
http://example.com/foo/bar
)'),
'proto-rel' => t('Protocol relative URL (//example.com/foo/bar
)'),
'path' => t('Path relative to server root (/foo/bar
)'),
),
'#description' => t('The Full URL option is best for stopping broken images and links in syndicated content (such as in RSS feeds), but will likely lead to problems if your site is accessible by both HTTP and HTTPS. Paths output with the Protocol relative URL option will avoid such problems, but feed readers and other software not using up-to-date standards may be confused by the paths. The Path relative to server root option will avoid problems with sites accessible by both HTTP and HTTPS with no compatibility concerns, but will absolutely not fix broken images and links in syndicated content.'),
'#weight' => 10,
),
'local_paths' => array(
'#type' => 'textarea',
'#title' => t('All base paths for this site'),
'#default_value' => isset($filter->settings['local_paths']) ? $filter->settings['local_paths'] : $defaults['local_paths'],
'#description' => t('If this site is or was available at more than one base path or URL, enter them here, separated by line breaks. For example, if this site is live at http://example.com/
but has a staging version at http://dev.example.org/staging/
, you would enter both those URLs here. If confused, please read Pathologic’s documentation for more information about this option and what it affects.', array('!docs' => 'http://drupal.org/node/257026')),
'#weight' => 20,
),
);
}
/**
* Pathologic filter callback.
*
* Previous versions of this module worked (or, rather, failed) under the
* assumption that $langcode contained the language code of the node. Sadly,
* this isn't the case.
* @see http://drupal.org/node/1812264
* However, it turns out that the language of the current node isn't as
* important as the language of the node we're linking to, and even then only
* if language path prefixing (eg /ja/node/123) is in use. REMEMBER THIS IN THE
* FUTURE, ALBRIGHT.
*
* The below code uses the @ operator before parse_url() calls because in PHP
* 5.3.2 and earlier, parse_url() causes a warning of parsing fails. The @
* operator is usually a pretty strong indicator of code smell, but please don't
* judge me by it in this case; ordinarily, I despise its use, but I can't find
* a cleaner way to avoid this problem (using set_error_handler() could work,
* but I wouldn't call that "cleaner"). Fortunately, Drupal 8 will require at
* least PHP 5.3.5, so this mess doesn't have to spread into the D8 branch of
* Pathologic.
* @see https://drupal.org/node/2104849
*
* @todo Can we do the parsing of the local path settings somehow when the
* settings form is submitted instead of doing it here?
*/
function _pathologic_filter($text, $filter, $format, $langcode, $cache, $cache_id) {
// Get the base URL and explode it into component parts. We add these parts
// to the exploded local paths settings later.
global $base_url;
$base_url_parts = @parse_url($base_url . '/');
// Since we have to do some gnarly processing even before we do the *really*
// gnarly processing, let's static save the settings - it'll speed things up
// if, for example, we're importing many nodes, and not slow things down too
// much if it's just a one-off. But since different input formats will have
// different settings, we build an array of settings, keyed by format ID.
$cached_settings = &drupal_static(__FUNCTION__, array());
if (!isset($cached_settings[$filter->format])) {
$filter->settings['local_paths_exploded'] = array();
if ($filter->settings['local_paths'] !== '') {
// Build an array of the exploded local paths for this format's settings.
// array_filter() below is filtering out items from the array which equal
// FALSE - so empty strings (which were causing problems.
// @see http://drupal.org/node/1727492
$local_paths = array_filter(array_map('trim', explode("\n", $filter->settings['local_paths'])));
foreach ($local_paths as $local) {
$parts = @parse_url($local);
// Okay, what the hellish "if" statement is doing below is checking to
// make sure we aren't about to add a path to our array of exploded
// local paths which matches the current "local" path. We consider it
// not a match, if…
// @todo: This is pretty horrible. Can this be simplified?
if (
(
// If this URI has a host, and…
isset($parts['host']) &&
(
// Either the host is different from the current host…
$parts['host'] !== $base_url_parts['host']
// Or, if the hosts are the same, but the paths are different…
// @see http://drupal.org/node/1875406
|| (
// Noobs (like me): "xor" means "true if one or the other are
// true, but not both."
(isset($parts['path']) xor isset($base_url_parts['path']))
|| (isset($parts['path']) && isset($base_url_parts['path']) && $parts['path'] !== $base_url_parts['path'])
)
)
) ||
// Or…
(
// The URI doesn't have a host…
!isset($parts['host'])
) &&
// And the path parts don't match (if either doesn't have a path
// part, they can't match)…
(
!isset($parts['path']) ||
!isset($base_url_parts['path']) ||
$parts['path'] !== $base_url_parts['path']
)
) {
// Add it to the list.
$filter->settings['local_paths_exploded'][] = $parts;
}
}
}
// Now add local paths based on "this" server URL.
$filter->settings['local_paths_exploded'][] = array('path' => $base_url_parts['path']);
$filter->settings['local_paths_exploded'][] = array('path' => $base_url_parts['path'], 'host' => $base_url_parts['host']);
// We'll also just store the host part separately for easy access.
$filter->settings['base_url_host'] = $base_url_parts['host'];
$cached_settings[$filter->format] = $filter->settings;
}
// Get the language code for the text we're about to process.
$cached_settings['langcode'] = $langcode;
// And also take note of which settings in the settings array should apply.
$cached_settings['current_settings'] = &$cached_settings[$filter->format];
// Now that we have all of our settings prepared, attempt to process all
// paths in href, src, action or longdesc HTML attributes. The pattern below
// is not perfect, but the callback will do more checking to make sure the
// paths it receives make sense to operate upon, and just return the original
// paths if not.
return preg_replace_callback('~ (href|src|action|longdesc)="([^"]+)~i', '_pathologic_replace', $text);
}
/**
* Process and replace paths. preg_replace_callback() callback.
*/
function _pathologic_replace($matches) {
// Get the base path.
global $base_path;
// Get the settings for the filter. Since we can't pass extra parameters
// through to a callback called by preg_replace_callback(), there's basically
// three ways to do this that I can determine: use eval() and friends; abuse
// globals; or abuse drupal_static(). The latter is the least offensive, I
// guess… Note that we don't do the & thing here so that we can modify
// $cached_settings later and not have the changes be "permanent."
$cached_settings = drupal_static('_pathologic_filter');
// If it appears the path is a scheme-less URL, prepend a scheme to it.
// parse_url() cannot properly parse scheme-less URLs. Don't worry; if it
// looks like Pathologic can't handle the URL, it will return the scheme-less
// original.
// @see https://drupal.org/node/1617944
// @see https://drupal.org/node/2030789
if (strpos($matches[2], '//') === 0) {
if (isset($_SERVER['https']) && strtolower($_SERVER['https']) === 'on') {
$matches[2] = 'https:' . $matches[2];
}
else {
$matches[2] = 'http:' . $matches[2];
}
}
// Now parse the URL after reverting HTML character encoding.
// @see http://drupal.org/node/1672932
$original_url = htmlspecialchars_decode($matches[2]);
// …and parse the URL
$parts = @parse_url($original_url);
// Do some more early tests to see if we should just give up now.
if (
// If parse_url() failed, give up.
$parts === FALSE
|| (
// If there's a scheme part and it doesn't look useful, bail out.
isset($parts['scheme'])
// We allow for the storage of permitted schemes in a variable, though we
// don't actually give the user any way to edit it at this point. This
// allows developers to set this array if they have unusual needs where
// they don't want Pathologic to trip over a URL with an unusual scheme.
// @see http://drupal.org/node/1834308
// "files" and "internal" are for Path Filter compatibility.
&& !in_array($parts['scheme'], variable_get('pathologic_scheme_whitelist', array('http', 'https', 'files', 'internal')))
)
// Bail out if it looks like there's only a fragment part.
|| (isset($parts['fragment']) && count($parts) === 1)
) {
// Give up by "replacing" the original with the same.
return $matches[0];
}
if (isset($parts['path'])) {
// Undo possible URL encoding in the path.
// @see http://drupal.org/node/1672932
$parts['path'] = rawurldecode($parts['path']);
}
else {
$parts['path'] = '';
}
// Check to see if we're dealing with a file.
// @todo Should we still try to do path correction on these files too?
if (isset($parts['scheme']) && $parts['scheme'] === 'files') {
// Path Filter "files:" support. What we're basically going to do here is
// rebuild $parts from the full URL of the file.
$new_parts = @parse_url(file_create_url(file_default_scheme() . '://' . $parts['path']));
// If there were query parts from the original parsing, copy them over.
if (!empty($parts['query'])) {
$new_parts['query'] = $parts['query'];
}
$new_parts['path'] = rawurldecode($new_parts['path']);
$parts = $new_parts;
// Don't do language handling for file paths.
$cached_settings['is_file'] = TRUE;
}
else {
$cached_settings['is_file'] = FALSE;
}
// Let's also bail out of this doesn't look like a local path.
$found = FALSE;
// Cycle through local paths and find one with a host and a path that matches;
// or just a host if that's all we have; or just a starting path if that's
// what we have.
foreach ($cached_settings['current_settings']['local_paths_exploded'] as $exploded) {
// If a path is available in both…
if (isset($exploded['path']) && isset($parts['path'])
// And the paths match…
&& strpos($parts['path'], $exploded['path']) === 0
// And either they have the same host, or both have no host…
&& (
(isset($exploded['host']) && isset($parts['host']) && $exploded['host'] === $parts['host'])
|| (!isset($exploded['host']) && !isset($parts['host']))
)
) {
// Remove the shared path from the path. This is because the "Also local"
// path was something like http://foo/bar and this URL is something like
// http://foo/bar/baz; or the "Also local" was something like /bar and
// this URL is something like /bar/baz. And we only care about the /baz
// part.
$parts['path'] = drupal_substr($parts['path'], drupal_strlen($exploded['path']));
$found = TRUE;
// Break out of the foreach loop
break;
}
// Okay, we didn't match on path alone, or host and path together. Can we
// match on just host? Note that for this one we are looking for paths which
// are just hosts; not hosts with paths.
elseif ((isset($parts['host']) && !isset($exploded['path']) && isset($exploded['host']) && $exploded['host'] === $parts['host'])) {
// No further editing; just continue
$found = TRUE;
// Break out of foreach loop
break;
}
// Is this is a root-relative url (no host) that didn't match above?
// Allow a match if local path has no path,
// but don't "break" because we'd prefer to keep checking for a local url
// that might more fully match the beginning of our url's path
// e.g.: if our url is /foo/bar we'll mark this as a match for
// http://example.com but want to keep searching and would prefer a match
// to http://example.com/foo if that's configured as a local path
elseif (!isset($parts['host']) && (!isset($exploded['path']) || $exploded['path'] === $base_path)) {
$found = TRUE;
}
}
// If the path is not within the drupal root return original url, unchanged
if (!$found) {
return $matches[0];
}
// Okay, format the URL.
// If there's still a slash lingering at the start of the path, chop it off.
$parts['path'] = ltrim($parts['path'],'/');
// Examine the query part of the URL. Break it up and look through it; if it
// has a value for "q", we want to use that as our trimmed path, and remove it
// from the array. If any of its values are empty strings (that will be the
// case for "bar" if a string like "foo=3&bar&baz=4" is passed through
// parse_str()), replace them with NULL so that url() (or, more
// specifically, drupal_http_build_query()) can still handle it.
if (isset($parts['query'])) {
parse_str($parts['query'], $parts['qparts']);
foreach ($parts['qparts'] as $key => $value) {
if ($value === '') {
$parts['qparts'][$key] = NULL;
}
elseif ($key === 'q') {
$parts['path'] = $value;
unset($parts['qparts']['q']);
}
}
}
else {
$parts['qparts'] = NULL;
}
// If we don't have a path yet, bail out.
if (!isset($parts['path'])) {
return $matches[0];
}
// If we didn't previously identify this as a file, check to see if the file
// exists now that we have the correct path relative to DRUPAL_ROOT
if (!$cached_settings['is_file']) {
$cached_settings['is_file'] = !empty($parts['path']) && is_file(DRUPAL_ROOT . '/'. $parts['path']);
}
// Okay, deal with language stuff.
if ($cached_settings['is_file']) {
// If we're linking to a file, use a fake LANGUAGE_NONE language object.
// Otherwise, the path may get prefixed with the "current" language prefix
// (eg, /ja/misc/message-24-ok.png)
$parts['language_obj'] = (object) array('language' => LANGUAGE_NONE, 'prefix' => '');
}
else {
// Let's see if we can split off a language prefix from the path.
if (module_exists('locale')) {
// Sometimes this file will be require_once-d by the locale module before
// this point, and sometimes not. We require_once it ourselves to be sure.
require_once DRUPAL_ROOT . '/includes/language.inc';
list($language_obj, $path) = language_url_split_prefix($parts['path'], language_list());
if ($language_obj) {
$parts['path'] = $path;
$parts['language_obj'] = $language_obj;
}
}
}
// If we get to this point and $parts['path'] is now an empty string (which
// will be the case if the path was originally just "/"), then we
// want to link to Dos grupos colombianos apoyados por el proyecto Hacktividad de la Fundación Gilberto Alzate Avendaño fueron ganadores del concurso La Ruta del Dinero. Hackatón Latinoamericano que busca crear prototipos para saber cómo se destinan los dineros públicos en distintas ciudades.
Una herramienta web que le sirve a la ciudadanía para saber cómo se gasta el dinero público en las entidades ambientales y una propuesta que busca crear una aplicación móvil con información de la contratación de obras públicas en el tema de movilidad, fueron dos iniciativas colombianas ganadoras del concurso La Ruta del Dinero. Hackatón Latinoamericano que se llevó a cabo el pasado 6 y 7 de junio en 14 ciudades del continente.
La sede en Colombia fue Plataforma Bogotá, laboratorio interactivo de arte ciencia y tecnología de la Fundación Gilberto Alzate Avendaño. La entidad acogió el evento en el marco de su convocatoria Hacktividad, espacio que busca apoyar e impulsar procesos tecnológicos al servicio social y comunitario a través de una convocatoria permanente para proyectos hacktivistas.
Hacktivista se entiende como las personas que experimentan con la tecnología y son apasionadas por la seguridad informática. Muchos de ellos utilizan sus conocimientos para hacer “hacktivismo” social y colaboran con personas que están involucradas en procesos comunitarios. El artista Gabriel Zea, explica que hacktividad es una práctica del hacktivismo que consiste en “la unión del activismo tradicional o el activismo de procesos sociales con prácticas tecnológicas”.
La comunidad Hacks/Hackers propuso y organizó la hackatón que se realizó de manera simultánea en Bogotá, Asunción, Buenos Aires, Guatemala, La Paz, Lima, Mendoza, México D.F, Montevideo, Rosario, San José de Costa Rica, San Pablo, Santiago de Chile y Miami, con las organizaciones International Center for Journalists, John S. y James L. Knight Foundation, World Bank Institute, el programa Knight Mozilla Open News y HackLabs.
Durante la jornada de dos días se desarrollaron y compitieron 25 proyectos por cinco estímulos de HackLabs que constan de $2.000 dólares para realizar la propuesta y una tutoría realizada por los becarios del programa Knight-Mozilla Open News, quienes asesoran a los ganadores vía internet a través de charlas electrónicas y revisiones periódicas para que culminen su proyecto satisfactoriamente.
En Plataforma Bogotá se reunieron más de 20 personas en tres grupos para lanzar ideas y crear en tiempo record la parte inicial del proyecto. De ahí nacieron las iniciativas “Dinero verde” y “Yo intervengo”, que finalmente se convirtieron en ganadores del concurso La ruta del dinero. La primera es una aplicación para periodistas y ciudadanía que permite visualizar cómo se gasta el dinero público en las entidades ambientales, así como generar un sistema de “alertas” tempranas del gasto público, importantes para verificar la transparencia y la implementación de los contratos y convenios.
La politóloga Mayra Báez Jimeno, miembro del grupo “Dinero Verde”, afirma que “la idea surgió como una preocupación personal de la manera como se maneja la información en el sector público. Esta aplicación es útil para la ciudadanía y para las mismas entidades del Estado, ya que permite tener los datos estandarizados y organizados”.
Por otro lado, “Yo intervengo” es una propuesta que busca crear una aplicación móvil con información de la contratación de obras públicas, específicamente en el tema de movilidad. Los ingenieros Jorge Ovalle y Roberto Sarmiento, miembros del grupo “Yo intervengo”, explican que esta aplicación puede “crear un sentimiento de pertenencia de las personas con su ciudad en el momento en que se involucran en el proceso de ejecución de obras públicas”.
La herramienta está organizada en tres secciones: (1) visualización de datos de contratación, (2) peticiones de la ciudadanía y (3) quejas sobre la ejecución o finalización de las obras. Cada una de ellas está enfocada en darle las herramientas al ciudadano para que interponga acciones legales concretas y para guiarlo en el proceso. Así, se busca darle transparencia al proceso de contratación y desarrollo de obras públicas por medio de la veeduría ciudadana.
William Valencia, Mayra Báez Jimeno, Daniel Amariles, Jorge Ovalle y Roberto Sarmiento, David Barinas y Daniel Olano, ganadores colombianos de “Dinero verde y “Yo intervengo”, ahora están en la etapa de desarrollo de sus proyectos para tener el producto final en septiembre y participar por un subsidio de 10.000 dólares con el fin de escalar su iniciativa.
Daniel Suarez, co-organizador de Hacks/hackers Bogotá, explica que estos proyectos “buscan que la ciudadanía pueda acceder a la información de manera fácil y eficiente. La democratización de los datos en pro de una ejecución del gasto público transparente”.
La hackatón fue un proceso donde se obtuvieron importantes resultados. Los grupos colombianos se destacaron por encima de otros países de la región con mayor trayectoria en estos proyectos como Brasil y Chile, y lograron fortalecer una comunidad que aboga por la transparencia del sector público colombiano desde la tecnología.