Note: in this manual the directory in which PEAR has
been installed is referred to with the $PEAR_HOME
notation,
while $HOME
is the user's home directory.
PEAR can be run through a couple of scripts
that may be found in the $PEAR_HOME/bin
directory: the Windows
version is run.bat
, while run.sh
is the UNIX/Linux
one. Both scripts load the modules found in $PEAR_HOME/lib
and set the Java Virtual Machine classpath so that any plug-in
will be looked for in the $PEAR_HOME/plugins
directory.
PEAR's standard ouput and error are redirected into two log file that
can be found in the $HOME/.pear
directory.
PEAR provides a graphic interface whose windows are dockable, that is they can be attached to the four sides of the main desktop through the Window menu, which also provides the automatic window disposal by means of the Dispose command.
Through the View menu the hidden windows corresponding to the selected items can be shown or taken to the front; the Validation Rules item is a submenu letting you choose whether show the actual window, or the inner elements: the source code of the validation rules, or the corresponding tree view.
The Container window shows the status of the containers and the substitution map during a validation session in step-by-step mode, Console shows the content of the log console, while Protocol Source holds a tab for each open protocol source.
Three tabs are placed below the desktop:
the desktop tab is the one that has been described since
now; console contains a text area showing the logger
output and the controls you can use to clear the text area and set the
logger actual verbosity; eventually the errors tab is
displayed when an unhandled error is detected.
Below the tabs you can find the status bar, showing the last message produced by the tool, while above them the tool bars are placed.
Four tool bars are available:
If you edit a protocol source and then you try to close it, an inner
panel will be shown up asking if you wish to save the modified source.
An analogous element is shown up when
the tag inference process is able to find one or more tagging suiting
to the given protocol: a component appears in the inner panel allowing
you to choose whether display the original source or the tagged ones.
The tag inference window provides controls above and below the current
source displaying area: the first control above allows you to choose whether
display the original source or the taggings found; the patience
slider allows you the select the number of equal taggings to be found
before acting on the stack of backtrack nodes; the two following check-boxes
enable or disable (if possible) the commit hints and the protocol reordering
algorithm (it improves performances).
The panel placed below the source displaying area contains four buttons
which provide the following actions respectively: run the tag inference
process, stop it, save the current tagging as a separate file, close the
tag inference window.
Operation and container plug-ins don’t require
configuration: if you wish to add a jar module, just put it in the $PEAR_HOME/plugins
directory; if you wish to add only a single class or resource, just place
it in $PEAR_HOME/plugins/it/unive/dsi/pear/plugins
. If you
are running PEAR through WebStart, you can put your plug-in jar modules
in $HOME/.pear
. Please note that in
this case you may provide only resource plugins./
plugins
Built-in resources may be specified trough two related
list files named
and
builtin-vrules.inibuiltin-protocols.ini
, respectively.
You can find the former in the builtin.jar
module, while
the latter is stored in the protocols.jar
module. Once properly
written the two lists, you'll have to make listed resources available
by treating them as standard plugins. Alternatively you may provide no
list, this way all the resources stored in any plugin module will be displayed
by the resource browser.