As an abstract base class, App::Info is not intended to be used directly. Instead, you'll use concrete
subclasses that implement the interface it defines. These subclasses each provide the meta data necessary
for a given software package, via the interface outlined above (plus any additional methods the class
author deems sensible for a given application).
This section describes the facilities App::Info provides for subclassing. The goal of the App::Info
design has been to make subclassing straight-forward, so that developers can focus on gathering the data
they need for their application and minimize the work necessary to handle unknown values or to confirm
values. As a result, there are essentially three concepts that developers need to understand when
subclassing App::Info: organization, utility methods, and events.
Organization
The organizational idea behind App::Info is to name subclasses by broad software categories. This
approach allows the categories themselves to function as abstract base classes that extend App::Info, so
that they can specify more methods for all of their base classes to implement. For example,
App::Info::HTTPD has specified the "httpd_root()" abstract method that its subclasses must implement. So
as you get ready to implement your own subclass, think about what category of software you're gathering
information about. New categories can be added as necessary.
UtilityMethods
Once you've decided on the proper category, you can start implementing your App::Info concrete subclass.
As you do so, take advantage of App::Info::Util, wherein I've tried to encapsulate common functionality
to make subclassing easier. I found that most of what I was doing repetitively was looking for files and
directories, and searching through files. Thus, App::Info::Util subclasses File::Spec in order to offer
easy access to commonly-used methods from that class, e.g., "path()". Plus, it has several of its own
methods to assist you in finding files and directories in lists of files and directories, as well as
methods for searching through files and returning the values found in those files. See App::Info::Util
for more information, and the App::Info subclasses in this distribution for usage examples.
I recommend the use of a package-scoped lexical App::Info::Util object. That way it's nice and handy when
you need to carry out common tasks. If you find you're doing something over and over that's not already
addressed by an App::Info::Util method, consider submitting a patch to App::Info::Util to add the
functionality you need.
Events
Use the methods described below to trigger events. Events are designed to provide a simple way for
App::Info subclass developers to send status messages and errors, to confirm data values, and to request
a value when the class cannot determine a value itself. Events may optionally be handled by module users
who assign App::Info::Handler subclass objects to your App::Info subclass object using the event handling
methods described in the "Event Handler Object Methods" section.
info
$self->info(@message);
Use this method to display status messages for the user. You may wish to use it to inform users that
you're searching for a particular file, or attempting to parse a file or some other resource for the data
you need. For example, a common use might be in the object constructor: generally, when an App::Info
object is created, some important initial piece of information is being sought, such as an executable
file. That file may be in one of many locations, so it makes sense to let the user know that you're
looking for it:
$self->info("Searching for executable");
Note that, due to the nature of App::Info event handlers, your informational message may be used or
displayed any number of ways, or indeed not at all (as is the default behavior).
The @message will be joined into a single string and stored in the "message" attribute of the
App::Info::Request object passed to info event handlers.
error
$self->error(@error);
Use this method to inform the user that something unexpected has happened. An example might be when you
invoke another program to parse its output, but it's output isn't what you expected:
$self->error("Unable to parse version from `/bin/myapp -c`");
As with all events, keep in mind that error events may be handled in any number of ways, or not at all.
The @erorr will be joined into a single string and stored in the "message" attribute of the
App::Info::Request object passed to error event handlers. If that seems confusing, think of it as an
"error message" rather than an "error error." :-)
unknown
my $val = $self->unknown(@params);
Use this method when a value is unknown. This will give the user the option -- assuming the appropriate
handler handles the event -- to provide the needed data. The value entered will be returned by
"unknown()". The parameters are as follows:
key The "key" parameter uniquely identifies the data point in your class, and is used by App::Info to
ensure that an unknown event is handled only once, no matter how many times the method is called. The
same value will be returned by subsequent calls to "unknown()" as was returned by the first call, and
no handlers will be activated. Typical values are "version" and "lib_dir".
prompt
The "prompt" parameter is the prompt to be displayed should an event handler decide to prompt for the
appropriate value. Such a prompt might be something like "Path to your httpd executable?". If this
parameter is not provided, App::Info will construct one for you using your class' "key_name()" method
and the "key" parameter. The result would be something like "Enter a valid FooApp version". The
"prompt" parameter value will be stored in the "message" attribute of the App::Info::Request object
passed to event handlers.
callback
Assuming a handler has collected a value for your unknown data point, it might make sense to validate
the value. For example, if you prompt the user for a directory location, and the user enters one, it
makes sense to ensure that the directory actually exists. The "callback" parameter allows you to do
this. It is a code reference that takes the new value or values as its arguments, and returns true if
the value is valid, and false if it is not. For the sake of convenience, the first argument to the
callback code reference is also stored in $_ .This makes it easy to validate using functions or
operators that, er, operate on $_ by default, but still allows you to get more information from @_ if
necessary. For the directory example, a good callback might be "sub { -d }". The "callback" parameter
code reference will be stored in the "callback" attribute of the App::Info::Request object passed to
event handlers.
error
The error parameter is the error message to display in the event that the "callback" code reference
returns false. This message may then be used by the event handler to let the user know what went
wrong with the data she entered. For example, if the unknown value was a directory, and the user
entered a value that the "callback" identified as invalid, a message to display might be something
like "Invalid directory path". Note that if the "error" parameter is not provided, App::Info will
supply the generic error message "Invalid value". This value will be stored in the "error" attribute
of the App::Info::Request object passed to event handlers.
This may be the event method you use most, as it should be called in every meta data method if you cannot
provide the data needed by that method. It will typically be the last part of the method. Here's an
example demonstrating each of the above arguments:
my $dir = $self->unknown( key => 'lib_dir',
prompt => "Enter lib directory path",
callback => sub { -d },
error => "Not a directory");
confirm
my $val = $self->confirm(@params);
This method is very similar to "unknown()", but serves a different purpose. Use this method for
significant data points where you've found an appropriate value, but want to ensure it's really the
correct value. A "significant data point" is usually a value essential for your class to collect meta
data values. For example, you might need to locate an executable that you can then call to collect other
data. In general, this will only happen once for an object -- during object construction -- but there may
be cases in which it is needed more than that. But hopefully, once you've confirmed in the constructor
that you've found what you need, you can use that information to collect the data needed by all of the
meta data methods and can assume that they'll be right because that first, significant data point has
been confirmed.
Other than where and how often to call "confirm()", its use is quite similar to that of "unknown()". Its
parameters are as follows:
key Same as for "unknown()", a string that uniquely identifies the data point in your class, and ensures
that the event is handled only once for a given key. The same value will be returned by subsequent
calls to "confirm()" as was returned by the first call for a given key.
prompt
Same as for "unknown()". Although "confirm()" is called to confirm a value, typically the prompt
should request the relevant value, just as for "unknown()". The difference is that the handler should
use the "value" parameter as the default should the user not provide a value. The "prompt" parameter
will be stored in the "message" attribute of the App::Info::Request object passed to event handlers.
value
The value to be confirmed. This is the value you've found, and it will be provided to the user as the
default option when they're prompted for a new value. This value will be stored in the "value"
attribute of the App::Info::Request object passed to event handlers.
callback
Same as for "unknown()". Because the user can enter data to replace the default value provided via
the "value" parameter, you might want to validate it. Use this code reference to do so. The callback
will be stored in the "callback" attribute of the App::Info::Request object passed to event handlers.
error
Same as for "unknown()": an error message to display in the event that a value entered by the user
isn't validated by the "callback" code reference. This value will be stored in the "error" attribute
of the App::Info::Request object passed to event handlers.
Here's an example usage demonstrating all of the above arguments:
my $exe = $self->confirm( key => 'shell',
prompt => 'Path to your shell?',
value => '/bin/sh',
callback => sub { -x },
error => 'Not an executable');
EventExamples
Below I provide some examples demonstrating the use of the event methods. These are meant to emphasize
the contexts in which it's appropriate to use them.
Let's start with the simplest, first. Let's say that to find the version number for an application, you
need to search a file for the relevant data. Your App::Info concrete subclass might have a private
method that handles this work, and this method is the appropriate place to use the "info()" and, if
necessary, "error()" methods.
sub _find_version {
my $self = shift;
# Try to find the revelant file. We cover this method below.
# Just return if we cant' find it.
my $file = $self->_find_file('version.conf') or return;
# Send a status message.
$self->info("Searching '$file' file for version");
# Search the file. $util is an App::Info::Util object.
my $ver = $util->search_file($file, qr/^Version\s+(.*)$/);
# Trigger an error message, if necessary. We really think we'll have the
# value, but we have to cover our butts in the unlikely event that we're
# wrong.
$self->error("Unable to find version in file '$file'") unless $ver;
# Return the version number.
return $ver;
}
Here we've used the "info()" method to display a status message to let the user know what we're doing.
Then we used the "error()" method when something unexpected happened, which in this case was that we
weren't able to find the version number in the file.
Note the "_find_file()" method we've thrown in. This might be a method that we call whenever we need to
find a file that might be in one of a list of directories. This method, too, will be an appropriate place
for an "info()" method call. But rather than call the "error()" method when the file can't be found, you
might want to give an event handler a chance to supply that value for you. Use the "unknown()" method for
a case such as this:
sub _find_file {
my ($self, $file) = @_;
# Send a status message.
$self->info("Searching for '$file' file");
# Look for the file. See App::Info:Utility for its interface.
my @paths = qw(/usr/conf /etc/conf /foo/conf);
my $found = $util->first_cat_path($file, @paths);
# If we didn't find it, trigger an unknown event to
# give a handler a chance to get the value.
$found ||= $self->unknown( key => "file_$file",
prompt => "Location of '$file' file?",
callback => sub { -f },
error => "Not a file");
# Now return the file name, regardless of whether we found it or not.
return $found;
}
Note how in this method, we've tried to locate the file ourselves, but if we can't find it, we trigger an
unknown event. This allows clients of our App::Info subclass to try to establish the value themselves by
having an App::Info::Handler subclass handle the event. If a value is found by an App::Info::Handler
subclass, it will be returned by "unknown()" and we can continue. But we can't assume that the unknown
event will even be handled, and thus must expect that an unknown value may remain unknown. This is why
the "_find_version()" method above simply returns if "_find_file()" doesn't return a file name; there's
no point in searching through a file that doesn't exist.
Attentive readers may be left to wonder how to decide when to use "error()" and when to use "unknown()".
To a large extent, this decision must be based on one's own understanding of what's most appropriate.
Nevertheless, I offer the following simple guidelines: Use "error()" when you expect something to work
and then it just doesn't (as when a file exists and should contain the information you seek, but then
doesn't). Use "unknown()" when you're less sure of your processes for finding the value, and also for any
of the values that should be returned by any of the meta data object methods. And of course, "error()"
would be more appropriate when you encounter an unexpected condition and don't think that it could be
handled in any other way.
Now, more than likely, a method such "_find_version()" would be called by the "version()" method, which
is a meta data method mandated by the App::Info abstract base class. This is an appropriate place to
handle an unknown version value. Indeed, every one of your meta data methods should make use of the
"unknown()" method. The "version()" method then should look something like this:
sub version {
my $self = shift;
unless (exists $self->{version}) {
# Try to find the version number.
$self->{version} = $self->_find_version ||
$self->unknown( key => 'version',
prompt => "Enter the version number");
}
# Now return the version number.
return $self->{version};
}
Note how this method only tries to find the version number once. Any subsequent calls to "version()" will
return the same value that was returned the first time it was called. Of course, thanks to the "key"
parameter in the call to "unknown()", we could have have tried to enumerate the version number every
time, as "unknown()" will return the same value every time it is called (as, indeed, should
"_find_version()". But by checking for the "version" key in $self ourselves, we save some of the
overhead.
But as I said before, every meta data method should make use of the "unknown()" method. Thus, the
"major()" method might looks something like this:
sub major {
my $self = shift;
unless (exists $self->{major}) {
# Try to get the major version from the full version number.
($self->{major}) = $self->version =~ /^(\d+)\./;
# Handle an unknown value.
$self->{major} = $self->unknown( key => 'major',
prompt => "Enter major version",
callback => sub { /^\d+$/ },
error => "Not a number")
unless defined $self->{major};
}
return $self->{version};
}
Finally, the "confirm()" method should be used to verify core pieces of data that significant numbers of
other methods rely on. Typically such data are executables or configuration files from which will be
drawn other meta data. Most often, such major data points will be sought in the object constructor.
Here's an example:
sub new {
# Construct the object so that handlers will work properly.
my $self = shift->SUPER::new(@_);
# Try to find the executable.
$self->info("Searching for executable");
if (my $exe = $util->first_exe('/bin/myapp', '/usr/bin/myapp')) {
# Confirm it.
$self->{exe} =
$self->confirm( key => 'binary',
prompt => 'Path to your executable?',
value => $exe,
callback => sub { -x },
error => 'Not an executable');
} else {
# Handle an unknown value.
$self->{exe} =
$self->unknown( key => 'binary',
prompt => 'Path to your executable?',
callback => sub { -x },
error => 'Not an executable');
}
# We're done.
return $self;
}
By now, most of what's going on here should be quite familiar. The use of the "confirm()" method is quite
similar to that of "unknown()". Really the only difference is that the value is known, but we need
verification or a new value supplied if the value we found isn't correct. Such may be the case when
multiple copies of the executable have been installed on the system, we found /bin/myapp, but the user
may really be interested in /usr/bin/myapp. Thus the "confirm()" event gives the user the chance to
change the value if the confirm event is handled.
The final thing to note about this constructor is the first line:
my $self = shift->SUPER::new(@_);
The first thing an App::Info subclass should do is execute this line to allow the super class to
construct the object first. Doing so allows any event handling arguments to set up the event handlers, so
that when we call "confirm()" or "unknown()" the event will be handled as the client expects.
If we needed our subclass constructor to take its own parameter argument, the approach is to specify the
same "key =" $arg> syntax as is used by App::Info's "new()" method. Say we wanted to allow clients of our
App::Info subclass to pass in a list of alternate executable locations for us to search. Such an
argument would most make sense as an array reference. So we specify that the key be "alt_paths" and allow
the user to construct an object like this:
my $app = App::Info::Category::FooApp->new( alt_paths => \@paths );
This approach allows the super class constructor arguments to pass unmolested (as long as we use unique
keys!):
my $app = App::Info::Category::FooApp->new( on_error => \@handlers,
alt_paths => \@paths );
Then, to retrieve these paths inside our "new()" constructor, all we need do is access them directly from
the object:
my $self = shift->SUPER::new(@_);
my $alt_paths = $self->{alt_paths};
SubclassingGuidelines
To summarize, here are some guidelines for subclassing App::Info.
• Always subclass an App::Info category subclass. This will help to keep the App::Info name space well-
organized. New categories can be added as needed.
• When you create the "new()" constructor, always call "SUPER::new(@_)". This ensures that the event
handling methods methods defined by the App::Info base classes (e.g., "error()") will work properly.
• Use a package-scoped lexical App::Info::Util object to carry out common tasks. If you find you're
doing something over and over that's not already addressed by an App::Info::Util method, and you
think that others might find your solution useful, consider submitting a patch to App::Info::Util to
add the functionality you need. See App::Info::Util for complete documentation of its interface.
• Use the "info()" event triggering method to send messages to users of your subclass.
• Use the "error()" event triggering method to alert users of unexpected conditions. Fatal errors
should still be fatal; use "Carp::croak()" to throw exceptions for fatal errors.
• Use the "unknown()" event triggering method when a meta data or other important value is unknown and
you want to give any event handlers the chance to provide the data.
• Use the "confirm()" event triggering method when a core piece of data is known (such as the location
of an executable in the "new()" constructor) and you need to make sure that you have the correct
information.
• Be sure to implement all of the abstract methods defined by App::Info and by your category abstract
base class -- even if they don't do anything. Doing so ensures that all App::Info subclasses share a
common interface, and can, if necessary, be used without regard to subclass. Any method not
implemented but called on an object will generate a fatal exception.
Otherwise, have fun! There are a lot of software packages for which relevant information might be
collected and aggregated into an App::Info concrete subclass (witness all of the Automake macros in the
world!), and folks who are knowledgeable about particular software packages or categories of software are
warmly invited to contribute. As more subclasses are implemented, it will make sense, I think, to create
separate distributions based on category -- or even, when necessary, on a single software package.
Broader categories can then be aggregated in Bundle distributions.
But I get ahead of myself...
Install Now
PNG Icons
