to \[img\]. preg_replace_callback callback.
*
* @param array $matches Array of matches
* @return string Substituted text
*
* @ignore
*/
function _img_tag_fixup(array $matches) : string
{
$params = trim($matches[1]);
if ($params != '') {
$params .= ' ';
}
$params .= trim($matches[3]);
if ($params != '') {
$params = ' ' . $params;
}
$params = str_replace('alt="', 'param="', $params);
$params = preg_replace('#style="[^"]*vertical-align: ([^;"]+)(;[^"]*)?;?"#i', 'align="${1}"', $params);
$params = str_replace(' class="c-img"', '', $params);
$extraneous = ['border', 'height', 'hspace', 'ismap', 'longdesc', 'usemap', 'vspace', 'width', 'id', 'class', 'title', 'style', 'lang'];
foreach ($extraneous as $ex) {
$params = preg_replace('# ' . $ex . '="[^"]*"#', '', $params);
}
$params = str_replace(' ismap', '', $params);
$params = preg_replace('#\sdata-[\w\-]+="[^"]*"#', '', $params);
/*$referer = post_param_string('http_referer', $_SERVER['HTTP_REFERER']);*/ // CKEditor allows us to specify the base, so we know get_base_url() is right
$caller_url = /*looks_like_url($referer) ? preg_replace('#/[^/]*$#', '', $referer) : */get_base_url();
if ((strpos($matches[2], '{$FIND_SCRIPT') === false) && (strpos($matches[2], '{$IMG') === false)) {
$new_url = qualify_url($matches[2], $caller_url);
} else {
$new_url = $matches[2];
}
return '[img' . rtrim($params) . ']' . $new_url . '[/img]';
}
/**
* Used by semihtml_to_comcode to turn fix URLs in ![]()
to be absolute. preg_replace_callback callback.
*
* @param array $matches Array of matches
* @return string Substituted text
*
* @ignore
*/
function _img_tag_fixup_raw(array $matches) : string
{
/*$referer = post_param_string('http_referer', $_SERVER['HTTP_REFERER']);*/ // CKEditor allows us to specify the base, so we know get_base_url() is right
$caller_url = /*looks_like_url($referer) ? preg_replace('#/[^/]*$#', '', $referer) : */get_base_url();
$matches[2] = html_entity_decode($matches[2], ENT_QUOTES);
if ((strpos($matches[2], '{$FIND_SCRIPT') === false) && (strpos($matches[2], '{$IMG') === false)) {
$new_url = qualify_url($matches[2], $caller_url);
} else {
$new_url = $matches[2];
}
$ret = '';
return $ret;
}
/**
* Used by semihtml_to_comcode to fix tag links. preg_replace_callback callback.
*
* @param array $matches Array of matches
* @return string Substituted text
*
* @ignore
*/
function _a_tag_link_fixup(array $matches) : string
{
$referer = post_param_string('http_referer', $_SERVER['HTTP_REFERER']);
$caller_url = looks_like_url($referer) ? preg_replace('#/[^/]*$#', '', $referer) : get_base_url();
$ret = '';
return $ret;
}
/**
* Used by semihtml_to_comcode to fix CSS colours away from RGB notation. preg_replace_callback callback.
*
* @param array $matches Array of matches
* @return string Substituted text
*
* @ignore
*/
function _css_color_fixup(array $matches) : string
{
$r = dechex(intval(trim($matches[2])));
if (strlen($r) == 1) {
$r = '0' . $r;
}
$g = dechex(intval(trim($matches[3])));
if (strlen($g) == 1) {
$g = '0' . $g;
}
$b = dechex(intval(trim($matches[4])));
if (strlen($b) == 1) {
$b = '0' . $b;
}
return $matches[1] . '#' . $r . $g . $b . $matches[5];
}
/**
* Used by semihtml_to_comcode to make it so inline CSS with quotes uses single quotes. preg_replace_callback callback.
*
* @param array $matches Array of matches
* @return string Substituted text
*
* @ignore
*/
function _css_quot_fixup(array $matches) : string
{
return str_replace('"', '\'', $matches[0]);
}
/**
* Apply temporary ad hoc-escaping to a CDATA area (we'll reverse convert later). preg_replace_callback callback.
*
* @param array $matches Array of matches
* @return string Substituted text
*
* @ignore
*/
function _cdata_protect(array $matches) : string
{
$new = $matches[2];
// We use a closing tag, as we can't just type these in the HTML normally (even in CDATA) - so they are safe unused strings
$new = str_replace(' ', '', $new);
$new = str_replace("\t", '', $new);
$new = str_replace("\n", '', $new);
$new = str_replace("\r", '', $new);
$new = str_replace('&', '', $new);
return $matches[1] . $new . $matches[3];
}
/**
* Apply temporary ad hoc-escaping to a code tags (we'll reverse convert later). preg_replace_callback callback.
*
* @param array $matches Array of matches
* @return string Substituted text
*
* @ignore
*/
function _codetag_protect(array $matches) : string
{
$new = $matches[2];
$new = str_replace('<', '___lt___', $new);
$new = str_replace('>', '___gt___', $new);
return $matches[1] . $new . $matches[3];
}
/**
* Remove temporary ad hoc-escaping to a code tags. preg_replace_callback callback.
*
* @param array $matches Array of matches
* @return string Substituted text
*
* @ignore
*/
function _codetag_unprotect(array $matches) : string
{
$new = $matches[2];
$new = str_replace('___lt___', '<', $new);
$new = str_replace('___gt___', '>', $new);
return $matches[1] . $new . $matches[3];
}
/**
* Reorder XHTML attributes alphabetically, so our regexp's match better. preg_replace_callback callback.
*
* @param array $matches Array of matches
* @return string Substituted text
*
* @ignore
*/
function _reorder_xhtml_attributes(array $matches) : string
{
$middle = trim($matches[2]);
$short = (substr($middle, -1) == '/');
if ($short) {
if (substr($middle, -2) != ' /') {
$middle = substr($middle, 0, strlen($middle) - 1);
} else {
$middle = substr($middle, 0, strlen($middle) - 2);
}
}
$bits = array_map('trim', preg_split('#\s(\w+=)\s*"#', ' ' . $middle, -1, PREG_SPLIT_DELIM_CAPTURE));
array_shift($bits);
$bits2 = [];
$cnt = count($bits);
for ($i = 0; $i < $cnt; $i++) {
if ($i % 2 == 0) {
$bits2[] = $bits[$i];
} else {
$bits2[intval($i / 2)] .= '"' . $bits[$i];
}
}
sort($bits2);
$middle = implode(' ', $bits2);
return '<' . $matches[1] . ' ' . $middle . ($short ? ' /' : '') . '>';
}
/**
* Reorder style properties alphabetically, so our regexp's match better. preg_replace_callback callback.
*
* @param array $matches Array of matches
* @return string Substituted text
*
* @ignore
*/
function _reorder_css_properties(array $matches) : string
{
$middle = $matches[2];
$bits = array_map('trim', explode(';', $middle));
sort($bits);
$middle = '';
foreach ($bits as $bit) {
if (trim($bit) == '') {
continue;
}
if ($middle != '') {
$middle .= '; ';
}
$middle .= trim($bit);
}
return $matches[1] . $middle . $matches[3];
}
/**
* Convert Semi-HTML into Comcode. Cleanup where possible. preg_replace_callback callback.
*
* @param array $matches Array of matches
* @return string Substituted text
*
* @ignore
*/
function _semihtml_to_comcode_wrap(array $matches) : string
{
$middle = semihtml_to_comcode($matches[2]);
if (substr($middle, 0, 10) == '[semihtml]') {
return substr($middle, 10, strlen($middle) - 21);
}
return $matches[1] . $middle . $matches[3];
}
/**
* Extract underlying Comcode from an editor Comcode-management button. preg_replace_callback callback.
*
* @param array $matches Array of matches
* @return string Substituted text
*
* @ignore
*/
function _debuttonise(array $matches) : string
{
return html_entity_decode($matches[1], ENT_QUOTES);
}
/**
* Extract underlying Comcode from an editor XML tag. preg_replace_callback callback.
*
* @param array $matches Array of matches
* @return string Substituted text
*
* @ignore
*/
function _detagonise(array $matches) : string
{
$tag = $matches[1];
$attributes = html_entity_decode(str_replace('"', '\"', isset($matches[2]) ? $matches[2] : ''), ENT_QUOTES);
$attributes = preg_replace('# id="[^"]*"#', '', $attributes); // IDs aren't a real Comcode attribute
return '[' . $tag . $attributes . ']';
}
/**
* Extract underlying Tempcode directive from an editor XML tag. preg_replace_callback callback.
*
* @param array $matches Array of matches
* @return string Substituted text
*
* @ignore
*/
function _dedirectiveise(array $matches) : string
{
$attributes_arr = [];
$matches_attributes = [];
$num_matches_attributes = preg_match_all('#\s+([\w\-]+)\s*=\s*"([^"]*)"#', $matches[1], $matches_attributes);
for ($i = 0; $i < $num_matches_attributes; $i++) {
$attributes_arr[$matches_attributes[1][$i]] = $matches_attributes[2][$i];
}
if (@cms_empty_safe($attributes_arr['params'])) {
return ''; // Should not happen
}
$directive_opener = html_entity_decode($attributes_arr['params'], ENT_QUOTES);
$directive_middle = html_entity_decode(str_replace('
', '', $matches[2]), ENT_QUOTES);
$directive_closer = '{+END}';
$directive = $directive_opener . $directive_middle . $directive_closer;
return $directive;
}
/**
* Extract underlying Tempcode symbol from an editor XML tag. preg_replace_callback callback.
*
* @param array $matches Array of matches
* @return string Substituted text
*
* @ignore
*/
function _desymbolise(array $matches) : string
{
return html_entity_decode($matches[1], ENT_QUOTES);
}
/**
* Cleanup HTML coming out of the WYSIWYG editor, converting represented Comcode back to proper Comcode.
*
* @param string $semihtml Semi-HTML
*/
function remove_wysiwyg_comcode_markup(string &$semihtml)
{
// Our invisible characters isolating the cms Keep markers from style run-off
$semihtml = str_replace('', '', $semihtml);
$array_html_preg_replace = [];
if (get_charset() == 'utf-8') {
$semihtml = str_replace("\u{200B}", '', $semihtml);
}
if (stripos($semihtml, ']*class="cms-keep-ui-controlled" [^>]*title="([^"]*)" [^>]*type="button" [^>]*value="[^"]*"[^>]*/?' . '>#siU', '_debuttonise', $semihtml);
} while ($semihtml != $semihtml_before);
}
// Our Comcode tag start/end markers
$array_html_preg_replace[] = ['#^]*class="(cms-keep|cms-keep-block)"[^>]*>(.*)$#siU', "\${2}"];
$semihtml = array_html_preg_replace('kbd', $array_html_preg_replace, $semihtml);
// Our wrapper tags
if (stripos($semihtml, ']*)?' . '>#', '_detagonise', $semihtml);
$semihtml = preg_replace('##', '[/' . $tag . ']', $semihtml);
}
}
if (stripos($semihtml, ']*)' . '>(.*)#Us', '_dedirectiveise', $semihtml);
}
// Our symbols as meta tags
$semihtml = preg_replace_callback('##', '_desymbolise', $semihtml);
}
/**
* Convert HTML headers to Comcode titles.
*
* @param string $semihtml Semi-HTML
* @param boolean $forceful Whether to force conversion on all header tags, even if they don't match Comcode-style/simple headers exactly
* @return string Semi-HTML, with headers converted to titles
*/
function convert_html_headers_to_titles(string $semihtml, bool $forceful) : string
{
if (stripos($semihtml, ']*>\s*(\s*)?\s*(.*?)\s*\s*(\s*)?\s*#si', '[title="1"]${2}[/title]' . "\n"];
$array_html_preg_replace[] = ['#\s*]*>\s*(\s*)?\s*(.*?)\s*\s*
(\s*)?\s*#si', '[title="1"]${2}[/title]' . "\n"];
$array_html_preg_replace[] = ['#\s*\s*(\s*)?\s*(.*?)\s*\s*
(\s*)?\s*#si', '[title="1"]${2}[/title]' . "\n"];
$array_html_preg_replace[] = ['#\s*]*>(\s*)?\s*(.*?)\s*
(\s*)?\s*#si', '[title="1"]${2}[/title]' . "\n"];
$array_html_preg_replace[] = ['#\s*]*>(\s*)?\s*(.*?)\s*
(\s*)?\s*#si', '[title="1"]${2}[/title]' . "\n"];
$array_html_preg_replace[] = ['#\s*]*>(\s*)?\s*(.*?)\s*
(\s*)?\s*#si', '[title="1"]${2}[/title]' . "\n"];
$array_html_preg_replace[] = ['#\s*(\s*)?\s*(.*?)\s*
(\s*)?\s*#si', '[title="1"]${2}[/title]' . "\n"];
if ($forceful) {
$array_html_preg_replace[] = ['#\s*]*>(\s*)?\s*(.*?)\s*(\s*)?\s*#si', '[title="1"]${2}[/title]' . "\n"];
}
$array_html_preg_replace[] = ['#\s*]+>\s*(.*?)\s*(\s*)?\s*#si', '[title="1"]${2}[/title]' . "\n"];
for ($i = 2; $i <= 6; $i++) {
$array_html_preg_replace[] = ['#\s*\s*(\s*)?\s*(.*?)\s*\s*(\s*)?\s*#si', '[title="' . strval($i) . '"]${2}[/title]' . "\n"];
$array_html_preg_replace[] = ['#\s*(\s*)?\s*(.*?)\s*(\s*)?\s*#si', '[title="' . strval($i) . '"]${2}[/title]' . "\n"];
if ($forceful) {
$array_html_preg_replace[] = ['#\s*]*>(\s*)?\s*(.*?)\s*(\s*)?\s*#si', '[title="' . strval($i) . '"]${2}[/title]' . "\n"];
}
}
foreach ($array_html_preg_replace as $bits) {
list($regexp, $replace) = $bits;
$semihtml = preg_replace($regexp, $replace, $semihtml);
}
}
return $semihtml;
}
/**
* Convert HTML-filled Comcode to cleaner Comcode.
*
* @param LONG_TEXT $comcode The messy Comcode
* @return LONG_TEXT The cleaned Comcode
*/
function force_clean_comcode(string $comcode) : string
{
$matches = [];
if (preg_match('#^\[semihtml\](.*)\[/semihtml\]$#s', $comcode, $matches) != 0) {
if ((strpos($matches[1], '[semihtml]') === false) && (strpos($matches[1], '[html]') === false)) {
return semihtml_to_comcode($matches[1], true);
}
}
if (preg_match('#^\[html\](.*)\[/html\]$#s', $comcode, $matches) != 0) {
if ((strpos($matches[1], '[semihtml]') === false) && (strpos($matches[1], '[html]') === false)) {
return html_to_comcode($matches[1], true);
}
}
return $comcode;
}
/**
* Strip down the contents of the media_set tag for easier WYSIWYG-editing.
*
* @param LONG_TEXT $semihtml The Semi-HTML to be converted
* @return LONG_TEXT The equivalent Comcode
*/
function wysiwygify_media_set(string $semihtml) : string
{
// Media set contents doesn't need any divs, which get left from native attachments
$i = 0;
do {
$media_set_start = strpos($semihtml, '[media_set', $i);
$media_set_end = strpos($semihtml, '[/media_set]', $i);
if (($media_set_start !== false) && ($media_set_end !== false) && ($media_set_end > $media_set_start)) {
$middle_before = substr($semihtml, $media_set_start, $media_set_end - $media_set_start);
$middle_after = preg_replace('#]*)?' . '>#', '', $middle_before);
$middle_after = preg_replace('#]*)? ' . '>.*#Us', '', $middle_after);
$semihtml = substr($semihtml, 0, $media_set_start) . $middle_after . substr($semihtml, $media_set_end);
$i = $media_set_end - (strlen($middle_before) - strlen($middle_after)) + 1;
}
} while (($media_set_start !== false) && ($media_set_end !== false) && ($media_set_end > $media_set_start));
return $semihtml;
}
/**
* Convert Semi-HTML into Comcode. Cleanup where possible.
*
* @param LONG_TEXT $semihtml The Semi-HTML to be converted
* @param boolean $force Whether to force full conversion regardless of settings
* @param boolean $quick Whether to trust the HTML is valid rather than cleaning it up (e.g. for software-generated HTML)
* @param ?MEMBER $member_id Member to do as (null: current member)
* @return LONG_TEXT The equivalent Comcode
*/
function semihtml_to_comcode(string $semihtml, bool $force = false, bool $quick = false, ?int $member_id = null) : string
{
if ($member_id === null) {
$member_id = get_member();
}
// Optimisations
$matches = [];
if (preg_match('#^\[semihtml\]([^\[\]<>]*)\[\/semihtml\]$#', $semihtml, $matches) != 0) { // Already have semihtml tags
return $matches[1];
}
if (preg_match('#^([^\[\]<>\{\}&]*)$#', $semihtml) != 0) { // Plain-text
return $semihtml;
}
$semihtml = trim($semihtml);
require_code('templates');
// Optimisation, not long enough to clean up
if (cms_trim($semihtml, strlen($semihtml) < 30) === '') {
return '';
}
$decoded = html_entity_decode($semihtml, ENT_QUOTES);
if ((strpos($decoded, '<') === false) && (strpos($decoded, '[') === false) && (strpos($decoded, '{') === false) && (strpos($decoded, '&') === false)) {
return $decoded;
}
require_code('crypt');
cms_ini_set('pcre.backtrack_limit', '10000000');
// Special clean up we always do regardless...
// Software markers
remove_wysiwyg_comcode_markup($semihtml);
// Links should be kept from being base URL-specific
$semihtml = reinstate_static_tempcode($semihtml);
// Empty comments
$semihtml = str_replace('#sU', '', $semihtml);
$semihtml = preg_replace('##sU', '', $semihtml);
$semihtml = str_replace(' $easy_replace)
$easy_replace = true;
$on_closer = true;
$pos = 0;
do {
$pos_opener_1 = strpos($semihtml, '<' . $element . '>', $pos);
$pos_opener_2 = strpos($semihtml, '<' . $element . ' ', $pos);
$pos_opener = (($pos_opener_1 !== false) && (($pos_opener_2 === false) || ($pos_opener_1 < $pos_opener_2))) ? $pos_opener_1 : $pos_opener_2;
if ($pos_opener === false) {
break;
}
if ($pos == 0) { // First iteration is just to find first opener
$pos = $pos_opener + 1;
continue;
}
$pos_closer_1 = strpos($semihtml, '', $pos);
$pos_closer_2 = strpos($semihtml, ']#', $semihtml, $matches, PREG_OFFSET_CAPTURE);
$tags = [];
for ($i = 0; $i < $count; $i++) {
$is_closer = ($matches[1][$i][0] == '/');
$tags[] = [
$is_closer ? -1 : 1, // Balancer
$matches[0][$i][1], // Offset
strlen($matches[0][$i][0]), // Length
];
}
$num_tags = count($tags);
foreach ($array as $index => $temp) {
list($pattern, $replacement) = $temp;
foreach ($tags as $i => $tag) {
if ($tag[0] == 1) {
$start = $tag[1];
// Find the matching end position
$end = null;
$balance = 0;
for ($j = $i; $j < $num_tags; $j++) {
$balance += $tags[$j][0];
if ($balance == 0) {
$end = $tags[$j][1];
$length = $tags[$j][2];
break;
}
}
if ($end === null) {
break;
}
// Process segment
$segment = substr($semihtml, $start, $end + $length - $start);
$before = substr($semihtml, 0, $start);
$after = substr($semihtml, $end + $length);
$subbed = cms_preg_replace_safe($pattern . 'A', $replacement, $segment);
$semihtml = $before . $subbed . $after;
if ($semihtml != $old_semihtml) {
break 2; // We need to start again now as the offsets have all changed
}
}
}
unset($array[$index]); // If we are going to recurse, we don't want extra work -- let's record that this one completed
}
} while (cms_preg_replace_safe('#(\s|]*>| )#i', '', $semihtml) != cms_preg_replace_safe('#(\s|]*>| )#i', '', $old_semihtml) && cms_preg_safety_guard_ok($safety_guard));
return $semihtml;
}