Xmlwiki.php

<php> <?php

1. xmlWiki - MediaWiki XML Hack
2. Nad - 2005-05-18

1. Maybe only use XML transforms
2. - use PHP via processing instructions

1. DTD's
2. - using xmlwiki:document, xmlwiki:image etc is for validation purposes
3. - xwTransformArticle is generic and cares only for xmlwiki:*

1. CSS STYLES
2. the are accessed via http, so their article url can be referred to locally in the html
3. - maybe make the default request 'raw' and must specifically request 'view'
4. - the default CSS's should deal with geshi colour/tab etc

1. LATER:
2. allow XIncludes to build large docs from others
3. allow docBook and DSSSL content

1. ---------------------------------------------------------------------------------------------------------------------- #
2. SCRIPT CONTENTS

1. INIT
2. - do admin and exit if running directly
3. - include geshi
4. - get wiki & request environment
5. - get article (as an object if possible)
6. - get user-info
7. - get object-specific info (perms, doc-root, doc-type)
8. - build env (if action=raw, env is empty)
9. INPUT
10. - deny access if !writable
11. - currently not doing anything else
12. OUTPUT
13. - trap MediaWiki output
14. - extract links (not used yet)
15. - apply security
16. - transform for raw, view, preview
17. - if saving, do publishing
18. - render output & messages
19. ARTICLE FUNCTIONS
20. - get article content
21. - get article type
22. - domificate article
23. - transform article
24. TRANSFORMS
25. - document transforms
26. - image transforms
27. - layout transforms
28. - database table transforms
29. - transform object to php-nav-tree
30. - geshi syntax highlighting
31. FUNCTIONS
32. - etxract hrefs from text
33. - message/log

1. ---------------------------------------------------------------------------------------------------------------------- #
2. INIT

1. If xmlwiki.php requested directly, return admin page

if (!eregi("index\\.php", $_SERVER['SCRIPT_NAME'])) die('xmlwiki.php must be included from index.php!');

1. Geshi syntax highlighting is now called from xmlWiki and should be removed from LocalSettings.php

require_once("extensions/GeshiSyntaxHighlight.php");

1. Otherwise set up xmlWiki global environment ready for input and output processing

$xwMessages = array(); $xwArticleCache = array(); $xwPostTransforms = array('geshi', 'default');

$xwEdit = isset($_REQUEST['action']) && ($_REQUEST['action'] == 'edit'); $xwView = (!isset($_REQUEST['action']) || (isset($_REQUEST['action']) && ($_REQUEST['action'] == 'view'))); $xwRaw = isset($_REQUEST['action']) && ($_REQUEST['action'] == 'raw'); $xwPreview = isset($_REQUEST['wpPreview']); $xwSave = isset($_REQUEST['wpSave']); $xwUserLinks = array(); $xwArticleLinks = array();

$xwArticleTitle = $wgTitle->getPrefixedURL(); $xwArticle = $wgArticle->getContent(false); $xwArticleMeta = xwArticleContent("xmlwiki:$wgTitle");

$xwUserName = $wgUser->mName; $xwUserMeta = xwArticleContent("xmlwiki:$xwUserName");

$xwSecurity = xwArticleContent("xmlwiki:Security"); # holds group membership info $xwReadable = true; $xwWritable = true;

1. Try to convert articles from XML to a DOM object

xwDomificateArticle($xwArticle); xwDomificateArticle($xwArticleMeta); xwDomificateArticle($xwUserMeta);

1. If successfully converted, get perms (?)

if (is_object($xwArticle)) { $xwUserGroups[] = $xwUserName;

# can't write to xmlwiki:security unless admin

$xwReadable = null; $xwWritable = null; }

1. ---------------------------------------------------------------------------------------------------------------------- #
2. INPUT

1. No input processing for now
2. - used later for custom edit forms (database table article will need it)

1. Parse and process input from forms

function xwProcessInput() {

global $xwArticle, $xwArticleTitle; global $xwReadable, $xwWritable, $xwMessages; global $xwSave;

# If article is being updated, if ($xwSave) { #global $_POST; #if (!$xwReadable || !$xwWritable) die("I'm sorry Dave, I'm afraid I can't do that."); }

}

1. ---------------------------------------------------------------------------------------------------------------------- #
2. OUTPUT

1. Generate new output to send back to browser

function xwProcessOutput() {

global $wgOut; global $xwArticle, $xwArticleTitle; global $xwView, $xwPreview, $xwSave, $xwRaw, $xwUserLinks, $xwArticleLinks; global $xwUserName, $xwUserMeta, $xwReadable, $xwWritable, $xwMessages;

# First trap Wiki output and extract content and links (should use $ver like Admin Page) ob_start(); $wgOut->output(); $out = ob_get_contents(); $xwUserLinks = preg_match('/id="p-personal"(.+?)<\/div>/s', $out, $m) ? xwExtractHrefs($m[1]) : array(); $xwArticleLinks = preg_match('/id="p-cactions"(.+?)<\/div>/s', $out, $m) ? xwExtractHrefs($m[1]) : array(); ob_end_clean();

# Security if (!$xwReadable) { # change output to access denied message if !readable } if (!$xwWritable) { # remove edit link }

# If raw, output consists only of transformed article if ($xwRaw) $out = xwTransformArticle();

# If normal view, article must appear within wiki content elseif ($xwView) { if (preg_match('/^(.+)(.+)(.+)$/ms', $out, $m)) { $out = xwTransformArticle(); } else xwMessage('xmlWiki couldn\'t match view-output!', 'red'); }

# If preview, article must appear in wiki content preserving wiki editing content elseif ($xwPreview) {

if (preg_match('/^(.+.+Remember.+saved!.+?<\\/p>.+?

)(.+?)(<\\/p>.+.+)$/s', $out, $m)) { # We have to transform the content in wikiOutput not the article itself $xwArticle = html_entity_decode($m[2]); xwDomificateArticle($xwArticle); $out = xwTransformArticle(); } else xwMessage('xmlWiki couldn\'t match preview-output!', 'red'); } # If saving, check the publishing list elseif (0&& $xwSave && is_object($xwArticle)) { $transformedArticle = xwTransformArticle(); # Get publish list $publist = $xwDocRoot->get_elements_by_tagname('publish'); if (count($publist)) $publist = $publist[0]->dump_mem(true); $publist = split(',', $publist); # Publish each item foreach ($publist as $url) { if ($FH = fopen($url, 'a')) { fwrite($FH, $transformedArticle); fclose($FH); xwMessage("Article published to $url"); } else xwMessage("Failed to publish to $url", 'red'); } } else { # This is probably a special page # - we'll just extract the content and do only post-default-transform if (preg_match('/^(.+)(.+)(.+)$/ms', $out, $m)) $out = xwTransform_default($m[2]); else xwMessage('xmlWiki couldn\'t match special-output!', 'red'); } # Add messages to output unless 'raw' requested if (!$xwRaw) $out = preg_replace('//', .join("\n", $xwMessages), $out); # Render final output and messages xwMessage('Rendering output'); print $out; }

1. ---------------------------------------------------------------------------------------------------------------------- #
2. ARTICLE FUNCTIONS

1. Retreive wiki-article

function xwArticleContent($articleTitle) { # Return with results if already cached global $xwArticleCache; if (isset($xwArticleCache[$articleTitle])) return $xwArticleCache[$articleTitle]; # Get wiki article content (or use global if no article passed) if ($article = new Article(Title::newFromText($articleTitle))) $article = $article->getContent(false); if ($article == '(There is currently no text in this page)') return false; # Return without caching if ($articleTitle) $xwArticleCache[$articleTitle] = $article; return $article; }

1. Decide kind of article from content and title

function xwArticleType($title, $article) { # Name based matches if (ereg('\\.php$', $title)) return 'php'; if (ereg('\\.css$', $title)) return 'css'; if (ereg('\\.as$', $title)) return 'actionscript'; if (ereg('\\.py$', $title)) return 'python'; if (ereg('\\.java$', $title)) return 'java'; if (ereg('\\.(cpp)|(h)$', $title)) return 'cpp'; if (ereg('\\.js$', $title)) return 'javascript'; if (ereg('\\.css$', $title)) return 'css'; # Content based matches if (preg_match('/^<\\?xml.+?\\?>\\s*<xsl:stylesheet/', $article)) return 'xslt'; if (ereg('^<\\?xml', $article)) return 'xml'; if (ereg('^<\\?html', $article)) return 'html4strict'; if (ereg('^#![/a-zA-Z0-9]+\\/perl', $article)) return 'perl'; if (ereg('^#![/a-zA-Z0-9]+sh', $article)) return 'bash'; }

1. Convert passed article to a DOM object
2. - Article is unchanged if not valid XML

function xwDomificateArticle(&$article) { ob_start(); if ($dom = domxml_open_mem($article)) $article = $dom; else { # Could not convert, extract error messages from output xwMessage("Domification failed:", 'red'); $err = preg_replace("/<.+?>/", "", ob_get_contents()); $err = preg_replace('/ in .+? on line.+?[0-9]+/', , $err); $err = preg_replace('/Warning.+?\\(\\): /', , $err); foreach (split("\n", $err) as $msg) if ($msg) xwMessage($msg, 'red'); } ob_end_clean(); }

1. Transform an article object
2. - non objects (strings) can't be transformed because they can't store any transform properties
3. - but the environment post-transforms (incl. geshi) will be applied

function xwTransformArticle($article = false) { global $xwArticle, $xwRaw, $xwPostTransforms; # Use global current article if none passed if ($article === false) $article = $xwArticle; if (is_object($article)) { # Get doc info $aType = $article->doctype(); $aType = $aType->name; $aRoot = $article->document_element(); xwMessage("Transforming doctype \"$aType\""); # If it's one of ours, loop thru transforms applying each # - the first transforms adjust the document while its still a DOM object # - but one of them will transform it to text, # - it may then have further text-based transforms if (ereg('^xmlwiki:', $aType)) { xwMessage('Applying object\'s transforms...'); foreach ($aRoot->get_elements_by_tagname('transform') as $tElement) { $tName = $tElement->get_content(); if (ereg('^(.+)\\(\\)$', $tName, $m)) { # Transform is "name()" format, try to execute from a local function declaration # - probably this is not needed soon, only accept article-transforms $tName = $m[1]; if ($tResult = @call_user_func("xwTransform_$tName", $article)) $article = $tResult; else xwMessage("Unknown transform \"$tName\"!", 'red'); } else { # Get transform-article and apply to $article if ($tText = xwArticleContent($tName)) { # Get transform type $tType = xwArticleType($tName, $tText); if ($tType == 'php') { # php, check perms and execute (trapping errors for reporting - unless fatal) if ($perms_check_out_ok) { ob_start(); eval($tText); $err = preg_replace("/<.+?>/", "", ob_get_contents()); foreach (split("\n", $err) as $msg) if ($msg) xwMessage($msg, 'red'); ob_end_clean(); } } elseif ($tType == 'xslt') { # - should probably trap output same as domify $tObject = domxml_xslt_stylesheet($tText); $tResult = $tObject->process($article); # If output-method is html, use XSLT's dump_mem to create an html-string if (preg_match('/<xsl:output +?method *= *"html"/i', $tText)) { xwMessage('htmlifying'); $tResult = $tObject->result_dump_mem($tResult); } if ($tResult) $article = $tResult; } elseif ($tType == 'css') { # include CSS ref in content } } else xwMessage("No such article \"$tName\"!", 'red'); } } #xwMessage(join(',',$xwTransforms)); # - add php-transform-articles,CSS's & XSLT's to $xwEnv } # If object hasn't been transformed into a string yet (or it isn't one of our kind)... if (!is_string($article)) { xwMessage('Dumping article-object to string'); # validate, report errors # apply stylesheet/XSLT if xml asks for one, # Convert DOM back to XML $article = $article->dump_mem(true); } } # Apply post transforms unless 'raw' requested xwMessage('Applying xmlWiki post-transforms...'); if (!$xwRaw) { foreach ($xwPostTransforms as $tName) if ($tResult = @call_user_func("xwTransform_$tName", $article)) $article = $tResult; else xwMessage("Unknown post-transform \"$tName\"!", 'red'); } return $article; }

1. ---------------------------------------------------------------------------------------------------------------------- #
2. TRANSFORMS
3. - these transforms are only temporarily here - they should be articles
4. - we should use DTD's for each kind: document, image, layout, table

1. This is the default postTransform which surrounds the page
2. - most transforms do not reference the global-current-article like this one
3. - this is a special transform, because it affects all pages, even special pages

function xwTransform_default($article) { global $xwArticle, $xwArticleTitle, $xwUserLinks, $xwArticleLinks; xwMessage("Hi defaultTransform here :-)"); return "

".join(' ',$xwArticleLinks)."

$xwArticleTitle $article

"; } function xwTransform_document_xxx() { # documents are built from parts, # the parts are a tree in its body # includes 'edit' link for each (like section, and also like table) }

1. A document object is finalised by transforming it to its body's text

function xwTransform_document_render() { global $xwArticle; $root = $xwArticle->document_element(); $body = $root->get_elements_by_tagname('body'); if (count($body)) $xwArticle = $body[0]->dump_mem(true); } function xwTransform_layout_xxx() { # for layout container style article } function xwTransform_image_xxx() { # object contains properties for loading and transforming an image # image needs no body # Get the global image on which imageTransforms work global $xwArticleImage; }

1. Converts the in-memory-working image into article content

function xwTransform_image_render() { global $xwArticle, $xwArticleTitle, $xwArticleImage, $xwRaw; if ($xwRaw) { # render article directly if 'raw' request imagePng($xwArticleImage); # Clear article or it will render as xml $xwArticle = ; # set header to image header("Content-type: image/png"); } else { # the raw image is not being requested, render image in-page with info instead $xwArticle = "<img src=\"$xwArticleTitle?action=raw\" alt=\"$xwArticleTitle\"/>"; } }

1. A table object is an interface to a database table
2. - this can't be implimented fully until the 'edit' hook is done
3. - the query result is a list in document body, so any normal document transforms can follow it

function xwTransform_table_xxx() { # - view lists the table/result with links to edit each (like a normal composition-doc) # - also form for create/edit/edlete # - result rows are in document body (just like a content-tree in a document would be) # NODE-LIKE-DB # - if all tables have GUID for pri-key, # - and all have time/duration (or cycle?), # - and we have master list of cols-to-table # - then we could create a meta-query (resulting in pri-key-list of all matches) }

1. Transform the body content to a xtml/javascript tree

function xwTransform_tree() { # import php-tree-component }

1. Include Geshi syntax highlighting
2. - Remember, Geshi should be removed from LocalSettings.php

function xwTransform_geshi($article) { xwMessage("Hi, Geshi here, sorry I'm not currently functional."); return $article; global $xwArticle, $xwArticleTitle; # First check for entire article highlighting # Highlight entire article if xwLang has been set if ($xwLang) { $xwGeshi = new GeSHi($xwArticle, $xwLang, "extensions/geshi/geshi"); $xwGeshi->set_tab_width(4); $geshi->set_keywords_style(1, 'font-weight: 100;', true); #$geshi->set_comments_style #$geshi->set_escape_char_style #$geshi->set_brackets_style #$geshi->set_strings_style #$geshi->set_numbers_style #$geshi->set_methods_style #$geshi->set_symbols_style #$geshi->set_regexps_style #$geshi->set_script_style #$xwArticle = $xwGeshi->parse_code(); } }

1. ---------------------------------------------------------------------------------------------------------------------- #
2. UTILITY FUNCTIONS

1. Return a list of the hrefs found in the passed content

function xwExtractHrefs($content) { preg_match_all('/(<a.+?<\/a>)/', $content, $match); return $match[1]; }

1. Add message to queue

function xwMessage($msg, $col = 'blue') { global $xwMessages; return $xwMessages[] = "

$msg

\n";

}

?> </php>