Order that things happen when returning from a function.

[jason.cipriani]

Let's say I have something like this, a class that wraps a "Mutex",
locks it on construction and unlocks it when it's destroyed (for
example):

class AutoMutex {
public:
AutoMutex (Mutex &) ... ;
~AutoMutex () ... ;
};

Now let's say I use it in a function that returns a copy of another
complex object:

class Something { ... };

Mutex TheMutex;
Something TheThing;

Something function () {
AutoMutex m(TheMutex);
return TheThing;
}

If I am using, in that example, "TheMutex" to protect access to
"TheThing", will "TheThing" be accessed before or after "m" goes out
of scope and unlocks the mutex?

Thanks,
Jason

 

[Helge Kruse]

Let's say I have something like this, a class that wraps a "Mutex",
locks it on construction and unlocks it when it's destroyed (for
example): [...]

class Something { ... };

Mutex TheMutex;
Something TheThing;

Something function () {
AutoMutex m(TheMutex);
return TheThing;
}

If I am using, in that example, "TheMutex" to protect access to
"TheThing", will "TheThing" be accessed before or after "m" goes out
of scope and unlocks the mutex?

The AutoMutex destructor is called after TheThing is accessed. If TheThing
is an struct or class, the assignemnt to the caller's object is called
before TheMutex is unlocked. If TheThing is a scalar (hold in a register),
the value is saved before calling Mutex destuctor.

/Helge

 

[jason.cipriani]

The AutoMutex destructor is called after TheThing is accessed. If TheThing
is an struct or class, the assignemnt to the caller's object is called
before TheMutex is unlocked. If TheThing is a scalar (hold in a register),
the value is saved before calling Mutex destuctor.

Thanks.

Jason