Author: kkolinko
Date: Fri Aug 14 08:57:18 2009
New Revision: 804132
URL: http://svn.apache.org/viewvc?rev=804132&view=rev
Log:
vote and comment
Modified:
tomcat/tc6.0.x/trunk/STATUS.txt
Modified: tomcat/tc6.0.x/trunk/STATUS.txt
URL:
http://svn.apache.org/viewvc/tomcat/tc6.0.x/trunk/STATUS.txt?rev=804132&r1=804131&r2=804132&view=diff
==============================================================================
--- tomcat/tc6.0.x/trunk/STATUS.txt (original)
+++ tomcat/tc6.0.x/trunk/STATUS.txt Fri Aug 14 08:57:18 2009
@@ -153,8 +153,31 @@
* Improve NIO connector shutdown time by doing shutdowns in parallel
and with a timeout
http://svn.apache.org/viewvc?view=rev&revision=791914
- +1: fhanik, markt
+ +1: fhanik, markt, kkolinko
-1:
+ kkolinko: (
+ Some nit-picking:
+ 1. I think that it is better to create the stopLatch in the
NioEndpoint#start()
+ method right before the array of Pollers is created, instead of
NioEndpoint#init().
+ And to pass a reference to it to the Poller (as an argument in Poller
+ constructor).
+ That is because otherwise if you call stop() start() the stopLatch won't
be in
+ its initial condition, and because the count of pollers can be changed
between
+ init() and start(). Just a theory, though.
+
+ 2. In NioEndpoint.Poller#run():
+ - if the Poller was looping with paused=true, is there a need to call
+ events() before exiting the loop on if (close) ?
+ - there are several ways to exit the loop: two if(close), one while
+ (running), and may be exceptions. The first if(close) does not call
+ selector.close(). And if you exit because of running = false there is
+ no call to timeout(0, false); and selector.close().
+ - there is special processing code for OutOfMemoryError, but inside the
+ loop, and in events() method, you catch any Throwable.
+ - it would be better to put stopLatch.countDown() inside a finally{} block
+ 3. In NioEndpoint.Poller#destroy() there is comment ("// Wait...").
+ Is it still relevant? I do not see any waiting there.
+ )
* Fix https://issues.apache.org/bugzilla/show_bug.cgi?id=47515
https://issues.apache.org/bugzilla/attachment.cgi?id=23973
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
