userguide.html

P2P NAP的C实现 P2P在网络应用中越来越流行

quit.

<h4><a name=4.8.></a>4.8. Napster vs. OpenNap</h4>

Since June 2001, Napster Inc. has modified their servers to use a new
client authentification scheme, and nap users can no longer connect to
Napster Inc.'s servers. Thus, nap is configured to connect to the
OpenNap servers by default. The OpenNap system is a network of napster
servers that are independently owned (not run by Napster, the
company). When starting up, nap will automatically download a list of
servers from www.napigator.com, and it will connect to the first
available one. You can change this behavior if you want; see <a
href=#7.3.>7.3. Meta-servers</a>.

<h4><a name=4.9.></a>4.9. News about new releases</h4>

At startup, nap will check online whether a newer release is
available. If this is the case, it will inform you of the new release
while starting up, as a gentle reminder to upgrade. If you never want
to be informed of new releases in this way, put <code>nonews=1</code>
in your config file or start nap with the option
<code>-ononews</code>. You can also check for news from the main
screen by typing <code>/news</code>.

<h4><a name=4.10.></a>4.10. If you are behind a firewall</h4>

Most firewalls allow only outgoing connections, not incoming ones. The
napster protocol can compensate for this to a certain extent, by
arranging things so that firewalled clients initiate connections to
non-firewalled clients, and not the other way around. In this way,
both uploads and downloads are possible between a firewalled and a
non-firewalled client. However, between two clients that are both
firewalled, neither uploads nor downloads are possible.

<p>

If you are behind a firewall, then nap needs to know about it, in
order to be able to compensate for it. However, nap can't find this
out on its own. The way to tell nap that you are behind a firewall is
to set <code>dataport = 0</code> in your configuration file (see <a
href=#7.>7. Configuration and user variables</a>).

<p>

If you are behind a firewall, you may also need to set up an http
proxy so that nap can read its server list; see <a
href=#7.4.>7.4. Http proxies</a>.

<!---------------------------------------------------------------------->
<h2><a name=5.></a>5. More on searching and downloading</h2>

<h4><a name=5.1.></a>5.1. Advanced searches <font size=3 color=#00ff00><b>[updated 1.5.1]</b></font></h4>

The <code>/search</code> command has several options. They are:
<p><pre>
        -m <i>n</i>            return at most <i>n</i> search results
        -l              local search: do not search on linked servers
        -p              toggle sending ping requests
        -f              list full filenames (only if noresultscreen=1)

        -b=<i>rate</i>, -b&lt;<i>rate</i>, -b&gt;<i>rate</i>   
                        bitrate equal to, at most, or at least <i>rate</i>.

        -c=<i>speed</i>, -c&lt;<i>speed</i>, -c&gt;<i>speed</i>   
                        link speed equal to, at most, or at least <i>speed</i>.

        -r=<i>freq</i>, -r&lt;<i>freq</i>, -r&gt;<i>freq</i>   
                        sample rate equal to, at most, or at least <i>freq</i>.

        -s=<i>size</i>, -s&lt;<i>size</i>, -s&gt;<i>size</i>   
                        file size equal to, at most, or at least
                        <i>size</i>, in bytes (opennap and slavanap only).

        -d=<i>duration</i>, -s&lt;<i>duration</i>, -s&gt;<i>duration</i>   
                        duration equal to, at most, or at least
                        <i>duration</i>, in seconds (opennap and slavanap only).
			<font size=3 color=#00ff00><b>[new 1.5.1]</b></font>

	-x <i>keyword</i> exclude results that match keyword (opennap and 
			slavanap only). Multiple -x options are possible.
			<font size=3 color=#00ff00><b>[new 1.5.1]</b></font>

        -t <i>filetype</i> search for the specified filetype, which must
	                be one of: audio, video, text, image,
	                application, mp3, any.
			<font size=3 color=#00ff00><b>[new 1.5.1]</b></font>
</pre>
The "-s" option only works on OpenNap servers; it has no effect on Napster.

<p>

The "-l" option will cause the search to only be performed on the
server you're connected to, and not on any other servers that might be
linked to it in a network. 

<p>

The "-m" option limits the number of results returned. However, even
if <i>n</i> is greater than 100, most servers will only return 100
results.

<h4><a name=5.2.></a>5.2. Pings</h4>

Pings are test packages that are sent to other clients to see how "far
away" they are in the internet. A ping result is the time, in
milliseconds, that it took the package to travel to the other client
and back. Normally, nap will try to collect ping results with each
search request. You can switch off pinging by giving the "-p" option
to the <code>/search</code> command, or by setting the user variable
<code>noping</code> to 1. 

<p>

Nap implements pings via a separate helper application called
"napping". The reason is that sending and receiving ping
packets requires raw network protocol access, which usually requires
root privileges on Linux. However, it would be dangerous to run nap
with root privileges. The program "napping", written by
Sebastian Zagrodzki, solves this problem. Napping is a very small
program which is called by nap to handle pings. Napping, unlike nap,
can be run "suid", i.e., with root privileges. This is safe because
napping drops root privileges immediately after opening the network
socket, which is in fact the first thing it does.

<p>

For napping to work, it must be installed somewhere in your $PATH, it
must be owned by root, and it must have "suid" permissions. Thus, the
permissions of napping should look like this:
<pre>
        -rwsr-xr-x    1 root     root        48417 Aug 29 17:54 napping*
</pre>
The "s" in the permissions is called the "suid" bit, and it means that
napping will be run with the effective user id of root, rather than
that of the user who actually runs it. You can set the right ownership
and permissions with the following commands (assuming that napping is
installed in /usr/local/bin):
<pre>
        chown root:root /usr/local/bin/napping
	chmod 4755 /usr/local/bin/napping
</pre>
You can specify an alternative name for the napping program by setting
the <code>napping</code> user variable. You can also use this to
specify an absolute filename, which is useful in case napping is
installed outside your $PATH (see <a href=#7.2.>7.2. User
variables</a>).

<h4><a name=5.3.></a>5.3. Advanced navigation of the search result screen</h4>

By default, the result screen displays the bitrate and length of each
song, as well as the connection speed and/or ping results. However,
you have full control over what information is displayed: The
following keys toggle each category of information on and off:
<pre>
        b - bitrate
	l - length
	m - megabytes
	f - frequency
	u - user name
	s - speed
	p - ping
</pre>
The "m" key also toggles between displaying the file size in megabytes
and in bytes. Moreover, if you press "a", all categories of
information will be turned on, and if you press "n", they will all be
turned off (i.e., you will only see the filenames).

<p>

The search results can also be sorted by various different
criteria. Pressing the uppercase equivalents of the above keys ("B",
"L", etc) will sort the search results by that particular category
(bitrate, length...). You can also press "N" (uppercase) to sort by
filename, and "D" to sort by directory name. Sorting by directory name
is useful if there are many files whose names start with "01", "02",
and so on. If you press a sort key more than once, the items will
alternately be sorted in ascending and descending order.


<p>

When you sort, the cursor will remain on the same item that it was
on. Thus, if your cursor is on an item shared by a particular user,
and you press "U", you will see other items of the same user
nearby. Notice that after sorting, the search results are re-numbered
starting from 1. Any subsequent <code>/get</code> commands must use
the new numbering. 

<p>

If you prefer an appearance of the search result screen different from
the default appearance, you can set the <code>sdefaults</code> user
variable to a string of key strokes executed when that screen is
initialized. The default is "blsp". You can set it, for instance, to
"lspmm" if you don't want to see the bitrate, but instead the file
size in bytes. You can also set it to something like "blspN" if you
like your results sorted by name, rather than by connection
speed. Note that ping results will only be displayed if they are
actually available, and connection speeds will not be displayed for
browse results.

<h4><a name=5.4.></a>5.4. Shortcut searches</h4>

A convenience features allows you to enter a new search directly from
the search result screen or download/upload screen. Just type "space",
and an input window will open in the status line with the prefix
"/search " entered for you. You can also type '/' to take you back to
the main screen and start a new command (a shortcut for 'q', '/').

<h4><a name=5.5.></a>5.5. Incomplete downloads and turds</h4>

Files that are currently downloading are kept in your "incomplete"
directory. They also have the suffix ".incomplete" added to their
filename (or alternatively the suffix defined by the user variable
"incompletesuffix", if any). When a download has been successfully
completed, the suffix is removed and the file is moved to your
"download" directory.

<p>

Nap will, by default, delete any incomplete files whose size is 100000
bytes or smaller. Nap considers such files "turds": they are too small
to be of any value, and just tend to clutter up your hard drive. You
can override the default cutoff of 100000 bytes by setting the user
variable "turdsize" (see <a href=#7.2.>7.2. User variables</a>). In
particular, if you set turdsize to 0, only empty files will be
deleted, and if you set it to -1, no files will be deleted at all.

<p>

Note that the turdsize only applies to <i>incomplete</i> files.  Nap
will never delete a completed file, even if the file size is very
small.

<p>

Note that the convention on incomplete files makes it possible to
choose the same directory as your download directory and your
incomplete directory (incomplete files will still be recognizable
because they have the suffix ".incomplete"). On the other hand, you
may choose to include your download directory in your upload path.
However, you should never include your incomplete directory in your
upload path, as people will probably not be too happy about
downloading your incomplete files.

<p>

If you try to download two files with identical names, nap will
automatically rename one of the files before saving them.

<h4><a name=5.6.></a>5.6. Upload and download limits and queueing</h4>

If <code>maxuploads</code> is set, it defined a limit on the total
number of simultaneous uploads. A per-user limit can be set in the
variable <code>maxupuser</code>. 

You can limit the maximal number of simultaneous uploads and
downloads, both globally and for each remote user.  To limit the
number of simultaneous uploads or downloads, set the user variables
<code>maxuploads</code> and <code>maxdownloads</code> to the desired
number. To limit the number of simultaneous uploads or downloads from
each user, use the user variables <code>maxupuser</code> and
<code>maxdownuser</code>. By default, there are no limits.

<p>

When the download limit is reached, nap will queue any additional
downloads, and it will start them as soon as space is available.
Queued downloads will be listed along with all other downloads on the
download/upload screen. If you want to force a queued download to be
started immediately, you can do so by pressing 'f' on the
download/upload screen.

<h4><a name=5.7.></a>5.7. Purging and retrying failed items</h4>

By default, items which are stopped (finished, interrupted, timed out,
or failed) are not automatically removed from the download list. They
remain on the list until you remove them explicitly. Individual items
can be removed from the download screen by pressing 'd'. You can also
press 'P' to purge all stopped items in the current section (i.e., all
uploads or all downloads). 

<p>

If you prefer stopped items to be automatically purged, you can set
the user variable <code>autopurge</code> to 0. You can also set
<code>autopurge</code> to a positive number, if you want items to be
purged after this many seconds. The variables <code>autopurgeup</code>
and <code>autopurgedown</code> allow you to set these parameters for
uploads and downloads separately.

<p>

A download which has been interruped, or which has timed out or
failed, can be retried by pressing 'r' on the item on the download
screen. You can also retry all such items with a single keystroke by
pressing 'R'.

<h4><a name=5.8.></a>5.8. Limiting bandwidth</h4>

It is possible to limit the bandwidth of up- and downloads. It is
possible to set a limit on the bandwidth per connection, and to set an
overall limit on the total bandwidth taken by all connections. This is
done separately for uploads and downloads. Up- and download bandwidth
limits are given in kilobytes per second, and they are set via the
user variables <code>bandwidthup</code>, <code>bandwidthup1</code>,
<code>bandwidthdown</code>, <code>bandwidthdown1</code>. See <a
href=#7.2.>7.2. User variables</a>.

<p>

It is possible to change the bandwidth limits at any time. Such
changes will immediately take effect, even for transfers that are
already in progress. Note that bandwidth limits for downloads only
affect the rate at which nap reads data, not the rate at which data
arrives from the network. If data arrives on the network at a higher
rate than nap reads it, then the operating system will usually buffer
such data. Thus, the actual network bandwidth may not go down until
the buffer is full.

<p>