diff options
Diffstat (limited to 'java/ChangeLog')
-rw-r--r-- | java/ChangeLog | 122 |
1 files changed, 0 insertions, 122 deletions
diff --git a/java/ChangeLog b/java/ChangeLog deleted file mode 100644 index 47a572138e8..00000000000 --- a/java/ChangeLog +++ /dev/null @@ -1,122 +0,0 @@ -Tue Aug 24 13:09:15 1999 Everett Anderson <eea1@cs.wustl.edu> - - * src/*, tests/*, netsvcs/*: Implemented major changes to the - package system. Now all files descend from the JACE directory - under the main archive level. All files are further grouped by - their subpackage names. This is the way Java usually wants it, - and makes JACE javadoc friendly. - - Rewrote and added javadoc comments throughout JACE. - - Changed all concurrency mechanisms to inherit from AbstractLock, - obeying all its requirements such as not releasing the lock - unless the accessing thread is the owner. - - Added the token service. - - Added and rewrote tests for Concurrency mechs and ServiceConfig. - - Rewrote network services to inherit from JACE.netsvcs.Server and - JACE.netsvcs.Handler, abstracting much of the work out to these - base classes. - - Debugging is now OFF by default. - -Wed Aug 18 19:43:23 1999 Everett Anderson <eea1@cs.wustl.edu> - - * src/GetOpt.java: Added another constructor to GetOpt with an - option to return args even if not in the optstring. This is - useful for the default case in a switch (for printUsage - and exit). - -Wed Aug 18 13:35:46 1999 Everett Anderson <eea1@cs.wustl.edu> - - * src/TimerQueue.java: Now when createInternalThread is specified - in the constructor, it is set to be a daemon thread. Thus, if a - TimerQueue thread is the only one executing, the VM exits. - -Wed Jul 14 11:27:47 1999 Everett Anderson <eea1@cs.wustl.edu> - - * src/StrategyAcceptor.java, AcceptStrategy.java - StrategyAcceptor now inherits from Acceptor. AcceptStrategy - can now delegate to an Acceptor, rather than having its own - SOCKAcceptor. Its default is to delegate to a generic - Acceptor instance (that uses a SOCKAcceptor). - -Fri Jul 09 13:58:42 1999 Everett Anderson <eea1@cs.wustl.edu> - - * src/Acceptor.java, TimeValue.java - Solved a few javadoc warnings. - -Fri Jul 09 11:08:32 1999 Everett Anderson <eea1@cs.wustl.edu> - - * src/TimedWait.java, Stream*.java, Token.java - Fixed a bug in TimedWait from the recent changes. Made - Streams (and Modules) use absolute times for timeouts. - -Thu Jul 08 14:54:36 1999 Everett Anderson <eea1@cs.wustl.edu> - - * src/Condition.java, TimedWait.java, EventHandler.java, - MessageQueue.java, ServiceObject.java, SvcHandler.java, - Task.java, Mutex.java, Semaphore.java, Token.java, - TimerQueue.java - - Changed the semantics of Condition, TimedWait, MessageQueue, - Task, Mutex, Semaphore, and Token to use absolute times for - their timeouts. - - Changed the semantics of EventHandler, ServiceObject, and - SvcHandler such that handleTimeout receives a TimeValue - representing when the event occured. - - Changed TimerQueue internally to reflect the above changes. - - Also made better checks in Token and Mutex so that non-owners - can call release without adverse effects. I plan to try to - do this with Semaphore and RWMutex as well. - -Thu Jul 08 11:20:17 1999 Everett Anderson <eea1@cs.wustl.edu> - - * src/TimeValue.java: Added relativeTimeOfDay(offsets) - methods which return TimeValues. These should help people - since many classes will be changed to use - absolute TimeValues shortly to correspond to ACE's versions. - -Wed Jul 07 22:42:27 1999 Everett Anderson <eea1@cs.wustl.edu> - - * src/Token.java: Fixed several bugs in tryacquire and renew. - Specifying a null TimeValue in acquire or renew now means to wait - until the lock is obtained. Specifying TimeValue.zero in acquire - means just to poll and don't call sleepHook(). In renew, it is - the same as specifying a requePosition of 0 -- doesn't give up - the token. - -Wed Jul 07 21:54:35 1999 Everett Anderson <eea1@cs.wustl.edu> - - * src/TimedWait.java: Fixed the semantics of - timedWait(TimeValue tv). Now when tv represents zero, it has - polling semantics, checking the condition once and returning - on success or throwing a TimeoutException on failure. Before, - it blocked until notified. (Also affects synch mechanisms) - -Wed May 20 14:37:48 CDT 1998 Everett Anderson <eea1@cs.wustl.edu> - -Reset the ChangeLog file for future users. If anyone wants more complete -file by file changes, they can be provided with the CVS log command. - -Updated the JACE page at http://www.cs.wustl.edu/~eea1/JACE.html to reflect -the Version 1.4.4 changes, which included: - -* Removed CUP and JLex from JACE. The java_cup directory is still in CVS, - but won't be included in future versions. This means that the option to - load C++ ACE ServiceConfigurator files is gone. - -* Changed the SOCKStream system so it uses DataInputStream and DataOutputStream - to maintain compatibility with other languages and remove the deprecation - warnings from using PrintStream. (Can't use Reader/Writers since they - write 16-bit chars not 8-bit bytes) - -* Made lots of modifications to comments and code to remove javadoc warnings - and deprecation warnings. They were all straight-forward changes such as - making sure a @throws line was in the comments for all exceptions a method - had, etc. |