reactos/rosapps/net/ncftp/sio/sio.html
Steven Edwards 86bda6b3d9 Added ncftp to rosapps
svn path=/trunk/; revision=2453
2001-12-30 06:31:25 +00:00

1086 lines
39 KiB
HTML

<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
<META NAME="Author" CONTENT="Mike Gleason">
<META NAME="GENERATOR" CONTENT="Mozilla/4.03 [en] (WinNT; I) [Netscape]">
<TITLE>sio: SocketIO Library</TITLE>
</HEAD>
<BODY BGCOLOR="#FFFFFF">
<H1>
SocketIO Library Documentation</H1>
<UL>
<LI>
<A HREF="#Introduction">Introduction</A></LI>
<LI>
<A HREF="#Installation">Installation</A></LI>
<LI>
<A HREF="#Functions">Function Reference</A></LI>
<LI>
<A HREF="#Sample">Sample Code</A></LI>
</UL>
The purpose of the <I>SocketIO Library</I> (<I>sio</I> for short) is to
provide a safe and consistent wrapper interface to a BSD Sockets implementation.&nbsp;
The library also takes steps to protect against a few common pitfalls that
plague poorly written BSD Sockets programs.&nbsp; Specifically, it was
designed to do the following:
<UL>
<LI>
Take care of reading and writing the full operation.
<P><FONT SIZE=-1>For example, a network <TT>write()</TT> for 50 bytes may
only write 30 bytes; <I>Sio</I> would continue writing until all 50 bytes
have been sent.</FONT>
<P></LI>
<LI>
<P>
Allow operations that would normally block can be set to timeout after
a customizable period of time.<P></LI>
<LI>
<P>
Catch the SIGPIPE signal which would cause an unexpecting process to exit.<P>
<P><FONT SIZE=-1>A frequent source of problems for beginning programmers
is the situation where their process is suddenly no longer running because
a write to a socket caused a broken pipe. This problem can be difficult
to diagnose if the process is running as a daemon process in the background.</FONT>
<P>
</LI>
<LI>
<P>
Resume operation when system calls are interrupted (operations errout
with <TT>EINTR</TT>).<P></LI>
<LI>
<P>
Cater to the internet socket interface (i.e. <TT>struct sockaddr_in</TT>).<P></LI>
<LI>
<P>
Simplified interface to name service.
<P><FONT SIZE=-1>Library routines can take a textual address specification
instead of you having to prepare a <TT>struct sockaddr_in</TT>.</FONT>
<P>
</LI>
</UL>
The library was written by Mike Gleason. The first incarnation dates
back to some time in 1992. The library may be used and distributed
freely, as long as you give due credit.
<P>
<H3>
<A NAME="Introduction"></A>Introduction</H3>
The reader is assumed to be familiar with BSD Sockets.&nbsp; An excellent
source of information is W. Richard Stevens' <U>UNIX Network Programming,
Volume 1, Second Edition: Networking APIs: Sockets and XTI</U>, Prentice
Hall, 1998, ISBN 0-13-490012-X.
<BR>&nbsp;
<H4>
<A NAME="Connection_modes"></A>Connection modes</H4>
A communications exchange between two end-points can be connection-oriented
or connectionless.&nbsp; An connection-oriented exchange is characterized
by a connection establishment, a series of messages, and a disconnection.&nbsp;
A connectionless exchange is characterized by one or more independent messages.&nbsp;
An analogy to a connection-oriented exchange would be a telephone call
where one party establishes a connection to another, and a conversation
ensues.&nbsp; A connectionless exchange analogy would be a letter sent
via the postal service -- even if a writer sends two letters to the same
recipient, each letter is transported independently.
<BR>&nbsp;
<H4>
<A NAME="Message_modes"></A>Message modes</H4>
For the <I>sio</I> library, all connection-oriented exchanges are data
streams using the <I>TCP/IP</I> protocol.&nbsp; After connection establishment,
the conversation consists of one long sequence of data bytes, which is
known as a <I>stream</I>.&nbsp; Therefore there is no concept of a record
boundary associated with a stream connection at the transport level, although
often there is a record associated over the stream at the application level.&nbsp;
The important thing here is to realize that there is no flow control associated
with the stream, so although the data bytes are guaranteed to arrive in
order, there is no guarantee that messages will be read in the same size
blocks as they were originally written.
<P>For connectionless exchanges, <I>sio</I> uses the <I>UDP/IP</I> protocol's
datagram messages.&nbsp; There is an implicit definition of a record boundary,
because each message is treated as a record. So, three writes results in
three separate messages.
<P>For example, let's say a sender writes three 50-byte messages.&nbsp;
If a receiver does a 20-byte read followed by 100-byte read followed by
a 50-byte read, the receiver would get the first 20 bytes of the first
message, 50 of 50 bytes of the second message, and 50 of 50 bytes of the
third message.&nbsp; It's important to understand that in the first read
that since UDP datagrams are message oriented that the remaining 30 bytes
of the first message are lost, and that although the second read is for
100 bytes, the read immediately returns as finished even though only 50
bytes were actually read.
<P>With UDP datagrams, there are also other important implications that
affect message ordering, duplication, and overall reliability.
<BR>&nbsp;
<H4>
<A NAME="Creating_and_disposing"></A>Creating and disposing sockets</H4>
A socket is an I/O descriptor that is associated with one side of a communications
exchange.&nbsp; A socket must be either a client or a server, although
after communications is established the difference is arbitrary.&nbsp;
A server socket is one that waits for contact to be initiated by a client
on a mutually agreed upon address and port number.&nbsp; A client socket
initiates the first communication by sending to an existing server socket.&nbsp;
Client and server sockets may of course be on different machines, and different
networks altogether.
<P>A stream server socket is created using <TT><A HREF="#SNewStreamServer">SNewStreamServer()</A></TT>,
which returns a socket file descriptor ready to accept new client connections.&nbsp;
After a socket descriptor is obtained, your server program can then use
<TT><A HREF="#SAccept">SAccept()</A></TT> to establish a connection with
a client.
<P>A stream client socket is created using <TT><A HREF="#SNewStreamClient">SNewStreamClient()</A></TT>.&nbsp;
After a socket descriptor is obtained, your client program can use <TT><A HREF="#SConnectByName">SConnectByName()</A></TT>
or <TT><A HREF="#SConnect">SConnect()</A></TT> to initiate a connection
to a server.
<P>A datagram server socket is created using <TT><A HREF="#SNewDatagramServer">SNewDatagramServer()</A></TT>.&nbsp;
After a socket descriptor is obtained, it is ready to receive messages
from clients using <TT><A HREF="#SRecvfrom">SRecvfrom()</A></TT> and reply
back with <TT><A HREF="#SSendto">SSendto()</A></TT>.
<P>A datagram client socket is created using <TT><A HREF="#SNewDatagramClient">SNewDatagramClient()</A></TT>.&nbsp;
After a socket descriptor is obtained, it is ready to communicate with
servers using <TT><A HREF="#SSendto">SSendto()</A></TT> and <TT><A HREF="#SRecvfrom">SRecvfrom()</A></TT>.
<P>All socket descriptors are disposed of with <TT><A HREF="#SClose">SClose()</A></TT>
when communication is finished.
<BR>&nbsp;
<H4>
<A NAME="SocketIO"></A>Socket I/O</H4>
A stream connection uses <TT><A HREF="#SRead">SRead()</A></TT> to receive
data from the stream and <TT><A HREF="#SWrite">SWrite()</A></TT> to send
data.&nbsp; A datagram socket should use <TT><A HREF="#SRecvfrom">SRecvfrom()</A></TT>
to read a message and <TT><A HREF="#SSendtoByName">SSendtoByName()</A></TT>
or <TT><A HREF="#SSendto">SSendto()</A></TT> to send a message.&nbsp; Since
each datagram communication is independent, these routines use an address
specifier along with each call.
<BR>&nbsp;
<H3>
<A NAME="Installation"></A>Installation</H3>
First, unpack the archive. If the package arrived as a '.tar.gz' or '.tgz'
file, you need to run gunzip, and then tar to extract the source files.
You can do that in one shot by doing "<TT>gunzip -c sio.tgz | tar xvf -</TT>".
If the package you have is a '.tar' file you can use "<TT>tar xvf sio.tar</TT>".
If the package you have is a '.zip' file you can use "<TT>unzip sio.zip</TT>"
or "<TT>pkunzip sio.zip</TT>".
<P>Now go to the "<TT>sio</TT>" directory you just made. There is a script
you must run which will checks your system for certain features, so that
the library can be compiled on a variety of U<FONT SIZE=-1>NIX</FONT> systems.
Run this script by typing "<TT>sh ./configure</TT>" in that directory.
After that, you can look at the <TT>Makefile</TT> it made if you like,
and then you run "<TT>make</TT>" to create the "<TT>libsio.a</TT>" library
file.
<P>Finally, install the library and headers. You can manually copy the
files, or you can run "<TT>make install</TT>" to copy the files for you.
If you choose to "<TT>make install</TT>" you may want to edit the <TT>Makefile</TT>
if you do not want to install to the <TT>/usr/local</TT> tree.
<BR>&nbsp;
<H3>
<A NAME="Functions"></A>Function Reference</H3>
<UL>
<table width="100%">
<tr>
<td><A HREF="#AddrStrToAddr">AddrStrToAddr</A></td>
<td><A HREF="#SConnectByName">SConnectByName</A></td>
<td><A HREF="#SSendtoByName">SSendtoByName</A></td>
</tr>
<tr>
<td><A HREF="#AddrToAddrStr">AddrToAddrStr</A></td>
<td><A HREF="#SListen">SListen</A></td>
<td><A HREF="#SWrite">SWrite</A></td>
</tr>
<tr>
<td><A HREF="#DisposeSReadlineInfo">DisposeSReadlineInfo</A></td>
<td><A HREF="#SNewDatagramClient">SNewDatagramClient</A></td>
<td><A HREF="#SelectR">SelectR</A></td>
</tr>
<tr>
<td><A HREF="#FlushSReadlineInfo">FlushSReadlineInfo</A></td>
<td><A HREF="#SNewDatagramServer">SNewDatagramServer</A></td>
<td><A HREF="#SelectSetAdd">SelectSetAdd</A></td>
</tr>
<tr>
<td><A HREF="#GetSocketBufSize">GetSocketBufSize</A></td>
<td><A HREF="#SNewStreamClient">SNewStreamClient</A></td>
<td><A HREF="#SelectSetInit">SelectSetInit</A></td>
</tr>
<tr>
<td><A HREF="#GetSocketLinger">GetSocketLinger</A></td>
<td><A HREF="#SNewStreamServer">SNewStreamServer</A></td>
<td><A HREF="#SelectSetRemove">SelectSetRemove</A></td>
</tr>
<tr>
<td><A HREF="#GetSocketNagleAlgorithm">GetSocketNagleAlgorithm</A></td>
<td><A HREF="#SRead">SRead</A></td>
<td><A HREF="#SelectW">SelectW</A></td>
</tr>
<tr>
<td><A HREF="#InitSReadlineInfo">InitSReadlineInfo</A></td>
<td><A HREF="#SReadline">SReadline</A></td>
<td><A HREF="#Sendto">Sendto</A></td>
</tr>
<tr>
<td><A HREF="#SAcceptA">SAcceptA</A></td>
<td><A HREF="#SRecv">SRecv</A></td>
<td><A HREF="#SendtoByName">SendtoByName</A></td>
</tr>
<tr>
<td><A HREF="#SAcceptS">SAcceptS</A></td>
<td><A HREF="#SRecvfrom">SRecvfrom</A></td>
<td><A HREF="#SetSocketBufSize">SetSocketBufSize</A></td>
</tr>
<tr>
<td><A HREF="#SBind">SBind</A></td>
<td><A HREF="#SRecvmsg">SRecvmsg</A></td>
<td><A HREF="#SetSocketLinger">SetSocketLinger</A></td>
</tr>
<tr>
<td><A HREF="#SClose">SClose</A></td>
<td><A HREF="#SSend">SSend</A></td>
<td><A HREF="#SetSocketNagleAlgorithm">SetSocketNagleAlgorithm</A></td>
</tr>
<tr>
<td><A HREF="#SConnect">SConnect</A></td>
<td><A HREF="#SSendto">SSendto</A></td>
</tr>
</table>
</UL>
<H5>
<A NAME="AddrStrToAddr"></A>AddrStrToAddr</H5>
<TT>int</TT> <TT>AddrStrToAddr(const char * const s, struct sockaddr_in
* const sa, const int defaultPort);</TT>
<P>This takes a textual internet address specification and converts it
to a <TT>struct sockaddr_in</TT>.&nbsp; An address string may be any of
the following forms:
<UL>
<LI>
<TT>hostname:port</TT></LI>
<LI>
<TT>service://hostname</TT></LI>
<LI>
<TT>service://hostname:port</TT></LI>
<LI>
<TT>service://hostname[/more/junk/which/is/ignored/]</TT></LI>
<LI>
<TT>port@hostname</TT></LI>
</UL>
Additionally, the <I>hostname</I> may be a name or an IP address string
(i.e. <TT>192.168.1.13</TT>).
<P>If the string contains a hostname but a port number does not appear
to be present and the <I>defaultPort</I> parameter is greater than zero,
then the address structure will use <I>defaultPort</I> as the port number.
<P>The function returns a negative number if the string could not be converted,
such as the case where the hostname was not found in the name service.
Name service is not used unless there is a name instead of an IP address,
so if you don't want to use name service, only use IP addresses.
<P><I>Example</I>:
<UL>
<PRE>struct sockaddr_in addr;
int result;
result = AddrStrToAddr("ftp.probe.net", &amp;addr, 21);
result = AddrStrToAddr("ftp.probe.net:21", &amp;addr, 21);
result = AddrStrToAddr("206.28.166.234:21", &amp;addr, 21);
result = AddrStrToAddr("ftp://ftp.probe.net", &amp;addr, 0);
result = AddrStrToAddr("21@ftp.probe.net", &amp;addr, 0);</PRE>
</UL>
<UL>
<PRE></PRE>
</UL>
<H5>
<A NAME="#AddrToAddrStr"></A>AddrToAddrStr</H5>
<TT>char *AddrToAddrStr(char *const dst, size_t dsize, struct sockaddr_in
* const saddrp, int dns, const char *fmt);</TT>
<P>This function takes a <TT>struct sockaddr_in</TT> and converts into
a readable internet textual address.
<P>The <I>dns</I> parameter specifies if name service should be used to
lookup the symbolic hostname from the raw address.&nbsp; If zero, it expresses
the raw IP address in the standard dotted-decimal format, like 192.168.1.13.
<P>The <I>fmt</I> parameter is a magic cookie string, with the following
cookies:
<UL>
<LI>
<TT>%h</TT> = hostname</LI>
<LI>
<TT>%p</TT> = port number</LI>
<LI>
<TT>%s</TT> = service name (based off of the port number)</LI>
</UL>
This lets you print the address in just about any way you like.&nbsp; The
<I>dst</I> parameter is the string to write the results to, and is always
null-terminated.&nbsp; The <I>dst</I> parameter is also returned as the
result of the function.
<P><I>Example</I>:
<UL>
<PRE>char str[128];
fputs(AddrToAddrStr(str, sizeof(str), &amp;sin, 1, "%h"), stdout);
fputs(AddrToAddrStr(str, sizeof(str), &amp;sin, 1, "%h:%p"), stdout);
fputs(AddrToAddrStr(str, sizeof(str), &amp;sin, 1, "%s://%h"), stdout);</PRE>
</UL>
<H5>
<A NAME="#DisposeSReadlineInfo"></A>DisposeSReadlineInfo</H5>
<TT>void DisposeSReadlineInfo(SReadlineInfo *srl);</TT>
<P>This function is used to dispose of a <TT>SReadlineInfo</TT> structure
that was created using <TT><A HREF="#InitSReadlineInfo">InitSReadlineInfo()</A></TT>.&nbsp;
You're required to use this to free dynamically allocated buffers it creates
unless you specified that you wanted to use your own buffer, in which case
it's optional (but recommended for clarity).
<H5>
<A NAME="#FlushSReadlineInfo"></A>FlushSReadlineInfo</H5>
<TT>void FlushSReadlineInfo(SReadlineInfo *srl);</TT>
<P>This rarely used function is used to reset and clear the buffer used
by <TT><A HREF="#SReadline">SReadline()</A></TT>.&nbsp; It acts similarly
to a <TT>fflush(stdin)</TT>.
<H5>
<A NAME="#GetSocketBufSize"></A>GetSocketBufSize</H5>
<TT>int GetSocketBufSize(int sockfd, size_t *const rsize, size_t *const
ssize);</TT>
<P>This utility routine returns the size of the socket's internal buffers,
as maintained by the kernel (or socket library).&nbsp; It does this by
calling <TT>getsockopt()</TT> with the <TT>SO_RCVBUF</TT> and <TT>SO_SNDBUF</TT>
options, if they are defined.&nbsp; In the event they aren't defined, it
returns a negative result code.
<P><I>Example</I>:
<UL>
<PRE>size_t rsize, ssize;
if (GetSocketBufSize(sockfd, &amp;rsize, &amp;ssize) == 0) ...</PRE>
</UL>
<H5>
<A NAME="#GetSocketLinger"></A>GetSocketLinger</H5>
<TT>int GetSocketLinger(const int fd, int *const lingertime);</TT>
<P>This utility routine returns whether linger mode has been turned on,
and also sets the amount of time in the <I>lingertime</I> parameter.
<P><I>Example</I>:
<UL>
<PRE>int lingertime;
if (GetSocketLinger(sockfd, &amp;lingtime) > 0)
&nbsp;&nbsp;&nbsp; /* linger is on... */ ...;</PRE>
</UL>
<H5>
<A NAME="#GetSocketNagleAlgorithm"></A>GetSocketNagleAlgorithm</H5>
<TT>int GetSocketNagleAlgorithm(const int fd);</TT>
<P>This utility routine returns whether the <I>Nagle Algorithm</I> is in
effect (<TT>TCP_NODELAY</TT> mode not on).
<H5>
<A NAME="#InitSReadlineInfo"></A>InitSReadlineInfo</H5>
<TT>int InitSReadlineInfo(SReadlineInfo *srl, int sockfd, char *buf, size_t
bsize, int tlen);</TT>
<P>This function is used to prepare a <TT>SReadlineInfo</TT> structure
for use with the <TT><A HREF="#SReadline">SReadline()</A></TT> function.&nbsp;
The <I>sockfd</I> parameter specifies the socket that will be used for
line buffering.&nbsp; The <I>sio</I> library does not open or close this
socket.
<P>The buf parameter is the area of memory to use for buffering; it should
be large enough to hold several lines of data.&nbsp; If <I>buf</I> is NULL,
the function will <TT>malloc()</TT> one of <I>bsize</I> bytes, otherwise
it is assumed that buf is maintained by you and is of size <I>bsize</I>
bytes.&nbsp; If you let <I>sio</I> <TT>malloc()</TT> this buffer, you must
also use <TT><A HREF="#DisposeSReadlineInfo">DisposeSReadlineInfo()</A></TT>
to free it when you're finished with it.
<P>The <I>tlen</I> parameter is the timeout value (in seconds) to use for
each call of <TT><A HREF="#SReadline">SReadline()</A></TT>.
<P><I>Example</I>:
<UL>
<PRE>SReadlineInfo sri;
char localbuf[2048];
if (InitSReadlineInfo(&amp;sri, sockfd, NULL, 512, 10) &lt; 0)
&nbsp;&nbsp;&nbsp; perror("malloc 512 bytes failed");
...or...
(void) InitSReadlineInfo(&amp;sri, sockfd, localbuf, sizeof(localbuf), 10);</PRE>
</UL>
<H5>
<A NAME="SAccept"></A>SAccept</H5>
<TT>int SAccept(int sfd, struct sockaddr_in *const addr, int tlen);</TT>
<P>This is does an <TT>accept()</TT>, with a timeout of <I>tlen</I> seconds
(<I>tlen</I> may be zero to mean block indefinitely).&nbsp; If no new connection
is accepted, <TT>kTimeoutErr</TT> is returned, otherwise a new socket or
(-1) is returned.&nbsp; The socket is still usable if a timeout occurred,
so you can call <TT>SAccept()</TT> again.
<H5>
<A NAME="#SBind"></A>SBind</H5>
<TT>int SBind(int sockfd, const int port, const int nTries, const int reuseFlag);</TT>
<P>This calls <TT>bind()</TT>, to the wildcard address with the specified
<I>port </I>(not in network byte order).&nbsp; The <I>nTries</I> parameter
tells how many attempts it tries before giving up (which is useful if other
processes are grabbing the same port number).&nbsp; If the <I>reuseFlag</I>
parameter is non-zero, <TT>SBind()</TT> tries to turn on the <TT>SO_REUSEADDR</TT>
(and <TT>SO_REUSEPORT</TT>, if available) socket options before binding.
<P>Normally you will not call this function directly, since <TT><A HREF="#SNewStreamServer">SNewStreamServer()</A></TT>
and <TT><A HREF="#SNewDatagramServer">SNewDatagramServer()</A></TT> do
this for you.
<H5>
<A NAME="#SClose"></A>SClose</H5>
<TT>int SClose(int sfd, int tlen);</TT>
<P>This is <TT>close()</TT> with a timeout of <I>tlen</I> seconds.&nbsp;
Normally you don't need to worry about <TT>close()</TT> blocking, but if
you have turned on linger mode, <TT>close()</TT> could block.&nbsp; <TT>SClose()</TT>
calls <TT>close()</TT> for up to <I>tlen</I> seconds, and if the timeout
expires, it calls <TT>shutdown()</TT> on the socket.
<H5>
<A NAME="#SConnect"></A>SConnect</H5>
<TT>int SConnect(int sfd, const struct sockaddr_in *const addr, int tlen);</TT>
<P>This is <TT>connect()</TT> with a timeout of <I>tlen</I> seconds (<I>tlen</I>
may be zero to mean block indefinitely).&nbsp; If it returns (-1) or <TT>kTimeoutErr</TT>,
the socket is no longer usable and must be closed.
<H5>
<A NAME="#SConnectByName"></A>SConnectByName</H5>
<TT>int SConnectByName(int sfd, const char * const addrStr, const int tlen);</TT>
<P>This is <TT>connect()</TT> with a timeout of <I>tlen</I> seconds (<I>tlen</I>
may be zero to mean block indefinitely).&nbsp; If it returns (-1) or <TT>kTimeoutErr</TT>,
the socket is no longer usable and must be closed. The difference between
<TT><A HREF="#SConnect">SConnect()</A></TT> is that this function takes
a textual address string instead of a <TT>struct sockaddr_in</TT>.
<P><I>Example</I>:
<UL>
<PRE>if (SConnectByName(sockfd, "http://www.probe.net", 15) == 0) ...</PRE>
</UL>
<H5>
<A NAME="#SListen"></A>SListen</H5>
<TT>int SListen(int sfd, int backlog);</TT>
<P>This isn't too useful at present, since it just does <TT>listen(sfd,
backlog)</TT>.&nbsp; And, you will not call this function directly, since
<TT><A HREF="#SNewStreamServer">SNewStreamServer()</A></TT> and <TT><A HREF="#SNewDatagramServer">SNewDatagramServer()</A></TT>
do this for you.
<H5>
<A NAME="#SNewDatagramClient"></A>SNewDatagramClient</H5>
<TT>int SNewDatagramClient(void);</TT>
<P>This returns a new datagram socket, which is ready to send (and then
receive) datagrams.&nbsp; This function is just <TT>socket(AF_INET, SOCK_DGRAM,
0)</TT>. If successful, it returns a non-negative socket descriptor.
<P><I>Example</I>:
<UL>
<PRE>int sockfd;
sockfd = SNewDatagramClient();</PRE>
</UL>
<H5>
<A NAME="#SNewDatagramServer"></A>SNewDatagramServer</H5>
<TT>int SNewDatagramServer(const int port, const int nTries, const int
reuseFlag);</TT>
<P>This function creates a new socket and binds it to the specified <I>port</I>
(not in network byte order) on the wildcard address.&nbsp; The <I>nTries</I>
and <I>reuseFlag</I> are used when it calls <TT><A HREF="#SBind">SBind()</A></TT>.
If successful, it returns a non-negative socket descriptor.
<P><I>Example</I>:
<UL>
<PRE>int sockfd;
sockfd = SNewDatagramServer(13, 3, 0);
if (sockfd >= 0)
&nbsp;&nbsp;&nbsp; /* ready to receive requests on the daytime port (13) */</PRE>
</UL>
<H5>
<A NAME="#SNewStreamClient"></A>SNewStreamClient</H5>
<TT>int SNewStreamClient(void);</TT>
<P>
This returns a new stream socket, which is ready to <TT><A HREF="#SConnect">SConnect()</A></TT>
to a server.&nbsp; This function is just <TT>socket(AF_INET, SOCK_STREAM,
0)</TT>.&nbsp; If successful, it returns a non-negative socket descriptor.
<P><I>Example</I>:
<UL>
<PRE>int sockfd;
sockfd = SNewStreamClient();</PRE>
</UL>
<H5>
<A NAME="#SNewStreamServer"></A>SNewStreamServer</H5>
<TT>int SNewStreamServer(const int port, const int nTries, const int reuseFlag,
int listenQueueSize);</TT>
<P>This function creates a new socket, binds it to the specified <I>port</I>
(not in network byte order) on the wildcard address, and turns it on for
listening.&nbsp; The <I>nTries</I> and <I>reuseFlag</I> are used when it
calls <TT><A HREF="#SBind">SBind()</A></TT>. The <I>listenQueueSize</I>
is for the <TT>listen()</TT> call.&nbsp; If successful, it returns a non-negative
socket descriptor.
<P><I>Example</I>:
<UL>
<PRE>int sockfd;
sockfd = SNewStreamServer(80, 3, 0);
if (sockfd >= 0)
&nbsp;&nbsp;&nbsp; /* ready to accept HTTP connections */</PRE>
</UL>
<H5>
<A NAME="#SRead"></A>SRead</H5>
<TT>int SRead(int sfd, char *const buf0, size_t size, int tlen, int retry);</TT>
<P>This is <TT>read()</TT> on a socket descriptor, with a timeout of <I>tlen</I>
seconds (<I>tlen</I> must be greater than zero).&nbsp; Like <TT>read()</TT>,
it can return 0, (-1), or the number of bytes read, but in addition, it
can also return <TT>kTimeoutErr</TT>.
<P>If <I>retry</I> is set to <TT>kFullBufferRequired</TT>, <TT>SRead()</TT>
does not return until <I>size</I> bytes has been read or EOF is encountered.&nbsp;
This is useful if you expect fixed-length records, and it doesn't do much
good until a complete record has been read. However, it is still possible
that an EOF is encountered after some bytes have been read, and with <I>retry</I>
set to <TT>kFullBufferRequired</TT>, you get EOF instead of that partial
record.&nbsp; If you set retry to <TT>kFullBufferRequiredExceptLast</TT>,
you get the partial record (and EOF on the next <TT>SRead()</TT>).&nbsp;
Otherwise, if you should set retry to kFullBufferNotRequired and SRead()
will return when there is some data read.
<P><I>Example</I>:
<UL>
<PRE>int nread;
char buf[256];
while (1) {
&nbsp;&nbsp;&nbsp; nread = SRead(sockfd, buf, sizeof(buf), 15, kFullBufferNotRequired);
&nbsp;&nbsp;&nbsp; if (nread &lt;= 0) {
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; if (nread == 0)
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; break;&nbsp;&nbsp;&nbsp; /* okay, EOF */
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; else if (nread == kTimeoutErr) {
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; fprintf(stderr, "timed-out\n");
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; break;
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; } else {
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; perror("read");
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; }
&nbsp;&nbsp;&nbsp; }
&nbsp;&nbsp;&nbsp; (void) write(1, buf, nread);
}</PRE>
</UL>
<H5>
<A NAME="#SReadline"></A>SReadline</H5>
<TT>int SReadline(SReadlineInfo *srl, char *const linebuf, size_t linebufsize);</TT>
<P>It is often desirable to process data from sockets line by line, however
this is cumbersome to do on a socket descriptor.&nbsp; The <TT>SReadline()</TT>
function allows you to do this, so you can do one call of the function
and get back a line of input.&nbsp; It accomplishes this much the same
way the standard library's I/O routines do this, using a buffer.&nbsp;
However, it is usually not possible to determine if the standard library
takes the same special I/O measures on socket descriptors that sio does,
so using the standard library function <TT>fdopen()</TT> with a socket
descriptor may or may not work the way you want.
<P><TT>SReadline()</TT> needs to maintain state information for each call,
so a data structure is required.&nbsp; You must first call <TT><A HREF="#InitSReadlineInfo">InitSReadlineInfo()</A></TT>
to initialize a <TT>SReadlineInfo</TT> structure.&nbsp; After that, you
may call <TT>SReadline()</TT> repeatedly until it returns 0 to indicate
EOF. The function returns the number of characters in the input line, including
a newline (however, carriage return characters are omitted).&nbsp; You
must call <TT><A HREF="#DisposeSReadlineInfo">DisposeSReadlineInfo()</A></TT>
if you chose to let <TT><A HREF="#InitSReadlineInfo">InitSReadlineInfo()</A></TT>
use <TT>malloc()</TT> to allocate your buffer.
<P><I>Example</I>:
<UL>
<PRE>SReadlineInfo sri;
char line[80];
int nread;
if (InitSReadlineInfo(&amp;sri, sockfd, NULL, 512, 10) &lt; 0) {
&nbsp;&nbsp;&nbsp; perror("malloc");
&nbsp;&nbsp;&nbsp; exit(1);
}
while (1) {
&nbsp;&nbsp;&nbsp; nread = SReadline(&amp;sri, line, sizeof(line));
&nbsp;&nbsp;&nbsp; if (nread &lt;= 0) {
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; if (nread == kTimeoutErr)
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; fprintf(stderr, "readline timed-out.\n");
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; else if (nread &lt; 0)
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; perror("readline");
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; break;
&nbsp;&nbsp;&nbsp; }
&nbsp;&nbsp;&nbsp; line[nread] = '\0';&nbsp;&nbsp;&nbsp; /* omit newline */
&nbsp;&nbsp;&nbsp; fprintf(stdout, "read [%s]\n", line);
}
DisposeSReadlineInfo(&amp;sri);
(void) SClose(sockfd, 3);</PRE>
</UL>
<H5>
<A NAME="#SRecv"></A>SRecv</H5>
<TT>int SRecv(int sfd, char *const buf0, size_t size, int fl, int tlen,
int retry);</TT>
<P>This is the corresponding wrapper for <TT>recv()</TT>, as <TT><A HREF="#SRead">SRead()</A></TT>
is for <TT>read()</TT>.&nbsp; You will never need this function, unless
you need the special receiving flags that <TT>recv()</TT> gives you.
<H5>
<A NAME="#SRecvfrom"></A>SRecvfrom</H5>
<TT>int SRecvfrom(int sfd, char *const buf, size_t size, int fl, struct
sockaddr_in *const fromAddr, int tlen);</TT>
<P>This is <TT>recvfrom()</TT> with a timeout of <I>tlen</I> seconds (<I>tlen</I>
must be greater than zero).&nbsp; Like <TT>recvfrom()</TT>, it can return
0, (-1), or the number of bytes read, but in addition, it can also return
<TT>kTimeoutErr</TT>. Upon a timeout, the socket is still valid for additional
I/O.
<P><I>Example</I>:
<UL>
<PRE>int nread;
char buf[80];
struct sockaddr_in remoteClientAddr;
nread = SRecvfrom(sockfd, buf, sizeof(buf), 0, &amp;remoteClientAddr, 15);
if (nread &lt;= 0) {
&nbsp;&nbsp;&nbsp; if (nread == kTimeoutErr)
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; fprintf(stderr, "recvfrom timed-out.\n");
&nbsp;&nbsp;&nbsp; else if (nread &lt; 0)
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; perror("recvfrom");
}</PRE>
</UL>
<H5>
<A NAME="#SRecvmsg"></A>SRecvmsg</H5>
<TT>int SRecvmsg(int sfd, void *const msg, int fl, int tlen);</TT>
<P>This is the corresponding wrapper for <TT>recvmsg()</TT>, as <TT><A HREF="#SRead">SRead()</A></TT>
is for <TT>read()</TT>.
<H5>
<A NAME="#SSend"></A>SSend</H5>
<TT>int SSend(int sfd, char *buf0, size_t size, int fl, int tlen);</TT>
<P>This is the corresponding wrapper for <TT>send()</TT>, as <TT><A HREF="#SWrite">SWrite()</A></TT>
is for <TT>write()</TT>.&nbsp; You will never need this function, unless
you need the special receiving flags that <TT>send()</TT> gives you.
<H5>
<A NAME="#SSendto"></A>SSendto</H5>
<TT>int SSendto(int sfd, const char *const buf, size_t size, int fl, const
struct sockaddr_in *const toAddr, int tlen);</TT>
<P>This is <TT>sendto()</TT> with a timeout of <I>tlen</I> seconds (<I>tlen</I>
must be greater than zero).&nbsp; Like <TT>sendto()</TT>, it can return
0, (-1), or the number of bytes sent, but in addition, it can also return
<TT>kTimeoutErr</TT>. Upon a timeout, the socket is still valid for additional
I/O.
<P>Since <TT>sendto()</TT> rarely blocks (only if the outgoing queue is
full), you probably do not want to use a timeout or bother with its associated
overhead; therefore <TT><A HREF="#Sendto">Sendto()</A></TT> would be a
better choice.
<H5>
<A NAME="#SSendtoByName"></A>SSendtoByName</H5>
<TT>int SSendtoByName(int sfd, const char *const buf, size_t size, int
fl, const char *const toAddrStr, int tlen);</TT>
<P>This is <TT><A HREF="#SSendto">SSendto()</A></TT>, only you can use
a textual internet address string instead of a <TT>struct sockaddr_in</TT>.
<H5>
<A NAME="#SWrite"></A>SWrite</H5>
<TT>int SWrite(int sfd, const char *const buf0, size_t size, int tlen);</TT>
<P>This is <TT>write()</TT> on a network socket with a timeout of <I>tlen</I>
seconds (<I>tlen</I> must be greater than zero).&nbsp; Like <TT>write()</TT>,
it can return 0, (-1), or the number of bytes sent, but in addition, it
can also return <TT>kTimeoutErr</TT>.
<H5>
<A NAME="SelectR"></A>SelectR</H5>
<TT>int SelectR(SelectSetPtr ssp, SelectSetPtr resultssp);</TT>
<P>This does a <TT>select()</TT> for reading with the file descriptors
in the specified <TT>SelectSet</TT>.&nbsp; Using a <TT>SelectSet</TT> ensures
that the first argument to select is always correct (the smallest correct
value, for speed) and <TT>SelectR()</TT> does not destroy the original
<TT>fd_set</TT> and <TT>timeval</TT> (it copies it to <I>resultssp</I>
before using <TT>select()</TT>).
<P><I>Example</I>:
<UL>
<PRE>SelectSet ss, selected;
SelectSetInit(&amp;ss, 10.0);
SelectSetAdd(&amp;ss, sockfd1);
SelectSetAdd(&amp;ss, sockfd2);
rc = SelectR(&amp;ss, &amp;selected);
if ((rc > 0) &amp;&amp; (FD_ISSET(sockfd2, &amp;selected.fds))) ...</PRE>
</UL>
<H5>
<A NAME="SelectW"></A>SelectW</H5>
<TT>int SelectW(SelectSetPtr ssp, SelectSetPtr resultssp);</TT>
<P>This does a <TT>select()</TT> for writing with the file descriptors
in the specified <TT>SelectSet</TT>.&nbsp; Using a <TT>SelectSet</TT> ensures
that the first argument to select is always correct (the smallest correct
value, for speed) and <TT>SelectW()</TT> does not destroy the original
<TT>fd_set</TT> and <TT>timeval</TT> (it copies it to <I>resultssp</I>
before using <TT>select()</TT>).
<P><I>Example</I>:
<UL>
<PRE>SelectSet ss, selected;
SelectSetInit(&amp;ss, 10.0);
SelectSetAdd(&amp;ss, sockfd1);
SelectSetAdd(&amp;ss, sockfd2);
rc = SelectW(&amp;ss, &amp;selected);
if ((rc > 0) &amp;&amp; (FD_ISSET(sockfd2, &amp;selected.fds))) ...</PRE>
</UL>
<H5>
<A NAME="#SelectSetAdd"></A>SelectSetAdd</H5>
<TT>void SelectSetAdd(SelectSetPtr const ssp, const int fd);</TT>
<P>This adds a descriptor to the set to select on.
<P><I>Example</I>:
<UL>
<PRE>SelectSet ss;
SelectSetInit(&amp;ss, 10.0);
SelectSetAdd(&amp;ss, sockfd);</PRE>
</UL>
<H5>
<A NAME="#SelectSetInit"></A>SelectSetInit</H5>
<TT>void SelectSetInit(SelectSetPtr const ssp, const double timeout);</TT>
<P>Before adding members to the <TT>SelectSet</TT> structure, it must be
initialized with this function.&nbsp; The timeout parameter initializes
the timeout to use for future <TT>Selects</TT>.
<H5>
<A NAME="#SelectSetRemove"></A>SelectSetRemove</H5>
<TT>void SelectSetRemove(SelectSetPtr const ssp, const int fd);</TT>
<P>This removes a descriptor from the set to select on.&nbsp; You will
need to do this just before you close a descriptor that was in the set.
<H5>
<A NAME="#Sendto"></A>Sendto</H5>
<TT>int Sendto(int sfd, const char *const buf, size_t size, const struct
sockaddr_in *const toAddr);</TT>
<P>This is a simple wrapper for <TT>sendto()</TT>, which only handles <TT>EINTR</TT>
for you.&nbsp; It does not worry about <TT>SIGPIPEs</TT>.
<H5>
<A NAME="#SendtoByName"></A>SendtoByName</H5>
<TT>int SendtoByName(int sfd, const char *const buf, size_t size, const
char *const toAddrStr);</TT>
<P>This is a simple wrapper for <TT>sendto()</TT>, which only handles <TT>EINTR</TT>
for you.&nbsp; In addition, you can use a textual internet address string
instead of a <TT>struct sockaddr_in</TT>.
<P><I>Example</I>:
<UL>
<PRE>if (SendtoByName(sockfd, msg, sizeof(msg), "elwood.probe.net:13") > 0) ...</PRE>
</UL>
<H5>
<A NAME="#SetSocketBufSize"></A>SetSocketBufSize</H5>
<TT>int SetSocketBufSize(int sockfd, size_t rsize, size_t ssize);</TT>
<P>This utility routine changes the size of the socket's internal buffers,
as maintained by the kernel (or socket library).&nbsp; It does this by
calling <TT>setsockopt()</TT> with the <TT>SO_RCVBUF</TT> and <TT>SO_SNDBUF</TT>
options, if they are defined.&nbsp; In the event they aren't defined, it
returns a negative result code.&nbsp; The operation is only performed if
the size is greater than zero, so if you only wanted to change the receive
buffer you could set <I>rsize</I> to greater than zero and <I>ssize</I>
to 0.
<H5>
<A NAME="#SetSocketLinger"></A>SetSocketLinger</H5>
<TT>int SetSocketLinger(const int fd, const int l_onoff, const int l_linger);</TT>
<P>This is an interface to the <TT>SO_LINGER</TT> socket option.&nbsp;
The <I>l_onoff</I> parameter is a boolean specifying whether linger is
on, and the <I>l_linger</I> parameter specifies the length of the linger
time if enabled.
<P><I>Example</I>:
<UL>
<PRE>if (SetSocketLinger(sockfd, 1, 90) == 0) ...</PRE>
</UL>
<H5>
<A NAME="#SetSocketNagleAlgorithm"></A>SetSocketNagleAlgorithm</H5>
<TT>int SetSocketNagleAlgorithm(const int fd, const int onoff);</TT>
<P>This utility routine enables or disables the <I>Nagle Algorithm</I>
(<TT>TCP_NODELAY</TT> mode is off or on).&nbsp; Generally you won't care
about this, unless you're writing an interactive application like <I>telnet</I>,
<I>talk</I>, or <I>rlogin</I>, where response time is more important than
throughput.
<P><I>Example</I>:
<UL>
<PRE>if (SetSocketNagleAlgorithm(sockfd, 0) == 0) ...</PRE>
</UL>
<H3>
<A NAME="Sample"></A>Sample Code</H3>
Here is an example client and server that uses the <I>sio</I> library routines.&nbsp;
The server simply takes data and uppercases any lower case data, and returns
the result to the client.&nbsp; To try it on port number 5123, you could
run "<TT>ucase_s 5123</TT>" in one window, and "<TT>ucase_c localhost:5123</TT>"
in another.
<BR>&nbsp;
<H4>
Server:</H4>
<PRE>/* ucase_s.c */
#include &lt;unistd.h&gt;
#include &lt;sys/types.h&gt;
#include &lt;sys/socket.h&gt;
#include &lt;sys/wait.h&gt;
#include &lt;netinet/in.h&gt;
#include &lt;arpa/inet.h&gt;
#include &lt;errno.h&gt;
#include &lt;stdio.h&gt;
#include &lt;string.h&gt;
#include &lt;stdlib.h&gt;
#include &lt;time.h&gt;
#include &quot;sio.h&quot;
static void
ServeOneClient(int sockfd, struct sockaddr_in *cliAddr)
{
char buf[32], cliAddrStr[64];
int nread, nwrote, i;
printf(&quot;subserver[%d]: started, connected to %s.\n&quot;, (int) getpid(),
AddrToAddrStr(cliAddrStr, sizeof(cliAddrStr), cliAddr, 1, &quot;&lt;%h:%p&gt;&quot;)
);
for (;;) {
nread = SRead(sockfd, buf, sizeof(buf), 15, kFullBufferNotRequired);
if (nread == 0) {
break;
} else if (nread &lt; 0) {
fprintf(stderr, &quot;subserver[%d]: read error: %s\n&quot;,
(int) getpid(), strerror(errno));
break;
}
for (i=0; i&lt;nread; i++)
if (islower(buf[i]))
buf[i] = toupper(buf[i]);
nwrote = SWrite(sockfd, buf, nread, 15);
if (nwrote &lt; 0) {
fprintf(stderr, &quot;subserver[%d]: write error: %s\n&quot;,
(int) getpid(), strerror(errno));
break;
}
}
(void) SClose(sockfd, 10);
printf(&quot;subserver[%d]: done.\n&quot;, (int) getpid());
exit(0);
} /* ServeOneClient */
static void
Server(int port)
{
int sockfd, newsockfd;
struct sockaddr_in cliAddr;
int pid;
sockfd = SNewStreamServer(port, 3, kReUseAddrYes, 3);
if (sockfd &lt; 0) {
perror(&quot;Server setup failed&quot;);
exit(1);
}
printf(&quot;server[%d]: started.\n&quot;, (int) getpid());
for(;;) {
while (waitpid(-1, NULL, WNOHANG) &gt; 0) ;
newsockfd = SAccept(sockfd, &amp;cliAddr, 5);
if (newsockfd &lt; 0) {
if (newsockfd == kTimeoutErr)
printf(&quot;server[%d]: idle\n&quot;, (int) getpid());
else
fprintf(stderr, &quot;server[%d]: accept error: %s\n&quot;,
(int) getpid(), strerror(errno));
} else if ((pid = fork()) &lt; 0) {
fprintf(stderr, &quot;server[%d]: fork error: %s\n&quot;,
(int) getpid(), strerror(errno));
exit(1);
} else if (pid == 0) {
ServeOneClient(newsockfd, &amp;cliAddr);
exit(0);
} else {
/* Parent doesn't need it now. */
(void) close(newsockfd);
}
}
} /* Server */
void
main(int argc, char **argv)
{
int port;
if (argc &lt; 2) {
fprintf(stderr, &quot;Usage: %s &lt;port&gt;\n&quot;, argv[0]);
exit(2);
}
port = atoi(argv[1]);
Server(port);
exit(0);
} /* main */
</PRE>
<H4>
Client:</H4>
<PRE>/* ucase_c.c */
#include &lt;unistd.h&gt;
#include &lt;sys/types.h&gt;
#include &lt;sys/socket.h&gt;
#include &lt;sys/wait.h&gt;
#include &lt;netinet/in.h&gt;
#include &lt;arpa/inet.h&gt;
#include &lt;errno.h&gt;
#include &lt;stdio.h&gt;
#include &lt;string.h&gt;
#include &lt;stdlib.h&gt;
#include &lt;time.h&gt;
#include &quot;sio.h&quot;
static void
Client(char *serverAddrStr)
{
char buf[256];
int nread, nwrote, sockfd;
sockfd = SNewStreamClient();
if (sockfd &lt; 0) {
fprintf(stderr, &quot;client[%d]: socket error: %s\n&quot;,
(int) getpid(), strerror(errno));
exit(1);
}
if (SConnectByName(sockfd, serverAddrStr, 15) &lt; 0) {
fprintf(stderr, &quot;client[%d]: could not connect to &lt;%s&gt;: %s\n&quot;,
(int) getpid(), serverAddrStr, strerror(errno));
exit(1);
}
printf(&quot;client[%d]: connected to &lt;%s&gt;.\n&quot;, (int) getpid(), serverAddrStr);
for (buf[sizeof(buf) - 1] = '\0';;) {
printf(&quot;client[%d]: Enter message to send -&gt; &quot;, (int) getpid());
if (fgets(buf, sizeof(buf) - 1, stdin) == NULL)
break;
buf[strlen(buf) - 1] = '\0'; /* Delete newline. */
if (buf[0] == '\0')
continue; /* Blank line. */
/* Send the request line to the server. */
nwrote = SWrite(sockfd, buf, strlen(buf), 15);
if (nwrote &lt; 0) {
fprintf(stderr, &quot;client[%d]: write error: %s\n&quot;,
(int) getpid(), strerror(errno));
break;
}
/* Wait for complete reply line */
nread = SRead(sockfd, buf, nwrote, 15, kFullBufferRequired);
if (nread == 0) {
fprintf(stderr, &quot;client[%d]: no reply received (EOF).\n&quot;,
(int) getpid());
break;
} else if (nread &lt; 0) {
fprintf(stderr, &quot;client[%d]: read error: %s\n&quot;,
(int) getpid(), strerror(errno));
break;
}
buf[nread] = '\0';
fprintf(stdout, &quot;client[%d]: received: %s\n&quot;,
(int) getpid(), buf);
}
(void) SClose(sockfd, 10);
printf(&quot;\nclient[%d]: done.\n&quot;, (int) getpid());
exit(0);
} /* Client */
void
main(int argc, char **argv)
{
int port;
if (argc &lt; 2) {
fprintf(stderr, &quot;Usage: %s &lt;host:port&gt;\n&quot;, argv[0]);
exit(2);
}
Client(argv[1]);
exit(0);
} /* main */
</PRE>
</BODY>
</HTML>