DO NOT REPLY [Bug 14059] - HTP Response does not respond correctly to flush()
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT http://nagoya.apache.org/bugzilla/show_bug.cgi?id=14059. ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND INSERTED IN THE BUG DATABASE. http://nagoya.apache.org/bugzilla/show_bug.cgi?id=14059 HTP Response does not respond correctly to flush() --- Additional Comments From [EMAIL PROTECTED] 2002-11-05 16:03 --- Remy, In Summary: There is a difference in behavior of when TC4.0.x and Tc4.1.12 sends the end of chunk message for a response. TC4.0.x sends an end of chunk immediately following the stream close() in a servlet response. TC4.1.12 does not send the end of chunk message until the servlet exits its DoPost method. On the client side, the sun java.net implementation does not return a stream from the getInputStream() method until it receives the end of chunk message on the wire. Hence, my issues. It looks like the browsers have access to intermediate chunks. Which is why you don't see any issue. In conclusion, TC behavior HAS changed; Is it a bug? Depends on who you ask... I think the behavior change should be at least clearly documented on the TC docs if it is NOT determined to be a bug. However, I do not think the close() is being handled correctly. It is not being handled in the manner it has been in previous TC releases. I would like to hear your view on this issue. Thanks, -j.p. -- To unsubscribe, e-mail: mailto:tomcat-dev-unsubscribe;jakarta.apache.org For additional commands, e-mail: mailto:tomcat-dev-help;jakarta.apache.org
DO NOT REPLY [Bug 14059] - HTP Response does not respond correctly to flush()
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT http://nagoya.apache.org/bugzilla/show_bug.cgi?id=14059. ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND INSERTED IN THE BUG DATABASE. http://nagoya.apache.org/bugzilla/show_bug.cgi?id=14059 HTP Response does not respond correctly to flush() [EMAIL PROTECTED] changed: What|Removed |Added Status|REOPENED|RESOLVED Resolution||WORKSFORME --- Additional Comments From [EMAIL PROTECTED] 2002-11-04 21:14 --- Testing with IE and your example does work fine. You should try to see what is handled differently which causes the different result. I don't think there's any bug here ... -- To unsubscribe, e-mail: mailto:tomcat-dev-unsubscribe;jakarta.apache.org For additional commands, e-mail: mailto:tomcat-dev-help;jakarta.apache.org
DO NOT REPLY [Bug 14059] - HTP Response does not respond correctly to flush()
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT http://nagoya.apache.org/bugzilla/show_bug.cgi?id=14059. ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND INSERTED IN THE BUG DATABASE. http://nagoya.apache.org/bugzilla/show_bug.cgi?id=14059 HTP Response does not respond correctly to flush() --- Additional Comments From [EMAIL PROTECTED] 2002-11-04 23:50 --- Remy, I'm open to looking at the Server code to see what is different. So the client code I am executing does not receive the first character in TC4.1.12 until the sleep() in the server has expired. As I'm thinking now, I will try IE and Mozilla again and trace the conversation on the wire to see if IE sends the request with different HTTP parameters. Something is different... Thanks for your time. -j.p. -- To unsubscribe, e-mail: mailto:tomcat-dev-unsubscribe;jakarta.apache.org For additional commands, e-mail: mailto:tomcat-dev-help;jakarta.apache.org
DO NOT REPLY [Bug 14059] - HTP Response does not respond correctly to flush()
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT http://nagoya.apache.org/bugzilla/show_bug.cgi?id=14059. ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND INSERTED IN THE BUG DATABASE. http://nagoya.apache.org/bugzilla/show_bug.cgi?id=14059 HTP Response does not respond correctly to flush() --- Additional Comments From [EMAIL PROTECTED] 2002-11-05 06:53 --- If you have doubts, try with a telnet. It will display very clearly which bytes are returned, and at what time. GET /myservlet HTTP/1.1 Host: localhost:8080 -- To unsubscribe, e-mail: mailto:tomcat-dev-unsubscribe;jakarta.apache.org For additional commands, e-mail: mailto:tomcat-dev-help;jakarta.apache.org
DO NOT REPLY [Bug 14059] - HTP Response does not respond correctly to flush()
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
http://nagoya.apache.org/bugzilla/show_bug.cgi?id=14059.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND
INSERTED IN THE BUG DATABASE.
http://nagoya.apache.org/bugzilla/show_bug.cgi?id=14059
HTP Response does not respond correctly to flush()
[EMAIL PROTECTED] changed:
What|Removed |Added
Status|RESOLVED|REOPENED
Resolution|INVALID |
--- Additional Comments From [EMAIL PROTECTED] 2002-11-01 12:49 ---
I have gone back and created a java client that reads the inputStream response a
char at a time:
logger.debug( opening buffered reader to read response );
InputStreamReader is = new InputStreamReader(
URLConnection.getInputStream() );
// Note: should read response to make sure no error.
String str;
int i;
StringBuffer sb = new StringBuffer();
while ( ( i = is.read() ) != -1 ) {
char c = (char)i;
logger.debug( Have a char: + c );
sb.append( c );
}
String s = sb.toString();
logger.debug( String returned is: + s );
if ( !s.startsWith( OK ) ) {
throw new RuntimeException( Server returned: + s );
}
is.close();
}
I pointed it at the Servlet test code on Tc4.0.6
Results:
2002-11-01 12:41:16,884 [GMT] DEBUG 10TestTestService [main] - Connecting
to
http://mocha.ko.kodak.com:8080/eFUSE/servlet/controller/Test?action=ReadRequestreadQueue=TestingMe
2002-11-01 12:41:16,934 [GMT] DEBUG 60TestTestService [main] - opening
buffered reader to read response
2002-11-01 12:41:16,944 [GMT] DEBUG 70TestTestService [main] - Have a char:O
2002-11-01 12:41:16,944 [GMT] DEBUG 70TestTestService [main] - Have a char:K
2002-11-01 12:41:16,954 [GMT] DEBUG 80TestTestService [main] - Have a char:
2002-11-01 12:41:16,954 [GMT] DEBUG 80TestTestService [main] - Have a char:
2002-11-01 12:41:16,954 [GMT] DEBUG 80TestTestService [main] - String
returned is: OK
TC 4.1.12 Results:
2002-11-01 12:38:48,050 [GMT] DEBUG 10TestTestService [main] - Connecting
to
http://arabica.ko.kodak.com:8080/eFUSE/servlet/controller/Test?action=ReadRequestreadQueue=TestingMe
2002-11-01 12:38:48,100 [GMT] DEBUG 60TestTestService [main] - opening
buffered reader to read response
2002-11-01 12:38:54,119 [GMT] DEBUG 6079 TestTestService [main] - Have a char:O
2002-11-01 12:38:54,119 [GMT] DEBUG 6079 TestTestService [main] - Have a char:K
2002-11-01 12:38:54,119 [GMT] DEBUG 6079 TestTestService [main] - Have a char:
2002-11-01 12:38:54,119 [GMT] DEBUG 6079 TestTestService [main] - Have a char:
2002-11-01 12:38:54,119 [GMT] DEBUG 6079 TestTestService [main] - String
returned is: OK
Note the times. Each TC instance was deployed out of the box, however is
there a directive in the server.xml that would account for this behavior difference?
Thank you for your time.
-j.p.
--
To unsubscribe, e-mail: mailto:tomcat-dev-unsubscribe;jakarta.apache.org
For additional commands, e-mail: mailto:tomcat-dev-help;jakarta.apache.org
DO NOT REPLY [Bug 14059] - HTP Response does not respond correctly to flush()
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT http://nagoya.apache.org/bugzilla/show_bug.cgi?id=14059. ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND INSERTED IN THE BUG DATABASE. http://nagoya.apache.org/bugzilla/show_bug.cgi?id=14059 HTP Response does not respond correctly to flush() --- Additional Comments From [EMAIL PROTECTED] 2002-10-31 18:36 --- I went back and checked the response times and behavior of the browser test. On 4.1.12 It looks like the OK data string comes back almost immediately, but the session does not release for a consistant 6.012 seconds - every time. On 4.0.6 Again the OK data string comes back immediately, the session release times are all over the map. .1 seconds to 5.01 seconds. But never the 6 seconds to match the sleep on the server. I also looked at the protocol n the wire and hereis what I see: 4.1.12 GET ... HTTP/1.1 200 OK Continuation Continuation (data frame contains OK) 4.0.6 GET ... HTTP/1.1 200 OK Continuation (data frame contains OK) Continuation It appears that the OK in the continuation frame indicates when the server actually closes the stream. if that is true, the TC4.1.12 close() is the operation that is not really occuring at the lines in the code. I am going to try reading the response in my client without buffering, but I'm not sure what will happen if I close the client stream, when the server still has it open. -j.p. -- To unsubscribe, e-mail: mailto:tomcat-dev-unsubscribe;jakarta.apache.org For additional commands, e-mail: mailto:tomcat-dev-help;jakarta.apache.org
DO NOT REPLY [Bug 14059] - HTP Response does not respond correctly to flush()
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT http://nagoya.apache.org/bugzilla/show_bug.cgi?id=14059. ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND INSERTED IN THE BUG DATABASE. http://nagoya.apache.org/bugzilla/show_bug.cgi?id=14059 HTP Response does not respond correctly to flush() [EMAIL PROTECTED] changed: What|Removed |Added Status|NEW |RESOLVED Resolution||INVALID --- Additional Comments From [EMAIL PROTECTED] 2002-10-30 08:28 --- The data flushed is immediately sent back to the client (and displayed in the client browser; IE 6 in my case). I don't know what's wrong with your testing, but it definitely works for me. The writer is also closed for the servlet perspective. However, the request processing is only actually terminated when the servlet returns. -- To unsubscribe, e-mail: mailto:tomcat-dev-unsubscribe;jakarta.apache.org For additional commands, e-mail: mailto:tomcat-dev-help;jakarta.apache.org
DO NOT REPLY [Bug 14059] - HTP Response does not respond correctly to flush()
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT http://nagoya.apache.org/bugzilla/show_bug.cgi?id=14059. ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND INSERTED IN THE BUG DATABASE. http://nagoya.apache.org/bugzilla/show_bug.cgi?id=14059 HTP Response does not respond correctly to flush() --- Additional Comments From [EMAIL PROTECTED] 2002-10-29 19:08 --- I am almost certain this is correctly implemented. I'll make sure of that with a telnet. -- To unsubscribe, e-mail: mailto:tomcat-dev-unsubscribe;jakarta.apache.org For additional commands, e-mail: mailto:tomcat-dev-help;jakarta.apache.org
