commons-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From robert burrell donkin <robertburrelldon...@blueyonder.co.uk>
Subject Re: [pool] synchronization issues in Pool
Date Mon, 24 Oct 2005 20:18:09 GMT
On Mon, 2005-10-24 at 14:32 -0400, Sandy McArthur wrote:
> On 10/23/05, robert burrell donkin <robertburrelldonkin@blueyonder.co.uk> wrote:
> > are there any pool developers out there with time to pick this up?
> >
> > otherwise, we could probably do with a volunteer to go through and
> > analyse these issues. anyone fancy taking a crack at this?
> 
> I'm not a pool dev but I'll make some time to implement fixes for the
> problems Mayur Naik found and submit patches.

cool

(in case you don't know the drill you might find some useful stuff in
http://www.apache.org/dev/,
http://jakarta.apache.org/site/getinvolved.html and
http://jakarta.apache.org/commons/patches.html. please create a bugzilla
for the patchs)

unless some one beats me to it, once the patch is available i'll add
myself to the dev list and review (as time permits).

> While I'm at it would it be desirable to transition to the privately
> head lock idiom to defend against intentionally malicious code? eg to
> defend against:
> 
> synchronized (aPool) {
>  Thread.sleep(Integer.Max_Value);
> }
> 
> by making anything that synchronized on this synchronize on a private field.

i'm not too bothered about (that kind of) malicious code: there are too
many ways the user of a library could subvert it. 

it might (however) stop a user doing something equivalent through
carelessness or naivity. so probably worth doing. would need to add a
note to the documentation since this would change the semantics.
opinions?

- robert


---------------------------------------------------------------------
To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-dev-help@jakarta.apache.org


Mime
View raw message