=head1 NAME
Prima::Clipboard - GUI interprocess data exchange
=head1 DESCRIPTION
Prima::Clipboard class is a descendant of Prima::Component. It serves as an
interface to the specific data storage, called clipboard, visible to all
clients of one GUI space. The system clipboard is intended for the exchange of
information of an arbitrary type between graphic applications.
=head1 SYNOPSIS
my $c = $::application-> Clipboard;
# paste data
my $string = $c-> text;
my $image = $c-> image;
my $other = $c-> fetch('Other type');
# copy datum
$c-> text( $string);
# copy data
$c-> open;
$c-> text( $string);
$c-> image( $image);
$c-> store( $image);
$c-> close;
# clear
$c-> clear;
=head1 USAGE
Prima::Clipboard provides access to the system clipboard data storage. For the
easier communication, the system clipboard has one 'format' field, that is
stored along with the data. This field is used to distinguish between data
formats. Moreover, a clipboard can hold simultaneously several data instances,
of different data formats. Since the primary usage of a clipboard is 'copying'
and 'pasting', an application can store copied information in several formats,
increasing possibility that the receiving application recognizes the data.
Different systems provide spectrum of predefined data types, but the toolkit
uses only three of these - ascii text, utf8 text, and image. It does not limit,
however, the data format being one of these three types - an application is
free to register its own formats. Both predefined and newly defined data
formats are described by a string, and the three predefined formats are
represented by C<'Text'>, C<'UTF8'>, and C<'Image'> string constants.
The most frequent usage of Prima::Clipboard is to preform two tasks - copying
and pasting. Both can be exemplified by the following:
my $c = $::application-> Clipboard;
# paste
my $string = $c-> text;
# copy
$c-> text( $string);
This simplistic code hides other aspects of Prima::Clipboard class.
First, the default clipboard is accessible by an implicit name call, as an
object named 'Clipboard'. This scheme makes it easily overridable. A more
important point is, that the default clipboard object might be accompanied by
other clipboard objects. This is the case with X11 environment, which defines
also 'Primary' and 'Secondary' system clipboards. Their functionality is
identical to the default clipboard, however. C<get_standard_clipboards()>
method returns strings for the clipboards, provided by the system.
Second, code for fetching and storing multi-format data is somewhat different.
Clipboard is viewed as a shared system resource, and have to be 'opened',
before a process can grab it, so other processes can access the clipboard data
only after the clipboard is 'closed' ( Note: It is not so under X11, where
there the clipboard locking is advisory, and any process can grab clipboard at
any time) .
C<fetch()> and C<store()> implicitly call C<open()> and C<close()>, but these
functions must be called explicitly for the multi-format data handling. The
code below illustrates the said:
# copy text and image
if ( $c-> open) {
$c-> clear;
$c-> store('Text', $string);
$c-> store('Image', $image);
$c-> close;
}
# check present formats and paste
if ( $c-> open) {
if ( $c-> format_exists('Text')) {
$string = $c-> fetch('Text');
}
# or, check the desired format alternatively
my %formats = map { $_ => 1 } $c-> get_formats;
if ( $formats{'Image'}) {
$image = $c-> fetch('Image');
}
$c-> close;
}
The clear() call in the copying code is necessary so
the newly written data will not mix with the old.
At last, the newly registered formats can be accessed
by a program:
my $myformat = 'Very Special Old Pale Data Format';
if ( $c-> register_format($myformat)) {
$c-> open;
$c-> clear;
$c-> store('Text', 'sample text');
$c-> store($myformat', 'sample ## text');
$c-> close;
}
=head2 Custom formats
Once registered, all processes in a GUI space can access the data by this
format. The registration must take place also if a Prima-driven program needs
to read data in a format, defined by an another program. In either case, the
duplicate registration is a valid event. When no longer needed, a format can
be de-registered. It is not a mandatory action, however - the toolkit cleans
up before exit. Moreover, the system maintains a reference counter on the
custom-registered formats; de-registering does not mean deletion, thus. If two
processes use a custom format, and one exits and re-starts, it still can access
the data in the same format, registered by its previous incarnation.
=head2 Unicode
In real life, application often interchange text in both ascii and utf8,
leaving the choice to reader programs. While it is possible to access both at
the same time, by C<fetch>'ing content of C<Text> and C<UTF8> clipboard slots,
widgets implement their own pasting scheme. To avoid hacking widget code, usage
of C<text> property is advised instead of indicating C<'Text'> and C<'UTF8'>
constants. This method is used in standard widgets, and is implemented so the
programmer can reprogram its default action by overloading C<PasteText>
notification of C<Prima::Application> ( see L<Prima::Application/PasteText> ).
The default action of C<PasteText> is to query first if C<'Text'> format is
available, and if so, return the ascii text scalar. If
C<Prima::Application::wantUnicodeInput> is set, C<'UTF8'> format is checked
before resorting to C<'Text'>. It is clear that this scheme is not the only
possibly needed, for example, an application may want to ignore ASCII text, or,
ignore UTF8 text but have C<Prima::Application::wantUnicodeInput> set, etc.
The symmetric action is C<CopyText>, that allows for a custom text conversion
code to be installed.
=head2 Images
Image data can be transferred in different formats in different OSes. The
lowest level is raw pixel data in display-based format, whereas GTK-based
applications can also exchange images in file-based formats, such as bmp, png
etc. To avoid further complications in the implementations, C<PasteImage>
action was introduced to handle these cases, together with a symmetrical
C<CopyImage>.
The default action of C<PasteImage> is to query first if C<'Image'> format is
available, and if so, return the Image object. This by default reads data from
raw image buffer, but if fails, on unix the logic proceeds by checking data in
formats 'image/bmp', 'image/png' etc. C<BMP> is checked first because the
corresponding codec is always compiled in Prima, it doesn't depend on external
libraries. Next is checked C<PNG> format, because it is lossless, then C<TIFF>,
then all others.
When storing the image on the clipboard, only the default format, raw pixel data
is used.
=head1 API
=head2 Properties
=over
=item image OBJECT
Provides access to an image, stored in the system clipboard.
In get-mode call, return C<undef> if no image is stored.
=item text STRING
Provides access to text stored in the system clipboard.
In get-mode call, return C<undef> if no text information is
present.
=back
=head2 Methods
=over
=item clear
Deletes all data from clipboard.
=item close
Closes the open/close brackets. open() and close() can
be called recursively; only the last close() removes the
actual clipboard locking, so other processes can use it as well.
=item deregister_format FORMAT_STRING
De-registers a previously registered data format.
Called implicitly for all not de-registered format
before a clipboard object is destroyed.
=item fetch FORMAT_STRING
Returns the data of FORMAT_STRING data format,
if present in the clipboard. Depending on FORMAT_STRING,
data is either text string for C<'Text'> format,
Prima::Image object for C<'Image'> format and a binary scalar
value for all custom formats.
=item format_exists FORMAT_STRING
Returns a boolean flag, showing whether FORMAT_STRING
format data is present in the clipboard or not.
=item get_handle
Returns a system handle for a clipboard object.
=item get_formats
Returns array of strings, where each is a format ID,
reflecting the formats present in the clipboard.
Only the predefined formats, and the formats registered
via C<register_format()> are returned. There is no
way to see if a format, not registered before, is present.
=item get_registered_formats
Returns array of strings, each representing
a registered format. C<Text> and C<Image>
are returned also.
=item get_standard_clipboards
Returns array of strings, each representing
a system clipboard. The default C<Clipboard>
is always present. Other clipboards are optional.
As an example, this function returns only C<Clipboard>
under win32, but also C<Primary> and C<Secondary>
under X11. The code, specific to these clipboards
must refer to this function first.
=item open
Opens a system clipboard and locks it for the process
single use; returns a success flag. Subsequent C<open>
calls are possible, and always return 1. Each C<open()>
must correspond to C<close()>, otherwise the clipboard
will stay locked until the blocking process is finished.
=item register_format FORMAT_STRING
Registers a data format under FORMAT_STRING string ID,
returns a success flag. If a format is already registered,
1 is returned. All formats, registered via C<register_format()>
are de-registered with C<deregister_format()> when a program is
finished.
=item store FORMAT_STRING, SCALAR
Stores SCALAR value into the clipboard in FORMAT_STRING
data format. Depending of FORMAT_STRING, SCALAR is treated as follows:
FORMAT_STRING SCALAR
------------------------------------
Text text string in ASCII
UTF8 text string in UTF8
Image Prima::Image object
other formats binary scalar value
NB. All custom formats treated as a binary data. In case
when the data are transferred between hosts with different byte orders
no implicit conversions are made. It is up to the programmer
whether to convert the data in a portable format, or leave it as
is. The former option is of course preferable. As far as the author knows,
the I<Storable> module from I<CPAN> collection provides the system-independent
conversion routines.
=back
=head1 AUTHOR
Dmitry Karasik, E<lt>dmitry@karasik.eu.orgE<gt>.
=head1 SEE ALSO
L<Prima>, L<Prima::Component>, L<Prima::Application>