📄 cescapes.html
字号:
<html><!-- Mirrored from c-faq.com/stdio/cescapes.html by HTTrack Website Copier/3.x [XR&CO'2008], Sat, 14 Mar 2009 08:02:58 GMT --><head><title></title></head><body><p>[Someone asked me whether it was or wasn't obvious that the way toprint a single <TT>%</TT> sign using <TT>printf</TT> was with the (correct)<TT>"%%"</TT> or the (incorrect) <TT>"\%"</TT>.This is part of my reply, edited slightly for this web page.]<p>From: Steve Summit<br>Subject: Re: printf("%")<br>In-Reply-To: your message <199703310107.RAA19240@mx2.eskimo.com><br> of Sun, 30 Mar 1997 17:05:57 -0800<br><!-- XXX need Date:, Message-ID: -->Cc: scs@eskimo.com<p>If you're used to thinking about how a C program really ``works'' -- how the compiler generates an executable program, and how functions get called at run time -- then it really can be obvious that <TT>\%</TT> can't possibly work, and that the solution has to be something like <TT>%%</TT>.<p>Suppose that you write this program:<p><pre> main() { char line[80]; fgets(line, 80, stdin); fputs(line, stdout); }</pre><p>Obviously, this program reads one line from standard input and prints it right back out again. Suppose that the line you type at it consists of the 8 characters<p><pre> abc\ndef</pre><p>What gets printed?The answer is,one line, consisting of the 8 characters<p><pre> abc\ndef</pre><p>You might think that it would instead print the two lines<p><pre> abc def</pre><p>(In fact, it occurs to me that ``Why <em>doesn't</em> it print two lines?'' is an FAQ which I've never added to the list.)<br>[But I was wrong; it's question 8.8.]<p>Why <em>doesn't</em> it print two lines? Why aren't the two characters <TT>\n</TT> interpreted as a single newline character, as they are in character and string constants in source code? Because those backslash escapes for various special characters are interpreted <em>only</em> in character and string constants in source code. Furthermore, they're interpreted by the compiler, at compile time, in the process of converting your source code into an executable program. None of the standard I/O functions (<TT>fgets</TT>, <TT>fputs</TT>, <TT>printf</TT>, etc.) ever do any interpretation of backslashes. As far as they're concerned, a backslash is just another characterbeing read or written.By the time these functions are running, any backslashes which had been in the source code (that is, in character or string constants which had been handed to these functions to work on) will already have been processed.<p>That is, when <TT>fgets()</TT> reads a <TT>\</TT> from the input file, it just places it into the array with all of the other characters. But when you write<p><pre> printf("Hello, world!\n");</pre><p><TT>printf</TT> receives a string consisting of 14 characters (not including the terminating <TT>\0</TT>). The 14th character in the string is a single newline character, however that's represented internally. <TT>printf</TT> does <em>not</em> receive the two characters <TT>\</TT> and <TT>n</TT>,nor does it have to translate them into a newline character.The compiler did that already.<p>If you understand this, it's pretty obvious that<p><pre> printf("\%");</pre><p>can't work. What will the compiler translate <TT>\%</TT> into?As it happens, <TT>\%</TT> isn't a legal special-character sequence, butby analogy with <TT>\'</TT>, <TT>\"</TT>, and <TT>\\</TT>, it's likely that the compiler will convert <TT>\%</TT> into a single <TT>%</TT> character. (It may also generate a warning message about the undefined escape sequence.)So <TT>printf</TT> will probably receive a string consisting of a single <TT>%</TT>, <em>exactly</em> as if you'd simply typed<p><pre> printf("%");</pre><p>Whenever printf sees a <TT>%</TT> character in its format string, it <em>always</em> looks at the next character(s) to decide what to print. But there is no following character, so <TT>printf</TT> is confused.<p>I understand why you expect <TT>\%</TT> to work -- in some contexts,at least, the <TT>\</TT> ``turns off'' the special interpretation of the following character. (That's what it's doing when you write <TT>'\''</TT>or <TT>"This string contains a \" and a \\"</TT>.) So when you want to print a single <TT>%</TT>, you want to ``turn off'' <TT>printf</TT>'s special interpretation of the <TT>%</TT> character, and the first thing you think of is the backslash. But wait: the backslash turns off special interpretation of characters <em>by the compiler</em>, while what you want is to turn off special interpretation of a character <em>by<TT>printf</TT></em>. The two contexts are completely separate, so the rules might be different, and in fact the rules <em>are</em> different.(The first context is ``string constants being processed by the compiler'', and in that context, the rule for turning off special interpretation of certain characters is to precede them with abackslash. The second context is ``format strings beinginterpreted by <TT>printf''</TT>, and in that context, the rule isthat if you want a single <TT>%</TT>, you write <TT>%%</TT>.)</body><!-- Mirrored from c-faq.com/stdio/cescapes.html by HTTrack Website Copier/3.x [XR&CO'2008], Sat, 14 Mar 2009 08:02:58 GMT --></html>⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -