On 27 September 2010 21:34, Ohad Shacham wrote:
> ConcurrentHashMap does not support null keys and entries.
Did not check that. However, the attributes Map field is protected and
not final, so it could in theory be replaced with a Map that allows
dupes.
That's perhaps another bug...
--
> From: ohad.shac...@gmail.com [mailto:ohad.shac...@gmail.com] On Behalf Of
> Ohad Shacham
> Subject: Re: Atomicity violation in removeAttribute
> I would use:
> if (attributes.containsKey(name)) {
> value = attributes.remove(name);
> if (value == null)
>
> If get() cannot return null, then I think one can just use:
> value = attributes.remove(name);
> if (value == null)
> return;
> If it is important to distinguish null entries from missing entries,
> then a different approach is needed - e.g. AtomicReference as
> suggested elsewhere
Concu
On 26 September 2010 15:00, Ohad Shacham wrote:
> Hi,
>
>
> In addition to the behavior we experienced in function setAttribute. We also
> experienced another atomicity violation in function removeAttribute.
>
> The following code is located at the beginning of function removeAttribute
> at class
Hello,
I'd rather use the compareAndSet method in AtomicReference for this
particular task.
On Sun, Sep 26, 2010 at 4:00 PM, Ohad Shacham wrote:
> Hi,
>
>
> In addition to the behavior we experienced in function setAttribute. We
> also
> experienced another atomicity violation in function rem
