forked from tectronics/JSDT
-
Notifications
You must be signed in to change notification settings - Fork 1
/
ChangeLog-1999
3281 lines (2373 loc) · 122 KB
/
ChangeLog-1999
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
/*
* Copyright (c) 1996-2005 Sun Microsystems, Inc.
* All Rights Reserved.
*
*/
JSDT Change History - 1999.
===========================
v2.1 - 20th December 1999.
* From: Scott Swann <[email protected]>
"Problem with the "multicast" implementation. When using multiple channels
and a proxy dies, all the proxies leave there channel. This also holds
true for tokens."
Created a new removedId() method in SessionServer.java. This now gets called
from the checkConnections() method in the CleanupConnections thread. When a
proxy is no longer pinging, it's clients are now cleaned up by that pinging
id, not by the multicast thread that it was using.
----
v2.1 - 15th December 1999.
* From: #CHEONG CHEE KIN# <[email protected]>
The Browser.html file should have been using:
"code=examples.browser.WebBrowser.class" and not:
"code=examples.browser.Student.class".
* From: Scott Swann <[email protected]>
"I have a server with 1 session and 2 channels. I start 1 proxy for each
channel. The first proxy connects and receives data without any problem,
but when the second proxy connects it begins throwing exceptions and
attempts to find the other channel."
In SessionProxyMessage.java, if the Channel with the given name cannot be
found, rather than put out an error message, it should silently ignore that
message (which wasn't for that proxy in the first place).
Similar changes needed for ByteArrays and Tokens too.
----
v2.1 - 15th December 1999.
* From: Scott Swann <[email protected]>
Getting packets delivered out of order when running the "multicast"
implementation at high data rates. Also getting an IndexOutOfBoundException
in the getData() method in .../impl/JSDTThread.java, because the
dataIn.read() call returned -1, which wasn't being checked for.
Adjusted the processEvent() method in the "multicast" implementation to
display an error message if showMessage/showStack are set true and an
LrmpEventHandler.UNRECOVERABLE_SEQUENCE_ERROR has occured. It will also
call the connectionFailed() method for any Connection listeners associated
with that JSDT application.
Adjusted the getData() method in .../impl/JSDTThread.java to check for a
return of -1 from the dataIn.read() call, and to throw an IOException if
found. This in turn will generate a ConnectionException in the calling method.
Added the following to the LrmpProfile associated with each LRMP connection:
lrmpProfile.setQoS(lrmpProfile.NoLoss, true, lrmpProfile.AdaptedThroughput);
Adjusted the stopMulticastSession() method to now take a boolean forceStop
parameter. This is true for the stopMulticastSession() in the processEvent()
method, and false otherwise.
Adjusted the sendWindowSize and rcvWindowSize for the LRMP profile to be
512 (the default is 64).
----
v2.1 - 9th December 1999.
* From: Pete Boysen <[email protected]>
Improvements to the phone example, to reduce the amount of "silent" data
send by each phone user. The algorithm checks each buffer for any bytes
above a certain threshold (BYTE_THRESHOLD) and sends the buffer if there
are any. Also, it only clamps the output if there are BUFFER_THRESHOLD or
more buffers of continuous silence.
This still results in transmitting 7.5-10K/sec which wouldn't work with
a 28.8 line. By using a Deflater during transmission and an Inflater
during reception I found that the sound portions were being sent at
2.8-3.2K with a lot of silence in between which should be acceptable.
Changes to .../phone/PhoneUser.java and .../phone/SoundPacket.java
----
v2.1 - 8th December 1999.
* Needed to adjust the writeMessageHeader() calls for the
Client/DestroyClient and Client/Invite message changes from yesterday,
to now wait for a reply. This now allows test3 on the Test Suite to work
with the "http" implementation for the HttpToPortSocketFactory.
----
v2.1 - 7th December 1999.
* From: Seppo Tarkiainen <[email protected]>
Registry manager authentication wasn't working (generated a null pointer
exception). The authThread in RegistryClient.java was being resumed before
it had ever been suspended.
Rewrote RegistryClient.java to use wait() and notifyAll() and added a
couple more states to its internal state machine.
* Test10 doesn't work with HTTP sockets:
port in use
at com.sun.media.jsdt.http.TCPSocketServer.<init>(TCPSocketServer.java:81)
at com.sun.media.jsdt.http.ClientServer.<init>(ClientServer.java:71)
at com.sun.media.jsdt.http.httpClient._createServer(httpClient.java:249)
at com.sun.media.jsdt.http.NamingProxy.bind(NamingProxy.java:414)
at com.sun.media.jsdt.Naming.bind(Naming.java:145)
at com.sun.media.jsdt.ClientFactory.createClient(ClientFactory.java:119)
at Test10.<init>(Test10.java:100)
at Test10.main(Test10.java:254)
This was caused because no response was sent to the HTTP request for the
Client/DestroyClient message in ClientServer.java. There was a similar
problem for the Client/Invite message in the same class.
----
v2.1 - 30th November 1999.
* From: Scott Swann <[email protected]>
Adjusted the removeConnectionListener() method in ConnectionThread.java,
for all three implementations, to throw a NoSuchListenerException if the
listener is null.
* From: Seppo Tarkiainen <[email protected]>
With .../JSDT-2.0/Test10, the Registry manager wasn't being called to
authenticate Session/Client creates and destroys.
Have revamped the startregistry() methods in RegistryFactory.java to
call a package protected startRegistry() method to do the actual registry
startup. This means the getRegistry() is only called once, and the manager
is attached to the proper registry instance.
* From: Scott Swann <[email protected]>
Running .../JSDT-2.0/Test10 generates the following exception:
NamingProxyMessage: handleMessage: null
java.io.EOFException
at java.io.DataInputStream.readUnsignedByte(DataInputStream.java)
at java.io.DataInputStream.readUTF(DataInputStream.java)
at java.io.DataInputStream.readUTF(DataInputStream.java)
at com.sun.media.jsdt.multicast.NamingProxyMessage.handleMessage(
NamingProxyMessage.java:97)
at com.sun.media.jsdt.multicast.NamingProxyThread.handleMessage(
NamingProxyThread.java:107)
at com.sun.media.jsdt.multicast.multicastThread.run(
multicastThread.java:563)
at java.lang.Thread.run(Thread.java)
Added a check in the handleMessage() method in NamingProxyMessage.java for
the multicast implementation, to see if the length of the resourceName
field was zero. If so, then no more fields are read, and a message in
not sent on to the Registry listeners.
----
v2.1 - 29th November 1999.
* Adjusted the versionString variable in JSDTObject so that it's no longer
final. This means that the version string will no longer be hard-wired
into any class files when it's compiled.
* From Scott Swann <[email protected]>
Adjusted the dataReceived() method in ChannelProxy.java (for all three
implementations), to first check if drq is null, before attempting to
add this message to the incoming queue. It's possible that the connection
might have been lost, and drq set to null.
----
v2.1 - 24th November 1999.
* From Fulvio Biondi <[email protected]>
Start the Registry as a separate application.
Start the server application with a Registry Listener.
Kill the Registry, and try to restart it as a thread inside the server
when the server's connectionFailed(RegistryEvent re) method gets called.
It hangs.
Three problems here:
1/ For the "socket" implementation, a check needed to be added to the
removeThread() method in SessionServer.java to see if the socket server
thread instance was not null.
2/ For the "socket" implementation, a check needed to be added to the
removeThread() method in SessionServer.java to see if the
SessionServerThread was an instance of TCPSocketThread before trying
to close the socket.
3/ The RegistryFactory class no longer keeps a static hashtable of
registry entries, but instead creates and destroys them as needed.
All the methods in RegistryFactory that call the getRegistry() method
have been adjusted to set the registry variable to null, so that it
will be garbage collected.
----
v2.1 - 22nd November 1999.
* From [email protected]
Reworked the way that threads are terminated when cleaning up
SessionServerThreads. No more calls to Thread.stop(). It now uses a
combination of setting a running boolean to false and interrupting the
thread to be terminated. It also closes the socket associated with that
SessionServerThread.
Changes applied to the "socket" implementation.
* Added a section to the introductory chapter in the Implementers Kit to
mention the two new methods that need to be supported by alternate socket
factory implementations.
* Moved several common methods from socket/SocketThread.java,
http/HttpThread.java and multicast/multicastThread.java into
impl/JSDTThread.java.
* For all three implementations, moved the create methods for ByteArrays,
Channels and Tokens with managers into .../impl/SessionImpl.java, and
adjusted the AbstractSessionProxy and SessionProxy classes accordingly.
Adjusted the template package for these changes, plus the Implementers
Guide.
----
v2.1 - 17th November 1999.
* From [email protected]
When all the Sessions associated with a SessionServerThread are destroyed,
that SessionServerThread was not being terminated.
Changes applied to the "socket" implementation.
See test program in /export/space/richb/jsdt/test/JSDT-2.1/Test5
* Removed the "synchronized" keyword from the finishReply() and
terminate() methods in .../impl/JSDTThread.java
* From Yanmu Zhou <[email protected]>
If a Channel has a manager, but not the Session within which it was created,
then this generates a NullPointerException when doing Client authentication.
The problem was due to always trying to send the Client/Authenticate message
over the manager thread associated with the Session. This has been adjusted
to now send the message over the manager thread associated with the managed
object.
Changes applied to all three implementations.
See test program in /export/space/richb/jsdt/test/JSDT-2.1/Test6
* Fixed up the synth example to work with the latest JavaSound sound.jar.
Changed code section in the initialize() method in MidiSynth.java to:
try {
synth = MidiSystem.getSynthesizer();
} catch (MidiUnavailableException mue) {
System.out.println("Could not get default synthesizer");
System.exit(1);
}
* Updated .../lib/sound/sound.jar with the latest version from the JavaSound
nightly builds.
----
v2.1 - 16th November 1999.
* From: Gregory Gray <[email protected]>
"Try running the chat example from 2.0 and killing the server process and
then bring it back up again. The first time the chat user program tries
to send a message, it should see that it is not connected (exception)
and reconnect. The next time it should send the message just fine.
Unfortunately this doesn't happen."
Needed to add a check to sendMulticastPacket in LRMPImpl to check if the
thread variable was null, in which case we should just return.
Also needed to set the static drq (data received queue) variable in
ChannelProxy.java to null, when cleanup up the DataReceivedThreads, and
recreate it when a new ChannelProxy (ie. when the user reconnects) is
created.
Similar changes applied to the "socket" and "http" implementations.
* With the "multicast" implementation, doing a "SignOff" with the chat
example, wasn't removing references to that user in the server. The
problem was that a call to proxyThread.terminate() in SessionProxy.java
was now occuring too early. This was moved to after the Session/Close
message had been sent to the server.
Similar changes applied to the "socket" and "http" implementations.
* From: Scott Swann <[email protected]>
The "multicast" implementation sometimes output the following error message
and stack:
NamingProxyMessage: handleMessage: null
java.io.EOFException
at java.io.DataInputStream.readUnsignedByte(DataInputStream.java:247)
at java.io.DataInputStream.readUTF(DataInputStream.java:526)
at java.io.DataInputStream.readUTF(DataInputStream.java:498)
at com.sun.media.jsdt.multicast.NamingProxyMessage.handleMessage(
NamingProxyMessage.java:95)
at com.sun.media.jsdt.multicast.NamingProxyThread.handleMessage(
NamingProxyThread.java:107)
at com.sun.media.jsdt.multicast.multicastThread.run(
multicastThread.java:820)
at java.lang.Thread.run(Thread.java:479)
Changed the handleMessage() method in NamingProxyMessage.java to handle
only the Registry/InformListener message types, not all InformListener
messages.
* The ConnectionThread and the NamingProxy thread are now only terminated
if there are no Session or Clients associated with that proxy, *AND* if
there are no connection listeners.
----
v2.1 - 15th November 1999.
* Moved the Message class for the three implementations to the impl package.
Adjust the three implementations to use this.
* Added an abstract JSDTThread class to the impl package. Adjusted the three
implementations to use this.
----
v2.1 - 12th November 1999.
* From: Scott Swann <[email protected]>
Got a connection failure with the latest version of the "multicast"
implementation. Two problems:
> One of the places where getPacket() was called in the getSocketMessage()
method in multicastThread.java was not then getting the new buffer from
that packet.
> Several memory leaks related to packets associated with messages that
had to be gragmented being left in hashtables.
See test program in /export/space/richb/jsdt/test/JSDT-2.1/Test4
* Adjusted all instances of Class.forName() throughout the JSDT code, to
instead use the new getClassForName() method in .../impl/Util.java. This
was done because it's possible that the Class.forName() call inside this
method will be replaced by:
Thread.currentThread().getContextClassLoader().loadClass(className);
at some later point in time, and this simplifies things. See entry from
Ted Neward in the TODO file on this.
* From [email protected]
Added a:
URLString Session.getURL();
method to Session.java, which returns the URLString used to create this
Session.
----
v2.1 - 10th November 1999.
* From: Marko Klopcic <[email protected]>
The JSDTSocketFactory doesn't allow you to create the unreliable
datagram sockets.
The factory has been adjusted to include two new methods:
/**
* <A NAME="SD_CREATEDATAGRAMSOCKET"></A>
* <EM>createDatagramSocket</EM> returns a socket for sending and receiving
* datagram packets, bound to any available port on the local host machine.
* This socket is configured using the socket options established for this
* factory.
*
* @exception SocketException if the socket could not be opened, or the socket
* could not bind to the specified local port.
*
* @see java.net.DatagramSocket
*
* @return a DatagramSocket on any available local port.
*/
DatagramSocket
createDatagramSocket() throws SocketException;
/**
* <A NAME="SD_CREATEDATAGRAMSOCKET"></A>
* <EM>createDatagramSocket</EM> returns a socket for sending and receiving
* datagram packets, at the given local port. This socket is configured using
* the socket options established for this factory.
*
* @param port the local port to use
*
* @exception SocketException if the socket could not be opened, or the socket
* could not bind to the specified local port.
*
* @see java.net.DatagramSocket
*
* @return a DatagramSocket on the given local port.
*/
DatagramSocket
createDatagramSocket(int port) throws SocketException;
The "socket" implementation code has been adjusted to create datagram
sockets through the provided factory.
Note that you can't use unreliable Channels with the SSLSocketFactory or
with any of the socket factory provided for the "http" implementation.
----
v2.1 - 9th November 1999.
* From: Craig Minton <[email protected]>
For the "socket" implementation, doing a RegistryFactory.stopRegistry()
wasn't terminating the RegistryServerThread and the NamingProxyThread
threads.
Similar fixes were applied to the other two implementations.
* From: Craig Minton <[email protected]>
The DataReceivedThread is used by the Channel proxy to handle incoming Data
sent over the Channel. A new one is created each time a new Channel is
created, upto "maxThreadPoolSize" of them (default is 5).
The code has been adjusted so that if the proxy application no longer has
any references to Sessions (ie. it's destroyed them all), then any
DataReceivedThreads that have been started, will be terminated.
This fix has been applied to all three implementations.
* From: Craig Minton <[email protected]>
The NamingProxy and ConnectionThread threads are two more proxy-side
threads. The former is used to listen for any Session or Client changes
happening in the Registry, and pass them onto any potential Registry
listeners. The latter is used to determine whether there has been any
kind of failure in the underlying connection.
The code has been adjusted so that if, the proxy application, when it's
tidying up after destroying a Session, no longer has any references to
any Sessions or special Clients, then these two threads are terminated.
This fix has been applied to all three implementations.
* From: Craig Minton <[email protected]>
TCPSocketServerThread is a server-side thread that has the server socket
running, and which accepts new connections for that particular port.
The code has been adjusted so that if there are no longer any server-side
sessions left, then this thread is terminated.
See test program in /export/space/richb/jsdt/test/JSDT-2.1/Test3
----
v2.1 - 8th November 1999.
* Removed the creation of the three client-side .jar files from the build.csh
script.
* Adjusted all references to LRMP to indicate it's now one of the possible
reliable multicast frameworks that can be used by the new "multicast"
implementation.
----
v2.1 - 5th November 1999.
* Got the "multicast" implementation of JSDT working with the RMF/RAMP RM
packages from TASC (downloadable from: http:https://www.tascnets.com/mist/RMF/)
Note that to prevent RMF/RAMP using 100% CPU, then you need to change the
RMF/Timer.Yield() method to look like:
public static void Yield() {
final String mn = "Yield";
try {
Thread.sleep(20);
} catch (InterruptedException ie) {
log(mn, "Interrupted exception.");
}
}
Added an rmf.jar file under the .../lib/rmf directory, and adjusted the
build environment to build against that.
* Optimised the LRMP implementation class so that only one LrmpImpl object
(and therefore thread) is created (and started). The packet handler
associated with that lrmp object now places a copy of each incoming LRMP
multicast packet into the vector of packets for all the interested threads.
This is a significant performance improvement (particularly on the
server-side).
----
v2.1 - 3rd November 1999.
* From [email protected]
JSDT has memory leaks. I have a simple test application that creates a
managed Session (with a Channel in it), and then destroys it. This
operation is repeated multiple times.
The problem was that there was a couple of threads hanging onto a reference
to the Session object that should have been garbage collected. I needed to
terminate the run() method for those threads, and then everything freed
itself. The slightly tricky (icky really) thing here was that these threads
were in a wait() state, waiting for the next incoming message, so I had to
do a notifyAll() to wake them up, and get them to check a boolean flag to
determine that they'd been terminated.
Changes applied to all three implementations.
See test program in /export/space/richb/jsdt/test/JSDT-2.1/Test2
* From Benjamin D. Engelsma <[email protected]>
It looks like the source for the synth example was left out of the JSDT
2.0 (FCS) distribution.
----
v2.1 - 2nd November 1999.
* The "send to others" case was not setting up the list of receiver names
correctly in the T_DataReceived message, in the send() method in
ChannelServer.java, for the "multicast" implementation.
* The checkForServer() method in NamingProxy.java, for all three
implementations needed to also check if the name in the Registry was a
Session URL (as well as having equal address and port).
----
v2.1 - 1st November 1999.
* Created an RMFImpl.java class for an RMF reliable multicast framework.
RMF is available from: http:https://www.tascnets.com/mist/RMF/index.html
Still needs some work.
* Set the "multicast" Registry version number to 2.1.
* For the "multicast" implementation, optimised the ChannelServer.send()
method to only send out one dataReceived message to the Channel proxies,
for each call to Channel.sendToAll() and Channel.sendToOthers().
This message now contains a list of the client names that should receive
this message. The code on the proxy-side has been adjusted to properly
handle this.
* The two calls to NamingProxy.cleanupSession() needs to be adjusted to pass
in a Session URL string which had been adjusted to contain the IP address
of the server (rather than the host). Change made to all three
implementations.
----
v2.1 - 28th October 1999.
* Added a "multicast" implementation, which replaces the "lrmp" implementation.
This uses a new "multicastImplClass" Connection property to determine the
reliable multicast implementation that should be used. There is currently
only one RM impl, and that's LRMP, so multicastImplClass is set to:
"com.sun.media.jsdt.multicast.LRMPImpl".
----
v2.1 - 27th October 1999.
* From: Joshua Fox <[email protected]>
The following generates a "port in use" exception:
> Create a Session, say a ChatSession, on a given port.
> From another java process, create a different Session (different
name, same port, same physical machine, different virtual machine).
This is no longer a problem. Changes applied to all three implementations.
See test program in /export/space/richb/jsdt/test/JSDT-2.1/Test1
* Started work on the JSDT 2.1 distribution:
- Copied the master JSDT 2.0 workspace and removed files that are no longer
needed.
- Made the various changes outlined in .../impl/BUILD.
- Changed the versionString variable in JSDTObject to "JSDT 2.1 (EA)".
---- Work finished on the JSDT 2.0 (FCS) release ----
v2.0 - 27th October 1999.
* Adjusted the install/solaris/script.txt file to properly create the .html
files for the Browser, Phone and Synth examples for the Solaris dist.
* Changed the versionString variable in JSDTObject to "JSDT 2.0 (FCS)".
----
v2.0 - 26th October 1999.
* For the http implementation, needed to start a ping thread in
SameVMSessionProxyThread and SameVMManagerProxyThread, if mustPing()
returned true.
----
v2.0 - 19th October 1999.
* From [email protected]
Included a better description of the httpURIName property in the JSDT
User Guide.
----
v2.0 - 8th October 1999.
* Updated the Release Notes to include in all the changes between beta and FCS.
----
v2.0 - 7th October 1999.
* Removed the transient keyword from the InputStream and OutputStream variables
in TCPSocketThread.java for the socket implementation. This was not the fix
for the RMI serialization problem.
* Added a waitForPermanentConnection() call in the authenticateClient()
method in NamingProxy.java for the http implementation, to be consistent
with the similar code in the authenticateClient() method in the
SessionProxy class.
* Needed a call to thread.finishMessage() in the authenticate() method in
RegistryClient.java for the http implementation, to correctly close the
socket connection when run with the HttpToPortSocketFactory class.
----
v2.0 - 6th October 1999.
* Adjusted the "http" implementation to (hopefully) properly close the
socket connection (both on the proxy and the server sides), after each
request/response, if the socket is not reusable.
* From [email protected]
Added in a new Connection property (httpURIName) for setting the relative
URI name associated with the HttpToPortSocketFactory class with the "http"
implementation of JSDT.
Updated the JSDT User Guide to reflect this change.
----
v2.0 - 4th October 1999.
* Adjusted the phone and synth examples to work with what will be the JavaSound
API for JDK 1.3 (FCS).
Added a .../lib/sound/sound.jar file to the JSDT distribution to allow it
to be successfully compiled against the new JavaSound classes.
* From: Vijay Bhandari <[email protected]>
Added the following method to RegistryListener and RegistryAdaptor:
public void
connectionFailed(RegistryEvent event);
For every Registry listener added, this method is called if the connection
to the Registry fails.
Two new convenience methods have been added to RegistryEvent:
public String
getAddress();
public int
getPort();
which return the host address and port number of the failed registry
connection.
The JavaDoc and JSDT User Guide have been updated to reflect these changes.
----
v2.0 - 30th September 1999.
* Synchronized the informListeners() method in the ConnectionThread class.
This means that for the socket implementation, one and only one call is
made to each connection listeners connectionFailed() method, irrespective
of the number of Sessions associated with the failed connection.
----
v2.0 - 29th September 1999.
* From: Vijay Bhandari <[email protected]>
Added a couple of paragraphs to the Connection section of chapter 2 of
the User Guide to describe what to do in the event of a connection failure.
* From: Vijay Bhandari <[email protected]>
For the "socket" implementation, if the server is terminated, the proxy
automatically detects this and cleans up any sessions associated with that
connection. These sessions are removed from a hashtable. Unfortunately this
is the same hashtable that a ConnectionThread uses to ping the servers.
So it's silently failing to generate a connectionFailed() call to any
connection listeners.
Adjusted the cleanupConnections() method in SessionProxyThread to inform
all connection listeners of the connection failure.
----
v2.0 - 28th September 1999.
* From: Vijay Bhandari <[email protected]>
The JavaDoc description of the port parameter for the
URLString.createClientURL() method was incorrect. This has been changed to:
@param <EM>port</EM> the port number the Client is running on. This is
a port on the local machine; not the port number of the Registry.
The ClientFactory section in the JSDT User Guide has been adjusted to better
explain this.
* From: John Ko <[email protected]>
Made the InputStream and OutputStream variables in the TCPSocketThread class
in the com.sun.media.jsdt.socket package transient, so that a Session is
now (hopefully) serializable over RMI.
* From: Vijay Bhandari <[email protected]>
Trying to destroy a special client caused a null pointer exception. Was
unable to reproduce this, but in order to prevent the null pointer exception,
have added a check to see if the proxy thread connection to the Registry
is null. If so, then no message is sent.
* From Ugo Cei <[email protected]>
Applied the same server-side cleanup changes for non-pinging clients,
to the "socket" implementation, as were applied to the "http" and "lrmp"
implementations.
----
v2.0 - 22nd September 1999.
* Changed all occurances of beta to FCS in the various JSDT files.
---- Work finished on the JSDT 2.0 (beta) release ----
v2.0 - 16th September 1999.
* From: Scott Swann <[email protected]>
For the "lrmp" implementation, the handleRequests() method in the Registry
has been adjusted to only process messages if the message type is T_Registry.
* From: Scott Swann <[email protected]>
For the "lrmp" implementation, the handleMessage() in ClientProxyMessage.java
should just silently have returned. There are no messages for the special
Client proxies.
----
v2.0 - 15th September 1999.
* The waitForReply() method (for each of the implementations), needed to be
adjusted to do a ">=" check (rather than just a ">" check) for the
millisecond difference.
----
v2.0 - 14th September 1999.
* Needed to remove the Debug() call in impl.Util.getBooleanProperty() to
provent a stack overflow.
* The previous timeout value wasn't been correctly reset after a
RegistryFactory.registryExists() call for the "lrmp" and "http"
implementations.
----
v2.0 - 9th September 1999.
* The tail line number value in .../install/solaris/script.txt was incorrect,
thus generating a Solaris distribution that wouldn't unpack properly.
* Added a check for "session != null" in the removeSession() method in
SessionProxy.java for all three implementations, to prevent a null pointer
exception.
* Adjusted the way that the sendMessage() method in ConnectionThread.java
works for each of the three implementations. If the proxy sends an IsAlive
message and gets a TimedOutException, a flag is now set. A comparison
is made between the current time and the last time it received a reply for
a ping message. If this is greater than the cleanup period, then a
connection failed event is sent, and the connectFailed() method gets called.
----
v2.0 - 8th September 1999.
* Removed the call to waitForReplyNoTimeout() from the join() method in
ManageableProxy.java for the "socket" and "lrmp" implementations, and
removed this method from the base thread class for all three impls.
This now totally reverts the authentication code back to what was in
JSDT 1.4.
----
v2.0 - 6th September 1999.
* For the "socket" and "lrmp" implementations, the JSDT 1.4 style
authentication code has been reinstated, which hopefully means
the "unexpected message type" problem has been fixed.
For those two implementations, it does reintroduce the following problem,
which has been added back to the TODO file:
* From: Ugo Cei <[email protected]>
When a Client does an operation that requires authentication, it's
authenticate() method is called. If that method did something like pop
up a dialog box asking the user for a password, then this would block
another Client from doing a similar thing until the first Client had
entered a password and dismissed that popup.
The "http" implementation should not have this problem.
----
v2.0 - 3rd September 1999.
* The following methods in the Impl classes needed to catch and throw
TimedOutException, rather than letting it get caught by JSDTException
clause:
ByteArrayImpl: addByteArrayListener()
ByteArrayImpl: removeByteArrayListener()
ChannelImpl: addChannelListener()
ChannelImpl: removeChannelListener()
SessionImpl: addSessionListener()
SessionImpl: removeSessionListener()
TokenImpl: addTokenListener()
TokenImpl: removeTokenListener()
----
v2.0 - 2nd September 1999.
* From Edwin Mol <[email protected]>
"Can still get the error : unknown type:
Message: session #: 1 id: -910417444 type: Client action: Authenticate
even with JSDT 2.0 (EA1)."
The problem was that the server was sending a reply to the managers
Client/Authenticate message before that manager had had a chance to call
waitForClientResponse(). Have adjusted waitForClientResponse() and the
run() method in SocketThread.java to now use a clientResponselock, and
calls to wait() and notifyAll(), rather than Thread.suspend() and
Thread.wait(). The if() test in the run() method now also specifically
looks for a thread instance of ManagerProxyThread or
SameVMManagerProxyThread when it's testing for Client/Authenticate messages.
Similar changes applied to the "http" and "lrmp" implementation.
----
v2.0 - 31st August 1999.
* The handling of the four different RegistryListener events needed to be
done in a separate thread, and surronded by a Throwable clause, just like
how all the other listener events are handled.
* Added the ConnectionMessage.java and RegistryMessage.java classes to the
client-side jar generation for each implementation in .../install/build.csh.
* From: Scott Swann <[email protected]>
If a connection failure is reported, then that connection listener is
automatically removed to prevent continual failure notifications.
* Added the following utility methods to impl/Util.java, and adjusted the
code accordingly:
public static Thread
startThread(Runnable target, String name, boolean isDaemon);
public static int
getIntProperty(String key, int default);
public static long
getLongProperty(String key, long default);
public static String
getStringProperty(String key, String default);
public static boolean
getBooleanProperty(String key, boolean default);
* From: Scott Swann <[email protected]>
The waitForReply() method needed to call finishReply() just before it
was about to throw a TimedOutException. This was needed, rather than the
existing code, so that other threads, currently in a wait() state, could
be notified that the state had changed.
----
v2.0 - 30th August 1999.
* Rather than having implementation specific variables set in JSDTObject,
the following five methods have been added to Connection.java:
public static Properties
getProperties();
public static String
getProperty(String key);
public static void
setProperties(Properties properties);
public static Object
setProperty(String key, String value);
public static Object
removeProperty(String key);
Each implementation would have specific properties such as:
showStack true
showMessage true
waitPeriod 5
which could be set and adjusted in this way. If no such property was found,
that the default value (still in JSDTObject) would be used.
The Implementations chapter of the JSDT User Guide has been updated to
reflect these changes, plus any JavaDoc that specifically mentioned
JSDTObject variables.
Note that explicitly setting the variables in JSDTObject will still
work, as long as there isn't an equivalent Connection property set.
* Added the following method to the URLString class:
public boolean
equals(Object anObject);
(see: ~/jsdt/test/JSDT-2.0/Test13).
* Adjusted @since references from 1.6 to 2.0.
* From: Scott Swann <[email protected]>
The handling of the ConnectionListener.connectionFailed() method call
needed to be in a separate thread, and surronded by a Throwable clause,
just like how all the other listener events are handled.
----
v2.0 - 24th August 1999.
* The server-side, when keeping information on the latest ping times for the
various proxy-side applications, should have been keying these in a hashtable
by id, not thread. When it had found one that needed to be removed, then it
needed to get the thread associated with that id.
----
v2.0 - 23rd August 1999.
* Added a section to the Introduction chapter in the Implementation Guide
to describe how to use alternate socket factory classes with the JSDT
"socket" implementation.
----
v2.0 - 22nd August 1999.
* Adjusted the three implementations and the template package so that each
class extends from impl.JSDTObject. Obsoleted socketJSDTObject,
lrmpJSDTObject, httpJSDTObject, templateJSDTObject.
----
v2.0 - 21st August 1999.
* Some adjustments to the way the server handles the cleaning up of clients
that are no longer pinging. When a proxy adds or removes a connection
listener, this now sends a message to each known session server thread,
which adds or removes an entry in ping[Threads, Ids], (and potentially
starts or stops the cleanup thread). The isAlive message now simply just
updates the hashtable entry for that thread/id with the current time.
This fixes the problem of the server incorrectly cleaning up a proxy