NAME Display - module to display fragments of text on the web and elsewhere SYNOPSIS #!/usr/bin/perl use Display; my $d = Display->new( root_dir => 'archives', url_root => '/display.pl?', # etc. ); print $d->handle(@ARGV); DESCRIPTION Display started life as a simple script to concatenate fragments of handwritten HTML by date. It has since haphazardly accumulated several of the usual weblog features (comments, lightweight markup, feed generation, embedded Perl, poetry tools, image galleries, and ill-advised dependencies), but the basic idea hasn't changed much. The module will work with FastCGI, if called from the appropriate wrapper script. If you use CGI::Fast, you can pass query objects directly to "handle()". By default, entries are stored in a simple directory tree under "root_dir". Like: archives/2001/1/1 archives/2001/1/1/sub_entry It is possible (although not yet as flexible as it ought to be) to redefine the directory layout. More about this after a bit. An entry may be either a plain text file, or a directory containing several files. If it's a directory, a file named "index" will be treated as the text of the entry, and all other lower-case filenames without extensions will be treated as sub-entries or documents within that entry, and displayed accordingly. Links to certain other filetypes will be displayed as well. Directories may be nested to an arbitrary depth, although it's probably not a good idea to go very deep with the current display logic. A PNG or JPEG file with a name like 2001/1/1.icon.png 2001/1/1/index.icon.png 2001/1/1/whatever.icon.png 2001/1/1/whatever/index.icon.png will be treated as an icon for the appropriate entry file. MARKUP Entries may consist of hand-written HTML (to be passed along without further interpretation), a supported form of lightweight markup, or some combination thereof. Actually, an entry may consist of any darn thing you please, as long as Perl will agree that it is text, but presumably you're going to be feeding this to a browser. Special markup is indicated by a variety of HTML-like container tags. Embedded Perl - evaluated and replaced by whatever value you return (evaluated in a scalar context): my $dog = "Ralph."; return $dog; This code is evaluated before any other processing is done, so you can return any other markup understood by the script and have it handled appropriately. Interpolated variables - actually keys to the hash underlying the Display object, for the moment: $self->title("About Ralph, My Dog"); return '';

The title is ${title}.

This will change. Embedded code and variables are intended for use in header and footer files, where it's handy to drop in titles or conditionalize aspects of a layout. You want to be careful with this sort of thing - it's useful in small doses, but it's also a maintainability nightmare waiting to happen. (WordPress, I am looking at you.) Several forms of lightweight markup: Wala::Markup, via Wala.pm - very basic wiki syntax Dean Allen's Textile, via Brad Choate's Text::Textile. An easy way to get properly broken lines plus -- en and em dashes --- for poetry and such. And a couple of shortcuts: filename.ext alt text, if any one list item another list item As it stands, freeverse, image, and list are not particularly robust. ARCHIVE LAYOUT default values entry_map(\%map) Takes a hashref which will dispatch entries matching various regexen to the appropriate output methods. The default looks something like this: nnnn/[nn/nn/]doc_name - a document within a day. nnnn/nn/nn - a specific day. nnnn/nn - a month. nnnn - a year. doc_name - a document in the root directory. You can re-map things to an arbitrary archive layout. Since the entry map is a hash, and handle() simply loops over its keys, there is no guaranteed precedence of patterns. Be extremely careful that no entry will match more than one pattern, or you will wind up with unexpected behavior. A good way to ensure that this does not happen is to use patterns like: qr( ^ # start of string [0-9/]{4}/ # year [0-9]{1,2}/ # month [0-9]{1,2] # day $ # end of string )x ...always marking the start and end of the string explicitly. METHODS For no bigger than this thing is, it gets a little convoluted. new() configure(param => 'value') Set specified parameters. walaconf(%options) Set parameters for Wala.pm. display($entry1, $entry2, ...) Return a string containing the given entries, which can be in the form of CGI query objects or date/entry strings. If no parameters are given, default to default_entry(). display() expands aliases ("new" and "all") and CGI query objects as necessary, collects input from handle($entry), and wraps the whole thing in header and footer files. handle($entry) Return the text of an individual entry. expand_query Expands a CGI query object (for example, one passed in from CGI::Fast) to an appropriate list of parameters. expand_option Expands/converts 'all' and 'new' to appropriate values. recent_month Tries to find the most recent month in the archive. If a year file is text, returns that instead. month_before Return the month before the given month in the archive. Very naive; there has got to be a smarter way. dir_list($dir, $sort_order, $pattern) Return a $sort_order sorted list of files matching regex $pattern in a directory. Calls $sort_order, which can be one of: alpha - alphabetical reverse_alpha - alphabetical, reversed high_to_low - numeric, high to low low_to_high - numeric, low to high year($year) List out the updates for a year. month($month) Prints the entries in a given month (nnnn/nn). entry($entry) Returns the contents of a given entry. Calls dir_list and icon_markup. Recursively calls itself. entry_wrapped Wraps entry() in entry_markup. entry_stamped Wraps entry() + a datestamp in entry_markup() icon_markup Check if an icon exists for a given entry if so, return markup to include it. Icons are PNG or JPEG image files following a specific naming convention: index.icon.[png|jp(e)g] for directories [filename].icon.[png|jp(e)g] for flat text files Calls image_size, uses filename to determine type. datestamp Returns a nice html datestamp for a given entry, including a wikilink for discussion and suchlike. fragment_slurp Read a text fragment, call line_parse to take care of funky markup and interpreting embedded code, and then return it as a string. Takes one parameter, the name of the file, and returns '' if it's not an extant text file. This might be the place to implement an in-memory cache for FastCGI or mod_perl environments. The trick is that the line_parse() results for certain files shouldn't be cached because they contain embedded code. eval_perl Evaluate embedded Perl in a string, replacing blocks enclosed with tags with whatever they return (well, evaluated in a scalar context). Modifies a string in-place, so be careful. Also handles simple ${variables}, replacing them from the keys to $self. month_name Turn numeric dates into English. root_locations($file) * Given a file/entry, return the appropriate concatenations with root_dir and url_root. local_path Return an absolute path for a given file. Called by root_locations. Arguably this is stupid and inefficient. feed_print Return an Atom feed of entries for a month. Defaults to the most recent month in the archive. Called from handle(), requires XML::Atom::SimpleFeed. SEE ALSO walawiki.org, Blosxom, rassmalog, Text::Textile, XML::Atom::SimpleFeed, Image::Size, CGI::Fast. AUTHOR Copyright 2001-2007 Brennen Bearnes Image sizing code (in image_size) derived from wwwis, by Alex Knowles and Andrew Tong. display.pl is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.