捕获简单表填充异常

红色应该在那里吗？GFXXXCHAR * GETLIGAGEXETCIN PTR（const gfxxxCHAR *（*text））



      以下为原文

    Should the red * be there?
GFX_XCHAR *getLanguageStringPtr(const GFX_XCHAR *(*text))

是的，它是一组指针。代码很好。我只是想弄清楚如何才能不通过一个坏指针。坏指针是我的错误。



      以下为原文

    Yes it is an array of pointers.
The code is good. I am just trying to figure out how to figure out how to no pass a bad pointer. The bad pointers are my bugs.

如果是一个指针表，那么



      以下为原文

    If it's a table of pointers, then maybe
if ( (text != NULL) && (*text != NULL) )
 

我会试试看。



      以下为原文

    I will try that.

指针可以指向任何地方。除了微不足道的NulLand对齐检查（或者检查物理内存的界限）之外，你不能真正判断指针是好还是坏。如果“文本[语言]”从一个随机位置抓取一个指针，在返回之后试图访问它的代码将崩溃。



      以下为原文

    The pointer may point anywhere. Except the trivial not-null and alignment checks (or perhaps checking the bounds of physical memory), you cannot really tell if the pointer is good or bad. If the "text[language]" grabs a pointer from a random place, the code that tries to access it after the return will crash.
 

你能给我们一个更完整的例子吗？指针的数组与指针指针的数组不一样。



      以下为原文

   
Can you give us a more complete example on how getLanguageStringPtr() is used? An array of pointers is not the same as pointer-to-pointer. 

我想我应该知道不要发布一个完整的例子。当我从指针表中拔出字符串数组并把它弄乱的时候，这个问题就发生了。所有的打印都是SNPrTfF（）和其他不会逃跑的函数。在函数调用所有函数时，垃圾更容易调试。



      以下为原文

   
 
#define GFX_XCHAR char;
 
#define NUMBER_OF_LANGUAGES 3 
 
 
 
 
 
static const GFX_XCHAR const *btnCaptionCancel [NUMBER_OF_LANGUAGES] = {"Cancel","",""};
 
 GFX_XCHAR *pointer; 
 
 
 
pointer = getLanguageStringPtr(btnCaptionCancel);
 


 
//******************************************************************************
// Function: getLanguageStringPtr()
//
// Description: returns a pointer to the string of the current language,
// it returns string [0] if the current language string is ""
//
// Parameter(s):
// text : a pointer to a an array of language text
//
// Return: pointer to the string
//
//******************************************************************************
GFX_XCHAR *getLanguageStringPtr(const GFX_XCHAR *(*text))
{
    uint16_t language;
    
    if(text != NULL)
    {
        // if language item is blank use English
        if(text[ext_eeprom.language][0] != '')language = ext_eeprom.language;
        else language = 0;
        if(language > NUMBER_OF_LANGUAGES) language = 0; // trap errors

        // return the selected string
        return (GFX_XCHAR *)text[language];
    }
    else
    {
        return (GFX_XCHAR *)""; // No string?
    }

}

 
I guess I should know better than to not post a full example.
The Issue happens when I am pulling the string array out of a pointer table and mess it up.
The prints are all snprintf() and other functions that will not run away. Garbage is easier to debug that an exception in a function everything calls.

显然，您不是通过捕获所有坏指针来修复这一点，而是通过不把错误指针传递给函数。单选空指针的情况并没有真正的帮助。真正简单的事情应该是足够的：你唯一需要做的就是确保ExtEXEPROM.Load永远不会超出0（到NoMyLoof Of LangeSe-1）范围，当你分配它时，你检查一次，而不是使用它的时候，我甚至会进一步简化它。



      以下为原文

    Obviously, you fix this not by trapping all bad pointers, but by not passing bad pointers to the function. Singling out the NULL pointer situation doesn't really help. Something really simple should be sufficient:
 
GFX_XCHAR *getLanguageStringPtr(const GFX_XCHAR **text) {
   if(text[ext_eeprom.language][0]) return text[ext_eeprom.language];
   return text[0];
}
 
The only thing you need to do is to make sure that ext_eeprom.language is never out of the 0 to (NUMBER_OF_LANGUAGES-1) range, which you check once when you assign it, not hundered times when you use it.
 
I would even simplify it further:
 
typedef const GFX_XCHAR * GFX_XSTRING;
 
static const GFX_XSTRING btnCaptionCancel [NUMBER_OF_LANGUAGES] = {"Cancel",NULL,NULL};
 
GFX_XSTRING getLanguageString(const GFX_XSTRING text[]) {
   return text[ext_eeprom.language]?text[ext_eeprom.language]:text[0];
}
 

我在跳，有一种方法来捕捉坏指针。这使得调试更加容易。语言索引的陷阱是由于一个bug而被覆盖。由于这是所有屏幕字符串的第一站，所以它是例外的所在。我越陷越容易。如果没有办法看它是否会引起异常。然后我需要在异常处理程序中处理它。设置一个标志，清除异常。然后回来？我会在星期一看你的建议。他们很难在电话上进行比较。此代码是在MPLABX 3天编写的。当IDE不能出现异常时。我不确定是否该加速了。



      以下为原文

    I was hopping there was a way to trap the bad pointer. This makes debugging easier.
The trap for the language index was due to it being overwritten due to a bug. Since this ends up being the first stop for all the screen strings, it is home to the exceptions. The more I trap the easier life is. If There is no way to see if it will cause an exception. Then I would need to handle it in the exception handler. Set a flag , clear the exception. and come back?

I will look at your suggestions on Monday. They are hard to compare on a phone.
This code was written in the MPLabX 3.00 Days. When the ide could not an exception.
I am not sure if it’s time to speed up yet.

      以下为原文

    (*text != NULL) appears to be the correct answer here.

I assume: ( (text != NULL) && (*text != NULL) ) would be better

如果文本为NULL，则第一个条件将阻止您对其进行转换。虽然*Text＝NULL可能是您的bug，但在捕捉明显的其他漏洞时没有坏处。



      以下为原文

    if text is null, the first condition will prevent you from casting it.  while *text == NULL may be the bug you have, there is no harm in catching the obvious other vulnerability.