Comments on: Hashtable and synchronization http://tech.puredanger.com/2008/02/21/hashtable-and-synchronization/ Alex Miller's technical blog Thu, 08 Jan 2009 22:22:26 +0000 http://wordpress.org/?v=2.0.2 by: Anonymous http://tech.puredanger.com/2008/02/21/hashtable-and-synchronization/#comment-32609 Fri, 14 Mar 2008 20:05:07 +0000 http://tech.puredanger.com/2008/02/21/hashtable-and-synchronization/#comment-32609 If you dont synchronize your Hashtable or Collections.synchronizedList when you iterate on it, you risk getting an ConcurrentModificationException if another thread mutate the data structure while you perform the iteration. Which I guess is the best case scenario, the worse case is getting corrupted data, such as the same data twice, or missing data. If you dont synchronize your Hashtable or Collections.synchronizedList when you iterate on it, you risk getting an ConcurrentModificationException if another thread mutate the data structure while you perform the iteration. Which I guess is the best case scenario, the worse case is getting corrupted data, such as the same data twice, or missing data.

]]> by: Robert Konigsberg http://tech.puredanger.com/2008/02/21/hashtable-and-synchronization/#comment-28714 Sat, 23 Feb 2008 07:00:42 +0000 http://tech.puredanger.com/2008/02/21/hashtable-and-synchronization/#comment-28714 Interestingly, your last example becomes more nefarious when the iterator is implicit, as in: for (Object o : list) { foo(o); } Who expected the need for those to be synchronized as well? But I am still unconvinced this is correct. (Not saying you're wrong, I'm just skeptical, and since the source is close by...) Interestingly, your last example becomes more nefarious when the iterator is implicit, as in:

for (Object o : list) { foo(o); }

Who expected the need for those to be synchronized as well? But I am still unconvinced this is correct. (Not saying you’re wrong, I’m just skeptical, and since the source is close by…)

]]> by: Alex http://tech.puredanger.com/2008/02/21/hashtable-and-synchronization/#comment-28495 Thu, 21 Feb 2008 13:45:16 +0000 http://tech.puredanger.com/2008/02/21/hashtable-and-synchronization/#comment-28495 Yes, I would in general agree, although there are some rare cases where I've used the synch wrappers. Yes, I would in general agree, although there are some rare cases where I’ve used the synch wrappers.

]]> by: Sam Halliday http://tech.puredanger.com/2008/02/21/hashtable-and-synchronization/#comment-28459 Thu, 21 Feb 2008 09:43:52 +0000 http://tech.puredanger.com/2008/02/21/hashtable-and-synchronization/#comment-28459 I'd argue that the <code>Collections.synchronized*</code> methods are only for dealing with legacy code as well... <code>ConcurrentHashMap</code> is a much more scalable solution. I’d argue that the Collections.synchronized* methods are only for dealing with legacy code as well… ConcurrentHashMap is a much more scalable solution.

]]>