I would like you to > do is to think how to construct test cases which involve multiple > threads that create the conditions for the priority inheritance > (and priority inversion if searching for that helps you understand > what is going on) that the new system call is attempting to fix. > > christos >
