java - 方法中的 `try catch//this is a workaround`。应该如何重写呢？

Question

项目源代码有一个用于 SQL 处理的 Java 方法。该方法确实有效，但它使用了一个有问题的解决方法:在方法的最后部分使用 try-catch block 来正常执行。正确的实现方法是什么？

    public void run() {
        if (running) {
            return;
        }
        running = true;            
        while(null == Common.server || null == Common.database || !ConnectionsPool.isInitialized()) {
            // Wait until the database is set before continuing...
            try {
                Thread.sleep(1000);
            }
            catch(Exception ex) {}
        }
        while(running) {
            final Connections cs = ConnectionsPool.getConnections();
            Connection  c  = null;
            while(!entries.isEmpty()) {
                if (null == c) {
                    c = cs.getConnection();
                }
                SQLLogEntry entry = entries.remove();
                if (null != entry) {
                    try {
                        write(entry, c); //find usages
                    }
                    catch (SQLException ex) {
                        writeLogFile("Could not write entry to SQL", ex);
                    }
                }
            }
            if (null != c) {
                try {
                    c.commit();
                }
                catch (SQLException ex) {
                    writeLogFile("Could commit to SQL", ex);
                    try {
                        c.rollback();
                    }
                    catch (SQLException ex1) {
                    }

                    // log
                    final StringWriter err = new StringWriter();                        
                    ex.printStackTrace(new PrintWriter(err));
                    EditorTransactionUtil.writeLogFile(err.toString());                        
                    // for user
                    final String msg = "Exception: " + EditorUtil.getErrorMessage(ex.getMessage());                        
                    try {
                        SwingUtilities.invokeAndWait(() -> {
                            JOptionPane.showMessageDialog(null, msg);
                        });
                    }
                    catch (Throwable ex1) {
                    }
                }
                finally {
                    cs.returnConnection(c);
                }
                c = null;
            }

            synchronized(entries) {
                try {
                    entries.wait(1000);
                }
                catch (InterruptedException ex) {
                    // This is a workaround to process this loop...
                }
            }
        }
        writeLogFile("SQLMsgLogger run loop stopping...");
    }

Answer 1

此代码的问题从这里开始。

If(running) return;
running=true;

这显然是为了确保只有一个线程执行。这是检查并发性的错误方法。当 if 检查结束时，第二步可能会立即开始，但分配尚未开始。您需要使用可同步的接口(interface)。

至于已释放的 try catch block - 正如 Konrad 指出的那样，如果没有 Thread.interrupt() 调用，它不会被执行。它可能是以前版本留下的死代码。