The ATS code uses that terminology. The C++ Standard lib has std::lock_guard . If you are saying we should not transition to the stdlib stuff, why?
The utility I'm proposing makes it a little more convenient to use a mutex with lock_guards in typical cases that a mutex is used. It also makes it very clear what the mutex is protecting, and makes accessing the protected data less error prone. To me, it's a big problem in the ATS source base that it's not clear how mutual exclusion to non-stack data is guaranteed. When you see a mutex in our code, it's generally not clear what all it's protecting. On Wed, Mar 20, 2019 at 4:17 PM Leif Hedstrom <zw...@apache.org> wrote: > Don’t we have scoped mutexes already ? > > — Leif > > > On Mar 20, 2019, at 12:08, Walt Karas <wka...@verizonmedia.com.invalid> > wrote: > > > > https://godbolt.org/z/6klEJn > > > > Useful for mutex-protected objects, where the mutex is only locked within > > code blocks. Could be used in this PR: > > > > https://github.com/apache/trafficserver/pull/5187 > >
