function &create(&$root, &$pipeline)
{
// Create contents of this inline box
if ($root->node_type() == XML_TEXT_NODE) {
$handler = get_css_handler('white-space');
return InlineBox::create_from_text($root->content, $handler->get());
} else {
$box =& new InlineBox();
// Initialize content
$child = $root->first_child();
while ($child) {
$child_box =& create_pdf_box($child, $pipeline);
$box->add_child($child_box);
$child = $child->next_sibling();
}
// Add fake whitespace box with zero size for the anchor spans
// We need this, as "reflow" functions will automatically remove empty inline boxes from the
// document tree
//
if ($box->is_null()) {
push_css_defaults();
pop_font_size();
push_font_size('0.01pt');
$whitespace = new WhitespaceBox();
$box->add_child($whitespace);
pop_css_defaults();
}
}
return $box;
}
function &create(&$root, &$pipeline)
{
$box =& new TableRowBox();
$child = $root->first_child();
while ($child) {
$child_box =& create_pdf_box($child, $pipeline);
$box->add_child($child_box);
$child = $child->next_sibling();
}
return $box;
}
function &create(&$root, &$pipeline)
{
$box =& new TableBox();
$box->readCSS($pipeline->getCurrentCSSState());
// This row should not inherit any table specific properties!
// 'overflow' for example
//
$css_state =& $pipeline->getCurrentCSSState();
$css_state->pushDefaultState();
$row =& new TableRowBox($root);
$row->readCSS($css_state);
$box->add_child($row);
$css_state->popState();
// Setup cellspacing / cellpadding values
if ($box->getCSSProperty(CSS_BORDER_COLLAPSE) == BORDER_COLLAPSE) {
$handler =& CSS::get_handler(CSS_PADDING);
$box->setCSSProperty(CSS_PADDING, $handler->default_value());
}
// Set text-align to 'left'; all browsers I've ever seen prevent inheriting of
// 'text-align' property by the tables.
// Say, in the following example the text inside the table cell will be aligned left,
// instead of inheriting 'center' value.
//
// <div style="text-align: center; background-color: green;">
// <table width="100" bgcolor="red">
// <tr><td>TEST
// </table>
// </div>
$handler =& CSS::get_handler(CSS_TEXT_ALIGN);
$handler->css('left', $pipeline);
// Parse table contents
$child = $root->first_child();
$col_index = 0;
while ($child) {
if ($child->node_type() === XML_ELEMENT_NODE) {
if ($child->tagname() === 'colgroup') {
// COLGROUP tags do not generate boxes; they contain information on the columns
//
$col_index = $box->parse_colgroup_tag($child, $col_index);
} else {
$child_box =& create_pdf_box($child, $pipeline);
$box->add_child($child_box);
}
}
$child = $child->next_sibling();
}
$box->normalize($pipeline);
$box->normalize_cwc();
$box->normalize_rhc();
$box->normalize_parent();
return $box;
}
function TableSectionBox(&$root, &$pipeline)
{
$this->GenericContainerBox();
// Automatically create at least one table row
if (count($this->content) == 0) {
$this->content[] =& new TableRowBox($root);
}
// Parse table contents
$child = $root->first_child();
while ($child) {
$child_box =& create_pdf_box($child, $pipeline);
$this->add_child($child_box);
$child = $child->next_sibling();
}
}
function IFrameBox(&$root, $pipeline)
{
$this->InlineBlockBox();
// If NO src attribute specified, just return.
if (!$root->has_attribute('src') || trim($root->get_attribute('src')) == '') {
return;
}
// Determine the fullly qualified URL of the frame content
$src = $root->get_attribute('src');
$url = $pipeline->guess_url($src);
$data = $pipeline->fetch($url);
/**
* If framed page could not be fetched return immediately
*/
if (is_null($data)) {
return;
}
/**
* Render only iframes containing HTML only
*
* Note that content-type header may contain additional information after the ';' sign
*/
$content_type = $data->get_additional_data('Content-Type');
$content_type_array = explode(';', $content_type);
if ($content_type_array[0] != "text/html") {
return;
}
$html = $data->get_content();
// Remove control symbols if any
$html = preg_replace('/[\\x00-\\x07]/', "", $html);
$converter = Converter::create();
$html = $converter->to_utf8($html, $data->detect_encoding());
$html = html2xhtml($html);
$tree = TreeBuilder::build($html);
// Save current stylesheet, as each frame may load its own stylesheets
//
$pipeline->pushCSS();
$css =& $pipeline->getCurrentCSS();
$css->scan_styles($tree, $pipeline);
$frame_root = traverse_dom_tree_pdf($tree);
$box_child =& create_pdf_box($frame_root, $pipeline);
$this->add_child($box_child);
// Restore old stylesheet
//
$pipeline->popCSS();
$pipeline->pop_base_url();
}
function &create(&$root, &$pipeline)
{
$state =& $pipeline->getCurrentCSSState();
$box =& new TableSectionBox();
$box->readCSS($state);
// Automatically create at least one table row
$row = new TableRowBox();
$row->readCSS($state);
$box->add_child($row);
// Parse table contents
$child = $root->first_child();
while ($child) {
$child_box =& create_pdf_box($child, $pipeline);
$box->add_child($child_box);
$child = $child->next_sibling();
}
return $box;
}
function &process($html, &$pipeline, &$media)
{
// Run the XML parser on the XHTML we've prepared
$dom_tree = TreeBuilder::build($html);
// Check if parser returned valid document
if (is_null($dom_tree)) {
readfile(HTML2PS_DIR . '/templates/cannot_parse.html');
error_log(sprintf("Cannot parse document: %s", $pipeline->get_base_url()));
die("HTML2PS Error");
}
/**
* Detect the base URI for this document.
*
* According to the HTML 4.01 p. 12.4.1:
* User agents must calculate the base URI according to the following precedences (highest priority to lowest):
*
* 1. The base URI is set by the BASE element.
* 2. The base URI is given by meta data discovered during a protocol interaction, such as an HTTP header (see [RFC2616]).
* 3. By default, the base URI is that of the current document. Not all HTML documents have a base URI (e.g., a valid HTML document may appear in an email and may not be designated by a URI). Such HTML documents are considered erroneous if they contain relative URIs and rely on a default base URI.
*/
/**
* Check if BASE element present; use its first occurrence
*/
$this->_scan_base($dom_tree, $pipeline);
/**
* @todo fall back to the protocol metadata
*/
/**
* Parse STYLE / LINK nodes containing CSS references and definitions
* This should be done here, as the document body may include STYLE node
* (this violates HTML standard, but is rather often appears in Web)
*/
$css =& $pipeline->getCurrentCSS();
$css->scan_styles($dom_tree, $pipeline);
if (!is_null($media)) {
// Setup media size and margins
$pipeline->get_page_media(1, $media);
$pipeline->output_driver->update_media($media);
$pipeline->_setupScales($media);
}
$body =& traverse_dom_tree_pdf($dom_tree);
$box =& create_pdf_box($body, $pipeline);
return $box;
}
function FrameBox(&$root, &$pipeline)
{
$css_state =& $pipeline->getCurrentCSSState();
// Inherit 'border' CSS value from parent (FRAMESET tag), if current FRAME
// has no FRAMEBORDER attribute, and FRAMESET has one
$parent = $root->parent();
if (!$root->has_attribute('frameborder') && $parent->has_attribute('frameborder')) {
$parent_border = $css_state->getPropertyOnLevel(CSS_BORDER, CSS_PROPERTY_LEVEL_PARENT);
$css_state->setProperty(CSS_BORDER, $parent_border->copy());
}
$this->GenericContainerBox($root);
// If NO src attribute specified, just return.
if (!$root->has_attribute('src')) {
return;
}
// Determine the fullly qualified URL of the frame content
$src = $root->get_attribute('src');
$url = $pipeline->guess_url($src);
$data = $pipeline->fetch($url);
/**
* If framed page could not be fetched return immediately
*/
if (is_null($data)) {
return;
}
/**
* Render only iframes containing HTML only
*
* Note that content-type header may contain additional information after the ';' sign
*/
$content_type = $data->get_additional_data('Content-Type');
$content_type_array = explode(';', $content_type);
if ($content_type_array[0] != "text/html") {
return;
}
$html = $data->get_content();
// Remove control symbols if any
$html = preg_replace('/[\\x00-\\x07]/', "", $html);
$converter = Converter::create();
$html = $converter->to_utf8($html, $data->detect_encoding());
$html = html2xhtml($html);
$tree = TreeBuilder::build($html);
// Save current stylesheet, as each frame may load its own stylesheets
//
$pipeline->pushCSS();
$css =& $pipeline->getCurrentCSS();
$css->scan_styles($tree, $pipeline);
$frame_root = traverse_dom_tree_pdf($tree);
$box_child =& create_pdf_box($frame_root, $pipeline);
$this->add_child($box_child);
// Restore old stylesheet
//
$pipeline->popCSS();
$pipeline->pop_base_url();
}
function FrameBox(&$root, &$pipeline)
{
// Inherit 'border' CSS value from parent (FRAMESET tag), if current FRAME
// has no FRAMEBORDER attribute, and FRAMESET has one
$parent = $root->parent();
if (!$root->has_attribute('frameborder') && $parent->has_attribute('frameborder')) {
pop_border();
push_border(get_border());
}
$this->GenericContainerBox($root);
// If NO src attribute specified, just return.
if (!$root->has_attribute('src')) {
return;
}
// Determine the fullly qualified URL of the frame content
$src = $root->get_attribute('src');
$url = $pipeline->guess_url($src);
$data = $pipeline->fetch($url);
/**
* If framed page could not be fetched return immediately
*/
if (is_null($data)) {
return;
}
/**
* Render only iframes containing HTML only
*
* Note that content-type header may contain additional information after the ';' sign
*/
$content_type = $data->get_additional_data('Content-Type');
$content_type_array = explode(';', $content_type);
if ($content_type_array[0] != "text/html") {
return;
}
$html = $data->get_content();
// Remove control symbols if any
$html = preg_replace('/[\\x00-\\x07]/', "", $html);
$converter = Converter::create();
$html = $converter->to_utf8($html, $data->detect_encoding());
$html = html2xhtml($html);
$tree = TreeBuilder::build($html);
// Save current stylesheet, as each frame may load its own stylesheets
//
global $g_css;
$old_css = $g_css;
global $g_css_obj;
$old_obj = $g_css_obj;
scan_styles($tree, $pipeline);
// Temporary hack: convert CSS rule array to CSS object
$g_css_obj = new CSSObject();
foreach ($g_css as $rule) {
$g_css_obj->add_rule($rule, $pipeline);
}
// TODO: stinks. Rewrite
//
$frame_root = traverse_dom_tree_pdf($tree);
$box_child =& create_pdf_box($frame_root, $pipeline);
$this->add_child($box_child);
// Restore old stylesheet
//
$g_css = $old_css;
$g_css_obj = $old_obj;
$pipeline->pop_base_url();
}
/**
* Create the child nodes of current container object using the parsed HTML data
*
* @param mixed $root node corresponding to the current container object
*/
function create_content(&$root, &$pipeline)
{
// Initialize content
$child = $root->first_child();
while ($child) {
$box_child =& create_pdf_box($child, $pipeline);
$this->add_child($box_child);
$child = $child->next_sibling();
}
}
function BoxPageMargin(&$pipeline, $at_rule)
{
$state =& $pipeline->get_current_css_state();
$state->pushDefaultState();
$root = null;
$at_rule->css->apply($root, $state, $pipeline);
$this->GenericContainerBox();
$this->readCSS($state);
$state->pushDefaultstate();
/**
* Check whether 'content' or '-html2ps-html-content' properties had been defined
* (if both properties are defined, -html2ps-html-content takes precedence)
*/
$raw_html_content =& $at_rule->get_css_property(CSS_HTML2PS_HTML_CONTENT);
$html_content = $raw_html_content->render($pipeline->get_counters());
if ($html_content !== '') {
// We should wrap html_content in DIV tag,
// as we treat only the very first box of the resulting DOM tree as margin box content
$html_content = html2xhtml("<div>" . $html_content . "</div>");
$tree = TreeBuilder::build($html_content);
$tree_root = traverse_dom_tree_pdf($tree);
$body_box =& create_pdf_box($tree_root, $pipeline);
$box =& $body_box->content[0];
} else {
$raw_content =& $at_rule->get_css_property(CSS_CONTENT);
$content = $raw_content->render($pipeline->get_counters());
$box =& InlineBox::create_from_text($content, WHITESPACE_PRE_LINE, $pipeline);
}
$this->add_child($box);
$state->popState();
$state->popState();
}
function &create(&$root, &$pipeline)
{
// Create contents of this inline box
if ($root->node_type() == XML_TEXT_NODE) {
$css_state =& $pipeline->getCurrentCSSState();
return InlineBox::create_from_text($root->content, $css_state->getProperty(CSS_WHITE_SPACE), $pipeline);
} else {
$box =& new InlineBox();
$css_state =& $pipeline->getCurrentCSSState();
$box->readCSS($css_state);
// Initialize content
$child = $root->first_child();
while ($child) {
$child_box =& create_pdf_box($child, $pipeline);
$box->add_child($child_box);
$child = $child->next_sibling();
}
// Add fake whitespace box with zero size for the anchor spans
// We need this, as "reflow" functions will automatically remove empty inline boxes from the
// document tree
//
if ($box->is_null()) {
$css_state->pushState();
$css_state->setProperty(CSS_FONT_SIZE, Value::fromData(0.01, UNIT_PT));
$whitespace = WhitespaceBox::create($pipeline);
$whitespace->readCSS($css_state);
$box->add_child($whitespace);
$css_state->popState();
}
}
return $box;
}
function TableBox(&$root, &$pipeline)
{
// Call parent constructor
$this->GenericContainerBox();
// Initialize line box
// $this->_current_x = 0;
// $this->_current_y = 0;
// List of column width constraints
$this->cwc = array();
// Initialize content
// $this->content = array();
// Automatically create at least one table row
// if (count($this->content) == 0) {
// This row should not inherit any table specific properties!
// 'overflow' for example
//
push_css_defaults();
$this->content[] =& new TableRowBox($root);
pop_css_defaults();
// }
// Setup cellspacing / cellpadding values
$handler =& get_css_handler('border-collapse');
if ($handler->get() == BORDER_COLLAPSE) {
$handler =& get_css_handler('padding');
$handler->css("0", $pipeline);
}
// Set text-align to 'left'; all browsers I've ever seen prevent inheriting of
// 'text-align' property by the tables.
// Say, in the following example the text inside the table cell will be aligned left,
// instead of inheriting 'center' value.
//
// <div style="text-align: center; background-color: green;">
// <table width="100" bgcolor="red">
// <tr><td>TEST
// </table>
// </div>
$handler =& get_css_handler('text-align');
$handler->css('left', $pipeline);
// Parse table contents
$child = $root->first_child();
$col_index = 0;
while ($child) {
if ($child->node_type() === XML_ELEMENT_NODE) {
if ($child->tagname() === 'colgroup') {
// COLGROUP tags do not generate boxes; they contain information on the columns
//
$col_index = $this->parse_colgroup_tag($child, $col_index);
} else {
$child_box =& create_pdf_box($child, $pipeline);
$this->add_child($child_box);
}
}
$child = $child->next_sibling();
}
$this->normalize();
$this->normalize_cwc();
$this->normalize_rhc();
$this->normalize_parent();
}
function &build(&$dom_tree, &$pipeline)
{
$body =& traverse_dom_tree_pdf($dom_tree);
$box =& create_pdf_box($body, $pipeline);
return $box;
}
