November 9 2006

最新推荐文章于 2025-10-21 22:09:11 发布

原创最新推荐文章于 2025-10-21 22:09:11 发布 · 810 阅读

0 ·

CC 4.0 BY-SA版权

文章标签：

#debugging #thread #function #database #string #sms

本文介绍了作者在进行调试任务中的经历，特别是针对一个用于时间同步的程序。该程序负责通知其他设备或主机当前的标准时间，并确保只有一个时间主设备存在。文章详细描述了程序的工作流程及调试过程中遇到的问题。

Today I go on my debugging task. This morning I finished the part of testing database codes.
Before 2 o,clock p.m. I finished the rest of debugging task.

The main task of my debugging program is to inform other devices or hosts of the standard time
now. Firstly, it must insure that there is only one time master, a host or a device. It inquired
the field "timeMaster" of a table in a database, and get the value of that field. The value must
be 1. That is to say, if there are time masters more than one, the program will strike.

There are two ways to launch it. One is to tell it a host name at command line; another
is just to type the program name and press "Enter" key. If the program find input a host name,
it will just create one thread for the host. The thread get the retry times and time out
from a CSV file, a configuration file; and get now time by calling the time() system call;
fulfill a message datagram; connect to the host with UDP; then send it out. If failure to
connect or to send, it retry the specified times.

If it is launched in another way, it will get how many SMS, a device, here are; and create
a thread respectively for every SMS. Every sub-thread get current system time and tell a SMS
through UDP. The steps of them is same to that of the running thread in the first way.

During debugging I found a strange code segment. I copied it as follow.

if( FALSE == getDataValue( DB_TYPE_U_CHAR, &result, &fieldType) ) {

          SGCMNsyslogErr("2006-11-07 LM Debug Print>>>>: Func %s Line %d FALSE == getDataValue( DB_TYPE_U_CHAR, &result, &fieldType). result=%s, fieldType=%u/n", __FUNCTION__, __LINE__, result, fieldType);

        if ( FALSE == getDataValue( DB_TYPE_STRING, &result, &fieldType) ) {
                SGCMNsyslogErr("getCLtimeMaster.c@parseCLtimeMasterDB():Read FieldType getDataValue() Error");
                ret = FALSE;
        }
}

This code segment tries to get the value of a field type. If can not get the DB_TYPE_U_CHAR type,
it tries to get DB_TYPE_STRING type. Let's look at the following code segment.

if (DB_TYPE_U_CHAR == fieldType) {
    SGCMNsyslogErr("2006-11-07 LM Debug Print>>>>: Func %s Line %d DB_TYPE_U_CHAR == fieldType./n", __FUNCTION__, __LINE__);
    //値の取得
    if( TRUE == getDataValue( fieldType, &result, timeMasterVal) ) {
        // 2006-11-07 add debug info.
        SGCMNsyslogErr("2006-11-07 LM Debug Print>>>>: Func %s Line %d timeMasterVal = %d/n", __FUNCTION__, __LINE__, *timeMasterVal);
    }
    else {
        SGCMNsyslogErr("getCLtimeMaster.c@parseCLtimeMasterDB():Read cal_no getDataValue() Error");
        ret = FALSE;
    }
}else {
    SGCMNsyslogErr("getCLtimeMaster.c@parseCLtimeMasterDB():Read cal_no TYPE Error");
    ret = FALSE;
}

You will find that the program can not acquire the value of timeMaster if the "fieldType"
is not "DB_TYPE_U_CHAR". These two code segments are both in one function. I don't understand
why get another field type value after failure to get one. If it get DB_TYPE_STRING value
successfully, the function still can not get the field value due to type error. So, whether
it is a redundant. It seems that only DB_TYPE_U_CHAR is valid. Why it check whether the
field type is DB_TYPE_U_CHAR after getting successfully.

Another thing I found is that these codes is easy to read. Sometimes, I am aware that
it was better and the binary program was smaller if programmers had applied some tricks.
For example, we can write a function to create a thread on that the way of creating a thread
is same. We just pass a thread handler and a name into that function. Whether one or more
than one threads, one "for" code segment can be up to do.