Stem Demonstration Scripts
Stem comes with several demonstration scripts and example configuration
files which are used by them. You can optionally install the executable
demonstrations and their configuration files. Note that the actual
demonstration scripts don't do anything special to Stem, they just
create windows and execute run_stem inside them with selected
configuration files. They also create telnet connections inside other
windows which you can use to interact with Stem. You can manually create
the windows and do the same commands, these scripts are just
conveniences. In fact, a good way to learn more about Stem is to copy
and modify the configuration files used in the demonstrations and run
them yourself.
When you run any of the demo scripts, the commands used to fork off an
xterm are printed. You can manually run those commands in your own
terminal windows if you want to experiment with or explore the Stem
configurations. If you kill the script (e.g. with ^C), all the created
xterm windows will be killed off leaving you with a cleaned up desktop.
There are 4 demonstration scripts that come with Stem. They are briefly
described here and in more detail in other files. They all have some
common features which are also described in detail below.
chat_demo and chat2_demo demonstrate a simple 4 user chat
server. chat_demo runs with a single Stem Hub and chat2_demo
uses 2 Hubs. Both bring up an xterm for each Stem Hub and 4 more
for the telnet sessions. Read DEMO_CHAT for the full details on
how to use this demo.
inetd_demo emulates the inetd Unix super-daemon. It runs a
single Stem Hub in an xterm and 4 telnet sessions each in their
own xterm. The server process it runs is proc_serv in the bin
directory. You can run it directly from the command to see how
it behaves (it is a simple command/response program). Read
DEMO_INETD for the full details on how to use this demo.
tail_demo monitors a typical log file and copies any new data it
finds there to a Stem Logical Log which writes it to a file and
optionally to other destinations. The status of the source file
is also sent to a Logical Log. Read DEMO_TAIL for the full
details on how to use this demo.
Using the console Cell Stem::TtyMsg
All of the demo configurations include the Stem::TtyMsg module which
allows you to enter command messages from the keyboard to a running Stem
Hub (process). This module is not required to run Stem but it is in the
demo configurations so you can interact with Stem and learn more about
it.
It reads lines from STDIN (using the Stem::AsyncIO module so the rest of
the Stem Hub continues to run), parses them and sends out a command
messages based on the lines. It also can set key/values in the local
Hub's environment (%Stem::Vars::Env) which is used to control many Stem
operations.
Command messages can generate response messages which will be sent back
to the TtyMsg Cell. These responses will be printed to STDOUT (again,
using the Stem::AsyncIO module). Any Cell can just send a data message
to the TtyMsg Cell (which is also registered with the class Cell name
tty) and its data will get printed.
The rules for parsing lines input to TtyMsg are very simple. There are
three kinds of command lines:
Direct commands
The only direct command is 'help' which has to be the
only token on the line. It causes the help text to be
printed.
Setting a Stem environment variable
Key/values in the local Hub's environment can be set
with lines of the form:
name=value
Token has to be only word characters ([a-zA-Z0-9_]) and
the string after the = is the value (stripped of
surrounding white space). The Hub environment variable
with the name token is set to the parsed value. The
token and value are printed.
You can also set any environment variable in any remote
Hub with the command message:
hub:var set_env name=value
Note that 'hub' must be the real name of that Stem Hub,
and var is already the registered class name of the
Stem::Vars class Cell.
See below for more on entering command messages and the
env_notes document in /Design.
Sending a Command Message
A command message line starts with a Cell address and
then must have a command token. The rest of the line is
optional data for the command message. A minimal Cell
address is just a Cell name. It can have an optional Hub
name prefix. Also a target name can be suffixed after a
trailing :. So the only legal Cell addresses look like
this:
cell
hub:cell
:cell:target
hub:cell:target
If the Hub is missing the message is destined for the
local Hub and if the Cell doesn't exist here, it is
routed to the DEFAULT Hub. Read the Cell and Message
technical note for more on this.
The next token on a message command line is the command
name and it is required. It will be the value of the
'cmd' field in the message. The rest of the line is used
as the data field of the message.
Some uses of command line messages are getting the
status of various Class Cells since almost all of them
have a status command. By listing all the registered
Cells you can see which ones you can send messages to.
This will print all of the Cells in this hub. The
listing shows all object Cell and Class Cells with their
aliases. Command line messages should use the aliases
for the Cell name as the class names have colons in
them.
This will print all of the Cells in the local hub.
reg status
This will print all of the Cells in the hub named remote.
remote:reg status
This will print all of the Portals in this hub. You can
use their hub names to send command messages to those
hubs.
port status
If you are running either chat demo you can change the
Switch maps which control which user get sent chat
messages. Here are some examples. The Switch Cell is
named sw, and the two Hubs in chat2_demo are named
chat_client and chat_server. Note that the 'sw' Cell is
only in the server Hub in chat2_demo, but since no 'sw'
Cell exists in the client Hub, any message sent to 'sw'
will still go to the server hub. So all of these map
commands can be issued from either Hub input and they
will work. The 'map' or 'status' token is the command
and the tokens after 'map' are data to the map
command. The first data token is the input map you are
setting and the rest of the tokens are the output maps
to send chat messages to.
Print the current maps.
sw status
Change the a map to just b.
sw map a b
Change the d map to all users.
sw map d a b c d