Author: markt
Date: Wed Aug 12 17:16:40 2015
New Revision: 1695582
URL: http://svn.apache.org/r1695582
Log:
Fix https://bz.apache.org/bugzilla/show_bug.cgi?id=58230
I can't see a reason for this to depend on available(). The dispatch should
only happen if there is data to read.
I can see good reasons for this not to depend on available(). The
AsyncStateMachine has different behaviour on some state changes if it thinks
async I/O is being used. With this test in place it was possible for some of
those actions to be skipped.
Modified:
tomcat/trunk/java/org/apache/coyote/http11/Http11Processor.java
Modified: tomcat/trunk/java/org/apache/coyote/http11/Http11Processor.java
URL:
http://svn.apache.org/viewvc/tomcat/trunk/java/org/apache/coyote/http11/Http11Processor.java?rev=1695582&r1=1695581&r2=1695582&view=diff
==============================================================================
--- tomcat/trunk/java/org/apache/coyote/http11/Http11Processor.java (original)
+++ tomcat/trunk/java/org/apache/coyote/http11/Http11Processor.java Wed Aug 12
17:16:40 2015
@@ -1723,9 +1723,7 @@ public class Http11Processor extends Abs
}
} else if (status == SocketStatus.OPEN_READ &&
request.getReadListener() != null) {
try {
- if (inputBuffer.available() > 0) {
- asyncStateMachine.asyncOperation();
- }
+ asyncStateMachine.asyncOperation();
} catch (IllegalStateException x) {
// ISE - Request/Response not in correct state for async read
if (log.isDebugEnabled()) {
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
