Unsafe: removing the monitorEnter/monitorExit/tryMonitorEnter methods

Martin Buchholz martinrb at google.com
Fri Feb 28 20:24:57 UTC 2014


I find myself in agreement with Heinz and David Loyd.

j.u.c. locks have one primary advantage over builtin locks: flexibility.
There used to be a performance advantage, but today it seems like builtin
locks use fewer resources and can be preferred if the added flexibility is
not required.  So why not add back in some of the flexibility when it
doesn't impose too much of a burden on the VM?  Check out the public
methods of ReentrantLock and see which are easy to support.  For the
purposes of the original usage of Unsafe in this thread. we want an
isLocked method on builtin monitors.


On Thu, Feb 27, 2014 at 7:11 AM, David M. Lloyd <david.lloyd at redhat.com>wrote:

> A "Monitors" class could possibly be introduced which simply has three
> (intrinsic) static methods: void lock(Object obj), void unlock(Object obj),
> boolean tryLock(Object obj).
>
> Could even add a void lockInterruptibly(Object obj) throws
> InterruptedException.
>
> Why not?  Monitors are not going anywhere, and they're considerably
> lighter than ReentrantLock in terms of memory usage and allocations in the
> (pretty common) case where you only need zero or one conditions. Avoiding
> something like this out of loyalty to new APIs is probably pretty misguided.
>
>
> On 02/27/2014 08:51 AM, Dr Heinz M. Kabutz wrote:
>
>> I have never used this "in the wild", but rather have moved over to the
>> ReentrantLock when I needed that particular functionality.
>>
>> However, I do see a place for this ability.  As I wrote in here:
>> http://www.javaspecialists.eu/archive/Issue194.html - sometimes you need
>> to modify code that is already using a particular locking mechanism.  To
>> then redo everything with ReentrantLock can introduce errors.
>>
>> If I had a say, I would vote to either keep that method in Unsafe (which
>> is where I think it belongs) or provide an alternative way to make the
>> tryMonitorEnter() mechanism available to those that might end up needing
>> it.
>>
>> Since I bought my Suzuki Jimny 7 years ago, I have not needed my spare
>> wheel even once.  So yeah, I could take it off and drive around without
>> it.  But chances are, the day after I take it off, I will need it.
>>
>> tryMonitorEnter() does no harm and it is out of reach of most programmers.
>>
>> Regards
>>
>> Heinz
>>
>
>
> --
> - DML
>



More information about the core-libs-dev mailing list