最近项目上调查printf语句不能正常格式化字符串的问题,做下总结。
以sprintf_s函数来说明问题的现象。
int sprintf_s( char *buffer, size_t sizeOfBuffer, const char *format [, argument] ... );
问题发生的条件
- 使用了setlocale
- format参数是utf8编码
- format的%是非ASC字符,已locale编码来解释的话,刚好%与前面的编码结合,被当初了一个新的文字。
例如我们出问题时,locale是".ACP",也就是ANSI(CP932),format参数是"日付%04d/%02d/%02d 時刻%02d:%02d:%02d.%03d",UTF编码如下:
00000003h: E6 97 A5 E4 BB 98 25 30 34 64 2F 25 30 32 64 2F ; 譌・莉・04d/%02d/
00000013h: 25 30 32 64 20 E6 99 82 E5 88 BB 25 30 32 64 3A ; %02d 譎ょ綾%02d:
00000023h: 25 30 32 64 3A 25 30 32 64 2E 25 30 33 64 ; %02d:%02d.%03d
其中日付对应的UTF8编码是
E6 97 A5 E4 BB 98
CP932环境下E6 97是一个独立的SJIS双字节文字【譌】。A5是一个单字节文字【・】。E4 BB是一个双字节文字【莉】,
98 与 25刚好满足SJIS双字节文字的编码要求,首字节在0x81~0x9f或者0xe0~0xef之间。所以对应print来说,第一个%s已经不存在了,自然不能被替换。
下面就来说说setlocale。
- 影响的函数
依据https://msdn.microsoft.com/en-US/library/x99tb11d(v=vs.80).aspx,setlocale会根据设定的category不同,产生不同的影响。
LC_COLLATE
The strcoll, _stricoll, wcscoll, _wcsicoll, strxfrm, _strncoll, _strnicoll, _wcsncoll, _wcsnicoll, and wcsxfrm functions.
LC_TIME
The strftime and wcsftime functions.
LC_MONETARY
Monetary-formatting information returned by the localeconv function.
//简单来讲,该类别就是用来输出本地化金融信息(如钱币符号)时用的,需要显示调用
LC_NUMERIC
Decimal-point character for the formatted output routines (such as printf), for the data-conversion routines, and for the non-monetary formatting information returned bylocaleconv. In addition to the decimal-point character, LC_NUMERIC also sets the thousands separator and the grouping control string returned by localeconv.
//与LC_MONETARY类似,但是需要注意,Decimal-point character (小数点符号)是不需要显示调用的,在printf等函数输出浮点数时,会隐式的使用locale相关的小数点。
比如默认的locale是C,C的小时点就是.。
LC_CTYPE
The character-handling functions (except isdigit, isxdigit, mbstowcs, and mbtowc, which are unaffected).
不得不吐槽的是msdn关于LC_CTYPE说得不全面+不正确。mbstowcs, and mbtowc,是受LC_CTYPE影响的。
http://www.cplusplus.com/reference/clocale/setlocale/?kw=setlocale的解释貌似是正确的。
Affects character handling functions (all functions of <cctype>, except isdigit and isxdigit), and the multibyte and wide character functions.
具体是哪些方法呢,参考<cctype>我们知道 character handling functions有:
- Check if character is alphanumeric (function )
- isalpha
- Check if character is alphabetic (function )
- isblank
- Check if character is blank (function )
- iscntrl
- Check if character is a control character (function )
- isgraph
- Check if character has graphical representation (function )
- islower
- Check if character is lowercase letter (function )
- isprint
- Check if character is printable (function )
- ispunct
- Check if character is a punctuation character (function )
- isspace
- Check if character is a white-space (function )
- isupper
- Check if character is uppercase letter (function )
- tolower
- Convert uppercase letter to lowercase (function )
- toupper
- Convert lowercase letter to uppercase (function )
另外,参考http://gsp.com/cgi-bin/man.cgi?section=3&topic=multibyte和https://msdn.microsoft.com/en-US/library/6y9se58z(v=vs.80).aspx,我们知道 the multibyte and wide character functions.有:
Function Description mblen 3 get number of bytes in a character mbrlen 3 get number of bytes in a character (restartable) mbrtowc 3 convert a character to a wide-character code (restartable) mbsrtowcs 3 convert a character string to a wide-character string (restartable) mbstowcs 3 convert a character string to a wide-character string mbtowc 3 convert a character to a wide-character code wcrtomb 3 convert a wide-character code to a character (restartable) wcstombs 3 convert a wide-character string to a character string wcsrtombs 3 convert a wide-character string to a character string (restartable) wctomb 3 convert a wide-character code to a character |