*: Regenerate.
[gcc.git] / libstdc++-v3 / doc / html / manual / fstreams.html
1 <html><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8"><title>File Based Streams</title><meta name="generator" content="DocBook XSL-NS Stylesheets V1.76.1"><meta name="keywords" content="
2 ISO C++
3 ,
4 library
5 "><meta name="keywords" content="
6 ISO C++
7 ,
8 runtime
9 ,
10 library
11 "><link rel="home" href="../index.html" title="The GNU C++ Library"><link rel="up" href="io.html" title="Chapter 13.  Input and Output"><link rel="prev" href="stringstreams.html" title="Memory Based Streams"><link rel="next" href="io_and_c.html" title="Interacting with C"></head><body bgcolor="white" text="black" link="#0000FF" vlink="#840084" alink="#0000FF"><div class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="3" align="center">File Based Streams</th></tr><tr><td width="20%" align="left"><a accesskey="p" href="stringstreams.html">Prev</a> </td><th width="60%" align="center">Chapter 13
12 Input and Output
13
14 </th><td width="20%" align="right"> <a accesskey="n" href="io_and_c.html">Next</a></td></tr></table><hr></div><div class="section" title="File Based Streams"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="std.io.filestreams"></a>File Based Streams</h2></div></div></div><div class="section" title="Copying a File"><div class="titlepage"><div><div><h3 class="title"><a name="std.io.filestreams.copying_a_file"></a>Copying a File</h3></div></div></div><p>
15 </p><p>So you want to copy a file quickly and easily, and most important,
16 completely portably. And since this is C++, you have an open
17 ifstream (call it IN) and an open ofstream (call it OUT):
18 </p><pre class="programlisting">
19 #include &lt;fstream&gt;
20
21 std::ifstream IN ("input_file");
22 std::ofstream OUT ("output_file"); </pre><p>Here's the easiest way to get it completely wrong:
23 </p><pre class="programlisting">
24 OUT &lt;&lt; IN;</pre><p>For those of you who don't already know why this doesn't work
25 (probably from having done it before), I invite you to quickly
26 create a simple text file called "input_file" containing
27 the sentence
28 </p><pre class="programlisting">
29 The quick brown fox jumped over the lazy dog.</pre><p>surrounded by blank lines. Code it up and try it. The contents
30 of "output_file" may surprise you.
31 </p><p>Seriously, go do it. Get surprised, then come back. It's worth it.
32 </p><p>The thing to remember is that the <code class="code">basic_[io]stream</code> classes
33 handle formatting, nothing else. In chaptericular, they break up on
34 whitespace. The actual reading, writing, and storing of data is
35 handled by the <code class="code">basic_streambuf</code> family. Fortunately, the
36 <code class="code">operator&lt;&lt;</code> is overloaded to take an ostream and
37 a pointer-to-streambuf, in order to help with just this kind of
38 "dump the data verbatim" situation.
39 </p><p>Why a <span class="emphasis"><em>pointer</em></span> to streambuf and not just a streambuf? Well,
40 the [io]streams hold pointers (or references, depending on the
41 implementation) to their buffers, not the actual
42 buffers. This allows polymorphic behavior on the chapter of the buffers
43 as well as the streams themselves. The pointer is easily retrieved
44 using the <code class="code">rdbuf()</code> member function. Therefore, the easiest
45 way to copy the file is:
46 </p><pre class="programlisting">
47 OUT &lt;&lt; IN.rdbuf();</pre><p>So what <span class="emphasis"><em>was</em></span> happening with OUT&lt;&lt;IN? Undefined
48 behavior, since that chaptericular &lt;&lt; isn't defined by the Standard.
49 I have seen instances where it is implemented, but the character
50 extraction process removes all the whitespace, leaving you with no
51 blank lines and only "Thequickbrownfox...". With
52 libraries that do not define that operator, IN (or one of IN's
53 member pointers) sometimes gets converted to a void*, and the output
54 file then contains a perfect text representation of a hexadecimal
55 address (quite a big surprise). Others don't compile at all.
56 </p><p>Also note that none of this is specific to o<span class="emphasis"><em>*f*</em></span>streams.
57 The operators shown above are all defined in the parent
58 basic_ostream class and are therefore available with all possible
59 descendants.
60 </p></div><div class="section" title="Binary Input and Output"><div class="titlepage"><div><div><h3 class="title"><a name="std.io.filestreams.binary"></a>Binary Input and Output</h3></div></div></div><p>
61 </p><p>The first and most important thing to remember about binary I/O is
62 that opening a file with <code class="code">ios::binary</code> is not, repeat
63 <span class="emphasis"><em>not</em></span>, the only thing you have to do. It is not a silver
64 bullet, and will not allow you to use the <code class="code">&lt;&lt;/&gt;&gt;</code>
65 operators of the normal fstreams to do binary I/O.
66 </p><p>Sorry. Them's the breaks.
67 </p><p>This isn't going to try and be a complete tutorial on reading and
68 writing binary files (because "binary"
69 covers a lot of ground), but we will try and clear
70 up a couple of misconceptions and common errors.
71 </p><p>First, <code class="code">ios::binary</code> has exactly one defined effect, no more
72 and no less. Normal text mode has to be concerned with the newline
73 characters, and the runtime system will translate between (for
74 example) '\n' and the appropriate end-of-line sequence (LF on Unix,
75 CRLF on DOS, CR on Macintosh, etc). (There are other things that
76 normal mode does, but that's the most obvious.) Opening a file in
77 binary mode disables this conversion, so reading a CRLF sequence
78 under Windows won't accidentally get mapped to a '\n' character, etc.
79 Binary mode is not supposed to suddenly give you a bitstream, and
80 if it is doing so in your program then you've discovered a bug in
81 your vendor's compiler (or some other chapter of the C++ implementation,
82 possibly the runtime system).
83 </p><p>Second, using <code class="code">&lt;&lt;</code> to write and <code class="code">&gt;&gt;</code> to
84 read isn't going to work with the standard file stream classes, even
85 if you use <code class="code">skipws</code> during reading. Why not? Because
86 ifstream and ofstream exist for the purpose of <span class="emphasis"><em>formatting</em></span>,
87 not reading and writing. Their job is to interpret the data into
88 text characters, and that's exactly what you don't want to happen
89 during binary I/O.
90 </p><p>Third, using the <code class="code">get()</code> and <code class="code">put()/write()</code> member
91 functions still aren't guaranteed to help you. These are
92 "unformatted" I/O functions, but still character-based.
93 (This may or may not be what you want, see below.)
94 </p><p>Notice how all the problems here are due to the inappropriate use
95 of <span class="emphasis"><em>formatting</em></span> functions and classes to perform something
96 which <span class="emphasis"><em>requires</em></span> that formatting not be done? There are a
97 seemingly infinite number of solutions, and a few are listed here:
98 </p><div class="itemizedlist"><ul class="itemizedlist" type="disc"><li class="listitem"><p><span class="quote"><span class="quote">Derive your own fstream-type classes and write your own
99 &lt;&lt;/&gt;&gt; operators to do binary I/O on whatever data
100 types you're using.</span></span>
101 </p><p>
102 This is a Bad Thing, because while
103 the compiler would probably be just fine with it, other humans
104 are going to be confused. The overloaded bitshift operators
105 have a well-defined meaning (formatting), and this breaks it.
106 </p></li><li class="listitem"><p>
107 <span class="quote"><span class="quote">Build the file structure in memory, then
108 <code class="code">mmap()</code> the file and copy the
109 structure.
110 </span></span>
111 </p><p>
112 Well, this is easy to make work, and easy to break, and is
113 pretty equivalent to using <code class="code">::read()</code> and
114 <code class="code">::write()</code> directly, and makes no use of the
115 iostream library at all...
116 </p></li><li class="listitem"><p>
117 <span class="quote"><span class="quote">Use streambufs, that's what they're there for.</span></span>
118 </p><p>
119 While not trivial for the beginner, this is the best of all
120 solutions. The streambuf/filebuf layer is the layer that is
121 responsible for actual I/O. If you want to use the C++
122 library for binary I/O, this is where you start.
123 </p></li></ul></div><p>How to go about using streambufs is a bit beyond the scope of this
124 document (at least for now), but while streambufs go a long way,
125 they still leave a couple of things up to you, the programmer.
126 As an example, byte ordering is completely between you and the
127 operating system, and you have to handle it yourself.
128 </p><p>Deriving a streambuf or filebuf
129 class from the standard ones, one that is specific to your data
130 types (or an abstraction thereof) is probably a good idea, and
131 lots of examples exist in journals and on Usenet. Using the
132 standard filebufs directly (either by declaring your own or by
133 using the pointer returned from an fstream's <code class="code">rdbuf()</code>)
134 is certainly feasible as well.
135 </p><p>One area that causes problems is trying to do bit-by-bit operations
136 with filebufs. C++ is no different from C in this respect: I/O
137 must be done at the byte level. If you're trying to read or write
138 a few bits at a time, you're going about it the wrong way. You
139 must read/write an integral number of bytes and then process the
140 bytes. (For example, the streambuf functions take and return
141 variables of type <code class="code">int_type</code>.)
142 </p><p>Another area of problems is opening text files in binary mode.
143 Generally, binary mode is intended for binary files, and opening
144 text files in binary mode means that you now have to deal with all of
145 those end-of-line and end-of-file problems that we mentioned before.
146 </p><p>
147 An instructive thread from comp.lang.c++.moderated delved off into
148 this topic starting more or less at
149 <a class="link" href="http://groups.google.com/group/comp.std.c++/browse_thread/thread/f87b4abd7954a87/946a3eb9921e382d?q=comp.std.c%2B%2B+binary+iostream#946a3eb9921e382d" target="_top">this</a>
150 post and continuing to the end of the thread. (The subject heading is "binary iostreams" on both comp.std.c++
151 and comp.lang.c++.moderated.) Take special note of the replies by James Kanze and Dietmar Kühl.
152 </p><p>Briefly, the problems of byte ordering and type sizes mean that
153 the unformatted functions like <code class="code">ostream::put()</code> and
154 <code class="code">istream::get()</code> cannot safely be used to communicate
155 between arbitrary programs, or across a network, or from one
156 invocation of a program to another invocation of the same program
157 on a different platform, etc.
158 </p></div></div><div class="navfooter"><hr><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="stringstreams.html">Prev</a> </td><td width="20%" align="center"><a accesskey="u" href="io.html">Up</a></td><td width="40%" align="right"> <a accesskey="n" href="io_and_c.html">Next</a></td></tr><tr><td width="40%" align="left" valign="top">Memory Based Streams </td><td width="20%" align="center"><a accesskey="h" href="../index.html">Home</a></td><td width="40%" align="right" valign="top"> Interacting with C</td></tr></table></div></body></html>