Hello Gene!

Just a lurker at this list, but what you've pointer out is very
impressive! :-(

How can this be fixed?
One may be bloody enough to write his own utility class
all like java.util.Date but with its own staticCal being
not static. Will it actually waste much memory?

Another solution would be to use some kind of pool for
these calendars, but this will require synchroniztion.

Thoughts?

BTW: if you're running on Sun's jvm you did not have to jad:
in their jdk distribution there's src.zip or src.jar that has
most of the sources :)

GC> Please bear with me, this may not seem like a ejb/j2ee question
GC> at first, but is very relevent to application container development.

GC> We are doing thread dump performance analyses of our production
GC> application servers and find the majority of our threads (say, 40
GC> out of 50) are in wait states, waiting on the constructor or an
GC> accessor (getYear, getMonth, etc..) of java.util.Date.

GC> After jad'ing the class, I discover all these methods and constructors
GC> are synchronized on a static member variable, Calendar staticCal.

GC> I am sure a lot of people use java.util.Date as a field in their data
GC> beans, or invoke JDBC ResultSet.getTimestamp() which eventually constructs
GC> a java.util.Date, and have seen this problem before, where in a multi-threaded
GC> environment the majority of their threads are blocked on Date.
GC>  Is there a way around it, short of spinning our own Date class?  Thanks.

GC> Gene


- Anton
[EMAIL PROTECTED]
[EMAIL PROTECTED]

===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff EJB-INTEREST".  For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".

Reply via email to