Thursday, December 18, 2008
游罢镇江，填词两首
游罢镇江，填词两首：）
南乡子·京口北固楼有怀
其一：
天下总堪忧，第一江山北固楼。多少干戈多少恨，休休，遍地繁华碎碎蹂。
浩荡望神州，满眼风光不肯收，落日烟波人醉后，中流，敢向惊涛叩吴钩。
其二：
天下可堪忧？第一江山北固楼。无限干戈多少恨，休休，遍地繁华碎碎蹂！
何必望神州！眼底风光不肯收，淡月轻云人醉后，温柔，漫卷馨香邀野舟！
南乡子·京口北固楼有怀
其一：
天下总堪忧，第一江山北固楼。多少干戈多少恨，休休，遍地繁华碎碎蹂。
浩荡望神州，满眼风光不肯收，落日烟波人醉后，中流，敢向惊涛叩吴钩。
其二：
天下可堪忧？第一江山北固楼。无限干戈多少恨，休休，遍地繁华碎碎蹂！
何必望神州！眼底风光不肯收，淡月轻云人醉后，温柔，漫卷馨香邀野舟！
Wednesday, May 07, 2008
Maybe this is a bug of JDK 1.6.03
My JDK version is 1.6.03
In the class ThreadPoolExecutor, the counter for completed tasks, long completedTaskCount, with the line 1007:
completedTaskCount += w.completedTasks;
will increase forever if the instance of ThreadPoolExecutor keep alive. It's never reset.
We know that the thread pool usually run as a long time task, so the variable long completedTaskCount is easy to overflow, especially for the heavy load application. We really don't know what will happen if it overflow.
So, this class should be rewritten for the more robust application.
In the class ThreadPoolExecutor, the counter for completed tasks, long completedTaskCount, with the line 1007:
completedTaskCount += w.completedTasks;
will increase forever if the instance of ThreadPoolExecutor keep alive. It's never reset.
We know that the thread pool usually run as a long time task, so the variable long completedTaskCount is easy to overflow, especially for the heavy load application. We really don't know what will happen if it overflow.
So, this class should be rewritten for the more robust application.
Thursday, March 20, 2008
Subscribe to:
Posts (Atom)