The ACPI specifications can be found on the net, also have a look at wikipedia.

If the classname argument is actually an existing alias name, the alias command will create a new alias that is bound to the same Java classname as the existing alias.

A command class (e.g. one whose name is given by an aliased classname) needs to implement an entry point method with one of the following signatures:

public void execute(CommandLine cmd, InputStream in, PrintStream out, PrintStream err);

This is the JNode preferred entry point. It allows stream redirection in any invoker / interpreter that supports this. In addition, the CommandLine object provides the alias and arguments in both String and Token form.

public static void main(String[] args);

This is the classic Java entry point signature. It does not allow the command class to find out the alias name that was used to invoke it. In JNode, this makes stream redirection problematic unless the command is executed in an isolate or a proclet.

If a command class has both execute and main methods, most invokers will use the former in preference to the latter. Ideally, a command class should extend org.jnode.shell.AbstractCommand.

When a host needs to comminutate with an unknown local IP address, it broadcasts an ARP request on the local network, asking for the MAC address for the IP address. The node with the IP address broadcasts a response giving the MAC address for the network interface corresponding to the IP address.

The ARP cache stores IP to MAC address mappings that have previously been discovered. This allows the network stack to send IP packets without repeatedly broadcasting for MAC addresses.

The arp command allows you to examine the contents of the ARP cache, and if necessary clear it to get rid of stale entries.

• POSIX basename

The bindkeys --add ... form of the command adds a new binding. The <action> argument specifies an input editing action; e.g. 'KR_ENTER' causes the input editor to append a newline to the input buffer and 'commit' the input line for reading by the shell or application. The <vkSpec> or <character> argument specifies an input event that is mapped to the <action>.

The recognized <action> values are listed in the output the no-argument form of the bindkeys command.

The <vkSpec> values are formed from the "VK_xxx" constants defined by the "java.awt.event.KeyEvent" class and "modifier" names; e.g. "Shift+VK_ENTER".

The <character> values are either single ASCII printable characters or standard ASCII control character names; e.g. "NUL", "CR" and so on.

The bindkeys --add ... form of the command removes a single binding or (if you leave out the optional <vkSpec> or <character> argument) all bindings for the supplied <action>.

The bindkeys command provides no online documentation for what the action codes mean / do.

• --interactive | -i start an interactive BeanShell.
• --file | -f <file> take BeanShell input from <file>. If --interactive is also given, the BeanShell will go into interactive mode after processing the file.
• --code | -c <code> execute <code> as a BeanShell code. If --interactive is also given, the BeanShell will go into interactive mode after processing the code.

If no arguments are given, --interactive is assumed.

• BZip2

• If there are no arguments, cat copies from standard input to standard output until it encounters "end of file" on the input stream. If standard input is coming from the console, CTRL-D is used to denote end of file.
• If the arguments consist of file names, the corresponding file's contents are copied to standard output in the order specified.
• If the first argument is --urls or -u, the arguments are URLs. The cat opens each URL in turn and copies the resulting data streams to standard output. Supported URL protocols include "file:", "http:" and "ftp:".

The name "cat" is borrowed from UNIX, and is short for "concatenate".

JNode currently changes the "current" directory by setting the "user.dir" property in the system properties object.

If the <url> argument ends with a '/', it will be interpreted as a base directory that may contain classes and resources. Otherwise, the argument is interpreted as the path for a JAR file. While "file:" URLs are the norm, protocols like "ftp:" and "http:" should also work.

The --level option allows you to select the optimization level. The --test option allows you to compile with the "test" compilers.

This command is primarily used for native code compiler development. JNode will automatically run the native code compiler on any class that is about to be executed for the first time.

The first form of the console command list all consoles registered with the console manager. The listing includes the console name and the "F<n>" code for selecting it. (Use ALT F<n> to switch consoles.)

The second form of the console command starts and registers a new console running a new instance of CommandShell. If the --isolate option is used with --new, the new console's shell will run in a new Isolate.

The last form of the console command starts a raw text console without a shell. This is just for testing purposes.

Printing date / time values as UTC is not supported.

This command will not help your love life.

Normally, the del command will only delete a directory if it is empty apart from the '.' and '..' entries. The -r option tells the del command to delete directories and their contents recursively.

The first form of the device command list all devices registered with the device manager, showing their device ids, driver class names and statuses.

The second form of the device command takes a device id given as the <device> argument. It shows the above information for the corresponding device, and also lists all device APIs implemented by the device. Finally, if the device implements the "DeviceInfo" API, it is used to get further device-specific information.

The last form of the device command performs actions on the device denoted by the device id given as the <device> argument. The actions are as follows:

• start - starts the device.
• stop - stops the device.
• restart - equivalent to stop followed by start.
• remove - performs a stop and then removes the device from the device manager.

DHCP allocates IP address dynamically. A DHCP server will often allocate the same IP address to a given machine, but this is not guaranteed. If you require a fixed IP address for your JNode machine, you should use bootp or ifconfig. (And, if you have a DHCP service on your network, you need to configure it to not reallocate your machine's staticly assigned IP address.)

• POSIX dirname

The --level option allows you to select the optimization level. The --test option allows you to compile with the "test" compilers.

This command is primarily used for native code compiler development. Note, contrary to its name and description above, the command doesn't actually disassemble the class method(s). Instead it runs the native compiler in a mode that outputs assembly language rather than machine code.

The edit command displays a screen with two parts. The top part is the menu section; press ENTER to display the file action menu. The bottom part is the text editing window. The TAB key selects menu entries, and also moves the cursor between the two screen parts.

If the -e or --env option is given, the env command prints out the current shell environment variables. At the moment, this only works with the bjorne CommandInterpreter and the proclet CommandInvoker.

In theory, it should not be necessary to use this command. The garbage collector should run automatically at the most appropriate time. (A modern garbage collector will run most efficiently when it has lots of garbage to collect, and the JVM is in a good position to know when this is likely to be.)

In practice, it is necessary to run this command:

• to cope with GC-related issues; e.g. before running startawt,
• when debugging the garbage collector, and
• when doing performance-related benchmarking.

• Posix grep
• GNU grep

• grep -E and grep -B do not really work, instead they default to grep -P.
• grep -o/--only-matching does not work

• GNU zip

Currently, help prints command usage information and descriptions that it derives from a command's old or new-style argument and syntax descriptors. This means that (unlike Unix "man" for example), the usage information will always be up-to-date.

No help information is printed for Java applications which have no JNode syntax descriptors.

The second form of the history command finds and executes a command from the history list and executes it. If an <index> "i" is supplied, the "ith" entry is selected, with "0" meaning the oldest entry, "1" the second oldest and so on. If a <prefix> is supplied, the first command found that starts with the prefix is executed. The --test (or -t) flag tells the history command to print the selected command instead of executing it.

When the shell executes a command, the history list gets reordered in a rather non-intuitive way.

The first form prints the MAC address, assigned IP address(es) and MTU for all network devices. You should see the "loopback" device in addition to devices corresponding to each of your machine's ethernet cards.

The second form prints the assigned IP address(es) for the given <device>.

The final form assigns the supplied IP address and associated subnet mask to the given <device>.

When you attempt to bind an address, the output shows the address as "null", irrespective of the actual outcome. Run "ifconfig <device>" to check that it succeeded

The kdb command turns this on and off. Kernel debugging is normally off when JNode boots, but you can alter this with a bootstrap switch.

The leed command understands the following control functions:

• CTRL-S saves the file.
• CTRL-Q quits the editor.

The levi command understands the following control function:

• q quits the viewer.

A JNode keyboard interpreter is identified by a triple consisting of a 2 character ISO country code, together with an optional 2 character ISO language code and an optional variant identifier. Examples of valid country codes include "US", "FR", "DE", and so on. Examples of language code include "en", "fr", "de" and so on. (You can use JNode completion to get complete lists of the codes. Unfortunately, you cannot get the set of supported triples.)

When you run "loadkeys <country> ..., the command will attempt to find a keyboard interpreter class that matches the supplied triple. These classes are in the "org.jnode.driver.input.i10n" package, and should be part of the plugin with the same identifier. If loadkeys cannot find an interpreter that matches your triple, try making it less specific; i.e. leave out the <language> and <variant> parts of the triple.

Note: JNode's default keyboard layout is given by the "org/jnode/shell/driver/input/KeyboardLayout.properties" file. (The directory location in the JNode source code tree is "core/src/driver/org/jnode/driver/input/".)

Loadkeys should allow you to set the keyboard interpreter independently for each connected keyboard.

Loadkeys should allow you to change key bindings at the finest granularity. For example, the user should be able to (say) remap the "Windows" key to "Z" to deal with a broken "Z" key. This would allow you to configure JNode to use a currently unsupported keyboard type. (It would also help those game freaks out there who have been pounding on the "fire" key too much.)

The first form of the log4j command list the currently defined Loggers and there explicit or effective logging levels. An effective level is typically inherited from the "root" logger, and is shown in parentheses.

The second and third forms of the log4j command reload the log4j configurations from a file or URL.

The final form of the log4j command allows you to manually change logging levels. You can use completion to see what the legal logging levels and the current logger names are. If no <logger> argument is given, the command will change the level for the root logger.

The second form of the command mounts a file system. The file system on <device> is mounted as <directory>, with <fsPath> specifying the directory in the file system being mounted that will be used as the root of the file system. Note that the mount point given by <directory> must not exist before mount is run. (JNode mounts the file system as the mount point, not on top of it as UNIX and Linux do.)

When you run the command with no options, the output report shows the heap usage for all classes. This is typically too large to be directly useful. If you are looking for statistics for specific classes, you can pipe the output to the grep command and select the classes of interest with a regex. If you are trying to find out what classes are using a lot of space, you can use the onheap command's options to limit the output as follows:

• --minCount <count> : this option filters out any classes that have less than <count> instances.
• --minTotalSize <size> : this option filters out any classes for which the total space usage is less than <size> bytes.
• --className <substring> : this option filters out any classes whose name doesn't contain <substring>.

The command uses keyboard input to control paging. For example, a space character advances one screen page and ENTER advances one line. Enter 'h' for a listing of the available pager commands and actions.

A number useful features supported by typical 'more' and 'less' commands have not been implemented yet.

The no argument form of the command lists all plugins known to the system, showing each one's status.

The one argument form lists a single plugin.

The --load, --unload and --reload options tell the plugin command to load, unload or reload a specified plugin. The --load and --reload forms can also specify a version of the plugin to load or reload.

The --addLoader option configures a new plugin loader that will load plugin from the location given by the <url>.

The System property space consists of the properties returned by "System.getProperty()". This space are currently isolate-wide, but there are moves afoot to make it proclet specific. The Shell property space consists of properties stored by each Shell instance. This space is is separate from an shell interpreter's variable space, and persists over changes in a Shell's interpreter.

The 'set' command is an alias for 'propset', but if you are using the 'bjorne' interpreter the 'set' alias is obscured by the POSIX 'set' builtin command which has incompatible semantics. Hence 'propset' is the recommended alias.

The --create form of the command creates a new RAM disk with a size in bytes given by the --size option The default size is 16K bytes. Note that the RAM disk has a notional block size of 512 bytes, so the size should be a multiple of that.

• --host | -h <host> - this option specifies the hostname or IP address of the host running the remote receiver.
• --port | -p <port> - this option specifies the port number for the remote receiver. The default port is 5612.
• --udp | -u - this option tells remoteout to use UDP. The default is to use TCP. UDP is inherently lossy, but TCP mode currently does not allow logger output to be captured; see Bugs.

Before you run remoteout on JNode, you need to start a TCP or UDP receiver on the relevant remote host and port. The JNode codebase includes a simple receiver application implemented in Java. You can run as follows:

    java -cp $JNODE/core/build/classes org.jnode.debug.RemoteReceiver &

Running the RemoteReceiver application with the --help option will print out a "usage" message.

Notes:

1. Before you run remoutout, you will need to configure JNode networking.
2. If you have problems getting remoteout to work, check that network access is not being blocked by a network firewall or local firewall software on the remote host. (TCP and UDP access on non-standard ports like 5612 is often blocked by default.)

Logger output redirection is disabled in TCP mode due to a bug that triggers kernel panics.

There is currently no way to turn off console/logger copying once it has been started.

Running remoteout and a receiver on the same JNode instance, may cause JNode to lock up in a storm of console output.

The zero argument form of resolver list the IP addresses of the DNS servers in the order that they are used.

The --add form adds a DNS aerver (identified by a numeric IP address) to the front of the resolver list.

The --del form removes a DNS server from the resolver list.

The route command manages the routing table. The no-argument form of the command lists the current routing table. The --add and --del add and delete routes respectively.

For more information on how to use route to configure JNode networking, refer to the FAQ.

#!<interpreter>

where <interpreter> is the name of a registered CommandInterpreter, the script will be run using the nominated interpreter. Otherwise, the script will be run using the shell's current interpreter.

There is more information on the JNode GUI page, including information on how to exit the GUI.

The built-in syntax for a command is typically specified in the plugin descriptor for a parent plugin of the command class. If there is no explicit syntax specification, a default one will be created on-the-fly from the command's registered arguments.

Note: not all classes use the new syntax mechanism. Some JNode command classes use an older mechanism that is being phased out. Other command classes use the classic Java approach of decoding arguments passed via a "public static void main(String[])" entry point.

• GNU tar

In "client mode", the tcpinout command opens a connection to the supplied <host> and <port>. This assumes that there is a service on the remote host that is "listening" for connections on the port.

In "server mode", the tcpinout command listens for an incoming TCP connection on the supplied <local port>.

The first form of the command traverses the ThreadGroup hierarchy, displaying information for each Thread that it finds. The information displayed consists of the Thread's 'id', its 'name', its 'priority' and its 'state'. The latter tells you (for example) if the thread is running, waiting on a lock or exited. If the Thread has died with an uncaught exception, you will also see a stacktrace.

If you set the --groupDump flag, the information is produced by calling the "GroupInfo.list()" debug method. The output contains more information but the format is ugly.

The second form of the thread command outputs information for the thread given by the <threadName> argument. No ThreadGroup information is shown.

The second form of the command should set a non-zero return code if it cannot find the requested thread.

There should be a variant for selecting Threads by 'id'.

• INFO-Zip

• POSIX wc

• INFO-Zip